Interference Rejection Thresholds 
of Consumer Digital Television Receivers 
Available in 2005 and 2006  
 
 
 
 
 
March 30, 2007 
 
 
 
 
 
 
 
Technical Research Branch 
Laboratory Division 
Office of Engineering and Technology 
Federal Communications Commission 
 
 
 
 
 
 
 
 
 
OET Report Prepared by: 
FCC/OET 07-TR-1003 Stephen R. Martin 
 
i 
 
ACKNOWLEDGEMENTS 
 
The author gratefully acknowledges the following contributions to this work.  Mark Hryszko of the 
Digital Television group of Advanced Micro Devices provided the “Muddy Waters” RF vector file 
(mathematically derived from an MPEG2 transport stream) that was used with the Wavetech WS2100 RF 
player to provide a “desired signal” for the final few tests performed for this report.  John Gabrysch of the 
Commission’s Media Bureau set up and evaluated the performance of the Rohde and Schwarz SFU DTV 
signal generator and developed channel assignment files for the instrument.  The work of Charles Rhodes 
on third-order intermodulation inspired the paired-signal tests in this project.  Gary Sgrignoli of MSW 
provided a fast-track DTV education to prepare the author for the prequel to this project.  The author 
thanks the following Commission employees for reviewing drafts of this document:  Alan Stillwell, 
Rashmi Doshi, and William Hurst (Office of Engineering and Technology); Doug Miller, John Raymond, 
and Steve DeSena (Enforcement Bureau); Richard Engelman and Sankar Persaud (International Bureau).  
The author also acknowledges the support of wife, Dr. Bonnie Dorr, through many long hours spent 
completing this work. 
 
ii 
 
TABLE OF CONTENTS 
 
EXECUTIVE SUMMARY ....................................................................................................................................... viii 
Tuner Type Tests .................................................................................................................................................. viii 
Interference Rejection Tests ....................................................................................................................................ix 
Interference From One Undesired Signal ................................................................................................................ix 
Interference From IM3-Generating Pairs of Undesired Signals ..............................................................................xi 
CHAPTER 1 INTRODUCTION...................................................................................................................................1 
Background...............................................................................................................................................................1 
Objectives .................................................................................................................................................................2 
Measurement Program..............................................................................................................................................3 
Overview...................................................................................................................................................................5 
CHAPTER 2 SCOPE AND APPROACH................................................................................................................. 2-1 
Scope of Testing ................................................................................................................................................... 2-1 
Interference Rejection Testing Approach ............................................................................................................. 2-4 
Presentation of Results.......................................................................................................................................... 2-6 
Test Summary....................................................................................................................................................... 2-8 
CHAPTER 3 TUNER TYPE TESTS ........................................................................................................................ 3-1 
Tuner Topologies.................................................................................................................................................. 3-1 
LO Measurement For Tuner Type Identification.................................................................................................. 3-2 
Interference Rejection Tests For Tuner Type Identification................................................................................. 3-3 
Summary............................................................................................................................................................... 3-4 
CHAPTER 4 INTERFERENCE REJECTION REFERENCE LEVELS AND TEST METHODOLOGY .............. 4-1 
Reference Levels for Interference Rejection Performance ................................................................................... 4-1 
Test Setup Requirements ...................................................................................................................................... 4-3 
Test Setup ............................................................................................................................................................. 4-4 
Test Setup Performance ........................................................................................................................................ 4-5 
Test Setup Configurations..................................................................................................................................... 4-7 
CHAPTER 5 REJECTION RESULTS ON CHANNEL 30 FOR SINGLE INTERFERERS WITH FULL 
CHANNEL WIDTH.................................................................................................................................................. 5-1 
Tests at ATSC-Specified Desired Signal Levels .................................................................................................. 5-1 
Selection of a Weaker Desired Signal for Tests.................................................................................................... 5-3 
Effect of Desired Signal Source............................................................................................................................ 5-5 
Tests with Desired Signal at D
MIN
 + 3 dB............................................................................................................. 5-7 
Summary of Fifth-Generation Receiver Results ................................................................................................... 5-8 
Earlier-Generation Receivers................................................................................................................................ 5-9 
Taboo Effects and Other Observations ................................................................................................................. 5-9 
CHAPTER 6 REJECTION RESULTS ON CHANNEL 51 FOR SINGLE 8-VSB INTERFERERS....................... 6-1 
Tests at ATSC-Specified Desired Signal Levels .................................................................................................. 6-1 
Comparison to Channel-30 Results....................................................................................................................... 6-2 
CHAPTER 7 TESTS WITH DIFFERENT SIGNAL TYPES AND SOURCES ...................................................... 7-1 
1-MHz Bandwidth Undesired Signal.................................................................................................................... 7-3 
Desired Signal Source:  SFU Versus ATSC997 ................................................................................................... 7-3 
Repeat Baseline Test............................................................................................................................................. 7-5 
Undesired Signal Type:  8-VSB Versus Bandlimited Gaussian Noise.................................................................7-6 
Undesired Signal Type:  DVB-H OFDM Versus Bandlimited Gaussian Noise ................................................... 7-7 
CHAPTER 8 THEORETICAL FRAMEWORK FOR OUT-OF-CHANNEL INTERFERENCE............................ 8-1 
Interference Mechanisms and Order..................................................................................................................... 8-1 
Threshold of Visibility of Picture Degradation Due to Interference..................................................................... 8-2 
Slopes of Log-Log Plots of D, U, and D/U........................................................................................................... 8-3 
Effect of AGC....................................................................................................................................................... 8-3 
Weak Signals ........................................................................................................................................................ 8-4 
IM3 With Paired Signals....................................................................................................................................... 8-5 
CHAPTER 9 THIRD-ORDER INTERMODULATION WITH PAIRED EQUAL-POWER INTERFERERS....... 9-1 
Spectra of Third-Order Intermodulation Distortion.............................................................................................. 9-1 
Channel-30 Results at Three Desired Signal Levels............................................................................................. 9-3 
Channel-51 Results at Three Desired Signal Levels............................................................................................. 9-6 
 
iii 
Estimating 3
rd 
Order Intercept Point (IP3) ............................................................................................................ 9-8 
Summary Data .................................................................................................................................................... 9-12 
CHAPTER 10 IM3 WITH PAIRED SIGNALS OF UNEQUAL AMPLITUDES ................................................. 10-1 
Measurements on Two Receivers ....................................................................................................................... 10-1 
Modeling and Extrapolation ............................................................................................................................... 10-2 
CHAPTER 11 SINGLE AND PAIRED REJECTION RATIOS VERSUS DESIRED SIGNAL LEVEL—A 
DETAILED EXAMPLE.......................................................................................................................................... 11-1 
General Observations.......................................................................................................................................... 11-2 
IM3 With Paired Undesired Signals ................................................................................................................... 11-3 
Single Undesired Signals .................................................................................................................................... 11-6 
Summary............................................................................................................................................................. 11-9 
CHAPTER 12 EXTENDING THE RESULTS TO LOWER SIGNAL LEVELS .................................................. 12-1 
Order of Interference Processes and Effect of AGC........................................................................................... 12-1 
Extrapolation to D = D
MIN
 + 1 dB....................................................................................................................... 12-2 
CHAPTER 13 COMBINING MEASURED AND EXTRAPOLATED RESULTS ............................................... 13-1 
CHAPTER 14 VERIFICATION TESTS USING ALTERNATIVE METHODS................................................... 14-1 
Alternative Test Setup With Bandpass Filter on Undesired Signal .................................................................... 14-1 
Broadband Notched Noise .................................................................................................................................. 14-2 
Screen-Room Test............................................................................................................................................... 14-5 
Summary............................................................................................................................................................. 14-5 
CHAPTER 15 SUMMARY AND CONCLUSIONS ..............................................................................................15-1 
DTV Tuner Type ................................................................................................................................................ 15-2 
Effect of Undesired Signal Type......................................................................................................................... 15-2 
Cliff Effect.......................................................................................................................................................... 15-2 
Interference from a Single Undesired Signal...................................................................................................... 15-3 
Interference From IM3-Generating Pairs of Undesired Signals ......................................................................... 15-4 
APPENDIX A TABULATED REJECTION PERFORMANCE RESULTS ........................................................... A-1 
Threshold D/U Statistics for 8 Fifth-Generation DTV Receivers........................................................................ A-2 
Threshold U Statistics for 8 Fifth-Generation DTV Receivers............................................................................ A-7 
APPENDIX B THEORETICAL BASIS FOR OUT-OF-CHANNEL INTERFERENCE ........................................B-1 
Single Interferer....................................................................................................................................................B-2 
IM3 With Paired Signals.......................................................................................................................................B-9 
Summary.............................................................................................................................................................B-12 
BIBLIOGRAPHY 
 
 
 
 
iv 
 
ILLUSTRATIONS 
 
Figure 2-1.  Spectra of the Four Undesired Signal Sources..................................................................................... 2-11 
Figure 2-2.  Spectra of the Four Undesired Signal Sources—Expanded Scale ....................................................... 2-11 
Figure 2-3.  Relationship Between Signal Excess and Coverage Area.................................................................... 2-12 
Figure 3-1.  Single-Conversion DTV Tuner Block Diagram Example...................................................................... 3-4 
Figure 3-2.  Double-Conversion DTV Tuner Block Diagram Example.................................................................... 3-4 
Figure 3-3.  Rejection Performance of Receivers with Undetected LO’s.................................................................. 3-5 
Figure 4-1.  Block Diagram of Interference Rejection Test Setup ............................................................................ 4-9 
Figure 4-2.  Leakage of U at N+2 into Channel N (30)........................................................................................... 4-10 
Figure 4-3.  Leakage of U pair at N-2/N-4 into Channel N (30) ............................................................................. 4-10 
Figure 4-4.  Leakage of U pair at N+1/N+2 into Channel N (30)............................................................................ 4-11 
Figure 5-1.  D/U of 8 Receivers at D = -68 dBm on Channel 30 ............................................................................ 5-12 
Figure 5-2.  D/U Statistics at D = -68 dBm on Channel 30 ..................................................................................... 5-12 
Figure 5-3.  D/U of 8 receivers at D = -53 dBm on Channel 30.............................................................................. 5-13 
Figure 5-4.  D/U Statistics at D = -53 dBm on Channel 30 ..................................................................................... 5-13 
Figure 5-5.  D/U of 8 receivers at D = -28 dBm on Channel 30.............................................................................. 5-14 
Figure 5-6.  D/U Statistics at D = -28 dBm on Channel 30 ..................................................................................... 5-14 
Figure 5-7.  Spectra of Two Desired Signal Sources............................................................................................... 5-15 
Figure 5-8.  Spectra of Two Desired Signal Sources on Expanded Scale ............................................................... 5-15 
Figure 5-9.  D/U of 8 receivers at D = D
MIN
 + 3 dB  on Channel 30 ....................................................................... 5-16 
Figure 5-10.  D/U Statistics at D = D
MIN
 + 3 dB on Channel 30 ............................................................................. 5-16 
Figure 5-11.  Median D/U of 8 receivers at Four Signal Levels on Channel 30...................................................... 5-17 
Figure 5-12.  Median Threshold U of 8 receivers at Four Signal Levels on Channel 30 ........................................ 5-17 
Figure 5-13.  2
nd
 Worst D/U of 8 receivers at Four Signal Levels on Channel 30 .................................................. 5-18 
Figure 5-14.  2
nd
 Worst Threshold U of 8 receivers at Four Signal Levels on Channel 30 ..................................... 5-18 
Figure 5-15.  Worst D/U of 8 receivers at Four Signal Levels on Channel 30 ........................................................ 5-19 
Figure 5-16.   Worst Threshold U of 8 receivers at Four Signal Levels on Channel 30.......................................... 5-19 
Figure 5-17.  D/U of Two Earlier-Generation Receivers at D = -68 dBm Relative to Fifth-Generation Results .... 5-20 
Figure 6-1.  D/U of 7 Receivers at D = -68 dBm on Channel 51 .............................................................................. 6-3 
Figure 6-2.  D/U Statistics at D = -68 dBm on Channel 51 ....................................................................................... 6-3 
Figure 6-3.  D/U of 7 Receivers at D = -53 dBm on Channel 51 .............................................................................. 6-4 
Figure 6-4.  D/U Statistics at D = -53 dBm on Channel 51 ....................................................................................... 6-4 
Figure 6-5.  D/U of 7 Receivers at D = -28 dBm on Channel 51 .............................................................................. 6-5 
Figure 6-6.  D/U Statistics at D = -28 dBm on Channel 51 ....................................................................................... 6-5 
Figure 6-7.  Comparison of Median D/U for 7 Receivers on Channels 30 and 51 at D = -68 dBm .......................... 6-6 
Figure 6-8.  Comparison of Median D/U for 7 Receivers on Channels 30 and 51 at D = -53 dBm .......................... 6-6 
Figure 7-1.  D/U With 1-MHz Undesired Signal Bandwidth .................................................................................... 7-8 
Figure 8-1.  Deviation of Log-U Versus Log-D From Straight Line For Linear Interference................................... 8-7 
Figure 8-2.  Deviation of Log-U Versus Log-D From Straight Line For 2
nd
-Order Interference.............................. 8-7 
Figure 8-3.  Deviation of Log-U Versus Log-D From Straight Line For 3
rd
-Order Interference .............................. 8-8 
Figure 8-4.  Deviation of Log-U Versus Log-D From Straight Line For Cross-Modulation .................................... 8-8 
Figure 9-1.  Third-Order  Intermodulation Distortion Spectra of Single and Paired Signals .................................. 9-15 
Figure 9-2.  Third-Order Intermodulation Distortion Spectra of Paired Signals Versus Amplitude ....................... 9-16 
Figure 9-3.  Paired-Signal D/U of 8 receivers at D = -68 dBm on Channel 30 ....................................................... 9-17 
Figure 9-4.  Paired-Signal D/U Statistics of 8 receivers at D = -68 dBm on Channel 30........................................ 9-17 
Figure 9-5.  Paired-Signal D/U of 8 receivers at D = -53 dBm on Channel 30 ....................................................... 9-18 
Figure 9-6.  Paired-Signal D/U Statistics of 8 receivers at D = -53 dBm on Channel 30........................................ 9-18 
Figure 9-7.  Paired-Signal D/U of 8 receivers at D = -28 dBm on Channel 30—1st-Adjacent Only ...................... 9-19 
Figure 9-8.  Paired-Signal D/U of 8 receivers at D = -28 dBm on Channel 30—1st-Adjacent Only ...................... 9-19 
Figure 9-9.  Paired-Signal D/U of Receiver A3 on Channel 30 with Summed D/U’s as Reference ....................... 9-20 
Figure 9-10.  Paired-Signal D/U of Receiver D3 on Channel 30 with Summed D/U’s as Reference ..................... 9-20 
Figure 9-11.  Paired-Signal D/U of Receiver I1 on Channel 30 with Summed D/U’s as Reference....................... 9-21 
Figure 9-12.  Paired-Signal D/U of Receiver J1 on Channel 30 with Summed D/U’s as Reference....................... 9-21 
Figure 9-13.  Paired-Signal D/U of Receiver M1 on Channel 30 with Summed D/U’s as Reference..................... 9-22 
Figure 9-14.  Paired-Signal D/U of Receiver N1 on Channel 30 with Summed D/U’s as Reference ..................... 9-22 
Figure 9-15.  Paired-Signal D/U of Receiver O1 on Channel 30 with Summed D/U’s as Reference ..................... 9-23 
 
v 
Figure 9-16.  Paired-Signal D/U of Receiver G4 on Channel 30 with Summed D/U’s as Reference ..................... 9-23 
Figure 9-17.  Paired-Signal D/U of 7 receivers at D = -68 dBm on Channel 51 .....................................................9-24 
Figure 9-18.  Paired-Signal D/U Statistics of 7 receivers at D = -68 dBm on Channel 51...................................... 9-24 
Figure 9-19.  Paired-Signal D/U of 7 receivers at D = -53 dBm on Channel 51 .....................................................9-25 
Figure 9-20.  Paired-Signal D/U Statistics of 6 receivers at D = -53 dBm on Channel 51...................................... 9-25 
Figure 9-21.  Paired-Signal D/U of 7 receivers at D = -28 dBm on Channel 51 .....................................................9-26 
Figure 9-22.  Paired-Signal D/U Statistics of 6 receivers at D = -28 dBm on Channel 51...................................... 9-26 
Figure 9-23.  Paired-Signal D/U of Receiver A3 on Channel 51 with Summed D/U’s as Reference ..................... 9-27 
Figure 9-24.  Paired-Signal D/U of Receiver D3 on Channel 51 with Summed D/U’s as Reference ..................... 9-27 
Figure 9-25.  Paired-Signal D/U of Receiver I1 on Channel 51 with Summed D/U’s as Reference....................... 9-28 
Figure 9-26.  Paired-Signal D/U of Receiver J1 on Channel 51 with Summed D/U’s as Reference....................... 9-28 
Figure 9-27.  Paired-Signal D/U of Receiver M1 on Channel 51 with Summed D/U’s as Reference..................... 9-29 
Figure 9-28.  Paired-Signal D/U of Receiver N1 on Channel 51 with Summed D/U’s as Reference ..................... 9-29 
Figure 9-29.  Paired-Signal D/U of Receiver O1 on Channel 51 with Summed D/U’s as Reference ..................... 9-30 
Figure 9-30.  Third-Order Intercept Point Parameter with Desired Signal = -68 dBm on Channel 30.................... 9-31 
Figure 9-31.  Third-Order Intercept Point Parameter with Desired Signal = -53 dBm on Channel 30.................... 9-31 
Figure 9-32.  Third-Order Intercept Point Parameter with Desired Signal = -68 dBm on Channel 51.................... 9-32 
Figure 9-33.  Third-Order Intercept Point Parameter with Desired Signal = -53 dBm on Channel 51.................... 9-32 
Figure 9-34.  Third-Order Intercept Point Parameter—Channel 30/51 Comparison............................................... 9-33 
Figure 10-1.  Threshold U for Paired, Unequal Undesired Signals on Receiver G4 ............................................... 10-6 
Figure 10-2.  Threshold U for Paired, Unequal Undesired Signals on Receiver M1............................................... 10-6 
Figure 10-3.  Threshold U for Paired, Unequal Undesired Signals on Receiver M1 at D = -68 dBm..................... 10-7 
Figure 10-4.  Modeled Thresholds for N-5/N-10 with D = -68 dBm ...................................................................... 10-7 
Figure 10-5.  Modeled Thresholds for N-4/N-8 with D = -68 dBm ........................................................................ 10-8 
Figure 10-6.  Modeled Thresholds for N-3/N-6 with D = -68 dBm ........................................................................ 10-8 
Figure 10-7.  Modeled Thresholds for N+2/N+4 with D = -68 dBm....................................................................... 10-9 
Figure 10-8.  Modeled Thresholds for N+3/N+6 with D = -68 dBm....................................................................... 10-9 
Figure 10-9.  Modeled Thresholds for N+4/N+8 with D = -68 dBm..................................................................... 10-10 
Figure 10-10.  Modeled Thresholds for N+5/N+10 with D = -68 dBm.................................................................10-10 
Figure 10-11.  Modeled Thresholds for N-5/N-10 with D = D
MIN
 + 3 dB............................................................. 10-11 
Figure 10-12.  Modeled Thresholds for N-4/N-8 with D = D
MIN
 + 3 dB............................................................... 10-11 
Figure 10-13.  Modeled Thresholds for N-3/N-6 with D = D
MIN
 + 3 dB............................................................... 10-12 
Figure 10-14.  Modeled Thresholds for N+2/N+4 with D = D
MIN
 + 3 dB............................................................. 10-12 
Figure 10-15.  Modeled Thresholds for N+3/N+6 with D = D
MIN
 + 3 dB............................................................. 10-13 
Figure 10-16.  Modeled Thresholds for N+4/N+8 with D = D
MIN
 + 3 dB............................................................. 10-13 
Figure 10-17.  Modeled Thresholds for N+5/N+10 with D = D
MIN
 + 3 dB........................................................... 10-14 
Figure 11-1.  D/U Versus D for Receiver D3 (Desired Signal on Channel 51)..................................................... 11-10 
Figure 11-2.  Threshold U Versus D for Receiver D3 (Desired Signal on Channel 51)........................................ 11-10 
Figure 11-3.  Slope of Log-D versus Log-U for Receiver D3 (Desired Signal on Channel 51)............................ 11-11 
Figure 11-4.  Slope of Log-U versus Log-D for Receiver D3 (Desired Signal on Channel 51)............................ 11-11 
Figure 11-5.  D/U Versus D for Receiver D3—Paired Signals And Their Constituents ....................................... 11-12 
Figure 11-6.  D/U Versus D for Receiver D3—Single Undesired Signals............................................................ 11-12 
Figure 11-7.  Threshold U Versus D for Receiver D3—Single Undesired Signals............................................... 11-13 
Figure 11-8.  Linear Undesired Power Versus Log-Desired Signal Power for N+6.............................................. 11-13 
Figure 12-1.  Slope of Threshold U Versus D from D = -68 dBm to -53 dBm ....................................................... 12-6 
Figure 12-2.  Slope of Threshold U Versus D from D =D
MIN
+3dB to -68 dBm...................................................... 12-6 
Figure 12-3.  Slope of Threshold D Versus U from D =D
MIN
+3dB to -68 dBm...................................................... 12-7 
Figure 12-4.  D/U of 8 Receivers at D = D
MIN
 + 1 dB on Channel 30 (Extrapolation)............................................ 12-7 
Figure 13-1.  D/U of Receiver A3 at Five Desired Signal Levels on Channel 30 ................................................... 13-3 
Figure 13-2.  Threshold U of Receiver A3 at Five Desired Signal Levels on Channel 30 ...................................... 13-3 
Figure 13-3.  D/U of Receiver D3 at Five Desired Signal Levels on Channel 30 ................................................... 13-4 
Figure 13-4.  Threshold U of Receiver D3 at Five Desired Signal Levels on Channel 30 ...................................... 13-4 
Figure 13-5.  D/U of Receiver I1 at Five Desired Signal Levels on Channel 30..................................................... 13-5 
Figure 13-6.  Threshold U of Receiver I1 at Five Desired Signal Levels on Channel 30........................................ 13-5 
Figure 13-7.  D/U of Receiver J1 at Five Desired Signal Levels on Channel 30..................................................... 13-6 
Figure 13-8.  Threshold U of Receiver J1 at Five Desired Signal Levels on Channel 30 ....................................... 13-6 
Figure 13-9.  D/U of Receiver M1 at Five Desired Signal Levels on Channel 30................................................... 13-7 
Figure 13-10.  Threshold U of Receiver M1 at Five Desired Signal Levels on Channel 30 ................................... 13-7 
Figure 13-11.  D/U of Receiver N1 at Five Desired Signal Levels on Channel 30 ................................................. 13-8 
 
vi 
Figure 13-12.  Threshold U of Receiver N1 at Five Desired Signal Levels on Channel 30 .................................... 13-8 
Figure 13-13.  D/U of Receiver O1 at Five Desired Signal Levels on Channel 30 ................................................. 13-9 
Figure 13-14.  Threshold U of Receiver O1 at Five Desired Signal Levels on Channel 30 .................................... 13-9 
Figure 13-15.  D/U of Receiver G4 at Five Desired Signal Levels on Channel 30 ............................................... 13-10 
Figure 13-16.  Threshold U of Receiver G4 at Five Desired Signal Levels on Channel 30 .................................. 13-10 
Figure 14-1.  Frequency Response of Channel-29 Bandpass Filter......................................................................... 14-6 
Figure 14-2.  Alternative Test Configuration With Bandpass Filter on Undesired Signal ...................................... 14-6 
Figure 14-3.  D/U Measurements on Receiver J1 with Primary Versus Alternative Test Setup ............................. 14-7 
Figure 14-4.  Spectrum of Undesired Signal for Broadband Notched Noise Test................................................... 14-7 
Figure 14-5.  Model Plot for “Plateau IM3” On Receiver G4 at D = -68 dBm with N-5/N-10............................... 14-8 
Figure 14-6.  Model Plot for “Plateau IM3” On Receiver D3 at D = -68 dBm with N+2/N+4............................... 14-8 
Figure 14-7.  Model Plot for “Plateau IM3” On Receiver G4 at D = D
MIN
+3dB with N+4/N+8 ............................ 14-9 
Figure 14-8.  Model Plot for “Plateau IM3” On Receiver G4 at D =-68 dBm with N-3/N-6.................................. 14-9 
Figure 15-1.  Median D/U of 8 Receivers at Five Signal Levels on Channel 30.....................................................15-8 
Figure 15-2.  Median Threshold U of 8 Receivers at Five Signal Levels on Channel 30........................................ 15-8 
Figure 15-3.  2
nd
 Worst D/U of 8 Receivers at Five Signal Levels on Channel 30.................................................. 15-9 
Figure 15-4.  2
nd
 Worst Threshold U of 8 Receivers at Five Signal Levels on Channel 30..................................... 15-9 
Figure 15-5.  Worst D/U of 8 Receivers at Five Signal Levels on Channel 30 ..................................................... 15-10 
Figure 15-6.  Worst Threshold U of 8 Receivers at Five Signal Levels on Channel 30 ........................................ 15-10 
Figure 15-7.  Median D/U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs............................................... 15-11 
Figure 15-8.  Median Threshold U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs.................................. 15-11 
Figure 15-9.  2
nd
 Worst D/U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs............................................ 15-12 
Figure 15-10.  2
nd
 Worst Threshold U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs ............................ 15-12 
Figure 15-11.  Worst D/U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs ............................................... 15-13 
Figure 15-12.  Worst Threshold U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs .................................. 15-13 
Figure 15-13.  Median D/U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs.......................................... 15-14 
Figure 15-14.  Median Threshold U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs............................. 15-14 
Figure 15-15.  2
nd
 Worst D/U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs....................................... 15-15 
Figure 15-16.  2
nd
 Worst Threshold U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs.......................... 15-15 
Figure 15-17.  Worst D/U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs ............................................ 15-16 
Figure 15-18.  Worst Threshold U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs ............................... 15-16 
Figure 15-19.  Median D/U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs.......................................... 15-17 
Figure 15-20.  Median Threshold U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs............................. 15-17 
Figure 15-21.  2
nd
 Worst D/U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs....................................... 15-18 
Figure 15-22.  2
nd
 Worst Threshold U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs.......................... 15-18 
Figure 15-23.  Worst D/U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs ............................................ 15-19 
Figure 15-24.  Worst Threshold U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs ............................... 15-19 
  
 
vii 
 
TABLES 
 
Table 2-1.  Measured Bandwidths of Undesired Signal Sources............................................................................... 2-2 
Table 2-2.  Relationship Between Excess Signal and Coverage Area....................................................................... 2-3 
Table 2-3.  DTV Receiver Samples........................................................................................................................... 2-4 
Table 2-4.  UHF Reception Examples....................................................................................................................... 2-7 
Table 2-5.  Summary of Tests.................................................................................................................................... 2-9 
Table 4-1.  Grand Alliance Receiver Interference Rejection Performance................................................................ 4-1 
Table 4-2.  OET-69 DTV-Into-DTV Interference Protection Criteria....................................................................... 4-1 
Table 4-3.  ATSC A/74 Recommended Thresholds for Receiver Interference Rejection ......................................... 4-2 
Table 4-4.  Measurement Limitations of the Test Setup............................................................................................ 4-7 
Table 4-5.  Test Setup Configurations and Settings .................................................................................................. 4-8 
Table 5-1.  Receiver Minimum Signal Level at TOV Versus Desired Signal Source............................................... 5-6 
Table 5-2.  Signal Quality Measurements on 8-VSB Sources................................................................................... 5-6 
Table 7-1.  Comparative Signal Test Summary......................................................................................................... 7-2 
Table 7-2.  D/U Ratio With SFU as Desired Signal Source Relative to That With Baseline Generator 
(ATSC997) ......................................................................................................................................... 7-4 
Table 7-3.  D/U Ratio for Repeat of Baseline Test Relative to Baseline................................................................... 7-5 
Table 7-4.  D/U Ratio With 8-VSB as Undesired Signal Relative to That with Bandlimited Gaussian Noise.......... 7-6 
Table 7-5.  D/U Ratio With DVB-H as Undesired Signal Relative to That with Bandlimited Gaussian Noise........ 7-7 
Table 8-1.  Slopes of Log-Log Plots of D, U, and D/U for Various Interference Mechanisms................................. 8-3 
Table 8-2.  Deviation in Threshold U from Straight-Line Projection as D approaches D
MIN
.................................... 8-5 
Table 9-1.  Parameters for Paired-Interferer Tests (N+K/N+2K).............................................................................. 9-2 
Table 9-2.  Statistics of Paired-Signal D/U’s Relative to Summed D/U’s for 8 Receivers on Channel 30 ............... 9-7 
Table 9-3.  IP3/SNR
R
1/2
 Based on Paired-Signal Rejection Measurements at Channel-30 ..................................... 9-10 
Table 9-4.  IP3/SNR
R
1/2
 Based on Paired-Signal Rejection Measurements at Channel-51 ..................................... 9-11 
Table 9-5.  IP3/SNR
R
1/2
 Statistics............................................................................................................................ 9-12 
Table 9-6.  Computed Threshold U Due to IM3 for Equal Paired Signals.............................................................. 9-13 
Table 9-7.  Computed Threshold D/U Due to IM3 for Equal Paired Signals ..........................................................9-14 
Table 10-1.  Range of Impact of IM3 from Pairs of Undesired Signals When D = -68 dBm.................................. 10-4 
Table 10-2.  Range of Impact of IM3 from Pairs of Undesired Signals When D =D
MIN
 + 3 dB............................. 10-5 
Table 12-1.  Characteristics of Log-Log Plots of Undesired Versus Desired Signal Power At TOV ..................... 12-1 
Table 12-2.  Process for Extrapolation from D = D
MIN
+3dB to D
MIN
+1dB............................................................. 12-4 
Table 12-3.  Error Test for Extrapolation From D = D
MIN
 + 3 dB to D
MIN
 + 1 dB .................................................. 12-5 
Table 14-1.  Broadband Notched Noise D/U Ratios at Threshold........................................................................... 14-3 
 
Table A-1.  Statistics of D
MIN
  for 8 Receivers on Channel 30................................................................................. A-1 
Table A-2.  D/U Statistics for 8 Receivers at D = D
MIN
 + 1 dB on Channel 30........................................................ A-2 
Table A-3.  D/U Statistics for 8 Receivers at D = D
MIN
 + 3 dB on Channel 30........................................................ A-3 
Table A-4.  D/U Statistics for 8 Receivers at D = -68 dBm on Channel 30 ............................................................. A-4 
Table A-5.  D/U Statistics for 8 Receivers at D = -53 dBm on Channel 30 ............................................................. A-5 
Table A-6.  D/U Statistics for 8 Receivers at D = -28 dBm on Channel 30 ............................................................. A-6 
Table A-7.  Threshold U Statistics for 8 Receivers at D = D
MIN
 + 1 dB on Channel 30 .......................................... A-7 
Table A-8.  Threshold U Statistics for 8 Receivers at D = D
MIN
 + 3 dB on Channel 30 .......................................... A-8 
Table A-9.  Threshold U Statistics for 8 Receivers at D = -68 dBm on Channel 30 ................................................ A-9 
Table A-10.  Threshold U Statistics for 8 Receivers at D = -53 dBm on Channel 30 ............................................ A-10 
Table A-11.  Threshold U Statistics for 8 Receivers at D = -28 dBm on Channel 30 ............................................ A-11 
Table B-1.  Slopes of Log-Log Plots of D, U, and D/U for Various Interference Mechanisms ................................B-4 
Table B-2.  Deviation in Threshold U from Straight-Line Projection as D Approaches D
MIN
..................................B-5 
Table B-3.  Slopes of Log-Log Plots of D, U, and D/U for Various Interference Mechanisms ..............................B-13 
Table B-4.  Deviation in Threshold U from Straight-Line Projection as D approaches D
MIN
.................................B-14 
 
 
 
 
viii 
EXECUTIVE SUMMARY 
This report presents the results of two categories of tests performed on consumer digital television 
receivers to assess their vulnerability to out-of-channel interference: 
• Tests of tuner type; and 
• Interference rejection threshold tests. 
 
The first category of test identifies the type of tuner (e.g., single-conversion versus double-conversion) in 
consumer DTV receivers.  A knowledge of tuner type aids in identifying the channel offsets at which a 
TV receiver is likely to be susceptible to interference.  The second type of test measures the ability of 
DTV receivers to reject interference from signals in TV channels other than the one to which the TV is 
tuned. 
 
The test results are intended to support assessments of interference to DTV reception from: 
• non-TV use of locally-unused spectrum within the TV broadcast spectrum, which is also often termed 
“white-space” use; 
• non-TV use of spectrum adjacent to or near TV broadcast spectrum (e.g., the TV channel 52 to 67 
spectrum that will be auctioned for other uses); and, 
• other DTV stations. 
 
The results also enable evaluation of the degree to which consumer DTV receivers comply with the 
voluntary standards contained in the receiver standards document, “ATSC Recommended Practice:  
Receiver Performance Guidelines”
*
 
 
It should be noted that this study characterizes susceptibility of consumer DTVs to interference in terms 
of signal and interference levels appearing at the antenna input terminal of the receiver.  Assessments of 
potential for interference to DTV reception will require additional analysis involving specific protection 
scenarios,
†
 propagation modeling, and antenna modeling, which are beyond the scope of the study. 
TUNER TYPE TESTS 
Measurements on the “Grand Alliance” prototype DTV receiver provided the basis for establishing 
interference protection criteria for the DTV channel allotment process.  That receiver used a double-
conversion tuner.  Consumer DTV receivers often use single-conversion tuners, which can be more 
susceptible to interference at certain channel spacings than a double-conversion tuner.  Most of the analog 
TV taboos that limited allotments of channels at certain spacings in a given local area were the result of 
interference vulnerabilities associated with single-conversion tuners. 
 
Tests were performed on thirty consumer DTV receivers to determine their tuner types.  All were found to 
have single-conversion tuners.  While the use of single-conversion tuners implies the possibility of 
interference susceptibilities at the same frequency offsets as those experienced by analog TV, it should be 
noted that such interference vulnerabilities are lower for digital TV than for analog TV because the ATSC 
DTV system is inherently more resistant to interference than the NTSC analog system. 
                                                      
*
 Advanced Television Systems Committee, “ATSC Recommended Practice:  Receiver Performance Guidelines”, 
<ATSC Receiver Guidelines>, ATSC Doc. A/74, 17 June 2004. 
†
 In this context, a protection scenario refers to a set of conditions that may be part of the analytic process of setting 
protection limits.  The conditions might include such factors as the distance from the interfering device to the 
antenna of the DTV receiver and the attenuation of any intervening walls. 
 
ix 
INTERFERENCE REJECTION TESTS 
The out-of-channel interference rejection threshold of consumer digital television (DTV) receivers was 
measured by supplying an ATSC 8-VSB
*
 signal (the desired signal), along with one or two undesired 
signals, to the antenna terminal of a DTV receiver and then adjusting the level of the undesired signal(s) 
to the point at which picture degradation begins to be observed—a point known as the threshold of 
visibility (TOV) of degradation.  In such testing, the DTV receiver is tuned to the channel number of the 
desired signal.  The undesired signal (the potential interferer) is placed on another channel, either above 
or below the desired channel. 
 
We refer to the desired signal power at the input to the DTV receiver as D and the undesired power when 
adjusted to the threshold (TOV) as U.  It is traditional to express interference rejection performance in 
terms of the ratio D/U.  When D/U is expressed in decibels (dB), it is typically a negative number for out-
of-channel interference; that is, the undesired signal level is greater than the desired signal at TOV.  Low 
values (i.e., large negative numbers) represent good rejection performance; high values (small negative 
numbers) represent high susceptibility to interference.  D/U ratios measured for this report ranged from 
below -74 dB to -20 dB.  Results in this report are presented both as D/U ratios and as threshold values of 
the undesired signal power U in order to meet the needs or preferences of different analysts. 
 
Most of the interference rejection tests were performed on eight consumer DTV receivers that were on the 
market in 2005 and 2006.  The eight receivers were identified as having fifth-generation multipath-
handling capability.  Limited tests were also performed on two earlier-generation receivers. 
 
It should be noted that the out-of-channel interference tests for this report were performed using a test 
setup that suppressed leakage of the undesired signal into the desired channel to a sufficient degree to 
make spectral leakage effects negligible.  Thus, the first-adjacent channel tests do not include the effects 
of spectral splatter representative of a DTV transmitter—an effect which drives the first-adjacent channel 
protection ratios used for DTV channel allotments. 
INTERFERENCE FROM ONE UNDESIRED SIGNAL 
Channel Selection 
All tests for this report were performed in the UHF TV band.  For the bulk of the testing, the TV was 
tuned to a desired signal on channel 30 (designated as channel “N”), and the undesired signal was 
sequenced through channels N-1 through N-16 and N+1 through N+16.  The earliest testing in the 
program was performed with the desired signal on channel 51 and the undesired signal placed on channels 
N+1 through N+16. 
 
The measurements on channel 51 generally matched those on channel 30 within about 4 dB—and in most 
instances much closer than that. 
 
Undesired Signals 
All tests were performed using undesired signals that were continuous, as opposed to occurring in bursts 
(as might be typical for packet-based communications). 
                                                      
*
 8-level Vestigial Side Band (8-VSB) is the over-the-air digital television (DTV) transmission format specified by 
the Advanced Television Systems Committee’s (ATSC) Digital Television Standard (A/53) and adopted by the FCC 
as the U.S. standard for terrestrial DTV broadcasting.  
 
x 
Signal Type 
For most tests the undesired signal was either an 8-VSB DTV signal or a white Gaussian noise signal that 
had been bandlimited to match the 3-dB width of the 8-VSB signal.  The selection between these two 
signal types was driven by equipment availability and spectrum shape requirements.
*
 
 
A smaller number of tests were performed to compare the interference effects of three signal types: 
• The 8-VSB DTV signal; 
• The Gaussian noise signal bandlimited to match the 8-VSB signal width (at the 3-dB points); 
• An OFDM signal (DVB-H) that was about 5-MHz wide. 
 
The results show that the DTV receivers are more vulnerable to out-of-channel interference from the 
Gaussian noise and OFDM signals than from the 8-VSB DTV signal by an amount just over 1 dB. 
Signal Bandwidth 
For most tests, the undesired signal filled most of the 6-MHz width of a TV channel.  Tests were also 
performed on one TV using a narrower signal—Gaussian noise bandlimited to a 3-dB bandwidth of 
1 MHz.  D/U ratios with this reduced-bandwidth signal were comparable to those for the wider undesired 
signals except in channel offsets where significant narrowband interference effects occurred.  One 
prominent such effect occurred with the interference on channel N+7, where an interference susceptibility 
peak was observable on 7 of the 8 DTV receivers that were tested with the broader signals.  Tests with the 
narrow signal showed that the N+7 interference effect is narrowly centered at the local oscillator 
frequency of the receiver (44 MHz above the center of the tuned channel). 
Signal Levels 
If D/U ratios were constant with changes in desired signal level, measurements at a single desired signal 
level would be sufficient to characterize the interference susceptibility of a TV receiver at a given channel 
offset; however, the tests showed that threshold D/U ratios vary with signal level, and do so in ways that 
vary among the different channel offsets on different TVs.  The nature of the variation also changes with 
signal level.  As a result, tests were performed at three desired signal levels specified by the ATSC (-28,   
-53, and -68 dBm) as well as one additional signal level, D
MIN
 + 3 dB, where D
MIN
 refers to the receiver-
specific desired signal level at TOV in the absence of interference (essentially, the minimum signal level 
at which a given DTV receiver can operate).  In addition, modeling was performed to extend the results 
down to D
MIN
 + 1 dB. 
 
Cliff Effect 
The digital television broadcast system can achieve flawless picture reception under interference 
conditions (D/U ratios) that would produce an unusable picture for analog broadcast TV; however, once 
an undesired signal reaches a level at which picture impairments become visible on a DTV receiver, the 
picture degrades extremely rapidly with further increases in undesired signal level.  This characteristic of 
DTV receiver performance is known as the “cliff effect”.  The interference rejection thresholds of the 
tested receivers exhibited a strong “cliff effect”.  In most cases, increasing interference level about 1 dB 
above the TOV level, at which picture impairments are first observed, caused complete loss of the TV 
picture. 
 
                                                      
*
 An 8-VSB signal was used for the first tests performed under this program (those on channel 51) because it nearly 
filled the 6-MHz width of a TV channel and it exhibited steep rolloff at the band edges—an essential characteristic 
for adjacent-channel interference tests.  After failure of that 8-VSB source, tests (on channel 30) were performed 
using the bandlimited Gaussian noise source.  Because the band-edge rolloff of that signal was not sufficiently steep 
to permit interference measurements on first-adjacent channels, those tests were postponed until procurement of a 
new 8-VSB signal source (after repair of the previous source was found to be impossible).  All tests involving two 
interferers made use of bandlimited Gaussian noise as the second source. 
 
xi 
Rejection Performance 
No receiver appeared to fully achieve the ATSC recommended guidelines for interference rejection 
performance—guidelines that are less stringent than the receiver performance assumptions on which 
current DTV interference protection criteria are based.
*
  After taking into account differences between the 
Gaussian-noise interferer used for most of the tests and the 8-VSB interferer specified by the ATSC, the 
best-performing receiver appears to fail the guidelines at only one channel offset, and there by only 1 dB.  
A second receiver failed to meet the voluntary guidelines by 1 to 2 dB at two channel offsets.  The 
remaining five receivers failed to meet the guidelines at two to 16 channel offsets; the worst failure for 
each of those receivers ranged from about 8 to 24 dB. 
 
The rejection performance of the receivers is summarized by plots in the final chapter of this report.  The 
following observations apply to the results. 
• In terms of absolute signal levels that can cause interference, the DTV receivers are at their most 
vulnerable when operating at low desired signal levels. 
• At low desired signal levels the DTV receivers are as susceptible to interference from the second-
adjacent channels (N-2 and N+2) as from first-adjacent channels (N-1 and N+1).  In terms of worst 
and second-worst performance, the receivers are actually more susceptible to interference from 
second-adjacent channels than from first-adjacent channels.  (This contradicts the assumptions of 
OET-69
†
 and the ATSC Receiver Guidelines.) 
• The receivers tend to be more susceptible to interference from N+2, N+1, N-1, N-2, N-3, N-4, and 
sometimes N-6 than from the mixer image channel offsets of N+14 and N+15. 
• At moderate desired signal levels, the receivers exhibit relatively high susceptibility to interference 
from channel N+7.  This interference threshold is nearly constant in terms of absolute power of the 
undesired signal necessary to cause interference at different levels of desired signals.  At lower 
desired signal levels, other channel offsets become more vulnerable. 
 
INTERFERENCE FROM IM3-GENERATING PAIRS OF UNDESIRED 
SIGNALS 
Pairs of undesired signals placed on channels N+K and N+2K, where K is a positive or negative integer, 
create an opportunity for third-order intermodulation (IM3) occurring in the DTV tuner to create 
spectral products that fall in the desired channel N.  These spectral products can interfere with TV 
reception.  Though the D/U ratios associated with paired-signal IM3 effects tend to decrease as the 
desired signal decreases, the test results show that IM3 can constitute a dominant interference mechanism 
even at desired signal levels very near the minimum signal threshold for the TV (D
MIN
).  
 
Measurements of interference thresholds were performed with equal-powered undesired signals on 
N+K/N+2K combinations for eight TVs on channel 30 and seven TVs on channel 51.  Tests were 
performed for K = -5 to 5 when N was 30 and for K = 1 to 8 when N was 51.  In both sets of tests, the 
desired signal levels were -68 dBm and -53 dBm.  Not all measured cases produced interference effects 
that were sufficiently higher than the single-channel effects to allow isolation of the IM3 effects.  For 
those that did, a third-order intercept point (IP3) was computed.   
 
The IP3 data was used to predict paired-signal IM3 interference effects at lower desired signal levels than 
the measurements and at unequal undesired signal levels—predictions that were tested by measuring two 
receivers using unequal undesired signals.  The results show that when such signal pairs occur, they 
                                                      
*
 The ATSC provides recommended guidelines for DTV receiver performance in its document, “ATSC 
Recommended Practice:  Receiver Performance Guidelines”, ATSC Doc. A/74, 17 June 2004. 
†
 “Longley-Rice Methodology for Evaluating TV Coverage and Interference”, Office of Engineering and 
Technology (OET) Bulletin No. 69, <OET-69>, Federal Communications Commission, 6 February 2004. 
 
xii 
represent the dominant interference mechanism for channel offsets from about N+4 to N+16 (with the 
exception of the mixer image at N+14 and N+15) and from about N-5 to N-10.  For spacings close to 
channel N (e.g., N+/-1 and N+/-2), the paired-signal effects appear less likely to dominate—at least if the 
pair has equal power levels.  No paired signal measurements were performed beyond N+16 and N-10, so 
it is not know how far out the effect continues; however, the effect is seen to diminish with increasing 
channel offset from the desired channel. 
 
Paired signals at IM3-generating spacings have the potential to create even greater interference 
susceptibilities if an existing undesired signal on one of the IM3-generating channels (e.g., a nearby DTV 
broadcast station when the receiver is tuned to a more distant station) exceeds the measured equal-power-
level threshold for paired signals.  In such a case, the presence of that signal can greatly increase 
susceptibility to interference on the other channel of the IM3-generating pair.  This situation generally 
creates the greatest vulnerabilities when the stronger undesired signal is on channel N+K and it exceeds 
the equal-power paired-signal threshold; in that case, the receiver susceptibility to interference on the 
N+2K channel increases by twice the N+K signal excess above the equal-power threshold (in dB). 
 
The ATSC Receiver Guidelines document (A/74) provides no recommended performance levels for 
rejection of paired-signal interference. 
 
1- 1
CHAPTER 1 
INTRODUCTION 
This report presents the results of laboratory tests of consumer digital television (DTV) receivers to 
determine: 
• Tuner type (e.g., single conversion); and 
• UHF out-of-channel interference rejection thresholds at the antenna input terminal of the receiver. 
 
The term, “out-of-channel” interference, as used in this report, includes any interference occurring outside 
of the 6-MHz width of the TV channel to which the receiver is tuned.  Thus, it includes:  first-adjacent 
channels; “taboo” channels associated with analog TV (i.e., TV channels more than one channel away 
from the desired channel at specific channel spacings associated with interference vulnerabilities of 
analog TV); other TV channels; and interference from sources outside of spectrum allocated to broadcast 
TV. 
 
The tuner type tests were performed on 30 consumer DTV receivers—29 of which were on the market in 
mid-to-late 2005 and the other of which was procured in 2006.  28 of the receivers had been subjected to 
other terrestrial reception performance testing during an earlier program conducted as part of the 
Commission’s compliance with the Satellite Home Viewer Extension and Reauthorization Act of 2004 
(SHVERA).
*
 
 
Extensive interference rejection tests were performed on eight of the receivers.  Limited interference 
rejection testing was performed on two other receivers as part of the tuner type tests. 
BACKGROUND 
The ability of television receivers to operate in the presence of interference is an important factor in 
determining whether consumers can receive TV broadcast service.  It is generally recognized that the 
ATSC DTV system (also known as 8-VSB)
†
 used for terrestrial broadcast digital TV in the United States 
is less susceptible to interference than the NTSC analog TV system that it is replacing.  This increased 
robustness has enabled a relaxation of the “taboo” rules that govern channel spacings of broadcast 
television stations in a local area—allowing part of the spectrum formerly assigned for broadcast TV use 
(UHF channels 52 through 69) to be made available for other uses after the completion of the DTV 
transition. 
 
The Federal Communications Commission’s (FCC) rules that are intended to prevent interference to 
broadcast DTV reception
‡
 are based on the interference rejection performance of the so-called “Grand 
Alliance” DTV prototype receiver.  These rules define the interference protection afforded to broadcast 
TV stations, and receiver manufacturers are expected to consider the signal protection ratios in the rules 
when designing their TV receiver products.  While the prototype DTV receiver used a double-conversion 
tuner configuration, some authors have surmised that many modern consumer DTVs employ single-
conversion tuners that may be more susceptible to out-of-channel interference at some frequencies than a 
                                                      
*
 Stephen R. Martin, “Tests of ATSC 8-VSB Reception Performance of Consumer Digital Television Receivers 
Available in 2005”, Report FCC/OET TR 05-1017, <SHVERA Study>, November 2, 2005.  
(www.fcc.gov/oet/info/documents/reports/TR-05-1017-ATSC-reception-testing.pdf) 
†
 8-level Vestigial Side Band (8-VSB) is the over-the-air digital television (DTV) transmission format specified by 
the Advanced Television Systems Committee’s (ATSC) Digital Television Standard (A/53) and adopted by the FCC 
as the U.S. standard for terrestrial DTV broadcasting. 
‡
 See 47 C.F.R. section 73.623(d) 
 
1- 2
double-conversion tuner.
*†
  In addition, other design differences between the “Grand Alliance” prototype 
DTV receiver and today’s consumer receivers could potentially affect the degree of a receiver’s ability to 
reject out-of-channel interference.   
 
Interference can be broadly divided into two categories: 
• Co-channel interference; and, 
• Out-of-channel interference. 
 
Co-channel interference results from undesired emissions at frequencies within the channel that a 
television receiver is currently attempting to receive.  Out-of-channel interference results from emissions 
occurring in a frequency range outside that of the TV channel to which the receiver is tuned.  Out-of-
channel emissions may come from another TV broadcast or a non-TV source. 
Co-Channel Interference Rejection 
Tests conducted by the FCC Laboratory in 2005 on 28 consumer DTV receivers demonstrated that the 
receivers differ very little in their immunity to broadband co-channel interference.  In those tests, 
receivers were found to produce pictures that were free of visual errors when the TV signal power 
exceeded broadband interference power within the same TV channel by a threshold ranging from 14.9 to 
15.8 dB, with the median threshold being 15.3 dB.
‡
  These results closely match the 15.2 dB threshold of 
the Grand Alliance receiver.
§
  
 
The small variation of co-channel interference rejection performance among the receivers and the close 
match to the older Grand Alliance results are in line with the expectation that co-channel interference 
rejection threshold is determined primarily by the structure of the ATSC DTV signal format adopted by 
the FCC. 
Out-of-Channel Interference Rejection 
Unlike co-channel interference rejection, out-of-channel interference rejection of a DTV receiver is 
expected to depend heavily on receiver characteristics, such as tuner selectivity, tuner image performance, 
automatic gain control (AGC) implementation, and tuner overload characteristics.   
 
Because of this dependence on receiver characteristics and because modern consumer DTV receivers may 
differ substantially in design from the Grand Alliance receiver, new measurements are required to 
characterize out-of-channel interference rejection of today’s DTV receivers. 
 
OBJECTIVES 
The DTV receiver measurements described in this report are intended to support assessments of 
interference to DTV reception from: 
• non-TV use of locally-unused spectrum within the TV broadcast spectrum, which is also often termed 
“white-space” use; 
• non-TV use of spectrum adjacent to or near TV broadcast spectrum (e.g., the TV channel 52 to 67 
spectrum that will be auctioned for other uses); and, 
• other DTV stations. 
                                                      
*
 Charles W. Rhodes, “The Challenge of Channel Election, TV Technology.com, January 19, 2005.  
(http://www.tvtechnology.com/features/digital_tv/Features_Rhodes-01.19.05.shtml) 
†
 Oded Bendov and others, “Planning Factors for Fixed and Portable DTTV Reception”, IEEE Transactions On 
Broadcasting, Vol. 50, No. 3, September 2004. 
‡
 Martin, <SHVERA Study>, 2005, chapter 3. 
§
 Federal Communications Commission Advisory Committee on Advanced Television Service, “Final Technical 
Report”, Oct 31, 1995, p.19. 
 
1- 3
 
The results also enable evaluation of the degree to which consumer DTV receivers comply with the 
voluntary standards contained in the receiver standards document, “ATSC Recommended Practice:  
Receiver Performance Guidelines”
*
 
 
A device that transmits radio energy on an otherwise unused TV channel (or in spectrum adjacent to a TV 
band) can interfere with DTV reception in one of two ways, or in a combination of the two: 
(1) The radio energy it creates in its intended band of operation may impact the DTV receiver through 
various receiver vulnerabilities caused by nonlinearities in the receiver, mixer images, or other 
mechanisms; 
(2) The device may unintentionally “splatter” enough energy into the channel to which the DTV receiver 
is tuned to interfere with TV operation. 
The first of these ways is true out-of-channel interference; the second is a form of co-channel 
interference.  Given that either of these factors—or a combination thereof—could determine interference 
rejection performance, one way to test DTV receivers would be to subject them to simulated interferers 
that contain both an intended emission on a channel outside of the desired channel and spectral splatter 
into the desired channel.  The resulting interference performance measurements would reflect the 
combination of the two interference effects.  However, given that there is debate within the technical 
community regarding both the intended emissions and the out-of-band (e.g., spurious) emissions for 
devices operating in the “white spaces”, we believe it is appropriate to deal with these two types of 
interference separately.  As was stated in the previous section, co-channel interference is relatively well 
understood and relatively constant among ATSC DTV receivers compared to out-of-channel interference; 
consequently, we chose to measure only true out-of-channel interference in this study.  As such, we 
applied filtering to the outputs of out-of-channel signal sources used in the tests in order to reduce their 
spectral splatter into the desired channel to negligible levels. 
 
It should be noted that this study characterizes susceptibility of consumer DTVs to interference in terms 
of signal and interference levels appearing at the antenna input terminal of the receiver.  Assessments of 
potential for interference to DTV reception will require additional analysis involving specific protection 
scenarios,
†
 propagation modeling, and antenna modeling, which are beyond the scope of the study. 
 
MEASUREMENT PROGRAM 
The tests were performed on subsets of the 28 DTV receivers that had been tested in 2005 in support of 
the Satellite Home Viewer Extension and Reauthorization Act of 2004 (SHVERA).
‡
  An additional 
receiver that was received too late for inclusion in that study, as well as one that was purchased in 2006, 
were also tested, bringing the total to 30 receivers.  Five of the 30 receivers were set top DTV tuners.  The 
remaining 25 were DTVs with integrated digital tuners. 
 
The measurement program described here consisted broadly of two parts:  (1) tuner type characterization; 
(2) interference rejection threshold tests. 
 
Tuner type characterization was intended to identify tuner topology used in consumer DTV receivers.  
The first phase of these tests involved searching for a local oscillator (LO) signal at the antenna inputs of 
each of the 30 DTV receivers while switching between two channels.  The tests demonstrated that at least 
                                                      
*
 Advanced Television Systems Committee, “ATSC Recommended Practice:  Receiver Performance Guidelines”, 
<ATSC Receiver Guidelines>, ATSC Doc. A/74, 17 June 2004. 
†
 In this context, a protection scenario refers to a set of conditions that may be part of the analytic process of setting 
protection limits.  The conditions might include such factors as the distance from the interfering device to the 
antenna of the DTV receiver and the attenuation of any intervening walls. 
‡
 Martin, <SHVERA Study>, 2005. 
 
1- 4
28 of the 30 receivers use single-conversion tuners with an intermediate frequency (IF) of 44 MHz.  For 
the two receivers that produced inconclusive results in the LO tests, interference rejection tests were used 
to identify any interference susceptibilities that would be a “signature” of single-conversion receivers 
with 44-MHz IF.  Those two receivers were also determined to use single-conversion tuners. 
 
Interference rejection threshold testing consists of measuring the maximum signal level at which an 
undesired (i.e., interfering) signal can be injected into the TV receiver’s antenna input without adversely 
affecting TV reception of a “desired” signal.  The behavior of DTV signal reception in this regard differs 
considerably from that of analog TV reception.  The DTV broadcast system can achieve flawless picture 
reception under interference conditions that would produce an unusable picture for analog broadcast TV; 
however, once an undesired signal reaches a level at which picture impairments become visible on a DTV 
receiver, the picture degrades extremely rapidly with further increases in undesired signal level—typically 
going from barely perceptible picture impairments to complete loss of picture with a span of about 1 dB.  
Similar degradation of analog reception occurs over a span as large as 30 dB, a difference that emphasizes 
the importance of these measurements for DTV. 
 
Because of the number of variables involved in interference rejection testing, measurements were 
performed on only ten DTV receivers.  Seven of these included one of each brand selected from among 
receivers exhibiting “fifth-generation” DTV demodulator performance in the SHVERA Study.  An eighth 
TV was procured in 2006.  The remaining two, on which only limited interference rejection testing was 
performed, were the two receivers for which the LO tests were inconclusive.  
 
Initial focus of the work reported herein was on the immunity of the receivers to interference from future 
radio services that will operate in the channel 52 to 67 spectrum after completion of the DTV transition.  
Consequently, the initial tests were performed with the receivers tuned to channel 51 and interference 
placed on channels 52 through 67.  Subsequent measurements, focused on the white-space use of the TV 
spectrum, were performed with the receivers tuned to channel 30—near the center of the UHF core TV 
spectrum—with interferers placed both above and below channel 30. 
 
Interfering sources used for most of the testing included an ATSC 8-VSB DTV signal and a pseudo-
random Gaussian noise signal bandlimited to the same spectral width as a DTV signal (5.38 MHz, 3-dB 
width).  Except for the small pilot tone of the DTV signal, both signals exhibit noise-like signal 
characteristics and relatively flat spectra similar to those of most modern communication systems using 
digital modulation.  A few tests were performed using an OFDM signal and using a narrower width 
Gaussian noise signal, for comparison. 
 
In accordance with standard industry practice, the interference rejection threshold of the receiver is 
expressed as the ratio of desired signal power to undesired signal power (D/U ratio) at the threshold of 
visibility (TOV) of degradation of the TV picture.  For a few interference mechanisms, the threshold D/U 
ratio remains constant as the desired signal level is varied except at low signal levels, but many other 
mechanisms are non-linear, resulting in variable D/U ratios that are further affected by the receiver’s 
automatic gain control (AGC) in a way that is not predictable without a detailed knowledge of the design 
of the each receiver; consequently, measurements were performed at multiple desired signal power levels. 
 
Interference rejection thresholds are also a function of the frequency spacing between the desired TV 
signal and the undesired interfering signal.  Single-conversion TV receivers are known to exhibit 
interference susceptibilities up to 90 MHz (15 TV channel widths) from the desired TV channel;
*
 
consequently, measurements with single interferers were performed at various frequency spacings within, 
and just beyond, this range. 
 
                                                      
*
 The mixer image for a single-conversion TV tuner is centered 88 MHz above the center frequency of the desired 
channel.  It occupies portions of channels N+14 and N+15, where N is the desired channel. 
 
1- 5
In addition, pairs of interfering signals at particular spacings can be expected to create interference 
through non-linear effects in a TV tuner; consequently, some tests were performed using pairs of 
interferers spaced at intervals that could cause third-order intermodulation distortion to fall in the desired 
channel. 
 
OVERVIEW 
Chapter 2 of this report describes the scope of the test program and the general testing approach. 
 
Chapter 3 describes the tests performed to determine DTV tuner type (single conversion versus double 
conversion or other types) and the results of those tests.   
 
All remaining chapters deal with interference rejection performance of the DTV receivers, as described 
below. 
 
Chapter 4 describes the test methodology employed for interference testing, the performance of the test 
setup, and some reference values for rejection performance from other documents. 
 
Chapter 5 presents the results of rejection performance measurements on eight DTV receivers tuned to 
channel 30 (N=30) for a single undesired signal placed on each of the channels from N-1 to N-16 and 
N+1 to N+16   The undesired signal was an 8-VSB signal for channels N-1 and N+1 and a white 
Gaussian noise signal bandlimited to match the 3-dB width of an 8-VSB signal for all other channels.  
The desired signal power levels for these tests were -28 dBm, -53 dBm, -68 dBm, and D
MIN
 +3 dB, where 
D
MIN
 refers to the individually measured threshold for each receiver when operated without interference. 
 
Chapter 6 presents results of rejection performance measurements on seven DTV receivers (a subset of 
the eight) tuned to channel 51 (N=51) for a single undesired 8-VSB signal placed on each of the channels 
from N+1 to N+16.  The desired signal power levels for these tests were -28 dBm, -53 dBm, and 
-68 dBm. 
 
Chapter 7 presents the results of measurements of the differences in interference effects of four undesired 
signal types: 
• White Gaussian noise bandlimited to the 3-dB width of an 8-VSB source; 
• 8-VSB; 
• DVB-H—an orthogonal frequency division multiplexing (OFDM) signal—set for a 5-MHz channel 
width; and 
• White Gaussian noise bandlimited to a 3-dB width of 1 MHz. 
The chapter also presents the effects of signal quality differences of two different 8-VSB signal 
generators that served as the desired source. 
 
Chapter 8 presents a theoretical framework that aids in interpreting and extending the out-of-channel 
interference measurements. 
 
Chapter 9 presents the results of interference rejection measurements performed using equal-power pairs 
of undesired signals spaced so as to place third-order intermodulation (IM3) products in the desired 
channel.  The tests were performed on eight DTV receivers tuned to channel 30 and seven DTVs tuned to 
channel 51.  Undesired signals were placed at channel pairs N+K/N+2K for K = -1 through -5 and +1 
through +5 for channel 30 and K = 1 through 8 for channel 51.  The undesired sources were the same as 
those described for Chapters 5 and 6.  Values of a third-order intercept parameter (IP3) were estimated 
based on some of the measurements. 
 
 
1- 6
Chapter 10 presents interference rejection measurements on two receivers (one N+K/N+2K channel offset 
pair for each) using a pair of undesired signals of unequal amplitudes.  It also presents a model that is 
then used to make predictions regarding interference rejection performance of the other tested DTV 
receivers for unequal, paired undesired signals.  The model makes use of measurements from Chapters 5 
and 9.   
 
Chapter 11 presents measurements of rejection performance for one DTV receiver over a wide, finely-
stepped desired-signal amplitude range.  The tests included single interferers, as well as paired interferers 
spaced to place IM3 products in the desired signal channel.  The measurements provide insight into the 
behavior of interference susceptibilities including their nonlinearities and the effects of receiver AGC. 
 
Chapter 12 extrapolates the channel-30, single-channel interference rejection measurements from 
Chapter 5 of this report to a lower desired signal level, D
MIN
 + 1 dB.  It also employs measured data from 
Chapter 11 to evaluate the extrapolation method. 
 
Chapter 13 combines the single-channel rejection performance measurements of the eight DTV receivers 
on channel 30 from Chapter 5 (at D = -28 dBm, -53 dBm, -68 dBm, and D
MIN
 + 3 dB) with the 
extrapolations of Chapter 12 (to D = D
MIN
 + 1 dB) to present combined results. 
 
Chapter 14 presents the results of tests and analyses performed to validate the test methodology and test 
setup used for the measurements in this report.  (Direct measurements of test setup performance are 
shown in Chapter 4.) 
 
Chapter 15 presents the summary and conclusions.  It also includes new graphs that integrate the results 
of some of the other chapters in terms of median, 2
nd
 worst, and worst performance among the eight 
receivers. 
 
2-1 
CHAPTER 2 
SCOPE AND APPROACH 
SCOPE OF TESTING 
Two types of tests were performed on DTV receivers in the measurement program reported herein:  
(1) frequency measurements of tuner local oscillator (LO) signals leaking out of the TV receiver antenna 
terminal; (2) signal levels of desired and undesired signals at the TV receiver antenna port with the 
undesired signal set to the threshold of visibility (TOV) of degradation of the received TV picture. 
 
Measurements of the LO, where possible, serve to characterize the type of receiver (single or double 
conversion) and intermediate frequency (IF) for single conversion receivers.   
 
The signal level measurements at TOV quantify the immunity of the receiver to out-of-channel 
interference and also can be used to aid in characterizing receiver type when LO leakage is not 
measurable.  These results are expressed as interference rejection ratio threshold—the ratio of desired 
signal power to undesired (interfering) signal power (D/U ratio) at TOV. 
 
Interfering Signal Types 
The objectives of the testing were to generate data that could be used to assess potential interference to 
DTV by future devices operating in TV white spaces, future services operating in channels 52 through 67 
after the completion of the DTV transition, and other DTV signals.  Ideally, measurements would have 
been performed using each potential type of interfering signal; however, white-space devices and signals 
to be transmitted in channels 52 through 67 are not fully defined, and, even if representative devices had 
been available, the number of tests would have been prohibitive. 
 
Most modern digital communications systems produce signals that are somewhat noise-like.  The spectra 
are often flat over most of their channel width; they have relatively high peak-to-average ratios and peak 
levels that are random, rather than fixed—as in the case of most analog transmissions.  A growing number 
of digital systems are using orthogonal-frequency-division-multiplexing (OFDM) signals.  Both the 8-
VSB DTV signal used for broadcast DTV in the United States and OFDM signals share these noise-like, 
flat-spectrum characteristics (except for the 8-VSB pilot). 
 
We initially expected that bandlimited white noise, 8-VSB DTV signals, and OFDM signals of equal 
bandwidths would be likely to exhibit similar interference effects on DTV reception.  (Later, testing 
described in Chapter 7 caused a small adjustment to this expectation.)  Initial single-interferer testing was 
performed using an 8-VSB DTV signal as the interferer and TV receivers tuned to channel 51.  For tests 
with pairs of interferers to create intermodulation distortion within the TV receivers, a Gaussian noise 
generator—bandlimited to the same 3-dB width as an 8-VSB signal—was used as the second interferer 
because an additional 8-VSB source was not available.  The measurements on channel 51 were terminated 
after failure of the 8-VSB source that had been used as an interferer. 
 
All further tests were performed with the TV receivers tuned to channel 30 because of the change in focus 
of the project from interference effects of services in channel 52 through 69 to interference effects of 
white-space devices, which would be limited to VHF and core UHF channels.  Channel 30 was selected 
for testing because it is a locally-unused channel near the center of the core UHF band. 
 
 
2-2 
The testing with TV receivers tuned to channel 30 was initiated using a Gaussian noise generator—
bandlimited to the 3-dB width of an 8-VSB signal—as the interferer, because the only available 8-VSB 
source was being used to generate the desired signal.  For intermodulation tests, two such signal sources 
were used as interferers.  The bandlimited Gaussian noise sources did not exhibit a band-edge rolloff 
steep enough to support tests of interference on first-adjacent-channels (i.e., N-1 or N+1 when the TV is 
tuned to channel N); consequently, first adjacent channel tests were postponed until a new 8-VSB DTV 
signal source was procured.  That source, which arrived late in the test program, was then used for all 
tests on first-adjacent channels for N=30. 
 
A limited number of tests were performed to compare the interference results of four interferer types: 
(1) 8-VSB DTV signal; 
(2) Gaussian noise bandlimited to match the 3-dB width of an 8-VSB signal; 
(3) an OFDM DVB-H signal; and 
(4) Gaussian noise bandlimited to a 3-dB width of 1 MHz. 
The OFDM DVB-H signal was generated using a commercial software package for a vector signal 
generator with parameters set for a 2k OFDM signal with 5-MHz channel width and 64 QAM 
modulation. 
 
Figures 2-1 and 2-2 show measured spectra of the four interferers at equal total power levels.  Table 2-1 
shows bandwidths of each signal. 
 
 
Table 2-1.  Measured Bandwidths of Undesired Signal Sources 
Signal 3-dB Bandwidth (MHz) 20-dB Bandwidth (MHz) 
8-VSB (Bandwidths, neglecting pilot) 5.38 5.90 
Gaussian noise (8-VSB width) 5.38 6.32 
Gaussian noise (1-MHz width) 1.00 1.18 
OFDM DVB-H 4.76 4.80 
 
 
Desired Signal Levels 
The ATSC specifies guidelines for interference rejection performance of DTV receivers at three desired 
signal levels:  -68 dBm, -53 dBm, and -28 dBm, which they designate as “weak”, “moderate”, and 
“strong”, respectively.  Our initial intent was to test only at these three signal levels; however, after tests 
demonstrated that D/U rejection ratios are by no means constant as a function of desired signal power, a 
decision was made to extend the measurements to a lower level that was receiver dependent.   
 
The motivation for the tests at lower desired signal levels, near D
MIN
, can be seen in Figure 2-3 and Table 
2-2, both of which show the relationship of desired signal levels to broadcast coverage area with flat 
terrain.
*
  The computations are based on the Egli propagation model,
†
 which has propagation loss 
proportional to the fourth power of distance.  If we view the outer boundary of a broadcast station’s 
coverage area to be determined by the point at which consistent reception is just barely possible with a 
given receiver and antenna system (i.e., no signal margin remains after normal amplitude variations of the 
signal due to fading), then the signal margin will be less than 1 dB in the outer 11 percent of the coverage 
area.  Similarly, the outer 29 percent of the coverage area has less than 3 dB of margin, and fully 84 
                                                      
*
 The actual pattern of TV desired signals can differ significantly from the flat terrain model due to the variability in 
terrain and other geographic features that are present in a local area; consequently, terrain-dependent models are 
normally used for coverage area calculations.  The simpler model used here is intended only to provide some insight 
regarding signal excess. 
†
 J. Egli, “Radiowave propagation above 40 Mc over irregular terrain”, Proceedings of the IRE, Vol. 45, Oct. 1957, 
pp.1383-1391. 
 
2-3 
percent of the coverage area would experience lower signal levels than the ATSC-designated “weak” 
signal level of -68 dBm (where the excess signal is nominally 16 dB).
*
   
 
Given the above analysis and the increased vulnerability to interference when the desired signal levels are 
low, a decision was made to extend the results—through both measurement and extrapolation—to signal 
levels close to the DTV receiver thresholds.  Specifically, the desired signal threshold (D
MIN
) for each TV 
was measured in the absence of interference.  Interference rejection tests were then conducted at a desired 
signal level 3 dB higher than this threshold—i.e., at D = D
MIN
 + 3 dB.  In addition, analytical work was 
performed to extrapolate the test results down to a desired signal level of D
MIN
 + 1 dB. 
 
 
Table 2-2.  Relationship Between Excess Signal and Coverage Area 
Excess Signal 
(dB) R/R
MAX
 
Percentage of Coverage Area Having 
Less Excess Signal Than That Shown 
0 1 0% 
1 0.94 11% 
3 0.84 29% 
10 0.56 68% 
16 0.40 84% 
R = range from the broadcast antenna to the TV reception antenna; 
R
MAX
 = R at the edge of coverage 
 
 
TV Receiver Samples 
A total of 30 receivers were available for this test program.  One TV was procured in 2006 and 29 were 
provided by the manufacturers for the Congressionally-mandated 2005 SHVERA Study (though only 28 
had arrived in time to be included in that study).  The 29 receivers from the SHVERA Study had been 
selected to be representative of consumer DTV receivers that were on the market in the summer of 2005.  
Five of the receivers were set top boxes and 25, including the one procured in 2006, were DTVs with 
built-in digital ATSC tuners.  The DTV receivers comprise 16 brand names and a wide range of prices, 
sizes, and display technologies, as shown in Table 2-3. 
   
All 30 receivers were tested to determine tuner type.  Local oscillator sensing was sufficient to identify all 
but two of the receivers as having single-conversion tuners with 44-MHz IF.  The remaining two were 
subjected to limited interference rejection performance tests to look for signatures of single-conversion 
tuners.  Such signatures were found in both receivers. 
 
Because of the complexity of interference testing, only eight of the receivers were selected for 
interference rejection tests (beyond the limited testing on two receivers mentioned in the preceding 
paragraph).  The selection process was as follows.  Of the 28 receivers that had been tested in the 
SHVERA Study, only ten had been found to exhibit “upper tier” (or “fifth generation”) multipath 
performance.  Those ten, which were among the more recently introduced receivers used for the 
SHVERA Study
†
, along with the receiver procured in 2006 (which also exhibited “upper tier” multipath 
performance), were taken to be representative of the new generation receiver technology.  The eleven 
                                                      
*
 We note that these percentages assume that the same TV antenna system (e.g., a high gain antenna on a 10-meter 
mast and a downlead having a given loss) is used throughout the coverage area.  If closer-in TV viewers choose less 
extensive antenna systems (lower gain, shorter mast, or indoor location), those customers may experience low signal 
excess even if far inside the maximum coverage range. 
†
 Martin, <SHVERA Study>, 2005, chapter 6. 
 
2-4 
comprised eight brands of receivers.  All were DTVs (i.e., none were set top boxes).
*
  One receiver of 
each brand from among the eleven—a total of eight receivers—was selected for interference rejection 
testing.  The eight included receivers from all price ranges. 
 
Specific receivers are identified in both this report and the SHVERA Study by two-character codes—a 
letter followed by a number—in order to avoid revealing specific brands or models.  Additional 
information regarding the samples can be found in the SHVERA Study.
†
 
 
 
Table 2-3.  DTV Receiver Samples 
Number of Samples 
Sample Type 
LO Tests 
Rejection 
Tests 
Display 
Size 
Display 
Aspect 
Ratio 
Display Technology
A 
Set-Top Box (STB) 5 0 N/A N/A N/A 
DTV with Integrated 
ATSC Digital 
Tuner
B
: 
     
• $370 - $1000 7 2 26” – 36” 4:3 or 16:9 Direct-View CRT 
• $1001 - $2000 8 2+1
C 
26” – 52” 16:9 Direct-View LCD, 
Plasma, 
CRT Rear Projection, 
DLP Rear Projection, 
LCD Rear Projection 
• $2001 - $4200 10 4+1
C 
32” – 62” 16:9 Direct-View LCD, 
Plasma, 
DLP Rear Projection, 
LCD Rear Projection 
TOTAL 30 8+2
C 
   
Note: 
A – Display Technologies 
? CRT = cathode ray tube (conventional picture tube) 
? DLP = digital light processing 
? LCD = liquid crystal display 
B -- Prices shown are market prices in August or September 2005. 
C – 1
st
 number represents receivers selected for test based on having “upper-tier” DTV demodulator 
performance; 2
nd
 number represents receivers tested to characterize receiver type where LO tests were 
inconclusive. 
 
INTERFERENCE REJECTION TESTING APPROACH 
TV Channel Selection 
Initial tests were performed with TV receivers tuned to channel 51 and interferers placed on channels 52 
through 67 in order to provide data on the receivers’ ability to reject interference from services that may 
operate in the spectrum of channels 52 through 67 after the DTV transition is complete. 
 
Subsequent testing was performed with TVs tuned to channel 30, to represent performance near the 
middle of the UHF core band. 
                                                      
*
 The absence of a set top box among the receivers is consistent with the fact that all set top boxes that were 
identified for inclusion in the SHVERA Study had been introduced to the market in or before November 2004—a 
date that was probably too early to have included “fifth generation” DTV reception technology. 
†
 Martin, <SHVERA Study>, 2005. 
 
2-5 
 
In order to provide comparative data using a different test method, one set of tests was performed with the 
interfering source fixed at channel 29 and the TV receiver tuned to various channels from 14 to 37. 
Filtering of Undesired Signals 
DTV interference rejection testing requires extremely high suppression of out-of-band emissions from the 
undesired signal that might otherwise spill into the desired channel.  Available signal generators do not 
provide sufficient suppression.   
 
The conventional approach to dealing with this issue is to place a bandpass filter around the undesired 
signal.  In that approach the undesired signal is typically placed on a fixed channel so that a fixed filter 
can be used.  The desired signal and the TV tuner are then switched to various channels to achieve the 
channel-spacings to be tested. 
 
For this report a different approach was taken to the problem.  The desired channel was fixed and the 
undesired channel was varied.  A fixed band-reject filter was used to suppress the out-of-band emissions 
of the undesired signal that fell within the desired channel.  The details of the technique are provided in 
Chapter 4.  Tests using the conventional approach are reported in Chapter 14 for comparison. 
Operation and Connection of Samples 
For receivers having multiple antenna inputs that could handle ATSC signals, only the input labeled 
“antenna A” or “antenna 1” was tested.  For receivers having a radio frequency (RF) output associated 
with the selected antenna input, the output was externally terminated in 75 ohms. 
 
Only one TV was turned on during any given test in order to avoid possible interference from emissions 
of other TV receivers. 
Identifying Interference Rejection Thresholds 
In determining interference thresholds, we are interested in picture degradation that is visible to the TV 
viewer.  With digital television, some data transmission errors are fully corrected by error correction 
algorithms—resulting in no errors in the video transport stream data.  Other transmission errors that 
cannot be corrected may, in some cases, be effectively masked by error concealment techniques used in 
the receiver’s video processor.  We are only interested in picture errors that will be perceived by the 
viewer.  The subjectivity of visual error detection could be eliminated through relationships that have 
been established between visible TOV and bit-error-rate (BER); however, such techniques cannot be 
applied in testing of consumer DTV receivers that do not provide access to bit streams; consequently, 
thresholds for this report were determined by visual observation of DTV pictures. 
 
In all interference rejection tests, the level of the desired signal D was adjusted as closely as possible to 
the intended value by using a step attenuator operating in 0.1-dB steps.  The level of the undesired 
(interfering) signal was then adjusted upward until picture errors were easily observed within a few 
seconds.  That level was then backed off and readjusted in 0.1-dB steps to determine the minimum 
undesired signal level at which one or more visible picture errors occurred in two consecutive 30-second 
intervals.  The power level of the undesired signal was then measured and this level was identified as the 
undesired power level U at TOV—except in rare cases as described below. 
 
The thresholds exhibited a strong “cliff effect”.  In most cases, the increasing interference level about 1 
dB above the TOV level identified by the method above caused complete loss of picture.  In some cases, 
picture loss did not occur until the undesired signal level rose as much as much as 3 dB and in one case, 
5 dB.  In a few cases, picture loss occurred concurrently with the appearance of errors or with only an 
additional 0.1 dB increase in interference—an extremely abrupt cliff! 
 
 
2-6 
When the picture was lost due to high interference levels, it was recovered in most cases by reducing the 
undesired signal level back to the TOV level that was identified by the procedure described above, though 
in some cases that recovery took 20-seconds or more.  In a few cases, it was observed that—after loss of 
picture due to either high interference levels or a channel change—the TV was unable to re-establish a 
picture without reducing the undesired signal level a few tenths of a dB below the apparent TOV point.  
In such cases, the threshold was rerecorded as one 0.1-dB step above the level necessary to permit picture 
recovery.  (i.e., inability to recover the picture was treated as an error.) 
 
D/U ratios were computed based on the actual measured value of the desired signal D rather than on the 
intended setting of D (though the difference was generally less than 0.05 dB). 
 
Signal Power Measurements 
All measurements of desired and undesired power levels were made by means of the band power 
integration function of a spectrum analyzer that was set to perform an RMS average of spectrum traces.  
The number of points in the spectrum sweep was set so that bin spacing matched the 30 kHz resolution 
bandwidth used for the measurements.  The spectrum analyzer’s internal preamp was used to ensure a 
sufficiently low instrumentation noise level (approximately -98 dBm in 6-MHz bandwidth + analyzer 
attenuation).  The analyzer was used in the automatic attenuation mode with the reference level set to the 
lowest multiple of 5 dBm that was at least 1 dB above the total signal power.  In cases where power levels 
below -70 dB were to be measured, the analyzer attenuation was manually set to 0 dB.  For measurements 
below -78 dBm (measurements of desired signal level at or near the receiver threshold), analyzer noise 
was measured separately and subtracted—in linear power units—from the measured values. 
PRESENTATION OF RESULTS 
Interference rejection performance measurements on TV receivers are typically presented in terms of the 
ratio of desired to undesired signal powers (D/U) at TOV.  Protection criteria designed to prevent 
interference are sometimes specified in terms of D/U ratios and at other times specified as absolute signal 
levels (e.g., of transmit power or field strength) permitted in a band. 
 
D/U rejection ratios would provide a particularly useful characterization of interference rejection 
performance of a receiver if those rejection ratios remained constant as signal levels were varied.  
Unfortunately, we found this not to be the case for DTV receivers.  Interference at many channel offsets is 
driven by nonlinear mechanisms that cause D/U to increase with increasing signal levels; such variability 
is particularly common at low desired signal levels where a TV is most susceptible to interference.
*
  Even 
for linear interference mechanisms, D/U increases as the desired signal level approaches D
MIN
 for a 
receiver. 
 
Since D/U rejection ratios of DTV receivers are not constant, interference assessment requires knowledge 
of the absolute levels of desired and undesired signals at the input to the receiver rather than just a 
knowledge of the ratio of the signal powers.  But, the absolute signal levels at the input to a DTV receiver 
can vary widely depending on the gain, height, or indoor-versus-outdoor placement of the antenna to 
which it is attached.  Table 2-4 shows UHF reception examples for three different antenna systems: 
• An outdoor antenna system with a mast-mounted preamp sufficient to overcome downlead loss; 
• An outdoor antenna system according to OET-69 planning factors; 
• A low-gain, indoor antenna. 
                                                      
*
 We show later in this report that a tuner’s automatic gain control (AGC) can “stabilize” the effects of nonlinear 
interference mechanisms resulting in a more constant value of D/U at higher signal levels. 
 
2-7 
The table shows that the signal level at the TV’s RF input can easily vary over a 26-dB range simply by 
changing from an indoor antenna to an outdoor, mast-mounted antenna.
*
  The span can be even wider 
(30-dB or more) if a mast-mounted preamp is used to minimize the effect of downlead attenuation. 
 
Knowledge of the receive antenna system and deployment (e.g., indoor versus outdoor) used by a given 
TV receiver is not generally available to an outside entity, whether that entity is a smart-radio device 
transmitting in locally unused TV spectrum or an analyst assessing potential for interference between two 
DTV broadcast stations.  Even if a potential interferer to DTV reception had access to complete, accurate 
information regarding desired and undesired signal field strengths in a given DTV reception area, there 
would be no way to know where within a 30-dB signal level span that a given DTV receiver is operating 
without knowing the gain, height, or indoor-versus-outdoor placement of its antenna.  Thus, for example, 
a given receiver could be operating with a 1-dB signal margin (at D = D
MIN
 + 1 dB, or about -83 dBm) or 
at the ATSC “moderate” signal level (D = -53 dBm), based only on changes in the antenna system. 
 
Table 2-4.  UHF Reception Examples 
 
Outdoor  
Reception 
w/Mast-
Mounted 
Preamp
1
 
Outdoor 
Reception 
(OET-69) 
Indoor 
Reception 
(Low Gain 
Antenna)
2
 
Antenna gain (dBd) 10 10    0.0 
Downlead loss (dB)   0   4    0.6 
Antenna height (m) 10 10    2.0 
Relative field strength due to height difference (dB)
3
   0   0 -14.0 
Building loss (dB)
4
   0   0     5.0 
Relative signal level at input to TV (dB) 10   6 -19.6 
Notes 
1
 Mast-mounted preamp is assumed to have gain sufficient to overcome downlead loss. 
2
 Downlead loss for indoor antenna is based on 2 meters of RG-59 at 573 MHz (geometric mean frequency 
between channel 14 and channel 51). 
3
 Signal-strength dependence on height is based on the Egli propagation model, in which received signal power 
is proportional to the square of antenna height (Egli, J., “Radiowave propagation above 40 Mc over irregular 
terrain”, Proceedings of the IRE, Vol. 45, Oct. 1957, pp.1383-1391) 
4
 Building loss attenuation shown is intended only as an example—not as an endorsement of a particular value 
 
 
Despite the variation of D/U rejection ratios of TV receivers with absolute signal amplitudes, the D/U 
ratio formulation is convenient to use in applications like DTV-into-DTV interference assessment 
because estimation of D/U ratios may be easier and more accurate than estimation of absolute levels 
where long-distance propagation is involved—especially if the broadcast stations are co-sited.  In such 
applications a change in antenna gain, height, or indoor-versus-outdoor placement are likely to affect the 
desired and undesired signal levels in the same way, so that the D/U ratio to which the TV is exposed 
remains constant with antenna changes (assuming that the undesired and desired signal sources both fall 
within the main response of the TV directional pattern).  The lack of knowledge of the reception antenna 
means that interference assessment might have to consider a range of rejection ratios that are possible for 
the receiver given the range of signal levels that could reach the TV RF input from the range of likely 
antenna systems to which the potential victim TV receiver might be attached. 
 
On the other hand, some may find absolute undesired signal level thresholds to be more useful for 
assessing shorter distance interference from low-power devices because the effects of TV antenna height 
                                                      
*
 This data was based on a simple, flat-terrain propagation model.  The results are intended only to illustrate that 
signal level at the input to a TV receiver can vary substantially with changes in antenna type and placement. 
 
2-8 
and placement on the undesired signal are likely to be different from their effects on the desired signal.  In 
such cases, it may be useful to assess interference by determining the absolute level of an undesired signal 
that could cause picture degradation under the assumption that the TV could be operating at a low signal 
margin. 
 
Because of these differences in approach, this report presents the interference rejection performance 
measurements in two ways:  as D/U ratios and as threshold values for the undesired signal level U.
*
 
TEST SUMMARY 
Table 2-5 summarizes the tests performed for this report, including local-oscillator sensing and over 2000 
measurements of D/U ratio. 
                                                      
*
 This is more than just a convenience.  For fixed desired signal levels, one can easily translate data between 
threshold U values and D/U ratios; however, for desired signals levels that are receiver dependent (e.g., D
MIN
+3db), 
the desired signal power necessary to convert between the two formats may be lost to the user for results that are 
presented as, for example, median across eight receivers or second-worst of eight receivers. 
 
2-9 
Table 2-5.  Summary of Tests  
Test 
Chap-
ter 
# of TV 
Receivers 
Tested 
TV 
Tuned 
Chan (N)
Interference 
Channel 
Interferer 
Type 
Desired 
Signal Levels 
# of D/U 
Measure-
ments 
Tuner Type Tests (e.g., Single/Double Conversion) 
Local-oscillator 
(LO) sensing 
3 
30 
(11,5
th
-G) 
51 & 53 N/A N/A N/A  
Interference 
rejection when 
not resolved by 
LO sensing 
3 
2 (4
th
 G or 
earlier) 
30 N+2 to N+16 WGN* N/A 30 
Interference Rejection—Single Interferer 
Channel 30 
tests 5 
8 (5
th
 G) 30 N-16 to N-1, 
N+1 to N+16
8-VSB: N±1.
WGN
1
:  N±2 
to N±16 
D
MIN
+3 dB, 
-68, -53, 
-28 dBm 
1024 
Channel-51 
tests 
6 
7 (5
th
 G) 51 N+1 to N+16 8-VSB -68, -53, 
-28 dBm
2
 
225 
Channel-51 
detailed tests 
of one TV 
11 
1 (5
th
 G) 51 N+1 to N+7, 
N+14, N+15 
8-VSB D
MIN
+1 dB, 
D
MIN
+3 dB, 
-78 to -8 dBm 
in 5-dB steps 
98 
Interference Rejection for Paired Signals (3
rd
-Order Intermodulation in TV Receiver) 
Channel-30 
tests 
9 
8 (5
th
 G) 30 N-5/N-10 to 
N-1/N-2 and 
N+1/N+2 to 
N+5/N+10 
8-VSB for 
N±1; WGN
1
 
for all others 
-68, -53, 
-28 dBm
3
 
176 
 
Channel-51 
tests 9 
7 (5
th
 G) 51 N+1/N+2 to 
N+8/N+16 
8-VSB for 
N+K; WGN
1
 
for N+2K 
-68, -53, 
-28 dBm
2
 
162 
Channel-51 
detailed tests 
of one TV 
11 
1 (5
th
 G) 51 N+1/N+2 to 
N+4/N+8 
8-VSB for 
N+1; WGN
1
 
for all others 
D
MIN
+1 dB, 
D
MIN
+3 dB, 
-78 to -8 dBm 
in 5-dB steps 
57 
Interference Rejection—Comparison Tests with Different Sources 
Repeat test for 
consistency 
7 
8 (5
th
 G) 30 N-6, N-4, N-3, 
N-2, N+2 
WGN
1
 -68 dBm 40 
OFDM 
Interference 
7 
8 (5
th
 G) 30 N-6, N-4, N-3, 
N-2, N+2 
DVB-H 
(5-MHz width)
-68 dBm 40 
8-VSB 
Interference 
7 
8 (5
th
 G) 30 N-6, N-4, N-3, 
N-2, N+2 
8-VSB -68 dBm 40 
1-MHz wide 
Interferer 
7 
1 (5
th
 G) 30 N-15 to N-1,
N+1 to N+15
Gaussian 
noise (1 MHz)
-68 dBm 88 
Higher quality 
desired signal 
7 
8 (5
th
 G) 30 N-6, N-4, N-3, 
N-2, N+2 
WGN
1
 -68 dBm 40 
Interference Rejection—To Evaluate Test Method 
Alternative test 
setup 
w/bandpass 
filter 
14 
1 (5
th
 G) 14, 15, 
21 to 37 
Channel 29 
(N-8 to N-1, 
N+1 to N+8, 
N+14, N+15)
8-VSB -68 dBm 18 
Broadband 
Notched Noise 14 
8 (5
th
 G) 30 Chan 2 -69 
w/notch for 
chan 29-31 
Gaussian 
noise + band-
reject filter 
-68, -53 dBm 16 
 
Screen-Room 14 1 (5
th
 G) 30 N+7 WGN
1
 -68 dBm 1 
TOTAL       2055 
Notes (see next page): 
 
2-10 
Notes from Table 2-5: 
5
th
 G” refers to receivers having multipath performance equivalent to that of 5
th
-generation 
Zenith demodulators; all others tested as having earlier-generation demodulators. 
D
MIN
 is the desired signal level corresponding to the threshold of visibility (TOV) of picture 
degradation in the absence of interference. 
1
 WGN = white Gaussian noise bandlimited to 5.38-MHz 3-dB width. 
2
 Tests of one of the seven receivers tested channel 51 were incomplete.  For that receiver:  all 
tests at D = -68 dBm were completed; at D = -53 dBm, the single-interferer tests were 
completed, but the paired-interferer tests were performed only for N+1/N+2; at D=  -28 dBm, 
the only tests performed were N+1 and N+1/N+2. 
3
 Channel-30 paired-signal tests at D = -28 dBm were limited to N+1/N+2 and N-1/N-2. 
 
 
2-11 
-60
-50
-40
-30
-20
-10
0
10
20
-4 -3 -2 -1 0 1 2 3 4
Frequency Difference From Channel Center (MHz)
Normali
zed Spectrum 
L
evel (dB 
in 10-kHz
 R
B
W)
8-VSB (SFU)
Gaussian Noise (8-VSB Width)
Gaussian Noise (1-MHz Width)
DVB-H
Spectra are shown at equal total power levels
and are shown in dB relative to the 8-VSB flat-top
 
Figure 2-1.  Spectra of the Four Undesired Signal Sources 
-20
-15
-10
-5
0
5
10
-4-3-2-101234
Frequency Difference From Channel Center (MHz)
Normali
zed Spectrum 
Level
 (dB
 in 1
0
-kHz
 
RBW)
8-VSB (SFU)
Gaussian Noise (8-VSB Width)
Gaussian Noise (1-MHz Width)
DVB-H
Spectra are shown at equal total power levels
and are shown in dB relative to the 8-VSB flat-top
 
Figure 2-2.  Spectra of the Four Undesired Signal Sources—Expanded Scale 
 
 
2-12 
0-dB Margin
1-dB Margin
3-dB Margin
16-dB Margin (D = -68 dBm)
31-dB Margin (D = -53 dBm)
56-dB Margin (D = -28 dBm)
11% of Coverage Area Has < 1 dB Margin
29% Has < 3 dB Margin
84% Has < 16 dB Margin
Signal margins 
refer to excess signal 
above that needed for 
consistent reception
R/R
MAX
1.00
0.94
0.84
0.40
0.17
0.04
 
R = range from the broadcast antenna to the TV reception antenna; 
R
MAX
 = R at the edge of coverage 
 
 
Figure 2-3.  Relationship Between Signal Excess and Coverage Area 
 
3-1 
CHAPTER 3 
TUNER TYPE TESTS 
This chapter presents the results of tests of 30 DTV receivers to determine their tuner topologies. 
TUNER TOPOLOGIES 
Two tuner topologies are known to have been used in ATSC DTV receivers:  single-conversion tuners 
and double conversion tuners. 
 
Both types of tuners are designed to shift a desired, 6-MHz wide TV signal from its original frequency 
(centered between 57 MHz for channel 2 and 803 MHz for channel 69) to a lower, fixed frequency range 
where it is feasible to implement a filter that passes the desired channel without significant distortion of 
its spectral shape while providing a high degree of rejection of adjacent-channel signals.  Typically such a 
filter is implemented at a center frequency of 44 MHz and passes frequencies from 41 to 47 MHz.  A 
single-conversion tuner performs the frequency shift in one operation.  A double-conversion tuner 
performs it in two steps. 
Single-Conversion Tuners 
Figure 3-1 shows a simplified block diagram of single-conversion tuner.  The input filter may be broad 
enough to pass an entire TV band, such as the UHF band containing channels 14 through 69.  A “tracking 
filter” in the RF section adjusts with the TV channel selection and passes the desired TV channel and 
perhaps several channels on either side.  A local oscillator (LO) at a frequency 44 MHz above the center 
of the desired TV channel (e.g., 739 MHz when the receiver is tuned to channel 51) is then non-linearly 
mixed with the amplified and filtered RF signal.  This mixing down-converts the desired TV channel 
from its original frequency (e.g., 695 MHz +/- 3 MHz, for channel 51) to 44 MHz +/-3 MHz, which can 
pass through the fixed-frequency IF filter that serves to perform the primary channel selection function.
*
 
 
In addition to the desired TV channel—centered 44 MHz below the LO frequency—incoming signals that 
are located 44 MHz +/- 3 MHz above the LO are also down-converted into the IF filter band.  For UHF 
channels, which occupy contiguous 6-MHz spectrum assignments, this corresponds to parts of the energy 
in TV channel numbers N+14 and N+15, where N is the desired channel number.  The presence of these 
image signals can interfere with reception of the desired signal.  One of the purposes of the tracking filter 
in a single-conversion TV tuner is to attenuate signals at the image frequencies before they reach the 
mixer in order to mitigate the interference potential. 
 
Certain other signal interactions in a single-conversion tuner can create interference sensitivities at other 
channel spacings.  These will be discussed later in this report. 
Double-Conversion Tuners 
Figure 3-2 shows a simplified block diagram of a double-conversion tuner—as implemented in the Grand 
Alliance receiver.
†
  This tuner configuration does not use a tracking filter.  Rather, the entire TV spectrum 
                                                      
*
 The down-conversion also reverses the direction of the frequency spectrum because the local oscillator frequency 
is above the frequency of the incoming signal. 
†
 The diagram omits automatic gain control elements and lumps the first IF filter into a single filter located after the 
first IF amplifier; the actual implementation included filters before and after the first IF amplifier.  A more detailed 
block diagram is available in the following reference: 
Advanced Television Systems Committee, “Recommended Practice:  Guide to the Use of the ATSC Digital 
Television Standard”, ATSC Doc. A/54A, 4 December 2003, Figure 9.2, p.86. 
 
3-2 
is up-converted in such a way as to center the desired TV signal at 920 MHz.  After filtering, the resulting 
signal is then down-converted by the second mixer so that the desired signal is centered at 44 MHz. 
Comparisons 
The double-conversion design results in image frequencies that are further separated from desired signal 
as compared to a single-conversion design.  This separation makes it easier to filter out those undesired 
signal components.  A double-conversion receiver is therefore less likely to have detectable image 
responses.
 *
 
†
 
 
On the other hand, the lack of a tracking filter in double-conversion designs means that the first mixer 
must process all received TV signals rather than just the few channels surrounding the desired signal.  
Non-linear interactions between these various signals can create other interference issues.
‡
 
§
 Additionally, 
achieving low noise figure and phase noise is more difficult in double-conversion than single-conversion 
receivers.
**
 
††
 
 
LO MEASUREMENT FOR TUNER TYPE IDENTIFICATION 
In a single-conversion tuner, the LO frequency is located within the TV bands except when tuning the 
upper channels of a band.  A small amount of the LO signal can leak out through the antenna port of the 
TV receiver.  If the LO is detectable at the antenna port, its presence and frequency can be used to 
identify the tuner as single conversion and to confirm the IF frequency. 
 
With double-conversion tuners, the LO is located above the UHF TV bands and is thus more easily 
filtered out and less likely to be detectable at the antenna port.
‡‡
 
§§
 
LO Frequency Test Methodology 
The antenna port of each of the 30 DTV receivers discussed in Chapter 2 was observed using a spectrum 
analyzer in search of LO emissions.  (For receivers having multiple antenna inputs that could handle 
ATSC signals, only the input labeled “antenna A” or “antenna 1” was tested.)  During these tests, no 
signal was supplied to the antenna port; however, prior to these tests, a channel scan was performed on 
each TV while simultaneously applying ATSC signals on UHF channels 51 and 53.  This step was 
necessary because many DTVs prevent selection of a given TV channel unless a valid signal was 
observed on that channel in a previous channel scan. 
 
To improve the detectability of very weak LO emissions, the spectrum analyzer was operated with 0 dB 
input attenuation, the internal preamp turned on, resolution bandwidth set to 10 kHz, and trace averaging 
enabled.  The analyzer was set to sweep a 20 MHz span that included the frequencies of interest.  Use of a 
                                                      
*
 N. Scheinberg and others, “A GaAs Up Converter Integrated Circuit for a Double Conversion Cable TV ‘Set-Top’ 
Tuner”, IEEE Journal of Solid-State Circuits, Vol. 29, No. 6, June 1994, p.688 
†
 Wayne Bretl and others, “VSB Modem Subsystem Design for Grand Alliance Digital Television Receivers”, IEEE 
Transactions on Consumer Electronics. Vol. 41, No. 3, August 1995, p.773. 
‡
 Scheinberg and others, 1994, p.688 
§
 Nick Cowley and Robert Hanrahan, “ATSC Compliance and Tuner Design Implications”, Electronic Engineering 
Times, May 1, 2006. (http://www.eetasia.com/ART_8800416208_480700_f6d4765f200605.HTM) 
**
 Yiyan Wu, “Performance Comparison of ATSC 8-VSB and DVB-T COFDM Transmission Systems for Digital 
Television Terrestrial Broadcasting”, IEEE Transactions on Consumer Electronics, Vol. 45, No.3, August 1999, 
p. 922. 
††
 Charles W. Rhodes, “Interference Between Television Signals due to Intermodulation in Receiver Front-Ends”, 
IEEE Transactions On Broadcasting, Vol. 51, No. 1, March 2005, p.36. 
‡‡
 John Henderson and others, “ATSC DTV Receiver Implementation”, Proceedings of the IEEE, Vol. 94, No. 1, 
January 2006, p.125. 
§§
 Wayne Bretl and others, 1995, p.773. 
 
3-3 
2001-point sweep enabled a 0.01-MHz bin-to-bin spacing.  Each TV was initially tuned to channel 51 and 
the spectrum was observed for the presence of a line at 739 MHz—the LO frequency expected for a 
single-conversion tuner with a 44-MHz IF when tuned to channel 51.  The TV was then changed to 
channel 53, the spectrum averaging was restarted, and the spectrum was observed to determine whether 
the line shifted upward by 12 MHz to the LO frequency expected for channel 53. 
LO Test Results 
For 28 of the tested 30 DTV receivers, LO signals were detected at frequencies consistent with a single-
conversion receiver with an IF frequency of approximately 44 MHz.  For 25 of those, each observed LO-
associated line was within 0.02 MHz of the value expected (44 MHz above the center frequency of the 
tuned channel) for each tested channel.  One of the receivers (designated O1 in the SHVERA Study) 
exhibited LO-associated lines that were 0.18 MHz above the expected frequencies.  The LO-associated 
emissions from two other receivers (designated D2 and D3) exhibited a hunting behavior around the 
expected frequency—extending as far as 0.25 or 0.26 MHz from the expected frequency. 
 
These results provide clear evidence that at least 28 of the 30 consumer DTV receivers that were tested 
have single-conversion tuners with an IF frequency at, or very near, 44 MHz. 
 
For two of the receivers (designated G3 and P1), no LO signal was observed in the expected frequency 
range.  Based only on these results, each of these two receivers could have had a different tuner topology, 
such as a double conversion tuner, or they could have had single-conversion tuners but either with a 
different IF frequency than was expected or with better control of LO leakage to the antenna port than the 
other receivers; consequently, a conclusion regarding topology of these two receivers required additional 
tests. 
INTERFERENCE REJECTION TESTS FOR TUNER TYPE 
IDENTIFICATION 
Interference rejection tests were performed for the receivers G3 and P1, the two receivers for which LO 
sensing was inconclusive.  These two DTVs would be classified as fourth-generation or earlier
*
 based on 
their multipath performance, which was tested as part of the SHVERA Study.  Both were on the market in 
2005, though P1 was actually introduced to the market in 2004.   
 
The measurements were performed using techniques to be described in the next chapter.  A desired signal 
of -68 dBm was applied to the receivers on channel 30 along with a white noise signal bandlimited to a 
3-dB width of 5.38 MHz on another channel.  The undesired signal level was adjusted to the TOV of 
degradation of the television picture.  The resulting D/U ratios are plotted in Figure 3.3.   
 
Both receivers exhibit a peak in sensitivity to interference at N+7.  This channel contains the LO 
frequency of a single-conversion tuner with 44-MHz IF.  Such a peak can be observed in D/U plots 
presented in a later chapter of this report for seven of the other eight single-conversion receivers tested.   
 
In addition, receiver P1 exhibits elevated sensitivity to interference at N+14 and N+15.  This corresponds 
to the mixer image for a single-conversion tuner with 44-MHz IF.  Mixer image peaks are seen in D/U 
plots presented later in this report for seven of the eight single-conversion receivers tested. 
 
Based on these observations, receivers G3 and P1 are judged to have single-conversion tuners with 
44-MHz IF.  No further testing was performed on these two receivers. 
 
                                                      
*
 The term “fifth generation” in this report refers to DTV receivers that exhibit multipath performance equivalent to 
that of Zenith fifth-generation demodulators.  These two TVs exhibited multipath performance well below that level. 
 
3-4 
SUMMARY 
All 30 tested receivers were judged to have single-conversion topologies with 44-MHz IF. 
 
Demodulator
Tracking
Filter
(at TV Channel 
Center)
Mixer
IF Filter
(Centered 
at 44 MHz)
Local Oscillator
(44 MHz above
TV Channel Center)
Input
Filter
RF
Amplifier
IF
Amplifier
 
Figure 3-1.  Single-Conversion DTV Tuner Block Diagram Example 
 
Demodulator
Mixer
IF Filter
(Centered 
at 44 MHz)
Local Oscillator
(876 MHz)
Mixer
IF Filter
(Centered 
at 920 MHz)
Local Oscillator
(920 MHz above
TV Channel Center)
Input
Filter
RF
Amplifier
1
st
IF
Amplifier
2
nd
IF
Amplifier
 
Figure 3-2.  Double-Conversion DTV Tuner Block Diagram Example 
 
 
3-5 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
N+
2
N+
3
N+
4
N+
5
N+
6
N+
7
N+
8
N+
9
N+
1
0
N+
1
1
N+
1
2
N+
1
3
N+
1
4
N+
1
5
N+
1
6
Interfering Channel
S
u
sc
epti
bil
i
t
y to I
n
ter
f
er
e
n
c
e
 
[
D
/U
 at
 
Thre
s
h
o
l
d (dB
)
]
Measurement Limit
G3
P1
A/74 Limit
 
Figure 3-3.  Rejection Performance of Receivers with Undetected LO’s 
 
4-1 
CHAPTER 4 
INTERFERENCE REJECTION REFERENCE LEVELS AND 
TEST METHODOLOGY 
This chapter describes the test methodology employed for interference testing, the performance of the test 
setup, and some reference values for rejection performance from other documents. 
 
Channel-to-channel signal level differences required for out-of-channel interference rejection testing 
present some challenges that require the use of specialized filters.  The amount of filtering depends on the 
threshold desired-to-undesired (D/U) power ratios that are to be measured. 
 
REFERENCE LEVELS FOR INTERFERENCE REJECTION 
PERFORMANCE 
Interference rejection performance is defined in terms of the ratio of desired signal power (D) to 
undesired signal power (U) at the point at which visible degradation begins to occur in the television 
picture. 
 
Grand Alliance Receiver Performance 
The interference rejection capability achieved by the Grand Alliance prototype DTV receiver for a DTV 
interferer is shown in Table 4-1. 
 
Table 4-1.  Grand Alliance Receiver Interference Rejection Performance 
D/U Ratio (dB) 
K 
N – K N + K 
1 -41.98 -43.17 
2 -60.52 -59.13 
3 -60.61 -61.53 
4 to 15 Not measurable Not measurable 
 
It can be seen that, for DTV interference, the receiver achieved rejection ratios of about -42 and -43 dB 
for the two first-adjacent channels (N+1 and N-1) and rejection ratios of about -60 dB or better at all other 
tested channel spacings. 
OET-69 Guidance for Evaluating TV Service Coverage and Interference 
OET Bulletin No. 69, “Longley-Rice Methodology for Evaluating TV Coverage and Interference”, 
defines the FCC methodology for evaluating TV service coverage and interference protection.  Though it 
does not directly define TV receiver characteristics or requirements, it does define D/U ratios that are to 
serve as protection criteria to avoid DTV-to-DTV interference.  Those protection criteria are shown in 
Table 4-2. 
 
Table 4-2.  OET-69 DTV-Into-DTV Interference Protection Criteria 
Interfering
Channel 
D/U Ratio (dB) 
for DTV-into-DTV Interference 
N-1 -28 
N+1 -26 
 
4-2 
 
 
For other channel offsets, OET-69 states the following. 
 
“The evaluation of service and interference in Appendix B of the Sixth Report and Order considered 
taboo channel relationships for interference into DTV. However, the D/U ratios (approximately -60 dB) 
were such that they rarely if ever had an effect on the results, and the FCC rules adopted in the 
Sixth Report and Order do not require attention to UHF taboo interference to DTV stations.”
*
 
  
Thus, the interference rejection threshold of -60 dB assumed for DTV receivers was considered adequate 
to protect against DTV-into-DTV interference for channel spacings beyond N+1 and N-1 based on the 
allotment scenarios that were evaluated. 
ATSC Recommended Performance 
The ATSC Receiver Guidelines
†
 recommend that DTV receivers achieve the interference rejection 
capabilities shown in Table 4-3.  The performance thresholds are specified at three different desired signal 
levels, which the ATSC designates as “weak”, “moderate”, and “strong”. 
 
Table 4-3.  ATSC A/74 Recommended Thresholds for Receiver Interference Rejection 
Threshold D/U for Specified Desired Signal Level (dB) 
Interfering 
Channel Number 
Weak 
(-68 dBm) 
Moderate 
(-53 dBm) 
Strong 
(-28 dBm) 
N+/-1 -33 -33 -20 
N+/-2 -44 -40 -20 
N+/-3 -48 -40 -20 
N+/-4 -52 -40 -20 
N+/-5 -56 -42 -20 
N+/-6 to N+/-13 -57 -45 -20 
N+/-14 to N+/-15 -50 -45 -20 
Notes 
Channel “N” is the channel number of the “desired” signal—to which the DTV receiver is tuned. 
Bold Italics denote D/U thresholds that correspond to an undesired signal level of -8 dBm. 
 
It should be noted that the ATSC-designated “weak” and “strong” levels do not bound the range of 
expected signal levels.  The document recommends that receivers be able to operate with DTV signals 
ranging from -83 dBm to -8 dBm in level. 
 
The ATSC document explains the basis for the -33 dB rejection ratio for first-adjacent channel 
interference in Table 4-3.  It points out that the OET-69 protection criteria for allotting DTV stations 
permits a D/U ratio as low as -26 and -28 dB for first-adjacent channel interference.  The recommended -
33 dB receiver threshold was obtained by subtracting 6 dB from the mean of these values.  Since the 
criteria in OET-69 are derived from receiver susceptibility to transmitter splatter into the first adjacent 
channel (based on the DTV emission mask), a receiver threshold of -33 dB, measured without splatter, 
ensures that the factor determining adjacent channel interference will be transmitter splatter rather than 
receiver performance. 
 
The ATSC document does not explain the basis of the other D/U values it recommends; however, those 
identified by red italics in the above table correspond to interference at the maximum expected DTV 
                                                      
*
 “Longley-Rice Methodology for Evaluating TV Coverage and Interference”, Office of Engineering and 
Technology (OET) Bulletin No. 69, <OET-69>, Federal Communications Commission, 6 February 2004, p.8. 
†
 <ATSC Receiver Guidelines>, ATSC Doc. A/74, p.13-14. 
 
4-3 
signal level of -8 dBm.  We understand the other values to be a result of negotiations that considered the 
performance levels that would likely be achievable by consumer-grade TV receivers. 
 
TEST SETUP REQUIREMENTS 
Measurement Requirements 
Goals for performance of the test setup were established based on the above reference levels.  For first-
adjacent channels (N+1 and N-1), our goal was to able to measure threshold D/U ratios at least as low as 
the -33 dB level recommended by the ATSC Receiver Guidelines.  For all other channel spacings, or goal 
was to be able to measure down to a D/U ratio of -60 dB.  
 
We also wished to be able to supply desired and undesired signals at levels up to -8 dBm. 
 
The Challenge 
Consider the case of a D/U ratio of -60 dB—meaning that the desired TV signal power is 60 dB below the 
power of the interferer.  The SHVERA Study demonstrated that the median DTV receiver requires that 
the desired signal be at least 15.3 dB above any co-channel noise in order for successful reception to 
occur.
*
  For testing at a D/U ratio of -60 dB, this means that the co-channel noise created by the test setup 
must be at least 60 + 15.3 dB below the power of the interferer; otherwise, the test setup itself will 
prevent successful DTV reception.  In fact, in order to make test setup noise relatively insignificant at a 
D/U ratio of -60 dB, we want an additional 10 dB margin—meaning that co-channel noise created by the 
test setup must be at least 85.3 dB below the power of the undesired signal when that signal is placed on 
any channel other than a first-adjacent one.   
 
No available signal sources met this requirement.  The source used to generate the “undesired” signal for 
the testing on channel 51 achieved only a 49-dB spread between signal power and power splattered into 
the second-adjacent channel (i.e., power splattered into channel N when the undesired source was placed 
at channel N+2).  The source used for most of the testing on channel 30 was somewhat better but still 
achieved only a 56 dB spread between signal power and splatter into the second-adjacent channel. 
 
The -33 dB D/U specification placed on first-adjacent channel interference requires a less formidable 
sounding dynamic range.  For that measurement, any noise created by the test setup must be at least 33 + 
15.3 + 10 dB = 58.3 dB below the power of the interferer.   This specification was not met in the first 
adjacent channel of any source available during most of the testing. 
 
To achieve the required test setup performance levels, a filter is needed to further reduce the out-of-band 
components of the undesired signal source. 
Solutions 
A typical solution to this filtering problem is to bandpass filter the undesired signal to reduce energy 
leakage into the desired channel.  Because variable filters tend to have poorer shape factors than high-
quality fixed filters, one approach would be to select a fixed interfering channel for the tests and to 
procure a fixed filter to shape the undesired signal on that channel.  Testing at various channel spacings 
between the desired and undesired signals could be accomplished by switching the desired channel 
number over some required range. 
 
While this approach would work well for tests with a single interferer, it creates a problem for testing 
against a pair of interferers.  Intermodulation effects in the DTV receiver are expected to be most 
                                                      
*
 Martin, <SHVERA Study>, 2005, chapter 3. 
 
4-4 
prominent for specific pairs of interfering signal channels, such as N+1 and N+2, N+2 and N+4, or N+3 
and N+6.  For tests of those effects, one interferer could operate on a fixed channel, but the other would 
have to be movable. 
 
To deal with this problem, the desired channel was fixed and the undesired channels were allowed to 
change.  The undesired signals were then subjected to a band-reject filter at the desired channel 
frequency.  This fixed band-reject filter served to reduce leakage of the undesired signals into the desired 
TV channel. 
 
In actuality, for a given desired channel, two custom filters were procured:  one for adjacent-channel tests 
and the other for interferers located beyond the first-adjacent channel.  Of the two filters, the filter for 
non-adjacent tests has greater rejection in the desired channel, but would cause unacceptable spectral 
distortion of an undesired signal on the adjacent channel (N+1).  The filter for adjacent tests avoids 
excessive distortion to a first-adjacent undesired signal; it has less rejection in the desired channel than the 
other filter, but the rejection is sufficient for measurements at the more modest D/U ratios required for the 
first-adjacent channel. 
 
In addition, the more conventional bandpass filter approach was implemented for one set of tests for 
comparison. 
 
TEST SETUP 
Figure 4-1 shows an overall block diagram of the test setup used for interference rejection tests. 
 
The top left portion of the diagram shows the desired DTV signal source and associated amplifiers, along 
with a step attenuator allowing signal level to be adjusted in 0.1-dB steps over an 81-dB range.  A 
Sencore ATSC997 8-VSB generator playing a built-in high-definition video stream of a football game 
was used for most of the testing.  A Rohde and Schwarz SFU generator, acquired relatively late in the test 
program, was used for tests at a low desired signal level (D
MIN
 + 3 dB) and for some comparative tests; a 
built-in high-definition video stream of a shark tank served as the video content for those tests.  In the 
final two weeks before the due date of this report, some adjacent-channel testing was performed using a 
newly acquired Wavetech WS2100 RF Player, combined with an external upconverter and a file 
containing an 8-VSB signal mathematically derived from an MPEG2 transport stream,
*
 to create a higher 
quality desired signal source than the Sencore ATSC997, while freeing the SFU to act as undesired signal 
on the adjacent channel. 
 
Up to two generators were used at any given time to create the undesired (i.e., interfering) signals.  The 
generators included a Sencore RFP910 RF Player playing a supplied recording (“Hawaii Reference A”), 
the Rohde and Schwarz SFU mentioned above, two Agilent E4437B vector signal generators used to 
generate band-limited white Gaussian noise, and an Agilent 4438C vector signal generator equipped with 
Signal Studio for DVB software to generate an OFDM DVB-H signal.  The Sencore RF Player failed near 
the end of the planned tests at channel 51 and was unrepairable.  The Rohde and Schwarz SFU was 
procured later in the testing period. 
 
The two undesired signals are combined and amplified by a 5-watt power amplifier which is operated at 
an output power of only 0.07 watts in order to limit third-order intermodulation distortion products, which 
would fall within the desired signal channel.  A step-attenuator (Atten-C) is used to adjust the input level 
of the amplifier. 
 
                                                      
*
 The file was provided by Mark Hryszko of the Digital Television group of Advanced Micro Devices. 
 
4-5 
The amplified, undesired signal pair is then passed through one of four band-reject filters to reduce out-
of-band splatter into the desired channel to which the TV is tuned.  The filter is followed by a fixed 
attenuator and a step attenuator that allows the undesired signal level to be adjusted in 0.1-dB steps over 
an 81-dB range.  No active devices (e.g., amplifiers) are included in the test setup beyond the filter output 
in order to avoid creation of intermodulation products. 
 
The desired signal is then combined with the filtered undesired signal pair.  The combined signals are 
split into two paths—one feeding the DTV receiver under test through an impedance-matching pad and 
the other feeding a spectrum analyzer used for all power measurements.  A total of about 9-dB of 
attenuation is provided in each splitter output path in order to reduce the impact of any reflections caused 
by impedance mismatches at the DTV receiver input.  The attenuators following the splitter were selected 
to provide a signal level match between the TV port and the measurement port to within 0.1 dB across all 
TV channels. 
 
Double-shielded cables were used throughout the test setup because of the wide range of signal levels 
present simultaneously.  A 50-ohm impedance was maintained throughout the setup, except at two 8-VSB 
sources and the consumer TV inputs, which were each specified to be nominally 75 ohms.  The 75-ohm 
devices were matched to the rest of the test setup through impedance-matching pads or—in the case of 
one of the 8-VSB sources—an impedance-matching transformer.  In addition to the impedance-matching 
pads, 50-ohm attenuator pads were used at various places throughout the test setups to reduce the effects 
of any impedance mismatches at places where such mismatches were considered likely or would be 
expected to have a significant impact, as well as to reduce third-order intermodulation in amplifiers A1 
and A2. 
 
The test setup is capable of delivering undesired signals to the TV receiver at a maximum level ranging 
from -7 to -1 dBm per interferer. 
 
TEST SETUP PERFORMANCE 
As was described in the Test Setup Requirements section, the test setup was required to suppress splatter 
from the undesired signals into the desired channel by as much as 85.3 dB.  No available spectrum 
analyzer had sufficient dynamic range to measure this degree of suppression.  Instead, the spectrum of the 
output of the test setup was first measured with the filter bypassed.  These measurements were performed 
with no desired signal present and for several different channel selections for the undesired signal with 
undesired signal set to a high level.  Separately, the frequency response of the test setup was then 
measured both with the filter in place and with the filter bypassed; the difference between these 
measurements represents the in-situ filter frequency response.  The filter frequency response was then 
applied to the spectrum measurements made with the filter bypassed—resulting in a computed value for 
the net output spectrum. 
 
Figure 4-2 shows the output spectrum of the test setup with channel 30 as the desired channel and a 
bandlimited white Gaussian undesired signal at channel N+2.  Integration of the blue—undesired-signal-
only curve shows that the power splattered by test setup into channel N is 99.3 dB below the total 
undesired signal power.  About 56 dB of this suppression is due to the performance of the undesired 
signal generator—degraded slightly by the amplifier that follows it.  The remaining 43 dB of suppression 
comes from the band-reject filter. 
 
In the figure, the spectrum of a desired DTV signal is plotted at a total power level 60 dB below the 
power of the undesired signal, i.e., at a D/U ratio of -60 dB.  Since a typical DTV requires that D be 
15.3 dB above any co-channel noise, the test setup noise is 24 dB below the point at which DTV 
 
4-6 
operation would fail.  This significantly exceeds the 10-dB margin that was considered essential for 
meaningful D/U measurements.
*
 
 
Figure 4-3 shows another example of the output spectrum of the test setup—again with channel 30 as the 
desired channel, but this time with a pair of interfering signals spaced to create intermodulation distortion 
in channel N.  In this case the unintended power leaked into the desired channel is 98.2 dB below the 
undesired signal power. 
 
Figure 4-4 shows an adjacent-channel example.  A pair of interferers is placed at N+1/N+2.  The 
interferer at N+1 (U
1
) is an 8-VSB source; some rounding of the left side of the 8-VSB signal by the 
band-reject filter can be seen.  The interferer at N+2 (U
2
) is a white-Gaussian-noise source bandlimited to 
match the 3-dB width of an 8-VSB signal.  Two different interferer types had to be used for this test:  
though two bandlimited noise generators were available, their spectrum rolloff is not steep enough for use 
as an adjacent channel (N+1) source because too much power would be spilled into the desired channel; 
and, only one 8-VSB source (besides the one used as a desired signal) was available. 
 
In the case of N+1/N+2 interference, there is no need to measure D/U ratios as low as -60 dB.  In this 
case, the desired signal is shown at a D/U ratio of -33 dB—the ATSC-recommended rejection 
performance for first-adjacent-channel interference.  Based on integration of the signal spectra, the total 
noise power leaked into the desired channel is 70.2 dB below the undesired signal power.  This provides a 
21.9 dB margin to the point of reception failure caused by the test setup.  (Actual margin is likely higher 
than this because, as can be seen from the plot, much of the undesired power that leaks into the desired 
channel N is at the band edges where filtering within the DTV receiver will reduce its effect.)  With 
10 dB being considered the minimum acceptable margin, receiver measurements could be made with this 
signal configuration down to a D/U ratio of about -45 dB per undesired signal. 
 
Table 4-4 summarizes the minimum D/U ratios that can be measured by the various test setup 
configurations used in this report based on the undesired signal leakage into the desired channel.  Two 
limitations are listed for adjacent-channel test configurations:  a limit based on total power leaked into the 
desired channel and a limit that includes the effect of a DTV receiver’s raised-cosine filter response on 
spectrum leakage at the edges of the desired channel.  The former limit (“worst case limit”) is displayed 
on plots as the “Measurement Limit” for measurements on channel 30.  The latter limit is used as a 
measurement limit for measurements at channel 51 to avoid unnecessarily excluding measurement results 
from further analysis.  (This decision was made because the channel-51 test setup had poorer performance 
than that for channel 30, in that it spilled more energy from the undesired signal into the edge of the 
desired channel.)  
 
In addition, minimum D/U ratio is limited by the maximum undesired signal level that the test setup can 
produce.  The maximum undesired signal level ranges from about -7 dBm to -1 dBm depending on the 
test setup configuration and channel spacing being tested.  Typically, at a desired signal level of -68 dBm 
the D/U measurement range is limited by leakage of undesired signal into the desired channel for 
measurements at N+1 or N-1 and by maximum undesired signal that the test setup can generate for all 
other channel spacings. 
 
                                                      
*
 The indirect measurement method used in generating the spectrum—measuring it without the filter, then adding in 
an in-situ measurement of filter response—would not identify any energy coupled by radiation into the test setup at 
a point after the filter or any spectral components created by intermodulation distortion occurring after the filter; 
however, the use of double-shielded cables throughout the test setup and the avoidance of using any active devices 
after the filter are expected to preclude significant degradation in test setup performance due to these factors. 
 
4-7 
TEST SETUP CONFIGURATIONS 
Table 4-5 summarizes the equipment configurations used for the various tests of interference rejection 
performance. 
 
Table 4-4.  Measurement Limitations of the Test Setup 
Desired 
Channel 
N Filter 
Undesired 
Source 1 
Undesired 
Source 2 
Applicable 
Interference 
Channels 
Cases 
examined
Worst-
Case 
Limit 
on D/U 
(dB) 
D/U Limit If 
the DTV’s 
Raised 
Cosine 
Filter Is 
Assumed in 
TV 
(dB) 
30 30N 
8-VSB 
(Rohde 
SFU) 
None N+1, N-1 Both -48.3 -59.4 
30 30W WGN None 
N+2 to N+16 
and 
N-2 to N-16 
N-16, N-8, 
N-2, N+2, 
N+8, 
N+16 
-74.0  
30 30N 
8-VSB 
(Rohde 
SFU) 
WGN 
Pairs:        
N+1/N+2 and
N-1/N-2 
Both -44.9 -52.5 
30 30W WGN WGN 
Pairs: 
N+2/N+4 to 
N+8/N+16; 
N-2/N-4 to  
N-8/N-16 
Pairs:      
N-8/N-16, 
N-4/N-8, 
N-2/N-4, 
N+2/N+4, 
N+4/N+8, 
N+8/N+16 
-71.7  
51 
P52-
53 
8-VSB 
(Sencore 
RFP) 
None N+1 N+1 -37.6 -43.9 
51 
P52-
53 
8-VSB 
(Sencore 
RFP) 
WGN N+1/N+2 N+1/N+2 -36.7 -42.6 
51 
P53-
56 
8-VSB 
(Sencore 
RFP) 
WGN 
N+2 to N+16 
and pairs:  
N+2/N+4 to 
N+8/N+16 
All pairs 
and N+9 
to N+16 
-71.0  
Note:  WGN refers to bandlimited white Gaussian noise from an Agilent E4437B vector signal generator 
 
 
4-8 
Table 4-5.  Test Setup Configurations and Settings 
N Interferer(s) Test 
Setup 
Desired 
Signal 
Source 
Undesired 
Source 1 
Undesired 
Source 2 
Step 
Atten-C 
(dB) 
Interference Rejection—Single Interferer 
30 Single adjacent Primary 8-VSB Src A 
@ 38dBmV 
8-VSB Src C 
@ +2 dBm 
 5 
30 Single adjacent 
w/D at D
MIN
+3 dB 
Primary 8-VSB Src D 8-VSB Src C 
@ +2 dBm 
 5 
30 Single non-
adjacent 
Primary 8-VSB Src A 
@ 38dBmV 
WGN Src A 
@ -7 dBm 
 2 
30 Single non-
adjacent w/D at 
D
MIN
+3 dB 
Primary 8-VSB Src C 
@+2dBm 
WGN Src A 
@ -7 dBm 
 2 
 
51 Single Primary 8-VSB Src A 
@ 38dBmV 
8-VSB Src B 
@ -30 dBm 
 2 
Interference Rejection for Paired Signals 
30 Paired adjacent 
(e.g., N+1/N+2) 
Primary 8-VSB Src A 
@ 38dBmV 
8-VSB Src C 
@ +2 dBm 
WGN Src 
A @ 
variable 
level 
5 
30 Paired non-
adjacent 
Primary 8-VSB Src C 
@+2dBm 
WGN Src B 
@ -7 dBm 
WGN Src 
A @ 
variable 
level 
8 
51 Paired Primary 8-VSB Src A 
@ 38dBmV 
8-VSB Src B 
@ -30 dBm 
WGN Src 
A @ 
variable 
level 
2 
Interference Rejection—Comparison Tests With Different Sources 
30 Single non-
adjacent DTV 
interferer 
Primary 8-VSB Src A 
@ 38dBmV 
8-VSB Src C 
@ +2 dBm 
 2 
30 Single non-
adjacent OFDM 
interferer 
Primary 8-VSB Src A 
@ 38dBmV 
OFDM Src A 
@ -6.9 dBm 
 2 
Interference Rejection—To Evaluate Test Method 
Variable (U 
at 29) 
Single adjacent & 
nonadjacent 
using Simplified 
Test Setup 
w/BPF on U 
Alt 8-VSB Src A 
@ variable 
output level 
8-VSB Src C 
@ variable 
output level 
 NA  
Test Setups 
• Primary: Figure 4-1 
• Alt.: Figure 14-2 
Signal sources 
• 8-VSB Source A: Sencore ATSC997 ATSC Source 
• 8-VSB Source B: Sencore RFP910 RF Player (“Hawaii_ReferenceA” file) 
• 8-VSB Source C: Rohde & Schwarz SFU 8-VSB Generator 
• 8-VSB Source D: Wavetech WS-2100 RF Player (“Muddy Waters” file) + Drake DUC860 Upconverter 
• AWGN Source A: Agilent E4437B Vector Signal Generator in AWGN mode 
• AWGN Source B: Agilent E4437B Vector Signal Generator in AWGN mode 
• OFDM Source A: Agilent E4438C Vector Signal Generator + Agilent Signal Studio for DVB software 
 
4-9 
AGILENT E4437B
VECTOR SIGNAL GEN
50-75 ohm
min loss pad
25-ft
Low-Loss
Coax
DTV Rcvr
50-ohm coax 
(25-ft cables are LMR-400-UF)
75-ohm BNC 90° elbow
+ BNC to F adapter
Direct connection
6-dB
PAD
SPECTRUM
ANALYZER
MLP#13
P#M
SENCORE 
RFP910
RF CAPTURE
PLAYER
(-30 dBm)
75-50 ohm
transformer
A3
Custom 
Band-
Reject 
Filter
SENCORE
ATSC997
@ 38 dBmV
STEP
ATTEN
-D
75-50 ohm
min loss pad
6-dB
PAD
A1 A2
6-dB
PAD
6-dB
PAD
COMBINER
STEP
ATTEN -C
A4
STEP 
ATTEN
-U
COMBINER SPLITTER
PtA
PtC
8-VSB Source A
8-VSB Source B
AWGN Source A
Select for <0.1 dB match
between PtC & PtE
Set to limit 
IM3 on A4
3-dB
PAD
3-dB
PAD#1
3-dB
PAD#2
3-dB
PAD
6-dB
PAD
PtD
6-dB
PAD
2 watt
rating
PtE
MLP#11
MLP#10
PtF
Rohde & Schwarz 
SFU 8-VSB 
GENERATOR
@ +2 dBm
8-VSB Source C
20-dB
PAD
AGILENT E4437B
VECTOR SIGNAL GEN
AWGN Source B
3-dB
PAD
6-dB
PAD
PtB
~17.8-18.8 dBm/signal
UND
ESI
R
E
D SI
G
N
A
L
 
S
O
U
RCE
S
DESIRED S
I
GN
A
L
25-ft
Low-Loss
Coax
Connection Alternatives
WAVETECH WS-2100
RF PLAYER
w/”Muddy Waters” File
DRAKE DUC860
Upconverter
8-VSB Source D
AGILENT E44338C
VECTOR SIGNAL GEN
OFDM Source A
UND
ESI
R
E
D SI
G
N
A
L
 
S
O
U
RCE
S
DESIRED S
I
GN
A
L
 
• Amplifiers 
? A1 = MiniCircuits ZFL-1000H (28 dB minimum gain; 20 dBm 1-dB compression) 
? A2 = MiniCircuits ZFL-1000VH (20 dB minimum gain; 25 dBm 1-dB compression) 
? A3 = HP8447B (22 dB gain; 400-1300 MHz) 
? A4 = Amplifier Research 5W1000 (37 dB gain; 500 kHz – 1000 MHz; 5 watts output) 
• Attenuators 
? Attenuators preceding A1 and A2 are selected to reduce IM3 to acceptable levels 
? Step Attenuator-C set to reduce IM3 of A4 to acceptable levels 
? Step Attenuators D & U: Alan Industries models 50V70 N, 50V10 N, and 50V1 N cascaded to provide 
0 - 81 dB in 0.1-dB steps (0.5W max power) 
• Combiners & Splitter:  MiniCircuits ZAPD-900-5W (100-900 MHz) 
• Custom Band-Reject Filters 
? “30N” = Tin Lee CE7-569(4.8)N50 
? “30W” = Microwave Filter Company model 16195 
? “P52-53” = Tin Lee CE7-692/697.4(20) N50 
? “P53-56” = Tin Lee CE7-692/698 N50 
• Impedance Matching 
? Minimum Loss Pads = Trilithic ZM-57 
? 75-50 ohm transformer = Trilithic ZMT-57 
• 25-ft coax = Times Microwave LMR-400-UF 
• Equipment settings 
? Agilent E4437B settings for bandlimited white Gaussian noise 
? AWGN mode w/length 1048576 
? Bandwidth setting = 4.686 MHz for 5.38-MHz 3-dB width 
? Bandwidth setting = 875 kHz for 1-MHz 3-dB width 
? Output setting = -7 dBm. (Higher could damage Step-Atten-U & raise IM3 of E4437B output) 
? Agilent E4438C vector signal generator using Agilent Signal Studio for DVB software 
? Signal Type:   DVB-H 
? Waveform parameters:  Size=2k; Modulation=64 QAM; Chan. width=5 MHz; Guard interval=1/8 
? Output setting = -6.9 dBm 
Figure 4-1.  Block Diagram of Interference Rejection Test Setup 
 
 
4-10 
 
-110
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
560 566 572 578 584 590
Frequency (MHz)
No
r
m
a
lized
 S
p
ec
tr
u
m
 L
e
v
e
l 
(d
B i
n
 3
0
 
kHz RBW
)
Undesired signal at N+2
Desired Signal with D/U = -60 dB
P
o
we
r
 L
e
a
k
e
d
 In
to
 Ch
an
n
e
l 
N
is
 99
.3
 d
B
 b
e
lo
w 
U
D/U
 
=
 -
6
0
 d
B
15.3 dB
Required CNR
24 dB Margin
Chan N Chan N+2
D
U
 
Figure 4-2.  Leakage of U at N+2 into Channel N (30) 
-110
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
536 542 548 554 560 566 572 578
Frequency Offset from Center (MHz)
N
o
r
m
alized
 S
p
ectr
u
m
 
Level (dB
 in 30-kHz
 RB
W
)
Pair of Undesired Signals at N-2/N-4
Desired Signal with D/U = -60 dB
Chan NChan N-2Chan N-4
D
U
1
U
2
 
Figure 4-3.  Leakage of U pair at N-2/N-4 into Channel N (30) 
 
4-11 
-110
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
560 566 572 578 584 590
Frequency Offset from Center (MHz)
N
o
r
m
alized
 S
p
ectr
u
m
 
Level (dB
 in 10-kHz
 RB
W
)
Pair of Undesired Signals at N+1/N+2
Desired Signal with D/U = -33 dB
Chan N
D
U
1
U
2
P
o
w
e
r
 Lea
ked
 
I
n
t
o
 C
h
a
nnel
 N
is
 7
0
.
2
 
dB b
e
lo
w 
U
D/
U =
 -
3
3
 d
B
15.3 dB
Required CNR
21.9 dB 
Chan N+1 Chan N+2
 
Figure 4-4.  Leakage of U pair at N+1/N+2 into Channel N (30) 
 
 
5-1 
CHAPTER 5 
REJECTION RESULTS ON CHANNEL 30 FOR SINGLE 
INTERFERERS WITH FULL CHANNEL WIDTH 
This chapter presents the results of interference rejection tests of eight “fifth-generation” DTV receivers 
tuned to channel 30.  The interferer for these tests was: 
• For channels N+1 and N-1, an 8-VSB signal; 
• For channels N+2 through N+16 and N-2 through N-16, a white Gaussian noise signal bandlimited to 
match the 3-dB width of an 8-VSB signal. 
Spectra of the sources were shown in Figures 2-1 and 2-2. 
 
A limited set of interference rejection tests was also performed on two earlier-generation receivers, 
designated G3 and P1, for purposes of identifying tuner type as described in Chapter 3.  This data was 
plotted in Figure 3-3, but it is repeated near the end of this chapter with an overlay showing the range of 
values measured for the fifth-generation receivers for comparison. 
 
The primary focus of these tests was to investigate the interference susceptibility of DTV receivers 
operating in the UHF band to interference from other occupants of the UHF TV spectrum, including other 
TV broadcasts as well as non-TV use of the “white spaces”.  As such, the desired channel (30) was 
selected as a locally unused channel near the center of the UHF core spectrum. 
 
Test results are presented in this chapter either as D/U ratios, in which case better performance 
corresponds to lower points on the graph, or as threshold values of the undesired signal level U, in which 
case better performance corresponds to points nearer the top of the graph. 
 
TESTS AT ATSC-SPECIFIED DESIRED SIGNAL LEVELS 
“Weak” Desired Signal (D = -68 dBm) 
Figure 5-1 shows measured values of D/U ratios at TOV for the eight DTV receivers for undesired signal 
channels ranging from N-16 to N+16.  (The case of co-channel interference, i.e., interference on channel 
N, is omitted.)  The desired signal power was set to -68 dBm, a signal level that the ATSC chose to 
designate as “weak”, although DTVs are assumed to operate down to a signal level of -84 dBm.
*
 
 
The shaded area at the bottom of the plot represents the measurement limitations imposed by the test 
setup—as described in Chapter 4.  Measurements falling in—or at the border of—this region are not 
valid; the actual performance of a receiver at these points is better (i.e., lower on the graph) than the 
plotted point indicates.  For N-1 and N+1 with D = -68 dBm, the limit is based on leakage from the 
undesired source into the desired channel.  For other offsets and higher desired signal levels, the 
measurements are limited by the maximum undesired signal power the test setup could inject into the 
DTV receiver. 
 
The ATSC-recommended DTV-into-DTV interference rejection thresholds are shown on the plot as a 
reference.  Those limits are defined for channels ranging from N-15 and N+15.  Compliance with those 
                                                      
*
 DTV allotment planning factors assume that a DTV receiver can operate at an input level of -84 dBm on UHF 
channels.  The ATSC Receiver Guidelines document recommends that receivers be able to operate with signal levels 
at least as low as -83 dBm.  Measurements on 28 DTV consumer receivers in the SHVERA Study showed a median 
capability of -83.9 dBm at channel 30. 
 
5-2 
voluntary limits would be indicated by all points on a measurement curve falling on or below the ATSC 
line.  It should be recognized, however, that the ATSC recommendations apply when the undesired signal 
is an 8-VSB DTV signal, which was the case only for the N-1 and N+1 points on each curve.  It can be 
seen that all eight receivers comply with the ATSC recommendation at N-1 and N+1. 
 
At the other channel offsets, N-2 through N-15 and N+2 through N+15, no receiver appears to fully 
comply with the recommended performance limit.  The best-performing receiver, designated G4, 
complied everywhere except at N+5 and N+6, where its performance failed to meet the limits by about 
2 dB and 1 dB, respectively.  On average, the receivers failed to meet the recommended performance on 
at about seven of the 30 channel offsets, with one receiver (D3) failing at twelve points.  The worst failure 
for each receiver ranged from about 2 dB to 25 dB. 
 
The above results cannot be viewed as definite failures to meet the performance guidelines because the 
tests (other than at N-1 and N+1) were performed using a bandlimited white noise source as the interferer, 
rather than an 8-VSB signal.  (An 8-VSB source was not available for most of the test period.)  Limited 
tests presented in Chapter 7 show an average performance improvement of 1.1 dB when the interference 
comes from an 8-VSB signal rather than from the white Gaussian noise source of the same 3-dB 
bandwidth.  However, even taking this difference into account, it is unlikely that any of the receivers 
would fully comply with the ATSC guidelines at every channel offset, though one or two would probably 
come close. 
 
Figure 5-2 summarizes the measurements that were shown in Figure 5-1.  The blue curve shows the 
median performance of the eight receivers.  Error bars show the best and worst performance among the 
receivers at each channel offset.  A dashed curve shows the performance of the second worst performing 
receiver at each channel offset.  On a median basis, the offsets that break the recommended limits by the 
largest amounts are N-4, N-6, and N+7. 
“Moderate” Desired Signal (D = -53 dBm) 
Figure 5-3 shows measured values of D/U ratios at TOV for the same eight DTV receivers with the 
desired signal power set to -53 dBm, a signal level that the ATSC designates as “moderate”.  Again, all 
eight receivers comply with the ATSC recommended performance on the first-adjacent channels (N+1 or 
N-1).  At the other channel offsets, only one receiver (G4) appears to fully comply with the recommended 
performance limit.  A second receiver (I1) appears to fail only at N-2 by 1 dB and at N-4 by a negligible 
amount.  Other receivers fail at from one to 16 points with worst-case failures ranging from three to 
18 dB.  The worst performing receiver was D3.  Again, these results cannot be viewed as definite failures 
to meet the guidelines because the tests were performed using a bandlimited white noise source as the 
interferer, rather than an 8-VSB signal.  Based on the differences in interference effect of the 8-VSB and 
Gaussian signals, it is likely that a second receiver would have complied with the guidelines at this 
desired signal level if an 8-VSB signal had been used as the interferer. 
 
Figure 5-4 shows the best, median, second worst, and worst performance at each channel offset.  On a 
median basis the only failure to satisfy the ATSC recommended performance is at N+7. 
“Strong” Desired Signal (D = -28 dBm) 
Figure 5-5 shows that, with the desired signal set to the level that the ATSC designates as “strong”, every 
receiver complied with the ATSC Recommended Guidelines at every point, with the exception of one 
receiver (G4) that appeared to fail by only 0.2 dB at N+1.  In most cases the test setup was not capable of 
generating strong enough undesired signals to cause visible degradation of the TV picture; consequently, 
most data points are plotted on the measurement limit line.  Figure 5-6 shows the best, median, second 
worst, and worst performance at each channel offset. 
 
 
5-3 
SELECTION OF A WEAKER DESIRED SIGNAL FOR TESTS 
Motivation 
The data in this chapter and later chapters of this report show that, as desired signal power at the input to a 
TV decreases, the amount of undesired power that the TV can tolerate on another channel also decreases.  
This means that a TV is most vulnerable to interference when operating at a low desired signal level—not 
a surprising result. 
 
One might also hypothesize that, as the desired signal power D at the input to the TV decreases, the 
undesired signal power U necessary to cause picture degradation would decrease at the same rate.  If this 
were true, the D/U ratio would remain constant, and interference rejection measurements would only need 
to be performed at a single power level to gain an understanding of the TV’s rejection performance.  
From the test results in this report, it is clear that this hypothesis is false.  In fact, the rate of change of 
threshold undesired signal with desired signal differs with channel offset between the desired and 
undesired channel and also varies with desired signal amplitude. 
 
The variation of D/U ratio with signal amplitude—often in unexpected ways—suggests the need for 
rejection performance measurements at a variety of amplitudes.  In Chapter 11, we subject a single TV to 
D/U measurements over a wide amplitude range to gain some insight into interference behavior.  
However, given that TV’s are most susceptible to interference at low desired signal levels, we focus here 
on weak signals. 
 
The range of desired signal levels that a DTV receiver can or should accommodate extends from: 
• D
MIN
 (nominally -84 dBm)  
— the desired signal level at which a DTV receiver begins to experience visible picture 
degradation; -84 dBm is the threshold assumed by OET-69 for edge-of-coverage 
UHF reception and also the median channel-30 threshold of 28 consumer DTV 
receivers measured under the SHVERA Study; to, 
• -8 dBm — the maximum DTV signal level anticipated by the ATSC Receiver Guidelines. 
 
In between are the three levels at which ATSC A/74 defines interference rejection performance guidelines 
for DTV receivers: 
• -68 dBm -- “weak” 
• -53 dBm -- “moderate” 
• -28 dBm -- “strong” 
The measurements of rejection performance presented thus far in this chapter were performed at these 
three levels. 
 
Note that the -68 dBm signal level designated by the ATSC as “weak” is 16 dB above the -84 dBm 
minimum signal level at which a typical DTV receiver can operate; that minimum signal level of -84 dBm 
is also the signal level assumed by OET-69 (FCC’s document for predicting coverage of a TV station) to 
be available to a DTV receiver at the edge of coverage of a TV broadcast station.
*
  Table 2-2 and Figure 
2-3 of Chapter 2 showed that fully 84 percent of the coverage area of a broadcast station may experience 
desired signal levels weaker than -68 dBm, assuming that the same type of antenna system is used at all 
locations in the viewing area.  This suggests a need to test at lower signal levels. 
 
                                                      
*
 Section 73.622E of the rules (CFR 47) establishes 41 dBuV/m as the UHF DTV field strength at the noise-limited 
edge-of-coverage contour.  This field strength is derived from a planning model in which a -84 dBm signal is 
delivered to the DTV input at the edge-of-coverage in order to enable at TV with a 7-dB noise figure and a 15-dB 
required signal-to-noise ratio to operate. 
 
5-4 
Signal Level Selection 
Selecting a signal level for interference tests that represent a near edge-of-coverage condition is not 
straight-forward.  Only about half of the 28 receivers tested for the SHVERA Study could produce an 
error-free picture with a channel-30 desired signal level matching the -84 dBm edge-of-coverage signal 
level assumed by OET-69, even without adding any interference, and a few TVs would produce no 
picture at all at this signal level.  Among those receivers, the TV with the least UHF sensitivity required 
an input signal level about 2.6 dB larger than the -84 dBm level, while the most sensitive TV could 
operate at a signal level 1.3 dB weaker than -84 dBm.  The spread between the most-sensitive and least-
sensitive receivers was larger in the VHF band—as high as 15 dB. 
 
If, in order to accommodate relatively insensitive receivers, a relatively high desired signal level is chosen 
as the measurement basis for rules to prevent interference to television, the resulting rules for a new 
service might permit interference levels that would limit the operation of the better receivers that can 
operate on weaker signals.  Hence a decision was made to also test at a receiver-dependent desired signal 
level near the threshold of visible picture degradation (TOV) that occurs in the absence of interference, a 
level we have designated as D
MIN
. 
 
Table 2-2 and Figure 2-3 of Chapter 2 showed that a TV located beyond 84 percent of the maximum 
reception range (R
MAX
) of the broadcast station (for a given antenna system) will receive signal levels at 
less than 3 dB of signal margin (i.e., signal level relative to that needed for consistent, clear reception of 
the broadcast signal).
*
  29 percent of the coverage area falls in this region of less than 3 dB signal excess.  
Similarly, and 11 percent of the coverage area will exhibit a signal excess of less than 1 dB.
†
 
 
The assumptions regarding coverage area are typically based on use of a high-gain outdoor antenna on a 
10-meter mast.  The percentages discussed above apply to such coverage area as long as the receiving 
system (including antenna gain and mast height) remains constant as one moves in to closer distances 
from the broadcast location.  Thus, for example, 29 percent of the receiving coverage area would operate 
with less than 3-dB signal excess for customers with that receiving system.  If, however, closer-in 
customers choose a lower-gain antenna, a lower mast height, or an indoor antenna location, those 
customers will operate with a shorter maximum possible range for consistent reception (R
MAX
).  Such 
customers would also operate with less than 3-dB excess signal whenever they are beyond 84% of the 
maximum reception range possible with that antenna system.  In essence, the region with less than 3-dB 
excess signal repeats itself as 29 percent of the area of a smaller coverage circle corresponding to use of 
the lesser antenna system.  Hence, the area where excess signal is less than 1 or 3 dB will be larger than 
the 11 and 29 percent values discussed here if closer-in customers choose lower-gain antennas or place 
those antennas at less optimal locations such as on a shorter outdoor mast or indoors.  (We note that, 
while the simple model used in the above analysis is useful for estimating the relative importance of 
various values for signal excess, terrain and man-made structures in real-world locations will cause the 
numbers to vary from those shown here.) 
 
Given that rejection performance tests are time consuming, a decision was made to make measurements at 
only one additional signal level beyond the three ATSC-specified levels.  We chose to make this level 
dependent on the threshold reception performance of the individual receivers.  Thus, we measured the 
desired signal level D
MIN
 at which TOV occurs for each receiver in the absence of interference.  
Interference rejection tests were then performed at a desired signal level 3 dB above this threshold.  The 
                                                      
*
 The repeated use of the number 84 is not a typographical error.  This number appears in three contexts in this 
discussion:  D
MIN
 is typically -84 dBm (a value also assumed in the rules); signal excess diminishes to 3 dB at 84 
percent of the maximum reception range; and, 84 percent of the coverage area that can be achieved by a nominal 
receiver (i.e. with D
MIN
 of -84 dBm) will exhibit signal levels at the TV input that are below the ATSC-specified 
“weak” signal level. 
†
 These percentages refer to fractions of the coverage area possible for a given TV receiver with a given antenna 
system and with flat local terrain. 
 
5-5 
desired signal level for these tests is designated as D
MIN
 + 3 dB.   In addition, we extrapolated the 
measurements to a desired signal power of D
MIN
 + 1 dB in Chapter 12 by means of a modeling approach 
developed from a theoretical framework presented in  Chapter 8.
*
 
 
EFFECT OF DESIRED SIGNAL SOURCE 
In preparing for the D
MIN
-referenced measurements, D
MIN
 was measured for three receivers that been 
tested for the SHVERA Study, but the new measurements were found to be about 1 dB higher than the 
earlier SHVERA measurements.
†
  While these differences could have been considered to be within 
measurement error, there was uncertainty about whether the difference could be attributed to differences 
between the SHVERA test setup and the current test setup or the use of different desired signal sources in 
the two test programs. 
 
The signal source used in the SHVERA Study had failed irreparably while being used as an undesired 
signal source for channel-51 tests in the current program; however, a new signal source (Rohde and 
Schwarz SFU) had been procured late in the current test program.  To evaluate the cause of 1-dB 
difference, the new SFU generator was substituted for the Sencore ATSC997 source that was used as the 
desired source for most of the testing described in this report.  The result of the switch was a 1-dB 
improvement in measured sensitivity of the receivers—to results that matched the SHVERA test results. 
 
Table 5-1 shows the observed differences in receiver threshold for various desired-signal configurations, 
with the SFU-based measurements as the reference.  In all cases, the measurements were made at the 
output of the test configuration that was shown in Figure 4-1.  The first measurement column shows the 
differences in measured D
MIN
 using the ATSC997 in the test setup as shown in Figure 4-1 (including 
amplifiers A1 and A2, as well as three 6-dB pads) instead of the SFU.  On average the ATSC997 in the 
“normal” test-setup configuration widely used in this report resulted in a receiver threshold about 0.9 dB 
higher than that when the SFU was used.  The second column represents measurements to determine 
whether the difference was due to the ATSC997 itself, or the amplifiers that followed it.  The third is a 
comparison to the SHVERA measurements. 
 
In general, all three tested TVs appeared to require an input DTV signal level about 0.9 dB higher when 
the signal was supplied by the Sencore ATSC997 than when it was supplied by either the Rohde and 
Schwarz SFU or, based on the SHVERA test results, the Sencore RF Player (playing the 
“Hawaii_ReferenceA” file).  This fact suggested that the signal quality of the ATSC997 was inferior to 
that of the other sources in a way that affected TV performance in a small, but measurable way.  While a 
0.9-dB discrepancy in TV receiver threshold may seem small, there was concern that it might be 
relevant when performing interference testing at levels only 3-dB above the receiver threshold.  
 
Measurements were performed on the desired signal at the output of the test setup in an attempt to 
determine the cause of the generator-dependence of receiver thresholds. 
 
Figures 5-7 and 5-8 show the spectra of the desired signal at the output of the test setup using each of the 
two remaining 8-VSB sources.  Figure 5-7 illustrates the higher noise floor of the ATSC997 relative to 
the SFU.  If the ATSC997’s noise floor extended through the desired signal band at a level roughly 40 dB 
below the desired signal, it would add to the TV receiver noise, which is generally expected to be about 
 
                                                      
*
 The decision to measure at D
MIN
 + 3 dB and extrapolate to D
MIN
 + 1 dB, rather than the other way around, was 
made because measurements become more sensitive to measurement error of the desired signal level and of D
MIN
 as 
the desired signal level approaches D
MIN
. 
†
 Martin, <SHVERA Study>, 2005, Chapter 4. 
 
5-6 
Table 5-1.  Receiver Minimum Signal Level at TOV Versus Desired Signal Source 
D
MIN
 Measured Using Signal Source Shown 
Relative to D
MIN
 Measured Using Rohde & Schwarz SFU (dB) 
TV 
Receiver 
Sencore ATSC997 (with 
amplifiers A1 & A2 as 
shown in Figure 4-1) 
Sencore ATSC997 
(bypassing external 
amplifiers) 
SHVERA Result 
(Sencore RF Player in a 
Different Test Setup) 
I1 0.77 0.87 -0.16 
M1 1.06 1.24 0.05 
N1 0.80 0.74 0.26 
Mean 0.87 0.95 0.05 
 
 
15 dB below the desired signal level at threshold.  This would cause the total noise seen by the receiver to 
rise, relative to receiver-noise-only, by 10 Log(1 + 10
-(40-15)/10
) = 0.014 dB.  Clearly this is not the cause of 
the 0.9-dB change in receiver threshold. 
 
The “flat top” of the ATSC997 signal exhibits a downward slope (visible in Figure 5-8) that projects to 
0.7 dB across a 6-MHz channel width—far greater than the slope of the SFU spectrum, which projects to 
0.04 dB across 6-MHz.  We are unaware of a way to predict the impact of the spectral slope. 
 
The pilot peak of the ATSC997 spectrum is 0.4 dB higher than that of the SFU—probably owing in part 
to the spectral slope. 
 
Table 5-2 shows some additional measurements that were made on the desired signal at the output of the 
test setup for each desired signal source using an Agilent 89441A vector signal generator.  Each 
measurement was an average of at least four successive readings that were obtained using the default 
settings of a control software package designed for DTV measurements.
*
 
 
 
Table 5-2.  Signal Quality Measurements on 8-VSB Sources 
Desired Signal Source--> ATSC997 SFU 
Vector Signal Analyzer Equalizer Setting--> OFF ON OFF ON 
Modulation Error Ratio (MER) (dB) 29.88 32.50 39.10 40.86
Phase Error (deg) 2.43 1.51 1.01 0.92
Pilot Level (dB) 0.59 0.30 0.04 0.08
 
 
The modulation error ratios (MER) were examined to determine whether MER differences could account 
for the observed receiver performance differences.  MER is definitely poorer for the ATSC997 source 
than for the SFU; however, even with the vector signal analyzer’s equalizer turned off, the MER of the 
ATSC997 is still quite respectable.  The ATSC states that the MER of a DTV transmitter should be 
greater than 27 dB in order to limit the impact on receiver threshold to about 0.25 dB.
†
  The even higher 
29.9 dB MER measured from the test setup using the ATSC997 source suggests that the impact on the 
receiver should be well under 0.25 dB.   
                                                      
*
 "Control Software for the HP89400 Vector Signal Analyzer for Measuring DTV and NTSC Signals", VSA5.BAS, 
Version 5.02, by Gary Sgrignoli.  Note that it was necessary to exit the control software to turn the instrument’s 
equalizer ON and OFF.  
†
 Advanced Television Systems Committee, “Transmission Measurement and Compliance for Digital Television”, 
ATSC Standard Doc. A/64 Rev A, 30 May 2000, p.5. 
 
5-7 
 
Eilers and Sgrignoli state that an MER of 27 dB will cause an increase in receiver threshold of 0.28 dB 
and explain that the increase can be computed “by converting the 27-dB transmitter figure of merit value 
and the 15-dB receiver threshold value to equivalent linear relative powers, adding them together, and 
converting back to a logarithmic value resulting in 0.28 dB (= 0.3 dB) of increased noise.”
*
  Applying 
this technique to an MER of 29.88 dB, yields a 0.14-dB predicted increase in receiver threshold—
nowhere near the observed 0.9-dB increase. 
 
The cause of the roughly 1-dB degradation in receiver threshold when the ATSC997 supplies the desired 
signal remains unresolved.  Similar results were obtained on repetition of the measurements—again 
swapping signal sources but changing nothing else in the test setup; this suggests that the difference was 
real rather than a result of measurement error.   
 
Selection of a Desired Signal Source for Tests at D
MIN
 + 3 dB 
Due to concern over possible impacts on interference rejection measurements to be made only 3-dB 
above the receiver threshold, a decision was made to make the “D
MIN
+3 dB” measurements using the SFU 
as the desired signal source.  The use of the SFU as the desired signal source for these measurements 
precluded testing on first-adjacent channel (N+1 and N-1) because neither the ATSC997 nor the 
bandlimited white noise source had adequate out-of-band spectral characteristics to support such testing.  
Hence, these measurements were performed (initially) only for an undesired signal (white Gaussian noise 
source bandlimited to match the 3-dB width of an 8-VSB signal) on channels N-2 through N-16 and N+2 
through N+16. 
 
Receipt of a Wavetech WS2100 RF Player shortly before the delivery deadline for this report enabled 
first-adjacent channel testing with a high-quality desired signal source at D
MIN
+3 dB.  The primary 
purpose of this player is playback of digitally-recorded RF captures of broadcast DTV signals received on 
TV antennas for evaluation of the multipath performance of DTV demodulators; however, the system is 
capable of performing as an unimpaired 8-VSB signal source if equipped with a recording of a high 
quality DTV signal.  The Digital Television group of Advanced Micro Devices (AMD) created such a 
“recording” by mathematically deriving an 8-VSB signal from an MPEG2 transport stream.  They 
provided this recording to us through Wavetech in the form of a file called “Muddy Waters”.  The 
WS2100, equipped with this file, was combined with an external upconverter to act as a desired signal 
source at a level of D
MIN
 + 3 dB for tests with the SFU as an undesired signal source on channels N-1 and 
N+1; those results are presented in the next section of this chapter. 
 
In tests of the three receivers indicated in Table 5-1 using the Wavetech and the SFU, alternately, as 
desired signal sources, D
MIN
 values obtained with the two sources matched within 0.1 dB for each TV, 
with an average difference of 0.0 dB.
†
  Thus, the TV receivers were unable to distinguish the signal 
quality of the Wavetech source from that of the SFU. 
 
TESTS WITH DESIRED SIGNAL AT D
MIN
 + 3 DB 
The results of the interference rejection measurements with the desired signal at D
MIN
 + 3 dB are shown in 
Figures 5-9 and 5-10.  Though the ATSC provides no performance guidelines for interference rejection 
with D = D
MIN
 + 3 dB, the performance guideline for D = -68 dBm is included on the plots as a reference.  
Note that the vertical scale of the plots was extended 5 dB lower than the earlier D/U plots, because the 
                                                      
*
 Carl Eilers and Gary Sgrignoli, “Digital Television Transmission Parameters—Analysis and Discussion”, IEEE 
Transactions on Broadcasting, Vol. 45, No. 4, Dec 1999, p.368. 
†
 Both the Wavetech and SFU results were an average of 0.2 dB higher than the four-months-earlier SFU results 
from Table 5-1. 
 
5-8 
measurement limit of the test setup extends lower at D
MIN
 + 3 dB.    The measurement limitations at D = 
D
MIN
 + 3 dB are based on leakage from the undesired source into the desired channel. 
 
Qualitatively, the interference behavior at the lower desired signal level is similar to that at -68 dBm, with 
at least one notable exception; the N+7 interference susceptibility that appeared so significant at -68 dBm 
is quite low at D
MIN
 + 3 dB. 
 
We also note that there was another of several discrepancies involving measurements of receiver G4.  
When D
MIN
 was measured for the N-1 and N+1 tests, the result was 5.5 dB higher than the two 
measurements performed three months earlier for the N+2 through N+16 and N-2 through N-16 tests.  A 
repeat of the tests the next morning produced a match to the older results.  Though receiver G8 was the 
best performing receiver among the eight in terms of interference rejection performance, measurements 
involving G8 were consistently inconsistent (a topic discussed further in Chapter 7). 
 
SUMMARY OF FIFTH-GENERATION RECEIVER RESULTS 
Figure 5-11 combines the measurements for each of the four desired signal levels—showing the median 
D/U ratio across the eight receivers at each channel offset and signal level.  The three upper solid black 
lines correspond to the measurement limits applicable to the three upper curves on the plot—for D = 
-28 dBm, -53 dBm, and -68 dBm.  Points falling on the respective black curves are at the measurement 
limit.  The black line at the top of the shaded region is the measurement limit corresponding the 
measurements at D = D
MIN
 + 3 dB; it is based on leakage from the undesired source into the desired 
channel. 
 
Figure 5-12 is similar to Figure 5-11, except the plot is in terms of undesired signal level (U) at the 
threshold.  This portrayal differs in two significant ways from the D/U plots: 
(1) Better performance is indicated by higher points on the plot rather than by lower ones; 
(2) It makes clear that the greatest susceptibility to interference occurs at low desired signal levels—even 
though D/U ratio is often lowest (i.e., best) at low desired signal levels. 
 
We note that the measurement limitation shown in Figure 5-12 is the maximum undesired signal power 
that the test setup could inject into the receiver.  The N-1 and N+1 offsets for D = -68 dBm and all of the 
offsets for D = D
MIN
 + 3 dB are subject to an additional limitation, shown only in the D/U plots, based on 
leakage of the undesired signal into the desired channel. 
 
Based on the D/U plots one might conclude that interference at N+7 is a problem area; however, in terms 
of absolute levels of undesired signals that can cause interference there are 14 other channel spacings that 
are more vulnerable (N-7 through N+5, N+14, and N+15) to interference.  
 
Figures 5-13 and 5-14 are repeats of the previous two plots except that they show the second-worst 
performance among the eight receivers, rather than the median.  Figure 5-15 and 5-16 show the worst 
performance among the eight receivers. 
 
Chapter 13 includes plots of these measurement results for the individual receivers along with 
extrapolations to D = D
MIN
 + 1 dB.  Chapter 15 presents plots of the measurements and extrapolations for 
median, second-worst, and worst performance among the receivers.  Appendix A includes tabulations of 
some of the data. 
 
 
5-9 
EARLIER-GENERATION RECEIVERS 
A limited set of interference rejection tests was also performed on two earlier-generation receivers for 
purposes of identifying tuner type, as described in Chapter 3.  The receivers are designated G3 and P1.   
These two DTVs would be classified as fourth-generation or earlier based on their multipath performance, 
which was tested as part of the SHVERA Study.  Both were on the market in 2005, though P1 was 
actually introduced to the market in 2004.   
 
Data from those measurements was plotted in Figure 3-7, but Figure 5-17 shows the same measurements 
overlaid with the median and range of measurements of the eight fifth-generation receivers, for 
comparison.  All measurements were made with a desired signal power of -68 dBm; the third/fourth 
generation measurements were limited to channel offsets of N+2 through N+16.   
 
It can be seen that the measurements on receiver P1 fall within the range of performance for the fifth-
generation receivers except at two points:  N+2, where the results for P1 are slightly better than any of the 
fifth-generation results, and N+10, where the results for P1 are worse than the fifth-generation results but 
are still within the ATSC A/74 guidelines.   
 
On the other hand, most of the measurements on receiver G3 correspond to worse performance than any 
of the tested fifth-generation receivers, and most fall well outside of the ATSC A/74 guidelines.  Given 
the small number of receivers tested, it is impossible to say whether similarly poor performance is 
common among earlier-generation receivers. 
 
TABOO EFFECTS AND OTHER OBSERVATIONS 
Some of the interference susceptibilities that can be seen in the plots were expected based on previous 
knowledge of the analog UHF taboos.
*
  For example, in Figures 5-1 and 5-9, an increased susceptibility to 
interference can be seen at channels N+14 and N+15 in seven of the eight fifth-generation receivers; a 
similar peak is seen for one of the two earlier generation receivers in Figure 5-17.  The mixer image 
frequency band,
†
 which is centered 88 MHz (14 
2
/
3
 channels) above the center of the desired channel, 
straddles channels N+14 and N+15—causing the increased susceptibility to interference from undesired 
signals on those channels.  
 
Similarly, a peak in susceptibility can be seen at N+7 in nine of the ten receivers.  The analog taboo at 
N+7 exists primarily because of the possibility that unintentional local oscillator radiation from one TV 
receiver might interfere with reception on another TV tuned seven channels higher.  Since all tests for this 
report were performed with only one TV turned on at a time, local oscillator radiation was not a factor.   
 
Another potential explanation for the N+7 peak was the possibility that an undesired signal on channel 
N+7 could beat with the desired signal—creating interference that could pass through the IF filter of the 
receiver.  This “IF beat” effect
‡
 has previously been recognized with analog TV receivers and was 
initially thought to be the explanation of the N+7 peak observed here—both because of its location (at 
N+7) and the fact that the susceptibility threshold exhibits more of a constant undesired signal level rather 
than a constant D/U ratio (see, for example, Figures 5-11 and 5-12).  The mixing product between the 
desired signal and an undesired signal would be expected to have a power level that is directly 
                                                      
*
 Taboos are channel spacings at which analog TV reception is vulnerable to interference.  The taboos limited the 
allotment of local analog channel assignments. 
†
 Mixer image for a single-conversion tuner occurs at a separation of twice the intermediate frequency (IF) above the 
center of the desired channel.  The IF for single-conversion TV receivers is 44 MHz. 
‡
 Young-Jun Chong, and others, “The design and implementation of TV tuner for digital terrestrial broadcasting.” 
Consumer Electronics, 2000. ICCE. 2000 Digest of Technical Papers. International Conference on Consumer 
Electronics (ICCE), 2000 Digest of Technical Papers, 13-15 June 2000, p. 40 – 41. 
 
5-10 
proportional to both the desired signal power and the undesired signal power.  The resulting interference 
passing through the receiver’s IF filter would thus increase in direct proportion to the desired signal 
power, so that once the undesired signal reaches a sufficient level to result in the interference coming 
within about 15 dB of the desired signal power, reception would no longer be possible; attempts to re-
establish reception by increasing the desired signal power would result in corresponding increases in the 
interference power. 
 
Tests results presented in Chapter 7 contradict the IF-beat explanation for the N+7 peak—suggesting 
instead some sort of direct interaction between the TV’s local oscillator and the undesired signal.  The 
absence of an N-7 peak also runs counter to the IF-beat explanation. 
 
Another analog taboo is associated with the “half IF” beat response.  This is the result of the second 
harmonic of an undesired signal, spaced above the desired channel by half the IF frequency, beating with 
the second harmonic of the TV’s local oscillator.  The result is an increased susceptibility to interference 
spaced 22 MHz (3 
2
/
3
 channels) above the center of the desired channel.  The susceptibility peak from this 
effect would be expected to be seen primarily at channel N+4.  Peaks at N+4 were observed on only two 
receivers:  A3 (Figure 5-3) and G3 (Figure 5-17). 
 
An unexpected sensitivity occurred at N-6 on some receivers.  Receiver J1 exhibits a large peak at N-6, as 
seen in Figure 5-1.  A smaller peak can be seen for receiver N1 in the same graph, and a third receiver 
(A3) exhibits a slight susceptibility peak in Figure 5-3.  The cause of this sensitivity is not known. 
 
Also unexpected is that two of the receivers (J1 and N1) are significantly more susceptible to interference 
on the second-adjacent channels (N-2 and N+2) than on the first-adjacent ones (N-1 and N+1), and one 
more receiver (O1) exhibits such behavior on the positive side (i.e., N+2 only).  (See Figure 5-1, for 
example.)  In fact, at low signal levels (D
MIN
 + 3 dB), the second adjacent channels are as susceptible to 
interference as the first adjacent channels even when viewed on a median basis across all eight TVs.  (See 
the lower curve on Figure 5-12, for example.) 
 
Some broader band effects can also been seen.  For example, in Figure 5-3, the susceptibility of receiver 
D3 to interference is seen to smoothly decrease from N-4 to N-16.  The absence of peaks at specific 
channels suggests that this is a cross-modulation effect.  Cross-modulation creates an interference power 
that is directly proportional to the desired signal power.  As a result, it is expected to exhibit a fixed 
undesired signal level threshold that does not change as the desired signal decreases (until desired signal 
approaches D
MIN
).  In Chapter 13 (Figure 13-4), we see that this is indeed the case at desired signal levels 
of -53 dBm and below.  The smooth decrease in the cross-modulation effect as the desired signal moves 
from N-4 to N-16 is likely to be a result of rolloff caused by the RF tracking filter in the receiver. 
 
This apparent cross-modulation effect on receiver D3 exhibits some rather unexpected behavior as the 
undesired signal moves from channel N-4 to N-3.  With a desired signal power of -53 dBm, the 
interference susceptibility, which had been rising smoothly as the undesired signal was moved closer to 
the desired channel, suddenly drops precipitously at channel N-3 by 20 dB or more
*
 and remains lower at 
channels N-2, N-1, N+1, and N+2 than it had been at N-4.  We believe that the reduction in susceptibility 
is due to the influence of the undesired signal on the receiver’s automatic gain control (AGC)—a topic 
that will be discussed further in Chapters 8 and 11. 
 
Though the “hardness” of the thresholds was not one of the measurement parameters for this study, we 
noted a strong “cliff effect” in most of the threshold measurements.  For example, in most cases, 
increasing interference level about 1 dB above the threshold of visibility (TOV)—the point at which 
picture degradation becomes perceptible—caused complete loss of picture.  In some cases picture loss 
didn’t occur until the undesired signal level rose as much as much as 3 dB and in one case, 5 dB (though 
                                                      
*
 The susceptibility drops below the measurement floor of the test setup. 
 
5-11 
picture errors occurred continuously in that case after only a 1.5 dB increase).  In a few cases, picture loss 
occurred concurrently with appearance of errors or with only an additional 0.1 dB increase in 
interference—an extremely abrupt cliff! 
 
5-12 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
S
u
sc
epti
bil
i
t
y to 
I
n
ter
f
er
e
n
c
e
 
[D
/
U
 at
 Th
res
h
ol
d (d
B
)
]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
G4
A/74 Limit
 
Figure 5-1.  D/U of 8 Receivers at D = -68 dBm on Channel 30 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
S
u
sc
epti
b
i
l
i
ty to
 
I
n
ter
f
er
ence 
[D
/
U
 at Thres
hol
d 
(dB
)
]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 5-2.  D/U Statistics at D = -68 dBm on Channel 30 
 
5-13 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
S
u
sc
epti
bil
i
t
y to 
I
n
ter
f
er
e
n
c
e
 
[D
/
U
 at
 Th
res
h
ol
d (d
B
)
]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
G4
A/74 Limit
 
Figure 5-3.  D/U of 8 receivers at D = -53 dBm on Channel 30 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
S
u
scepti
b
i
l
i
ty to Inter
f
er
e
n
ce
 
[D
/
U
 at Thresh
old (dB
)
]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 5-4.  D/U Statistics at D = -53 dBm on Channel 30 
 
5-14 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
S
u
sc
epti
bil
i
t
y to 
I
n
ter
f
er
e
n
c
e
 
[D
/
U
 at
 Th
res
h
ol
d (d
B
)
]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
G4
A/74 Limit
 
Figure 5-5.  D/U of 8 receivers at D = -28 dBm on Channel 30 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Su
s
c
e
p
t
i
b
ility
 to
 In
te
rfe
r
en
c
e
 
[D/U at 
Th
re
sh
old
 (d
B)
]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 5-6.  D/U Statistics at D = -28 dBm on Channel 30 
 
5-15 
-70
-60
-50
-40
-30
-20
-10
0
10
20
565 566 567 568 569 570 571 572 573
Frequency (MHz)
No
r
m
al
ized
 Sp
e
c
t
r
u
m
 L
e
v
e
l
 
(
d
B in
 
1
0
-
k
Hz 
RBW)
ATSC997
SFU
 
Figure 5-7.  Spectra of Two Desired Signal Sources 
-2
-1
0
1
565 566 567 568 569 570 571 572 573
Frequency (MHz)
N
o
r
m
a
l
i
z
e
d
 S
p
ectr
u
m
 
Le
ve
l
 (dB
 i
n
 
10-k
H
z
 
RB
W
)
ATSC997
SFU
 
Figure 5-8.  Spectra of Two Desired Signal Sources on Expanded Scale 
 
5-16 
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susceptibil
i
ty to
 Interference 
[D/
U
 at 
Thres
hold (d
B)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
G4
A/74 Limit (for D = -68 dBm)
 
Figure 5-9.  D/U of 8 receivers at D = D
MIN
 + 3 dB  on Channel 30 
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Su
s
c
e
p
t
i
b
ility
 to
 In
te
rfe
r
en
c
e
 
[D/U at 
Th
re
sh
old
 (d
B)
]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 5-10.  D/U Statistics at D = D
MIN
 + 3 dB on Channel 30 
 
5-17 
 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susce
ptibility
 to Inte
rfere
n
ce
 
[D/U
 at 
Thr
e
shold (
d
B)
]
Meas. Limit at
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Meas. Limits
MLiit
Median of 8 DTV Receivers
 
Figure 5-11.  Median D/U of 8 receivers at Four Signal Levels on Channel 30 
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dBMedian of 8 DTV Receivers
 
Figure 5-12.  Median Threshold U of 8 receivers at Four Signal Levels on Channel 30 
 
5-18 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susc
eptib
ility to 
Interference
 
[D
/U a
t
 Thre
s
h
old
 (dB)]
Meas. Limit at
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Meas. Limits
MLiit
2nd Worst of 8 DTV Receivers
 
Figure 5-13.  2
nd
 Worst D/U of 8 receivers at Four Signal Levels on Channel 30 
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB2nd Worst of 8 DTV Receivers
 
Figure 5-14.  2
nd
 Worst Threshold U of 8 receivers at Four Signal Levels on Channel 30 
 
5-19 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
c
e
p
tibility
 to Inte
r
f
er
e
n
ce
 
[
D
/U 
at Th
r
e
shold (
d
B)
]
Meas. Limit at
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Meas. Limits
MLiit
Worst of 8 DTV Receivers
 
Figure 5-15.  Worst D/U of 8 receivers at Four Signal Levels on Channel 30 
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dBWorst of 8 DTV Receivers
 
Figure 5-16.   Worst Threshold U of 8 receivers at Four Signal Levels on Channel 30 
 
5-20 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
+
1
N+
2
N+
3
N+
4
N+
5
N+
6
N+
7
N+
8
N+
9
N+
1
0
N+
1
1
N+
1
2
N+
1
3
N+
1
4
N+
1
5
N+
1
6
Interfering Channel
S
u
sceptib
ility to
 In
ter
f
er
en
ce 
[
D
/
U
 
a
t
 Thre
s
h
ol
d (dB
)
]
Measurement Limit
G3
P1
ATSC A/74 Recommendation
Median (w/Best & Worst)
 
Figure 5-17.  D/U of Two Earlier-Generation Receivers at D = -68 dBm Relative to Fifth-Generation 
Results 
 
 
6-1 
CHAPTER 6 
REJECTION RESULTS ON CHANNEL 51 FOR SINGLE 8-VSB 
INTERFERERS 
This chapter presents the results of interference rejection tests of seven DTV receivers tuned to channel 
51.  The interferer for these tests was an 8-VSB signal placed sequentially on channels N+1 through 
N+16.  The tested receivers comprise seven of the eight receivers that were tested for Chapter 5.  Tests on 
one of the seven receivers were not completed due to equipment failure.  (Note that presentation of 
detailed test results showing the variation of D/U with desired signal level for one TV receiver at channel 
51 are postponed until Chapter 11.) 
 
The channel-51 tests for this chapter occurred chronologically before those on channel-30, described in 
Chapter 5.  The primary focus of these tests was to investigate interference susceptibility of DTV 
receivers operating in the upper UHF core channels to interference from future services to be offered in 
spectrum currently assigned to television channels 52 through 67.  Because of this focus, tests were 
limited to undesired signals above the desired channel. 
 
Notwithstanding the above focus, an 8-VSB source was used as the undesired signal for all of these tests 
(unlike the tests in Chapter 5, which used an 8-VSB source only at N-1 and N+1 and used a bandlimited 
Gaussian noise source at all other channel offsets).
*
  That choice was based on the steep rolloff of the 8-
VSB source at its channel edges and the lack of availability of surrogate devices representative of signals 
planned for use in channels 52 through 67. 
 
TESTS AT ATSC-SPECIFIED DESIRED SIGNAL LEVELS 
“Weak” Desired Signal (D = -68 dBm) 
Figure 6-1 shows measured values of D/U ratios at TOV for seven DTV receivers for undesired signal 
channels ranging from N+1 to N+16 with the desired signal power set to -68 dBm, the signal level that 
the ATSC designates as “weak”.  The configuration of the plots is similar to those in Chapter 5, except for 
the single-sided range of channel offsets, so the reader is referred to the corresponding section of that 
chapter for further explanation of the format. 
 
The ATSC-recommended DTV-into-DTV interference rejection thresholds are shown on the plot as a 
reference.  Compliance with those voluntary limits would be indicated by all points on each measurement 
curve falling on or below the ATSC line.  For the desired signal level of -68 dBm, only one of the 
receivers (M1) fully complies with the guidelines.  Receiver D3 fails to comply at seven of the sixteen 
tested channel offsets.  The other receivers fail to comply at from one to three of the channel offsets.  The 
N+7 channel offset seems to offer the most challenge in that five of the receivers failed to comply with 
the guidelines at that offset.  We note that these measurements are directly comparable to the ATSC 
performance guidelines since they were made with an 8-VSB signal as the undesired signal. 
 
The measurements on the seven receivers for the first adjacent channel (N+1) are closely clustered, and 
all satisfy the ATSC guidelines.  The second-adjacent channel results (N+2) are much more scattered—
with three of the receivers failing to meet the ATSC guidelines and actually exhibiting more susceptibility 
to interference at N+2 than at N+1. 
                                                      
*
 Failure of the 8-VSB source that had served as the undesired signal for these tests forced the use of the bandlimited 
Gaussian noise source for subsequent tests on channel 30. 
 
6-2 
 
Figure 6-2 shows the best, median, second worst, and worst performance at each channel offset.  On a 
median basis the only failure to satisfy the ATSC recommended performance is at N+7. 
 
“Moderate” Desired Signal (D = -53 dBm) 
Figure 6-3 shows measured values of D/U ratios at TOV for the same seven DTV receivers with the 
desired signal power set to -53 dBm, the signal level that the ATSC designates as “moderate”.  None of 
the receivers comply with the performance guidelines at all channel offsets.  Again, N+7 appears to be the 
most challenging channel offset in that six of the seven receivers failed to comply at that offset.  Receiver 
D3 was, again, the poorest performer—failing to comply with the guidelines at five of the sixteen tested 
channel offsets.  The other receivers failed to comply at one or two channel offsets. 
 
Figure 6-4 shows the best, median, second worst, and worst performance at each channel offset.  On a 
median basis the only failure to satisfy the ATSC recommended performance is at N+7. 
 
“Strong” Desired Signal (D = -28 dBm) 
Figure 6-5 shows that, with the desired signal set to the level that the ATSC designates as “strong”, every 
receiver complied with the ATSC guidelines.  Since the measurement limit was close to the ATSC 
guidelines, almost all of the measurements fell at the measurement limit (i.e., the test setup was not 
capable of generating a strong enough undesired signal to cause interference.) 
 
Figure 6-6 shows the corresponding statistical data. 
COMPARISON TO CHANNEL-30 RESULTS 
Figures 6-7 and 6-8 compare the median D/U measurements on channel 30 with those on channel 51 for 
desired signal levels of -68 dBm and -53 dBm, respectively, for the seven DTV receivers that were tested 
on channel 51.  (Note that the medians for channel 30 shown in these two charts are for only seven 
receivers—in order to match the channel-51 test group.  The channel-30 data in Chapter 5 includes an 
eight receiver.)  The largest differences among points that were measurable were 5 dB at N+15 (mixer 
image frequency) and 4 dB at N+2; both of those measurements indicate better performance at channel 51 
than at channel 30 (perhaps surprisingly).  For all other measurements, the match between channels was 
within about 2 dB or less. 
 
 
6-3 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+
1
N
+
2
N+
3
N+
4
N+
5
N+
6
N+
7
N+
8
N+
9
N+
1
0
N+
1
1
N+
1
2
N+
1
3
N+
1
4
N+
1
5
N+
1
6
Interfering Channel
Susceptib
ilit
y to
 Int
e
r
f
erence 
[D
/U
 at Thr
e
shold 
(dB)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
A/74 Limit
 
Figure 6-1.  D/U of 7 Receivers at D = -68 dBm on Channel 51 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
S
u
sc
ep
t
i
bil
i
t
y
 t
o
 I
n
t
e
r
f
er
en
ce
 
[D
/U
 at T
h
res
h
old 
(dB)]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 6-2.  D/U Statistics at D = -68 dBm on Channel 51 
 
6-4 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+
1
N
+
2
N+
3
N+
4
N+
5
N+
6
N+
7
N+
8
N+
9
N+
1
0
N+
1
1
N+
1
2
N+
1
3
N+
1
4
N+
1
5
N+
1
6
Interfering Channel
Susceptib
ilit
y to
 Int
e
r
f
erence 
[D
/U
 at Thr
e
shold 
(dB)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
A/74 Limit
 
Figure 6-3.  D/U of 7 Receivers at D = -53 dBm on Channel 51 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
S
u
sc
ep
t
i
bil
i
t
y
 t
o
 I
n
t
e
r
f
er
en
ce
 
[D
/U
 at T
h
res
h
old 
(dB)]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 6-4.  D/U Statistics at D = -53 dBm on Channel 51 
 
6-5 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+
1
N
+
2
N+
3
N+
4
N+
5
N+
6
N+
7
N+
8
N+
9
N+
1
0
N+
1
1
N+
1
2
N+
1
3
N+
1
4
N+
1
5
N+
1
6
Interfering Channel
Susceptib
ilit
y to
 Int
e
r
f
erence 
[D
/U
 at Thr
e
shold 
(dB)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
A/74 Limit
 
Figure 6-5.  D/U of 7 Receivers at D = -28 dBm on Channel 51 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
S
u
sc
ep
t
i
bil
i
t
y
 t
o
 I
n
t
e
r
f
er
en
ce
 
[D
/U
 at T
h
res
h
old 
(dB)]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 6-6.  D/U Statistics at D = -28 dBm on Channel 51 
 
6-6 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
cepti
bility
 to Inte
rference
 
[D/
U
 at
 Thresho
l
d 
(dB)]
Measurement Limit
ATSC A/74 Recommendation
Channel 51 Median
Channel 30 Median
 
Figure 6-7.  Comparison of Median D/U for 7 Receivers on Channels 30 and 51 at D = -68 dBm 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
cepti
bility
 to Inte
rference
 
[D/
U
 at
 Thresho
l
d 
(dB)]
Measurement Limit
ATSC A/74 Recommendation
Channel 51 Median
Channel 30 Median
 
Figure 6-8.  Comparison of Median D/U for 7 Receivers on Channels 30 and 51 at D = -53 dBm 
 
7-1 
CHAPTER 7 
TESTS WITH DIFFERENT SIGNAL TYPES AND SOURCES 
This chapter presents the results of tests performed to determine the influence of test signal sources and 
signal types on the interference rejection performance of DTV receivers.  Specifically, the tests compare 
the following: 
• Effect of two different 8-VSB signal generators as the desired signal; 
• Effect of several types of undesired signals, including 
? White Gaussian noise bandlimited to the 3-dB width of an 8-VSB source; 
? 8-VSB; 
? DVB-H—an orthogonal frequency division multiplexing (OFDM) signal—set for a 5-MHz 
channel width; and 
? White Gaussian noise bandlimited to a 3-dB width of 1 MHz. 
In addition, the tests include a repeat of the baseline test conditions (described below) to determine the 
amount of variation that may have been due to repeatability or equipment changes (spectrum analyzer).  
All of the tests discussed in this section were performed with the desired signal on channel 30 and set to a 
level of -68 dBm. 
 
Spectra and bandwidths of the various sources can be found in Chapters 2 and 5.  Figures 2-1 and 2-2 
show spectra of the four undesired signals; bandwidths characteristics are shown in Table 2-1.  
Figures 5-7 and 5-8 show spectra of the two desired signal sources. 
 
The baseline for comparison is: 
• Desired signal source—Sencore ATSC997; 
• Undesired signal type—white Gaussian noise (from an Agilent E4437B vector signal generator) 
bandlimited to match the 3-dB width of an 8-VSB signal. 
 
The 1-MHz bandwidth tests were performed on one DTV receiver (N1) for channel offsets from N-16 to 
N+16.  All other tests were performed on all eight DTV receivers, but were limited to the five non-
adjacent channel offsets that exhibited the most interference potential among the receivers at low signal 
levels (Figure 5-16):  N+2, N-2, N-3, N-4, and N-6.  (Note that channels N-1 and N+1 were not tested 
because the Gaussian source did not have adequate band-edge rolloff to permit testing on first-adjacent 
channels.) 
 
The baseline tests, a subset of the measurements presented in Chapter 5, were performed between 
August 30 and October 23, 2006.  The comparative tests presented here, including the “repeat baseline” 
test, were performed between January 31 and February 6, 2007.
*
 
 
The results for all tests except the 1-MHz bandwidth tests are summarized in Table 7-1.  Individual results 
are presented in each section.  Results for the 1 MHz tests are in the next section of this chapter. 
 
Table 7-1 shows the means and standard deviations for the differences in D/U measurements between a 
“comparative test” configuration and the baseline configuration.  The statistics are based on 35 measured 
values (7 receivers X 5 channel offsets), except in the case of the DVB-H signal, where 34 measurements 
were used for reasons explained in the DVB-H section below.  The standard deviations shown are for the 
individual differences; standard deviation of the mean would be reduced from this by a factor of square-
                                                      
*
 Dates are presented to address issues of whether observed differences were caused by possible changes in test 
setup performance or equipment.  (A different spectrum analyzer was used for the comparative signal tests than for 
the baseline tests.)  Though the baseline tests were performed about four months before the comparative tests, the 
“repeat baseline” test was performed as part of the comparative tests and was performed after the “SFU as D” test. 
 
7-2 
root of 34 or 35 (5.8 or 5.9) if the individual differences can be considered independent.  Results are 
shown in hundredths of a dB to reduce round-off error, although D/U measurement resolution was 0.1 dB 
since the measurements were made by stepping the undesired signal in increments of 0.1 dB to locate the 
receiver’s TOV. 
 
 
Table 7-1.  Comparative Signal Test Summary 
  Signals 
D/U Ratio Relative to 
Baseline (dB) 
Test Case Test Dates 
Desired Signal 
Source 
Undesired 
Signal Mean 
Standard 
Deviation 
Baseline 
8/30/2006 – 
10/23/2006 
ATSC997 WGN 0 N/A 
Repeat Baseline 
2/02/2007 – 
2/05/2007 
ATSC997 WGN -0.12 0.33 
SFU as D 
1/31/2007 – 
2/1/2007 
SFU WGN -1.14 0.40 
8-VSB as U 
2/06/2007 – 
2/07/2007 
ATSC997 8-VSB -1.28 0.68 
DVB-H OFDM 
as U 
2/05/2007 – 
2/06/2007 
ATSC997 DVB-H 
-0.40 
(-0.25*) 
1.09 
(0.66*) 
Notes: 
• WGN = White Gaussian noise bandlimited to 5.38 MHz 3-dB width. 
• The means and standard deviations were across five channel offsets and seven receivers—a total of 
35 measurements, except results marked “*” are for 34 measurements.  Measurement results for 
receiver G4 were not included in these statistics for reasons discussed in the section of this chapter 
entitled, “Desired Signal Source:  SFU Versus ATSC997” 
 
 
The “repeat baseline” test results differed from the baseline test results by an average of only 0.12 dB—a 
reassuringly small difference given the four-month delay and spectrum analyzer change between the 
baseline and “repeat baseline”.  Though the results in the table are presented relative to the original 
baseline tests, we can use the newer “repeat baseline” as the point of reference by subtracting -0.12 dB 
from the means.  Doing so, we find the following. 
• Use of the SFU as the desired signal source resulted in D/U ratios that were 1.0 dB lower (better) than 
were achieved with the ATSC997 (the source used for most of the tests presented in this report).
*
  
Thus the DTV receivers could tolerate about 1.0 dB more interference when the SFU was the desired 
signal source.  This suggests that the ATSC997 may have had a slightly degraded performance that 
affected the results, at topic that was discussed in Chapter 5. 
• An 8-VSB interferer results in D/U ratios that are 1.2 dB lower than those measured with a 
bandlimited white Gaussian noise interferer.  Thus, the TV’s are 1.2 dB less susceptible to out-of-
band interference from an 8-VSB DTV signal than from a Gaussian-noise signal of comparable 
bandwidth. 
• The OFDM DVB-H signal causes an interference effect comparable to that of bandlimited Gaussian 
noise based on the mean listed in parentheses in the table; this mean omits one measurement (in 
addition to those of receiver G4) for reasons discussed in the DVB-H section of this chapter. 
 
                                                      
*
 The ATSC997 was the desired signal source for all tests except as follows.  The SFU was a desired source for: 
(1) some tests to identify effect of the desired signal source (here and in Chapter 5, “Effect of Desired Signal 
Source”), and (2) tests with D = D
MIN
 + 3 dB with non-adjacent interference.  The Wavetech WS-2100 with an 
external Drake upconverter served as the desired signal source for tests with D = D
MIN
 + 3 dB with the undesired 
signal at N-1 or N+1. 
 
7-3 
The results of tests with the reduced-bandwidth interferer are presented in the next section followed by 
detailed results of the tests described above. 
 
1-MHZ BANDWIDTH UNDESIRED SIGNAL 
Most tests performed for this report involved undesired signals that filled, or nearly filled, a 6-MHz wide 
TV channel.  Interference rejection tests were also performed on one DTV receiver (receiver N1) using a 
reduced-bandwidth source to determine the effect of a narrower undesired signal spectrum.  Specifically, 
a Gaussian noise signal with a 1-MHz 3-dB width was created using the same vector signal generator that 
was used to generate 5.38-MHz wide Gaussian noise signal.  Spectral plots were shown in Figures 2-1 
and 2-2. 
 
The interference rejection measurements were performed with the 1-MHz bandwidth undesired signal 
centered on each channel from N-16 to N+16, except for channel N.  In addition, where the 8-VSB-width 
source had identified interference vulnerabilities, tests were performed with the 1-MHz wide source 
stepped in 1-MHz or 2-MHz steps to look for finer frequency dependence of the vulnerability.  The 
desired signal was set to -68 dBm. 
 
Figure 7-1 shows the threshold D/U ratio measurements for the 1-MHz wide Gaussian noise source along 
with the results (from Chapter 5) for the 5.38-MHz-wide baseline source.  (The baseline source was a 
Gaussian noise source bandlimited to match the 3-dB width of an 8-VSB signal, except that for N-1 and 
N+1, an actual 8-VSB source was used in order to achieve band-edge characteristics adequate for 
adjacent channel testing.)  Though measurements were made from N-16 to N+16, the plot is limited to a 
range over which the 1-MHz width tests yielded D/U ratios that were above the measurement limit; 
consequently, the plot extends from channel N-9 through the center point of channel N+16.  Vertical 
gridlines correspond to the boundaries between TV channels. 
N+7 Interference 
In general the two plots in Figure 7-1 track each other reasonably well except where narrowband 
vulnerabilities are apparent.  One point of interest is the interference susceptibility peak at N+7.  The 
1-MHz data shows high susceptibility when the undesired signal is centered at 44 MHz, but no 
measurable interference susceptibility at 43 or 45 MHz (or elsewhere within channel N+7).  Based on the 
bandwidth ratio of the two signal sources (5.38 MHz / 1 MHz), one would expect the 1-MHz wide signal 
to have a 7.3 dB greater power spectral density than the 5.38 MHz wide signal used for the baseline tests 
when the power levels are the same.  This closely matches the 6.7 dB difference in D/U peaks in N+7 for 
the two signal bandwidths. 
 
It appears that N+7 interference is seen only when the undesired signal overlaps the receiver’s local-
oscillator frequency, 44 MHz above the center of the channel to which the DTV receiver is tuned.  Our 
initial hypothesis for the cause of the N+7 peak in the previous D/U plots was that the undesired signal 
was acting as a noise-like local oscillator—beating with the desired signal and creating a mixer product 
that falls within the IF frequency of the TV.  Such a mechanism would not result in the narrow sensitivity 
spike observed here.  Rather, the interference mechanism apparently involves interaction between the 
DTV receiver’s local oscillator and the incoming undesired signal when the spectrum of the incoming 
signal overlaps the local oscillator frequency. 
DESIRED SIGNAL SOURCE:  SFU VERSUS ATSC997 
Test results described in Chapter 5, in the section entitled “Effect of Desired Signal Source”, demonstrate 
that, with no undesired signals present, DTV receivers could operate at a lower desired signal level when 
the desired signal was supplied by a newer, higher-quality 8-VSB signal source (the Rohde and Schwarz 
 
7-4 
SFU acquired by the FCC late in this measurement program) than when it was supplied by the Sencore 
ATSC997, the source that supplied the desired signal for most of the measurements presented in this 
report (and the only 8-VSB signal source available at the Laboratory during most of the channel-30 test 
period).  That section of Chapter 5 also presents the results of signal quality tests, such as modulation 
error ratio, on each of the two sources. 
 
The tests presented here look at the effect of the desired signal source on susceptibility to interference.  
Specifically, interference rejection measurements were performed with a desired signal level of -68 dBm 
on channel 30 supplied by each of the two signal generators. 
 
Table 7-2 shows the D/U ratios measured with the SFU as the desired signal source relative to the 
baseline measurements made with the ATSC997 as the source.  The table also includes means and 
standard deviations calculated across channel offsets (statistics computed on five measurements) for each 
DTV receiver, across DTV receivers except for receiver G5 (statistics computed on seven measurements), 
and across both the channel offsets and receivers (statistics computed on 35 measurements). 
 
 
Table 7-2.  D/U Ratio With SFU as Desired Signal Source Relative to That With Baseline Generator 
(ATSC997) 
 
D/U Ratio Relative to Baseline (dB) 
Channel 
Offset 
A3 D3 I1 J1 M1 N1 O1 G4 
Mean 
(dB) 
(Excluding 
G4) 
Standard 
Deviation 
(dB) 
(Excluding 
G4) 
N-6 -1.37 -0.01 -1.54 -1.42 -1.39 -1.96 -1.29 -6.29 -1.28 0.60 
N-4 -1.15 -0.11 -1.00 -1.44 -1.16 -1.51 -1.03 -5.16 -1.06 0.46 
N-3 -1.03 -0.88 -0.98 -0.57 -1.10 -1.50 -1.06 -5.70 -1.02 0.28 
N-2 -1.21 -0.56 -0.96 -1.63 -1.14 -1.42 -0.90 -4.82 -1.12 0.35 
N+2 -1.28 -1.50 -1.27 -1.44 -0.75 -1.30 -1.12 -4.49 -1.24 0.25 
Mean -1.21 -0.61 -1.15 -1.30 -1.11 -1.54 -1.08 -5.29 -1.14  
Std Dev 0.13 0.61 0.25 0.42 0.23 0.25 0.14 0.72  0.40 
Note: 
The overall means and standard deviations (lower right corner of the chart) omitted data for receiver G4 
for reasons discussed in the text. 
 
 
Overall, the SFU measurements differed from the baseline measurements by an average of about -1.1 dB 
(excluding receiver G4), indicating that the receivers were slightly less susceptible to out-of-channel 
interference when looking at a desired signal from the SFU as opposed to the ATSC997.  This difference 
reduces to -1.0 dB if the newer “repeat baseline” mean is subtracted.  A difference of 1 dB in an 
individual measurement having a 0.4 dB standard deviation is large enough to suggest a statistically 
significant difference.  If the measurement differences were deemed independent of one another, the 
standard deviation of the mean would be reduced by a factor of square-root of 35 (i.e., 5.9) to less than 
0.1 dB.  The size of the 1-dB mean difference relative to this computed standard deviation provides strong 
evidence that this is a real difference rather than a statistical artifact.  Other measurements, presented in 
Chapter 5, support the notion that there is a real difference between the signal sources that influences the 
performance of DTV receivers.   
 
Except for the tests with receiver G4, there appears to be no major trend across the receivers or the 
channel offsets.   
 
 
7-5 
The measurements for receiver G4 were dramatically different from those for the other receivers.  The 
change in signal sources appeared to have a 5 dB effect on D/U ratios for that receiver; however, when 
the “repeat baseline tests” were performed (the next section of this chapter), the measurements on receiver 
G4 were performed twice with results differing by as much as 6 dB for one of the channel offsets.  
Furthermore, when receiver threshold without interference (D
MIN
) had been measured as part of the 
D
MIN
+3 dB measurements presented in Chapter 5, D
MIN
 measurements on consecutive days yielded results 
differing by 5.4 dB.  Later, in measuring intermodulation effects with pairs of undesired signals of 
unequal power, another discrepancy of about 6 dB arose for that receiver.  No other receiver exhibited 
such variations.  Although receiver G4 is the best-performing receiver tested (in terms of interference 
rejection capabilities), there appears to be something intermittent or variable in its performance.  
Consequently, results for G4 were omitted from the overall mean and standard deviation data presented in 
this chapter. 
 
The results suggest that degraded signal quality from the ATSC997 reduces the receiver’s available 
margin to handle interference.  Signal quality measurements presented in Chapter 5 did show that the 
signal from the ATSC997 is inferior to that from the ATSC997, but we are unable to identify a signal 
quality measurement low enough to explain the performance differences of the TV receivers.  
REPEAT BASELINE TEST 
We wanted to determine whether the 1-dB change in D/U ratio discussed in the previous section was 
actually related to the change in signal sources or whether it might have been caused by some aspect of 
the test setup.  A different spectrum analyzer was used for the comparative tests than for the baseline 
tests.  Also, the baseline tests had been performed about four months earlier than the comparative tests, so 
it was plausible that some other unintended change in the equipment setup or performance might have 
contributed to the observed change. 
 
To rule out such equipment and test-setup issues the baseline tests were repeated.  Table 7-3 shows the 
results.  The repeat test produced results that, on average, differed from the original measurements by 
only about 0.1 dB.  The standard deviation of those differences was 0.33 dB.  This agreement between the 
four-month old baseline data and new measurements is considered quite good.  
 
 
Table 7-3.  D/U Ratio for Repeat of Baseline Test Relative to Baseline 
 
D/U Ratio Relative to Baseline (dB) 
Channel 
Offset 
A3 D3 I1 J1 M1 N1 O1 G4 
G4 
again
Mean 
(dB) 
(Excluding 
G4) 
Standard 
Deviation 
(dB) 
(Excluding 
G4) 
N-6 -0.73 0.01 0.00 0.02 -0.48 -0.27 -0.11 -6.23 -0.23 -0.22 0.29
N-4 -0.69 0.11 0.50 0.13 -0.36 -0.22 -0.01 -3.37 -0.09 -0.08 0.39
N-3 -0.47 -0.68 0.15 0.06 -0.26 -0.44 0.10 0.14 0.63 -0.22 0.33
N-2 -0.31 -0.21 0.33 0.00 -0.04 -0.15 0.12 0.68 0.19 -0.04 0.22
N+2 0.06 -0.90 -0.03 0.15 0.44 -0.01 0.13 -0.28 0.36 -0.02 0.42
Mean -0.43 -0.33 0.19 0.07 -0.14 -0.22 0.05 -1.81 0.17 -0.12 
Std Dev 0.32 0.44 0.22 0.07 0.36 0.16 0.10 2.93 0.34  0.33
Note: 
The overall means and standard deviations (lower right corner of the chart) omitted data for receiver G4 
for reasons discussed in the section of this chapter entitled, “Desired Signal Source:  SFU Versus 
ATSC997”. 
 
 
 
7-6 
Results for receiver G4 were omitted from the statistics presented in the previous paragraph for reasons 
described in the previous section.  The measurements shown in the “G4” column were made on a Friday.  
G4 was measured again (“G4 again” column) on the following Monday with significantly different 
results. 
UNDESIRED SIGNAL TYPE:  8-VSB VERSUS BANDLIMITED 
GAUSSIAN NOISE 
Table 7-4 shows the effect of using an 8-VSB signal instead of bandlimited white Gaussian noise as the 
undesired signal.  On average, the TVs are 1.3 dB less susceptible to interference from a DTV 8-VSB 
signal than from white Gaussian noise bandlimited to the same 3-dB bandwidth.  The difference reduces 
to 1.2 dB when the newer baseline measurements are used as the reference.  Given that the standard 
deviation of the mean is expected to be much less than the 0.68 dB value for individual measurements (by 
a factor of 5.9, if individual differences were statistically independent), the observed difference appears to 
be real, rather than a statistical artifact. 
 
The reason for the difference could be related to the amplitude statistics of the respective waveforms.  An 
8-VSB waveform is likely to exhibit fewer and smaller amplitude extremes (i.e., less time spent at levels 
far above the average power) than a Gaussian noise signal.  For linear interference mechanisms, the 
interference signal (within the TV) is linearly related to the incoming undesired signal, and the 
interference power (within the TV) is related to the mean-square of that signal (i.e., the second-order 
moment).  For a third-order interference mechanism on the other hand, the interference signal in the TV is 
related to the cube of the incoming undesired signal, and the interference power in the TV is related to the 
6
th
-order moment of the undesired signal.  We consider it likely that the high order moments of an 8-VSB 
signal are lower than those for a Gaussian noise signal (or for an OFDM signal) of the same power. 
 
 
Table 7-4.  D/U Ratio With 8-VSB as Undesired Signal Relative to That with Bandlimited Gaussian Noise  
 
D/U Ratio Relative to Baseline (dB) 
Channel 
Offset 
A3 D3 I1 J1 M1 N1 O1 G4 
Mean (dB) 
(Excluding 
G4) 
Standard 
Deviation 
(dB) 
(Excluding 
G4) 
N-6 -1.32 -1.25 -1.61 -0.73 -1.85 -0.87 -1.76 -3.72 -1.34 0.43 
N-4 -1.31 -1.43 -0.92 -1.07 -1.66 -1.07 -0.87 -1.94 -1.19 0.29 
N-3 -0.79 -1.85 -2.23 -1.73 -1.50 -2.38 -1.47 -0.26 -1.71 0.53 
N-2 -0.76 -1.94 -2.31 -0.53 -1.31 -0.91 -0.71 -0.89 -1.21 0.68 
N+2 0.02 -2.99 -1.94 -0.48 -0.04 -0.67 -0.40 0.50 -0.93 1.12 
Mean -0.83 -1.89 -1.80 -0.91 -1.27 -1.18 -1.04 -1.26 -1.28  
Std Dev 0.55 0.68 0.56 0.51 0.72 0.69 0.56 1.64  0.68 
Note: 
The overall means and standard deviations (lower right corner of the chart) omitted data for receiver G4 
for reasons discussed in the section of this chapter entitled, “Desired Signal Source:  SFU Versus 
ATSC997”. 
 
 
The variability among these measured differences in D/U ratios for the change in undesired signal type is 
somewhat larger than that associated with the change in the desired signal source.  There may be patterns 
in the data.  For example, the undesired signal type appears to have a greater effect at N-3 than at N+2.  
Whether these differences are associated with the order of the associated interference mechanisms-in 
conjunction with moments of the undesired signals (as discussed in the previous paragraph) is uncertain 
 
7-7 
because the order of the interference mechanisms is often masked by AGC operation—a topic to be 
addressed in subsequent chapters.   
UNDESIRED SIGNAL TYPE:  DVB-H OFDM VERSUS BANDLIMITED 
GAUSSIAN NOISE 
Tests were performed with an OFDM undesired signal and compared to the tests with bandlimited 
Gaussian noise.  The OFDM signal was produced by an Agilent E4438C vector signal generator using 
Agilent Signal Studio for DVB software.  The signal type was selected as DVB-H, an OFDM signal 
format designed for mobile video application.  The waveform parameters were as follows: 
• Size 2k 
• Modulation 64 QAM 
• Channel width 5 MHz 
• Guard interval 1/8 
 
Plots of the signal spectrum were shown in Figures 2-1 and 2-2.  The bandwidth measurements were 
shown in Table 2-1. 
 
Table 7-5 shows the D/U ratio measurements made with the DVB-H source as the undesired signal 
relative to those for the bandlimited Gaussian noise baseline.  The table includes one point that is greatly 
inconsistent with the others (by about 5 dB).  That measurement, for receiver N1 with the undesired 
signal on channel N-3, indicates that the receiver was significantly less susceptible to the DVB-H 
interference than to the bandlimited Gaussian noise signal.   
 
 
Table 7-5.  D/U Ratio With DVB-H as Undesired Signal Relative to That with Bandlimited Gaussian 
Noise 
D/U Ratio Relative to Baseline (dB) 
Channel 
Offset 
A3 D3 I1 J1 M1 N1 O1 G4 
Mean (dB) 
(Excluding 
G4) 
Standard 
Deviation 
(dB) 
(Excluding 
G4) 
N-6 -0.87 -0.08 0.24 0.17 -0.78 -0.23 -0.54 -1.47 -0.30 0.44 
N-4 -0.94 -0.1 1.92 0.04 -0.74 -0.43 -0.54 -0.87 -0.11 0.96 
N-3 -0.53 -0.53 0.09 -1.13 -0.52 -5.44 -0.53 0.49 -1.23 1.89 
N-2 -0.54 -0.15 0.2 -0.09 -0.05 -0.08 0.00 -0.08 -0.10 0.22 
N+2 0.12 -2.37 -0.57 0.26 0.43 0.11 0.24 -1.53 -0.25 0.99 
Mean -0.55 -0.65 0.38 -0.15 -0.33 -1.21 -0.27 -0.69
-0.40 
(-0.25)  
Std Dev 0.42 0.98 0.92 0.56 0.52 2.37 0.37 0.88  1.09 (0.66) 
Note: 
The overall means and standard deviations (lower right corner of the chart) exclude data for receiver G4 
for reasons discussed in the section of this chapter entitled, “Desired Signal Source:  SFU Versus 
ATSC997”.  The means and standard deviations shown in parentheses also exclude the value associated 
with receiver N1 at N-3 boxed bold italics  for reasons discussed below. 
 
 
Since the 4.8 MHz width of the DVB-H waveform, combined with its extremely steep spectrum rolloff on 
each side, left a gap of about 0.6 MHz between each side of the waveform’s spectrum and the channel 
edges, we wondered whether receiver N1 might have a narrowband interference susceptibility within 
channel N-4 that falls outside of the DVB-H waveform bandwidth.  To test this theory, the measurement 
 
7-8 
was repeated twice, once with the waveform shifted 0.5 MHz downward and once with the waveform 
shifted 0.5 MHz upward.  The measured D/U ratios relative to the baseline were as follows: 
 
• Frequency shift = -0.5 MHz D/U relative to baseline = -5.97 dB 
• Frequency shift = +0.5 MHz D/U relative to baseline = +1.95 dB 
 
The results suggest that this receiver had a narrowband susceptibility between the upper edge of the DVB-
H waveform, when it was centered on the channel, and the upper edge of the channel.  The tests 
conducted with a 1-MHz wide interferer (Figure 7-1) happen to have been performed with the same 
receiver.  Those tests do reveal a rapidly increasing susceptibility to interference as the undesired signal 
approaches the upper band edge for channel N-3. 
 
The mean and standard deviation shown in the table were computed in the same manner those for the 
other comparison tests in this chapter.  In addition, a second computation was performed omitting the 
measurement associated with N-4 for receiver N1 because the reduced width of the OFDM signal had 
apparently caused a narrowband susceptibility to be missed. 
 
Except for the one aberrant point caused by placement of the DVB-H waveform spectrum, there is no 
obvious trend in the data. 
 
Using the parenthetical value of mean from the table (-0.25 dB) and subtracting the mean for the “repeat 
baseline” test shows a mean difference of -0.1 dB between the OFDM results and the Gaussian results.  
The interference effect of the DVB-H waveform appears to be essentially identical to that of bandlimited 
Gaussian noise. 
 
 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-57 -51 -45 -39 -33 -27 -21 -15 -9 -3 3 9 15 21 27 33 39 45 51 57 63 69 75 81 87 93
Interferer Center Frequency Separation from Desired Channel Center (MHz)
S
u
sc
epti
bil
i
t
y to 
I
n
ter
f
e
r
e
n
ce
 [D/
U
 a
t
 T
h
resh
old (d
B
)
]
1-MHz Bandwidth Interferer
5.38-MHz Bandwidth Interferer
1-MHz Measurement Limit
NN+15N+5 N+10N-5N-9
Receiver N1
 
Figure 7-1.  D/U With 1-MHz Undesired Signal Bandwidth 
 
8-1 
CHAPTER 8 
THEORETICAL FRAMEWORK FOR OUT-OF-CHANNEL 
INTERFERENCE 
Before presenting more measurement results, we devote a chapter to establishing a theoretical framework 
that will be used for interpreting or extending the measurement results in most of the remaining chapters.  
We apply this theoretical basis to some of the common interference mechanisms that apply to TV 
reception.  A more detailed derivation is included in Appendix B. 
 
When a DTV receiver operates in the presence of white Gaussian co-channel interference, the threshold 
of visibility (TOV) of picture degradation occurs when the desired signal power D exceeds the co-channel 
interference by about 15 dB.
*
  This number may vary somewhat for noise having other statistical 
properties, and may be much lower if the noise is heavily concentrated at a band edge where filtering in 
the DTV provides additional rejection; nonetheless, one expects that, as signal power D varies, the 
undesired signal power at threshold will vary linearly with it—resulting a constant D/U ratio as D or U 
are varied.  This relationship holds whenever the co-channel interference is high enough that the effect of 
internal noise in the receiver becomes insignificant. 
 
For most out-of-channel interference mechanisms, the DTV receiver unintentionally converts a small 
portion of the out-of-channel power into co-channel power.  If one knows the amount of conversion into 
co-channel interference, one can treat the problem as a co-channel interference problem, which is 
relatively well understood, as described above.  In this formulation of the problem, measuring the desired 
signal power D at the TOV provides an indirect method of measuring the co-channel power created 
internal to the receiver, since we know that the co-channel power will be about 15 dB below the measured 
value of D. 
 
The conversion process by the DTV receiver from out-of-channel interference to co-channel interference 
may be linear or nonlinear.  If it is linear, then the internally-created co-channel interference will vary 
linearly with the out-of-channel interference power U causing the value of the desired signal power D at 
threshold to vary linearly with U.  The result will be that threshold D/U ratio will be constant as D or U is 
varied.  If the conversion process is nonlinear, then the relationship between D and U will be nonlinear 
and the D/U ratio will vary with D and U. 
INTERFERENCE MECHANISMS AND ORDER 
We will assume that the co-channel interference power created by the DTV receiver in response to an 
out-of-channel undesired signal power U will be proportional to D
L
U
M
, where L and M are integer 
constants that define the order of the interference mechanism.  For most interference mechanisms, L will 
be zero, so only the U
M
 term exists.  The following are among the interference mechanisms that can be 
modeled by this formulation. 
• Linear interference:  L=0, M=1.  Creates co-channel interference proportional to U.   
? Example:  mixer image.  The mixer in a TV receiver converts the spectrum of the intended 
channel of the received signal to an intermediate frequency (IF) where it can be filtered more 
precisely to pass the desired channel while rejecting the undesired frequencies.  Unfortunately, in 
single-conversion tuners a second a 6-MHz wide portion of the input spectrum centered 88 MHz 
above the desired channel is also converted to that same IF.  Filtering prior to the mixer strongly 
diminishes—but doesn’t fully extinguish—this unintended signal. 
                                                      
*
 The SHVERA Study test results on 28 receivers showed that D must exceed U by amounts ranging from 14.9 to 
15.8 dB, with a median value of 15.3 dB. 
 
8-2 
? Example:  leakage of the adjacent channel signal through the channel selection filter of the DTV 
receiver would also constitute a linear interference mechanism. 
• Second-order interference:  L=0, M=2.  Creates co-channel interference proportional to U
2
. 
? Example:  “half-IF” taboo.  The second harmonic of an undesired signal 22 MHz above the 
desired signal beats with the second harmonic of the receiver’s local oscillator, creating a 
difference frequency that falls within the IF band of the receiver. 
• Third-order interference:  L=0, M=3.  Creates co-channel interference proportional to U
3
. 
? Example:  third-order intermodulation (IM3) of a single, adjacent-channel undesired signal.  IM3 
creates spectral components that spill into each adjacent channel.  
? Example:  third-order intermodulation (IM3) of a pair of undesired channels placed at channels 
N+K and N+2K where N is the desired channel.  In this case, the interference power created in 
channel N is proportional to U
N+K
2
U
N+2K
.  The result is a process that is second-order in terms of 
U
N+K
 and linear in terms of U
N+2K
; however, if the two undesired signals are set to equal powers 
and varied in amplitude together, the resulting interference is third order. 
• Cross-modulation:  L=1, M=2.  Creates co-channel interference proportional to DU
2
. 
? Cross-modulation is essentially a third-order effect, but the co-channel interference created is 
proportional to D and to U
2
.  As a result, increasing the desired signal power does not improve the 
signal-to-interference ratio. 
THRESHOLD OF VISIBILITY OF PICTURE DEGRADATION DUE TO 
INTERFERENCE 
As stated earlier, the TOV for a DTV receiver occurs when the ratio of the desired signal to the co-
channel interference-plus-noise exceeds the required signal-to-noise ratio SNR
R
 for the DTV receiver.    
We will assume that the co-channel interference-plus-noise power includes two components:  receiver 
noise power N
R
 and the co-channel interference that was created by linear or nonlinear interference 
mechanisms operating on the out-of-channel undesired signal (c D
L
U
M
, where c is a constant).  Note that 
the both the co-channel receiver noise and the co-channel interference power are generated within the 
DTV receiver.  The levels we refer to here are expressed in terms of equivalent input power levels to the 
TV. 
 
Thus we can say 
 
D / (c D
L
U
M
 + N
R
) = SNR
R
 
 
For 8-VSB DTV receivers the SNR
R
 is about 33.9 (i.e., 15.3 dB converted to a linear power ratio) if the 
noise and interference have white spectra and Gaussian statistics.  Note that, while the receiver noise is 
likely to be white and Gaussian, the interference may not be—in which case a different SNR would apply 
to it, but we will neglect that difference for the discussion here. 
 
We note that the threshold desired signal power in the absence of interference (D
MIN
) is given by 
 
D
MIN
 = SNR
R
 N
R
  
 
Thus we can write, 
 
D = SNR
R
 c D
L
U
M
 + D
MIN
   
 
 
8-3 
SLOPES OF LOG-LOG PLOTS OF D, U, AND D/U  
We first consider the case in which the desired signal at threshold is much larger than D
MIN
.  We can then 
write, 
 
D ? SNR
R
 c D
L
U
M
 
 
Depending on the values of L and M, a plot of D versus U (in power units, such as microwatts) may be 
linear or nonlinear; however, a plot of log(D) versus log(U) will always be linear.  Thus, plots of D versus 
U in units of dBm will be straight lines, since decibels are a logarithmic unit.  The slope of such a log-log 
plot indicates the order of the interference mechanism.  For example, for a third-order interference 
mechanism (L=0, M=3), the slope of log-D versus log-U will be 3 dB/dB and the log-U versus log-D will 
be 0.333 dB/dB; similarly, a plot of D/U versus D on a log-log scale will have a slope of 0.667 dB/dB.  
The expected slopes of log-log plots for various interference mechanisms are summarized in Table 8-1.  
The final row of the table will be explained in the next section. 
 
 
Table 8-1.  Slopes of Log-Log Plots of D, U, and D/U for Various Interference Mechanisms 
Interference Mechanism 
Slope of 
Log (D) 
Versus 
Log (U) 
in dB/dB 
Slope of 
Log (U) 
Versus 
Log (D) 
in dB/dB 
Slope of 
Log (D/U) 
Versus 
Log (D)  
in dB/dB Characterization 
Linear (M = 1) 1 1 0 Constant D/U 
Second order (M = 2) 2 0.5 0.5  
Third order (including third-order 
intermodulation of a pair of equal-
power interferers) (M = 3) 
3 0.333 0.667  
Cross modulation (M = 2, L = 1) Infinite 0 1 Constant U 
AGC-Stabilized Nonlinear
1
 1 1 0 Constant D/U 
Notes: 
1
 See next section (“Effect of AGC”) 
 
EFFECT OF AGC 
Television receivers incorporate an automatic gain control (AGC) function that adjusts the gain of one or 
more stages in the tuner in order to maintain acceptable signal levels.  The gain may be constant (at its 
maximum) when signal levels are low.  As signal levels rise, a point is reached at which the AGC begins 
to reduce the gain to avoid overdriving the tuner circuitry.  The AGC control function may be based on 
the level of the desired signal, or on the combined power of the desired and undesired signals at some 
point in the tuner (typically the mixer), or both.
*
  Typically, the AGC will act to reduce the gain of both 
RF and IF amplifier stages, but not necessarily at the same signal levels.  As input signal levels increase, 
the AGC may act to reduce IF gain first—waiting for higher signal levels before reducing RF amplifier 
gain (delayed AGC).   
 
If we consider an interference mechanism that is caused by a nonlinearity at a given point in the tuner, the 
expected nonlinear behavior described in the previous section will exist only for signal levels up to the 
point at which AGC begins to reduce gain prior to the point of the nonlinearity.  For example, if a given 
                                                      
*
 O. Bendov and C. B. Patel, “Television Receiver Optimization in the Presence of Adjacent Channel Interference”, 
IEEE Transactions On Broadcasting, Vol. 51, No. 1, March 2005, p.38-39. 
 
8-4 
interference mechanism is caused by nonlinearity in the receiver’s mixer, AGC gain reductions in the IF 
amplifier will not affect it, but gain reductions in the RF amplifier will. 
 
Under the assumption that AGC operation tends to adjust gain so as to maintain either the desired signal 
or the undesired signal at a constant level
*
 at the point of nonlinearity, then the interference begins to 
behave as if it were linear (with the possible exception of cross-modulation when AGC operates on 
undesired signal power).
†
  That is, further increases in D result in corresponding increases in threshold U, 
so D/U remains constant.  This occurs because the amplitudes at the nonlinearity remain constant in spite 
of further increases in input signal levels at the receiver’s antenna input terminal.  The results are derived 
in Appendix B. 
 
This suggests that we may see interference that behaves like third-order interference, for example, at low 
signal levels, but switches to linear behavior when a certain signal threshold associated with the AGC is 
exceeded.  The last row of Table 8-1 refers this effect. 
WEAK SIGNALS 
As stated above, both the linear and the nonlinear interference mechanisms will plot as straight lines on a 
log-log plot (e.g., plotting U, D, and D/U in decibels).  However, this relationship is true only at desired 
input signal levels high enough to make the receiver’s own internal noise an insignificant contributor to 
performance. 
 
As signal level approaches the threshold for the receiver in absence of interference, the receiver becomes 
increasingly more susceptible to interference than such straight-line projections would predict.   
 
Referring to the earlier equation defining the interference threshold, we had 
 
D = SNR
R
 c D
L
U
M
 + D
MIN
   
 
The presence of the quantity D
MIN
 causes the log-log plots to deviate from straight-line behavior as D 
approaches D
MIN
.  Figures 8-1 through 8-4 depict the deviation of threshold U versus D from a straight 
line on a log-log plot for linear, second-order, third-order, and cross-modulation interference, 
respectively.  It can be seen that for the first three cases, the distance in the D direction from the straight 
line projection is 3 dB and 6.9 dB when D = D
MIN
 + 3 dB and D
MIN
 + 1 dB, respectively.  Distances in the 
U direction vary with slope of the line as shown in the illustrations and in Table 8-2.  Thus, for example, 
when desired signal level drops to a point 1 dB above the receiver’s threshold (D
MIN
), the DTV receiver 
                                                      
*
 AGC will generally act on the basis of either desired signal power or total power at a point in the tuner.  Broadband 
AGC is an example of the latter.  If operating on the basis of total power, the relative contributions of the desired 
and undesired signals at the AGC control point may differ from their relative levels at the receiver input due to 
filtering within the receiver.  In the case of AGC operation based on total power, the two bounding cases—desired 
signal being dominant and undesired signal being dominant—were considered in the analysis. 
†
 For the case in which the AGC acts to maintain the desired signal at a constant power at the point of tuner 
nonlinearity that is responsible for the dominant interference, Appendix B shows the math behind the transition to 
linear behavior for each of the four types of interference discussed in this chapter.  For the case in which the AGC 
acts to maintain the undesired signal at a constant level, Appendix B shows the math only for three of the four 
interference categories.  The case of cross-modulation with AGC acting on undesired signal was not completely 
solved.  When desired signal power is much greater than D
MIN
, cross-modulation causes TOV to occur at a fixed 
threshold undesired signal power (U), independent of the desired signal power, for cases in which gain is constant.  
If the undesired signal rises to a point at which the AGC begins reducing gain prior to the point of nonlinearity, the 
model suggests the following:  (1) if that AGC threshold occurs before the TOV is reached, then the subsequent gain 
reductions will prevent TOV from being reached as U increases further (until the range of gain reduction for the 
AGC is exceeded or another nonlinearity becomes significant); (2) if the AGC threshold occurs after TOV is 
reached, then the TV will remain in a degraded picture state with further increases in undesired signal power. 
 
8-5 
will be sensitive to interference from undesired signals that are lower than a straight-line projection would 
predict, by amounts ranging from 2.3 to 6.9 dB, depending on the order of the interference mechanism. 
 
In terms of deviation from the straight-line projection, the effect of AGC varies according to whether the 
AGC is driven primarily by desired signal power or by undesired signal power.  In the former case, the 
deviation from a straight-line projection matches that of the underlying interference mechanism.  In the 
latter, the deviation matches that of a linear process. 
 
 
Table 8-2.  Deviation in Threshold U from Straight-Line Projection as D approaches D
MIN
 
 
Deviation in Threshold U 
from Straight-Line Projection (dB) 
Interference Mechanism 
D/D
MIN
1
  
= 16 dBm 
D/D
MIN
 
= 3 dB 
D/D
MIN
 
= 1 dB 
D/D
MIN
= 0 dB 
Linear (M = 1) -0.1 -3.0 -6.9 Infinite 
Second order (M = 2)  -1.5 -3.4 Infinite 
Third order (including third-order 
intermodulation of a pair of 
equal-power interferers) (M = 3) 
 
-1.0 -2.3 Infinite 
Cross modulation (M = 2, L = 1)  -1.5 -3.4 Infinite 
AGC-Stabilized Nonlinear w/U 
driving AGC
2
 
-0.1 -3.0 -6.9 Infinite 
Note: 
1
 For the nominal D
MIN
 value of -84 dBm, D/D
MIN
 = 16 dB when D = -68 dBm 
2
 With desired signal driving AGC, deviation from straight-line projection matches that of 
the original nonlinear process, except in the case of cross-modulation, which was not 
completely solved.
 
 
IM3 WITH PAIRED SIGNALS 
When the interference is caused by a pair of signals located at channels N+K and N+2K, the receiver 
creates third-order intermodulation (IM3) products in the desired channel N. 
 
In this case, Appendix B shows that the interference equation becomes: 
 
D = (SNR
R
 / IP3
2
) U
N+K
2
U
N+2K
 + D
MIN
. 
 
We will use the term IP3 / SNR
R
1/2
 to quantify the IM3 properties of the receiver by computing it from 
measurements of D and U at threshold with D >> D
MIN
 and U = U
N+K
 = U
N+2K
. 
 
IP3 / SNR
R
1/2
 = (U
3
 / D)
1/2
 
 
or, in decibel units 
 
(IP3 / SNR
R
1/2
)|
dB
 = 1.5 U|
dB
 - 0.5 D|
dB
  
 
where |
dB
 means conversion to decibels: 
 
X|
dB
 = 10 log(X) 
 
To compute interference susceptibilities knowing (IP3 / SNR
R
1/2
)|
dB
, we will use the following. 
 
8-6 
 
U
N+K
|
dB
 = (IP3 / SNR
R
1/2
)|
dB
 + (D|
dB
 - U
N+2K
|
dB
)/2 
 
or, 
 
U
N+2K
|
dB
 = 2(IP3 / SNR
R
1/2
)|
dB
 + D|
dB
 - 2 U
N+K
|
dB
]  
 
See Appendix B for details the derivation. 
 
8-7 
-8
-6
-4
-2
0
2
4
6
-6-4-20246810
Desired Signal Power Relative to Dmin (dB)
No
rmalized
 
Un
d
e
sired
 Sign
al Po
wer (d
B)
Linear Interference Mechanism
6.9 dB
3 dB
3 dB
6.9 dB
D = Dmin + 3 dB
D = Dmin + 1 dB
 
Figure 8-1.  Deviation of Log-U Versus Log-D From Straight Line For Linear Interference 
-8
-6
-4
-2
0
2
4
6
-6-4-20246810
Desired Signal Power Relative to Dmin (dB)
No
rmalized
 
Un
d
e
sired
 Sign
al Po
wer (d
B)
2nd Order Interference Mechanism
6.9 dB
3 dB
1.5 dB
3.4 dB
D = Dmin + 3 dB
D = Dmin + 1 dB
 
Figure 8-2.  Deviation of Log-U Versus Log-D From Straight Line For 2
nd
-Order Interference 
 
8-8 
-8
-6
-4
-2
0
2
4
6
-6-4-20246810
Desired Signal Power Relative to Dmin (dB)
No
rmalized
 
Un
d
e
sired
 Sign
al Po
wer (d
B)
3rd Order Interference Mechanism
6.9 dB
3 dB
1 dB
2.3 dB D = Dmin + 3 dB
D = Dmin + 1 dB
 
Figure 8-3.  Deviation of Log-U Versus Log-D From Straight Line For 3
rd
-Order Interference 
-8
-6
-4
-2
0
2
4
6
-6-4-20246810
Desired Signal Power Relative to Dmin (dB)
No
rmalized
 
Un
d
e
sired
 Sign
al Po
wer (d
B)
Cross-Modulation Interference Mechanism
1.5 dB
3.4 dB
D = Dmin + 3 dB
D = Dmin + 1 dB
 
Figure 8-4.  Deviation of Log-U Versus Log-D From Straight Line For Cross-Modulation 
 
9-1 
CHAPTER 9 
THIRD-ORDER INTERMODULATION WITH PAIRED EQUAL-
POWER INTERFERERS 
Rhodes and Sgrignoli have raised the issue of third-order intermodulation (IM3) distortion occurring 
within a TV receiver between pairs of undesired input signals as a potentially significant interference 
mechanism for DTV reception.
*
 
†
 
‡
  This chapter presents the results of interference rejection 
measurements performed using pairs of undesired signals spaced so as to place IM3 products into the 
desired channel.  Thus, signal pairs were placed on channels N+K and N+2K, where K is a positive or 
negative integer.  For this chapter the two undesired sources were maintained at equal power levels.  
Additional test results with paired signals are contained in Chapters 10 (unequal sources) and 11 (detailed 
measurements on one TV). 
 
In all tests, the undesired signal placed farthest from the desired channel (i.e., at N+2K) was a white 
Gaussian noise signal bandlimited to match the 3-dB width of an 8-VSB signal.  For tests with the desired 
channel N = 51, the closer undesired signal (at N+K) to the desired channel N was an 8-VSB signal.  For 
tests with the desired channel N = 30, the closer undesired signal was 8-VSB only if it was on a first-
adjacent channel (N-1 or N+1); otherwise, the bandlimited Gaussian noise signal was used as the closer 
undesired signal source.  The reason for the use of bandlimited Gaussian noise instead of 8-VSB was that, 
throughout most of the test period, only one 8-VSB source was available and it was needed as a desired 
signal source.  Tests involving first adjacent channels were postponed until the procurement of another 
8-VSB generator because the edge-of-band rolloff of the Gaussian source was not sufficient to support 
rejection tests on first adjacent channels.   
 
Table 9-1 summarizes the test parameters.  Only the test results from the first three rows of the chart are 
presented in this chapter.  The fourth row describes tests at unequal undesired signal levels, which are 
presented in Chapter 10.  The last row of the table identifies very detailed tests of the variation of D/U 
with desired signal level, which are presented in Chapter 11. 
 
The purpose of these tests was to determine the extent to which pairs of undesired signals create an 
interference effect—through IM3 occurring within a TV receiver—that exceeds the effects of the 
individual signals. 
SPECTRA OF THIRD-ORDER INTERMODULATION DISTORTION 
IM3 creates signal components at frequencies that were not present in the input spectrum, but are in the 
same general frequency range as the input signals.  The IM3 of interest in this chapter is created within 
the tuners of consumer DTV receivers.  Since test points are not available within consumer receivers to 
show the actual signal effects that occur, spectra of IM3 components created by a laboratory 
instrumentation amplifier were measured to illustrate the concepts.
§
 
 
 
                                                      
*
 Charles W. Rhodes, “Interference Between Television Signals due to Intermodulation in Receiver Front-Ends”, 
IEEE Transactions On Broadcasting, Vol. 51, No. 1, March 2005, p.31-37. 
†
 Charles W. Rhodes, and Gary J. Sgrignoli, “Interference Mitigation for Improved DTV Reception”, IEEE 
Transactions on Consumer Electronics, Vol. 51, No. 2, May 2005, p. 463-470. 
‡
 Charles W. Rhodes, “DTV interference could be mitigated by receivers,” TV Technology Magazine, vol. 22, no. 
17, p.21-23, Aug. 18, 2004. 
§
 An HP8447E amplifier rated for 0.1 watts (20 dBm) output served as the IM3 producer.  For the first illustration it 
was driven with a signal sufficient to create an output level of 8.1 dBm on each of one or two TV channels. 
 
9-2 
Table 9-1.  Parameters for Paired-Interferer Tests (N+K/N+2K) 
Chapter 
Desired 
Channel 
N 
Desired 
Signal 
Power K 
Undesired 
Signal at 
N+K 
Undesired 
Signal at 
N+2K 
Relative 
Power of 
U
N+K
 and 
U
N+2K
 
Number 
of DTV 
Receivers 
Tested 
9 30 
-68 dBm, 
-53 dBm, 
-28 dBm 
1, -1 8-VSB WGN
1
 Equal 
8 
9 30 
-68 dBm, 
-53 dBm 
2 through 5
and 
-2 through 
-5 
WGN
1
 WGN
1
 Equal 8 
9 51 
-68 dBm, 
-53 dBm, 
-28 dBm 
1 through 8 8-VSB WGN
1
 Equal 7
2
 
10 30 
D
MIN
+3 dB,  
-68 dBm, 
-53 dBm 
2 or 3 WGN
1
 WGN
1
 Variable 2 
11 51 
D
MIN
+1 dB, 
D
MIN
+3 dB, 
-78 to 
-8 dBm in 
5-dB steps 
1 through 4 8-VSB WGN
1
 Equal 1 
Notes: 
1
 WGN = white Gaussian noise signal bandlimited to match the 3-dB width of an 8-VSB signal. 
2
 For one of the seven TVs tested at channel 51, the complete set of measurements was performed only 
for D = -68 dBm; for the other two levels, only N+1/N+2 tests were performed. 
 
 
Figure 9-1(a) shows the spectrum of an 8-VSB signal on channel 34 after passing through the amplifier at 
a level 12 dB below the maximum rated output power of the amplifier.  The “shoulders” that can be seen 
emerging from the sides of the 8-VSB spectrum about 33-dB below the main signal spectrum level are 
IM3 products.  It can be seen that they extend through most of the width of each adjacent channel 
(channels 33 and 35), but not beyond.  Thus, IM3 generated with a DTV tuner in response to an undesired 
signal on a single TV channel could influence reception of a desired signal on a first-adjacent channel.  
Put another way, if the TV is tuned to channel N, its reception there could be adversely impacted by IM3 
caused by an undesired signal on channel N-1 or on channel N+1. 
 
Figure 9-1(b) shows the effect of adding a signal of equal power on a second channel—in this case 
channel 38.  The addition of the second signal causes the shoulders around the first signal to increase.  A 
similar pair of shoulders is also created around the second signal.  The spectral shoulders occupy channels 
33, 35, 37, and 39—the channels adjacent to each of the input signals.  In addition to the shoulders around 
each input signal, the presence of the second signal causes two other bumps in the spectrum to occur.  
One is centered on channel 30 and the other on channel 42, but they also spill over into the channels 
adjacent to those two channels.  Thus, intermodulation products can now be found in channels 29, 30, 31, 
33, 35, 37, 39, 41, 42, and 43. 
 
If one imagines that a TV is attempting to receive a signal on channel 30 in the presence of the two 
undesired signals creating (within the TV tuner) intermodulation distortion shown in Figure 9-1(b), we 
would have the case of interference to channel N from an undesired signal pair, N+4/N+8.  This is one of 
the cases examined in this chapter, as is the case of N-4/N-8, which would occur if the receiver were 
tuned to channel 42.  (In test results shown in this chapter, the TV was always tuned to either channel 30 
or channel 51, and the undesired signal channels were shifted appropriately.) 
 
 
9-3 
Figure 9-2(a) shows the effect of changing the amplitudes of both signals by 5 dB.  The IM3 products 
change by 15 dB, as is expected since this is a third order process.  From this we can conclude that if the 
undesired signal levels should rise, the interference effects created within the TV will rise even faster—at 
three times the rate of the rise in input signal level. 
 
If only one of the two signals changes in amplitude, the results are somewhat more complex.  
Figure 9-2(b) shows the effect of a 5-dB change in amplitude of the signal at channel 34, while the one at 
channel 38 remains constant.  The result is a 10-dB change in the amplitude of the IM3 bump closest the 
changed signal (i.e., the bump at channel 30) and a 5-dB change in amplitude of the IM3 bump closest to 
the unchanged signal (i.e., the bump at channel 42).  
 
The tests in this chapter are configured so that undesired signals are placed on channels N+K and N+2K 
and the TV is tuned to channel N.  Thus, channel N is centered on one of the outer spectral bumps.  
Figure 9-2(b) confirms the prediction shown in Chapter 8 that the interference power created at channel N 
is proportional to U
N+K
2
 U
N+2K
, where U
N+K
 and U
N+2K
 represent the respective power levels of the 
undesired signals.  Thus, the interference created at channel N is a second order function of U
N+K
 and a 
linear function of U
N+2K
. 
CHANNEL-30 RESULTS AT THREE DESIRED SIGNAL LEVELS 
“Weak” Desired Signal (D = -68 dBm) 
Figure 9-3 shows measured D/U ratios at TOV for eight DTV receivers for pairs of equal-level undesired 
signals on channel pairs N+K/N+2K for K = -5, -4, -3, -2, -1, 1, 2, 3, 4, and 5 and a desired signal power 
of -68 dBm.  The measurements were made with the TVs tuned to channel 30 as the desired channel N.  
The desired signal power was set to -68 dBm.  In computing the D/U ratio, U was taken as the power of 
each signal in the signal pair rather than the combined power of the two signals; thus,  
 
 U = U
N+K
 = U
N+2K
 
 
The shaded area at the bottom of the plot represents the measurement limitations imposed by the test 
setup—as described in Chapter 4.  None of these measurements were limited by the measurement system. 
 
Though the ATSC A/74 has no recommended performance limits related to interference by a pair of 
signals, the red dashed curve labeled “A/74 Max of Pair” is provided as a reference based on single-
channel ATSC interference rejection guidelines.  For each point, N+K/N+2K, the value is equal to the 
higher (least restrictive) of the ATSC-recommended DTV-into-DTV interference rejection thresholds for 
single-channel DTV interferers.  All of the TVs exhibited higher D/U ratios (poorer performance) at most 
points than is indicated by this A/74-based curve. 
 
Figure 9-4 summarizes the measurements that were shown in Figure 9-3.  The solid curve shows the 
median performance of the eight receivers.  Error bars show the best and worst performance among the 
receivers at each channel offset.  A dashed curve shows the performance of the second worst performing 
receiver at each channel offset. 
 
Since the IF filter in a receiver is expected to greatly reduce signal levels available for creation of 
intermodulation products after the filter, one would expect that intermodulation between pairs of out-of-
channel interferers would occur prior to the IF filter—probably in the mixer.  Prior to the mixer, a typical 
single-conversion TV receiver would include a “tracking filter” that passes the desired channel but 
provides a gradually increasing attenuation to signals in other channels based on separation from the 
desired channel.  The filter is expected to provide substantial attenuation at N+14 and N+15 in order to 
reduce the mixer image response.  One would expect that such a tracking filter—if placed before the point 
of nonlinearity that created the intermodulation products—would cause interference effects of a 
 
9-4 
N+K/N+2K signal pair to diminish as |K| increases.  Such a trend is evident in some, but not all of the 
receivers, based on the measured data, which extends to |K| = 5. 
“Moderate” Desired Signal (D = -53 dBm) 
Figure 9-5 shows D/U-ratio measurements at TOV for the same eight receivers at a higher desired signal 
level of -53 dBm.  Figure 9-6 shows the median and range of the measurements, as well as the second 
worst performance among the eight receivers.   
 
The expected rolloff with increasing |K| is even less evident than in the signal measurements that were 
made with a desired signal power of -68 dBm.  In some cases, the influence of AGC operation on the 
N+1/N+2 and N-1/N-2 data may be partly responsible for the lack of increase at low absolute values of K.  
This topic is discussed more in Chapter 11. 
“Strong” Desired Signal (D = -28 dBm) 
Figure 9-7 shows the D/U measurements for a desired signal level of -28 dBm.  At this signal level, the 
measurements were performed for only two channel pairs—N-1/N-2 and N+1/N+2.  The maximum signal 
levels that could be generated by the test setup (about -7 dBm for each channel for adjacent-channel tests) 
were not high enough to create picture errors for most of the receivers.  As a result, threshold 
measurements were possible on only three of the receivers for N-1/N-2 and only two of the receivers for 
N+1/N+2.  The other measured points are shown at the measurement-limit line. 
 
Figure 9-8 presents information regarding the best, median, second worst, and worst performance for each 
of the two channel pairs measured.  Since there were only two points per curve, the data is presented in 
tabular form.  The table shows the threshold values for the worst and second worst performers.  The 
median and best performing values are shown only as “< [value]” because most of the data points were 
beyond the measurement limit of the test setup. 
Combined Results With Single-Channel Reference Values 
The plots shown in the previous three subsections provide no indication of whether the DTV picture 
errors are caused by IM3 interactions between the pair of undesired signals or whether they are caused by 
each signal independently. 
 
To determine whether IM3 is the cause, it would be useful to show the interference thresholds for the 
individual undesired signals as well.  Thus, we could plot the interference threshold for N+K/N+2K as 
well as the interference threshold for N+K alone and that for N+2K alone.  An even better reference 
would be one that combined the effects of N+K interference and N+2K interference under the assumption 
that the interference mechanisms are independent and not the result of intermodulation.  We attempt to 
take the latter approach—partly to reduce the number of curves that must be plotted so that results can be 
combined into a manageable number of graphs for this report. 
 
The approach taken is to recognize that the observed interference phenomena are the result of 
mechanisms at work within the TV that convert out-of-channel undesired signals into co-channel 
interference at some point within the TV.  The desired signal at threshold is then—in essence—a measure 
of the power level of that internal co-channel interference level, since the desired signal at threshold is 
expected to be about 15 dB above the co-channel interference.  This concept was discussed in Chapter 8. 
 
We will define D
N+K
 as the desired signal at threshold resulting from interferer U
N+K
, and D
N+2K
 as the 
desired signal at threshold resulting from interferer U
N+2K
.  If we were to take measurements of D
N+K
 and 
D
N+2K
 at equal undesired signal levels (U
N+K
=U
N+2K
=U), then D
N+K
 + D
N+2K
 would be the expected desired 
signal level threshold if both undesired signals were applied simultaneously and there were no interaction 
between them.  (Summing D
N+K
 and D
N+2K
 is equivalent to summing the co-channel interference powers 
created within the receiver from each of the two undesired signals individually.)  If the undesired signal 
powers are equal, this is equivalent to summing the D/U ratios: 
 
9-5 
 
(D/U)|
N+K/N+2K
 = (D
N+K
+D
N+2K
)/U = (D/U)|
N+K
 + (D/U)|
N+2K
 , if there is no interaction between the two 
undesired signals. 
 
This provides a reference point for determining whether the paired-signal D/U is caused by the combined 
individual effects of the two undesired signals or a by nonlinear interaction between them (i.e., IM3).  
Summing the D/U ratios to provide this reference point is strictly valid only if the measurements were 
performed at equal undesired signal levels (and if the AGC state of the tuner is the same for each 
measurement).  Our measurements were performed with equal D values rather than equal U values; 
however, it would still be valid to apply this technique if D/U for the individual interferers were constant 
with variations in D.  This is true of some cases but not for others.  Even when it is not true, very little 
error is made in summing the D/U ratios if the interference effect (D/U ratio) for one interferer is much 
higher for the other, since, in that case the sum is approximately equal the D/U of the dominant interferer. 
 
Based on the above we will plot the summed D/U ratio of the individual interferers along with the 
measured D/U ratio for the pair of interferers to provide an indication of whether IM3 effects are at work.  
(Note that the summing is performed on the direct power ratios, not on their values in decibels.) 
 
Figures 9-9 through 9-16 show the measured D/U ratios for the paired interferers along with the summed 
D/U ratios for the corresponding individual interferers.  Each graph presents the measurements for one 
DTV receiver.  The two solid lines on each graph show the paired-signal D/U ratios measured at desired 
signal levels of -68 dBm and -53 dBm.  The dashed lines show the summed D/U’s for those signal levels.  
Note that if one or both of the individual D/U’s was beyond the measurement limit of the test setup, then 
the sum was computed using the measurement limit; this fact may cause a summed D/U to exceed the 
paired-signal D/U if one element of the sum was at the measurement limit and the other was near the 
limit. 
 
Where the solid lines are closely matched by the dashed summed-D/U lines, the interference effect of the 
pair of equal-powered undesired signals is primarily due to one of the individual signals—or to the 
combined effect of both—rather than to an IM3 interaction between the two interferers.  Where a solid 
line is significantly higher than a dashed line, there is a significant IM3 effect from the pair of signals. 
 
Notably, most of the receivers exhibit very little evidence of IM3 interaction from the N-1/N-2 and 
N+1/N+2 pairs.  For a desired signal level of -68 dBm, the paired-signal D/U’s exceed the summed signal 
D/U’s by no more than 1.2 dB (and in most cases less than 0.5 dB) with one exception.  For receiver G4 
the difference is 4 dB at N-1/N-2.  This indicates that, for equal-power paired undesired signals, any IM3 
effects resulting from a signal pair on the first-adjacent channels (N-1/N-2 and N+1/N+2) of most 
receivers are insignificant relative to the interference sensitivities associated with the individual channels, 
at least for a desired signal power of -68 dBm.  The one exception to this—one case out of 16 (8 
receivers, upper and lower channel pairs)—yields a 4-dB increase in susceptibility as a result of IM3.  At 
a desired signal power of -53 dBm, three of the 16 combinations exhibit an IM3 effect that is about 2 dB 
above the summed D/U’s and one exceeds the summed D/U’s by 3.6 dB. 
 
Even at N-2/N-4 and N+2/N+4 about half of the receivers exhibit little or no evidence of an IM3 effect 
for paired equal-power undesired signals.  Among the other half, some exhibit very pronounced IM3 
effects up to 11 dB above the summed D/U’s. 
 
At larger channel spacings, there is evidence of IM3 from the signal pairs significantly exceeding the 
interference effect of the individual interferers—by amounts up to at least 17 dB.  (In cases where 
summed D/U’s are computed from measurements that were at the limit of the measurement setup, the 
actual summed D/U’s are lower than those shown, so the true amount of the excess is greater than that 
shown.) 
 
 
9-6 
An unexpected and seemingly impossible behavior can be seen for receiver D3 (Figure 9-10) at N-2/N-4 
for a desired signal level D = -53 dBm.  The D/U for the pair of interferers operating together is actually 
significantly lower than the summed D/U’s for the individual interferers.  Examination of the single-
channel data shows that the receiver is far more sensitive to interference at N-4 (D/U = -32.3) than at N-2 
(D/U = -49.0 dB); however, the D/U ratio for the pair of signals (with U referring to the undesired power 
on each of the two channels) is only -38.8 dB, 6.5 dB less than the D/U for N-4 alone.  Put another way, 
an undesired signal level of -20.8 dBm on channel N-4 causes picture errors on the TV, but if a second 
undesired signal is placed on channel N-2, the TV can tolerate a higher interference level of -14.3 dBm on 
each of the channels simultaneously! 
 
This result is highly counter-intuitive until one examines detailed test results for that TV—presented in 
the Chapter 11.  Though those measurements are limited to cases with interferers on higher channel 
numbers than the desired signal (N+K/N+2K with positive values of K), the data suggests that the 
presence of an undesired signal on channel N+2 any higher than approximately -35 dBm causes the DTV 
receiver’s AGC to reduce the RF gain, whereas an undesired signal at N+4 causes no such gain reduction 
(or, if it does, the reduction occurs at a higher signal level).  In the case described above, the TV is very 
susceptible to interference from an undesired signal placed at N-4, probably due to a nonlinearity in the 
tuner; but, adding in an undesired signal at N-2 probably causes the AGC to reduce the RF gain of the 
tuner (as was the case at N+2), which reduces the signal levels at the point of the nonlinearity.  A 
theoretical basis for understanding the effect of AGC on interference phenomena was presented in 
Chapter 8 (with more details in Appendix B).  More discussion about the effect of AGC will be presented 
in Chapter 11 on the basis of detailed measurement data for Receiver D3. 
 
Table 9-2 shows statistics for the difference (in dB) between the paired-signal D/U’s and the summed 
D/U’s for the eight TV receivers that were tested.  Looking across all of the data, we see that a pair of 
appropriately-spaced equal-power undesired signals can create an intermodulation effect that enhances 
the interference potential of the signals.  The combined signals can cause TV picture degradation (or loss) 
at signal levels at least as much as 17 dB lower than the levels that would be required to cause picture 
degradation based on the combined individual effects of each signal, i.e., without intermodulation 
between the pair.  (The actual difference may have been much larger.  See the note in table.)  At the other 
extreme, paired signals applied to some TV receivers on some N+K/N+2K channel pairs cause no 
increase in interference effect above the summed effects of the individual signals, and in one case 
(discussed in the preceding paragraph), AGC action causes the paired signal combination to have less of 
an interference effect than that of one of the individual signals. 
CHANNEL-51 RESULTS AT THREE DESIRED SIGNAL LEVELS 
In this section we present the results of paired-signal rejection measurements made with the TVs tuned to 
channel 51 as the desired channel N.  The two undesired signals were placed on N+K and N+2K for K = 
1 through 8.  The reason for limiting these tests to positive values of K is that the focus was on potential 
for interference to upper UHF channels from other radio services that are expected to operate in channels 
52 through 69 after completion of the DTV transition.   
 
For these tests, the undesired signal on channel N+K was an 8-VSB DTV signal.  The undesired signal on 
channel N+2K was a white Gaussian noise source bandlimited to match the 3-dB width of an 8-VSB 
DTV signal.   
 
 
 
 
 
9-7 
Table 9-2.  Statistics of Paired-Signal D/U’s Relative to Summed D/U’s for 8 Receivers on Channel 30 
Excess of Paired-Signal D/U Above Summed D/U’s (dB) 
D = -68 dBm D = -53 dBm 
Undesired 
Channel 
Pair 
Min Mean Median Max Min 
Lower 
Bound 
on 
Mean
1
 
Lower 
Bound 
on 
Median
1
 
Lower 
Bound 
on 
Max
1
 
N-5/N-10 1.8 9.8 9.8 15.7 3.4 11.7 11.7 16.9
N-4/N-8 0.4 5.7 5.1 12.2 3.4 9.5 9.8 16.3
N-3/N-6 -1.2 5.2 3.1 15.1 -0.2 5.9 6.4 11.3
N-2/N-4 -1.9 -0.2 -0.6 3.0 -6.6 0.5 0.2 5.6
N-1/N-2 -0.3 0.7 0.2 4.0 -0.3 0.7 0.1 3.6
N+1/N+2 -1.5 0.0 0.3 0.6 -1.3 0.6 0.5 2.3
N+2/N+4 -1.5 2.7 0.1 11.3 -1.2 4.0 3.9 11.4
N+3/N+6 -0.2 7.2 6.8 14.1 2.1 9.8 9.9 14.7
N+4/N+8 1.5 8.2 7.0 16.4 0.7 8.7 8.9 13.8
N+5/N+10 0.0 7.9 9.1 12.8 -0.3 7.1 7.3 13.7
 -1.9 4.7 4.1 16.4 -6.6 5.9 6.8 16.9
Note 
1
 For D = -53 dBm, the actual means, medians, and maxima for channels N+3/N+6 and beyond on the 
positive side and N-4/N-8 and beyond on the negative side are underestimated because most of the 
individual measurements on which the summed D/U’s are based were for measurement conditions in 
which TOV for the receivers was not reached due to limitations on maximum signal that the test setup 
could generate; consequently, the values shown in red italics should be viewed as lower bounds on the 
actual values. 
 
 
“Weak” Desired Signal (D = -68 dBm) 
Figure 9-17 shows measured values of D/U ratios at TOV for seven DTV receivers for pairs of equal-
level undesired signals with the desired signal power was set to -68 dBm.  The seven receivers are a 
subset of the eight that were tested for the previous major section of this chapter.  The reader is referred to 
the channel-30 section of this chapter for more information on the plot format. 
 
Figure 9-18 summarizes the measurements that were shown in Figure 9-17.  The solid curve shows the 
median performance of the eight receivers.  Error bars show the best and worst performance among the 
receivers at each channel offset.  A dashed curve shows the performance of the second worst performing 
receiver at each channel offset. 
 
For the most part the D/U ratios exhibit an expected drop (indicating less susceptibility to interference) as 
the channel spacing from the desired channel increases.  There are some exceptions.  The increase in D/U 
for receiver D3 as the undesired signal pair is moved from N+1/N+2 to N+2/N+4 is believed to be the 
result of an AGC gain reduction occurring when the undesired signal is placed on N+1.  This will be 
discussed further in Chapter 11.  The peak exhibited by some TVs at N+7/N+14 is the result of single-
channel effects at N+7 (the local oscillator frequency) or N+14 (mixer image frequency). 
 
“Moderate” Desired Signal (D = -53 dBm) 
Figure 9-19 shows D/U-ratio measurements at TOV for the same seven receivers at a higher desired 
signal level of -53 dBm.  We note that only one measurement was performed on receiver N1 at this 
 
9-8 
desired signal level—at N+1/N+2.
*
  Figure 9-20 shows the median and range of the measurements, as 
well as the second worst performance among the receivers; except for N+1/N+2, the data are for only six 
of the receivers.   
 
The expected rolloff with increasing |K| is less evident than in the “weak”-signal measurements.  In some 
cases, the influence of AGC operation on the N+1/N+2 and N-1/N-2 data may be partly responsible. 
 
“Strong” Desired Signal (D = -28 dBm) 
Figure 9-21 and 9-22 show corresponding plots of D/U measurements of seven receivers for a desired 
signal level of -28 dBm.  As in the case of -53 dBm, one of the receivers was measured only at N+1/N+2.  
Most of the D/U ratios fall outside the measurement range of the test setup. 
 
Combined Results With Single-Channel Reference Values 
Figures 9-23 through 9-29 show the measured D/U ratios for the paired interferers along with the summed 
D/U ratios for the corresponding individual interferers.  Each graph presents the measurements for one 
DTV receiver.  The solid colored lines on each graph show the paired-signal D/U ratios measured at 
desired signal levels of -68 dBm, -53 dBm, and -28 dBm. 
 
The dashed lines on each graph show the summed D/U’s for each of the three desired signal levels.  The 
summed D/U’s represent the summed interference effects of the two undesired signals in the absence of 
any intermodulation products generated by nonlinear interactions of one signal with the other.  Note that 
if one or both of the individual D/U’s was beyond the measurement limit of the test setup, then the sum 
was computed using the measurement limit.  This occurred for many of the summed D/U values for D = 
-68 dBm and most of those for D = -53 dBm and D = -28 dBm.  Where it occurred, the actual TOV levels 
for the individual undesired signals are unknown, and the actual differences between the paired-signal 
D/U’s and the summed D/U’s are greater than those shown in the charts.  Because of the number of data 
points affected by this limitation, the channel-51 data were not tabulated as the channel-30 data were.  We 
also note that the use of measurement limit values sometimes caused a summed D/U value on the plots to 
artificially exceed the paired-signal D/U when one element of the sum was at the measurement limit and 
the other was near the limit—a condition that occurred for most of the measurements for D = -28 dBm. 
 
Where the solid lines are closely matched by the dashed “summed D/U” lines, the interference effect of 
the pair of interferers is primarily due to one of the individual interferers—or the combined effect of 
both—rather than an IM3 interaction between the two interferers.  Where the solid lines significantly 
exceed the dashed lines, there is a significant IM3 effect from the pair of signals. 
 
ESTIMATING 3
RD 
ORDER INTERCEPT POINT (IP3) 
The measurements presented in this chapter can be used to determine the DTV receiver’s third-order 
intercept point (IP3)—a property that quantifies the nonlinearity of the receiver and allows computation 
of IM3 interference effects from undesired signal amplitudes that differ from the levels actually tested.  
Our use of the term IP3 here, while similar to its traditional use in characterizing amplifiers, differs from 
the traditionally defined IP3 in two ways: 
• The measurements here were made with broadband, noise or noise-like signals rather than with the 
CW (continuous wave) sinusoids usually used for IP3 measurements; since the process is a third-
order one, one would expect the effective interference power to be related to the sixth-order moments 
of the input signals; those moments are expected to be higher for the noise-like waveforms than for a 
sinusoid of the same input power; 
                                                      
*
 This test series was terminated by equipment failure, followed by a change in focus of the test program. 
 
9-9 
• As Figure 9-2 showed, the IM3 power in each side “bump” of the spectrum is split across three TV 
channels, of which only the center one (where most of the power is concentrated) is “measured” by 
the TV in our tests; thus, a portion of the IM3 power is excluded from the measurement. 
 
There are three limitations on such IP3-based assessment of DTV interference: 
(1) IP3 can be quantified from the paired-signal measurements only if the observed interference is 
primarily due to IM3 between the signal pair; 
(2) While IP3 is usually treated as a constant for a given amplifier, it will, in fact, vary with AGC 
operation; and, 
(3) IP3 computation depends on a knowledge of the signal-to-noise ratio necessary for the receiver to 
reach TOV because we are not measuring IM3 effects directly, but rather are inferring them from 
measurements of U and D at TOV. 
 
Regarding the first limitation, we will compute IP3 only when the D/U ratio for the pair of signals 
exceeds the summed D/U ratios of the individual signals (plotted as reference curves in Figures 9-9 
through 9-16 and 9-23 to 9-29) by at least 4 dB. 
 
Regarding the second limitation, we will recognize that IP3 will be constant only when desired and 
undesired signal levels are low enough not to cause AGC gain reductions prior to the point of the 
nonlinearity that causes the IM3.  We will attempt to use changes in measured IP3 to determine whether 
or not the AGC is active in the vicinity of a paired-signal measurement. 
 
Regarding the third limitation, we could choose to substitute a value such as 33.9 (i.e., 15.3 dB converted 
to a linear power ratio) for the required signal-to-noise ratio (SNR
R
) since all of the receivers have 
required signal-to-noise ratios close to this value (according to measurements presented in the SHVERA 
Study
*
); however, that value was measured using white Gaussian noise as the interferer, and, strictly 
speaking, the SNR required to overcome IM3 interference may differ from that of white Gaussian noise 
due to differing spectral shape or perhaps statistical properties.  Consequently, we choose to incorporate 
the SNR
R
 term into the quantity calculated to quantify the receiver’s nonlinearity. 
 
Thus, instead of computing IP3, we choose to compute the quantity, IP3 / SNR
R
1/2
, which was shown in 
Chapter 8 to be related to desired and undesired signal powers at threshold for paired-signal IM3 
interference.  In linear power units, the quantity can be computed by  
 
IP3 / SNR
R
1/2
 = (U
3
 / D)
1/2
 
 
In decibel units, it is given by 
 
(IP3 / SNR
R
1/2
)|
dB
 = 1.5 U|
dB
 - 0.5 D|
dB
  
 
The values of IP3 / SNR
R
1/2
 computed here will be applied to the case of unequal undesired signal levels 
in the next chapter. 
Channel 30 
Table 9-3 and Figures 9-30 and 9-31 present computed values of IP3 / SNR
R
1/2
 (in dB) based on the 
channel-30 measurements of rejection ratios for paired signals. Values were computed only when the 
paired-signal D/U ratio exceeded the summed single-signal D/U ratios by at least 4 dB, as stated above.  
The blank cells represent measurements that did not meet this condition.  See the note at the bottom of the 
table for conversion to IP3. 
  
                                                      
*
 Martin, <SHVERA Study>, 2005, chapter 3. 
 
9-10 
Table 9-3.  IP3/SNR
R
1/2
 Based on Paired-Signal Rejection Measurements at Channel-30 
  IP3 / SNR
R
1/2
 (dB) 
 KÆ -5 -4 -3 -2 -1 1 2 3 4 5 
Receiver 
D 
(dBm) 
N-5/ N-
10 
N-4/ N-
8 
N-3/ N-
6 
N-2/ N-
4 
N-1/ N-
2 
N+1/ 
N+2 
N+2/ 
N+4 
N+3/ 
N+6 
N+4/ 
N+8 
N+5/ 
N+10
A3 -68   0.7 3.9 6.8
D3 -68 -5.5 -9.4 -13.9 -24.7 -17.1 
G4 -68 -1.3 -0.1 -3.4 -10.4 -7.1 -4.8 -1.3
I1 -68 -12.2   5.2 3.7
J1 -68 -3.9  -6.4 -4.3 2.0
M1 -68 -5.4 -5.0 -5.5 -5.2 -6.1 0.6 5.1
N1 -68 6.7 7.6
O1 -68 -5.2 -7.1 -5.4  8.2 7.7
A3 -53 5.4 2.2 -2.8 0.5 -0.8 -0.5 2.7 5.6
D3 -53   -14.9 -21.7  
G4 -53 1.8 -0.7 7.6 5.7 -5.8 -8.2 -1.7 3.2
I1 -53 -4.3 0.4 6.0 5.4 3.8
J1 -53 -1.4 -3.3 -3.1 -1.3 -1.5
M1 -53 -4.7 -4.3 0.9 1.9 -3.2 -2.5 4.7
N1 -53 1.5 4.7 5.2 7.0 6.8
O1 -53 -3.9 -1.6 12.4 9.3 8.8 7.8
Notes: 
•  Outlines  represent receivers and channel offsets subjected to further measurement in the next 
chapter. 
• IP3 can be estimated by adding SNR
R
|
dB
/2 (nominally 15.3 / 2 = 7.6 dB) to the values of 
(IP3 / SNR
R
1/2
)|
dB
 shown. 
 
 
In the table, bold italics with shading indicates values that are believed to correspond to full gain 
operation of the receiver through the mixer (i.e., no RF AGC operation); this judgment was made on the 
basis of the change in computed IP3 as desired signal level changed from -68 dBm to -53 dBm.  (Note the 
lack of shading does not necessarily indicate that AGC did operate in a given region; rather, it may 
indicate that the measurements were insufficient to make a judgment.)  If the AGC was inactive 
throughout the range, one would expect a constant IP3 value.  If the AGC operated throughout the range 
to maintain constant signal levels at the point of nonlinearity, one would expect IP3 to increase by 15 dB 
(same change as D).  The observed IP3 changes from D = -68 dBm to D = -53 dBm within the same 
receiver ranged from -5 to +18 dB.  AGC was judged to have operated throughout the range if the IP3 
change was greater than 10 dB; in that case, corresponding cells for both -68 dBm and -53 dBm are 
marked as having been influenced by AGC.  Pairs with lower IP3 changes were judged to have had AGC 
operation through no more than part of the range, so the low end of the range (D = -68 dBm) was judged 
not to have been influenced by AGC.  Where IP3 change was less than 3.5 dB, we judged that both ends 
of the range were free of AGC operation. 
 
Measurable IM3 effects were not observed on the first-adjacent channel cases (N+1/N+2 and N-1/N-2) 
and many of the second-adjacent channel cases (N+2/N+4 and N-2/N-4).  This is believed to be a result of 
two factors:  (1) AGC-induced gain reductions in each receiver’s front end reduced the amplitudes of IM3 
effects for those channel offsets; and (2) those channel offsets exhibited higher single-channel 
interference effects that would have tended to mask the IM3 effects.   
 
 
9-11 
In Figure 9-30, which presents IP3 / SNR
R
1/2
 values for D = -68 dBm, the expected increase in IP3 with 
increasing offset from channel N due to filtering in the receiver front end is clearly evident for most 
receivers on the right half of the graph where there are more data points.  Such a tendency is seen at D = 
-53 dBm (Figure 9-31) for fewer receivers. 
 
Channel 51 
Table 9-4 and Figures 9-32 and 9-33 present computed values of IP3 / SNR
R
1/2
 (in dB) based on the 
channel-51 measurements of rejection ratios for paired signals. As with the channel-30 measurements, 
values were computed only when the paired-signal D/U ratio exceeded the summed single-signal D/U 
ratios by at least 4 dB.  The blank cells represent measurements that did not meet this condition.  The 
reader is referred to the discussion in the previous section (for channel 30) for information regarding the 
shading in the table.  See the note at the bottom of the table for conversion to IP3. 
 
For D = -68 dBm (Figure 9-32) most receivers exhibit an increase in IP3 with increased spacing between 
the undesired channels and the desired channel.  Such a trend is not clear for D = -53 dBm (Figure 9-33); 
in fact, receiver I1 exhibits a very flat IP3. 
 
 
Table 9-4.  IP3/SNR
R
1/2
 Based on Paired-Signal Rejection Measurements at Channel-51 
  IP3 / SNR
R
1/2
 (dB) 
 KÆ 1 2 3 4 5 6 7 8 
Receiver 
D 
(dBm) 
N+1/ 
N+2 
N+2/ 
N+4 
N+3/ 
N+6 
N+4/ 
N+8 
N+5/ 
N+10 
N+6/ 
N+12 
N+7 
/N+14 
N+8/ 
N+16 
A3 -68   -2.5 1.7 5.3 8.8   
D3 -68 -25.6 -17.8      
I1 -68    4.8 4.7 4.0  4.1 
J1 -68   -3.6 -3.5 2.7 6.1 7.9 9.2 
M1 -68 -3.4 -2.0 3.0 8.2 13.4   
N1 -68   4.9 0.2 0.4 3.1  5.0 
O1 -68     8.7 7.9 7.8 
A3 -53  1.1 -2.5 1.8 5.3    
D3 -53 -18.7       
I1 -53   5.8 5.4 4.8 4.6 4.3 5.3 
J1 -53 4.5 -1.1 0.2 3.6   
M1 -53  1.0 0.2 2.8     
N1 -53 NM NM NM NM NM NM NM 
O1 -53         
Notes: 
• Bold italics with shading indicates probable operation at signal level low enough to avoid AGC 
operation; that assessment is based on the observed change in IP3 / SNR
R
1/2
 as desired signal 
level changes from -68 dBm to -53 dBm.  (See text for details.) 
• NM (for receiver N1) indicates that no measurements of paired-signal D/U ratio were made due to 
equipment failure.  The lack of measurements for receiver N1 at D = -53 dBm precluded the 
opportunity to estimate AGC operation; consequently, none of the N1 measurements at D = 
-68 dBm could be judged to be free of AGC gain changes. 
• IP3 can be estimated by adding SNR
R
|
dB
/2 (nominally 15.3 / 2 = 7.6 dB) to the values of 
(IP3 / SNR
R
1/2
)|
dB
 shown. 
 
 
9-12 
Comparison of Channels 30 and 51 
Figure 9-34 combines the IP3 / SNR
R
1/2
 data measured on channels 30 and 51 for a desired signal power 
of -68 dBm.  The measurements include up to four channel pairs in common.  For most of the TV 
receivers, there is a close match between the measurements in the overlap region. 
SUMMARY DATA 
The IP3 data tables are incompletely filled because of our restriction that we computed the IP3 parameter 
only when the paired-signal D/U ratio exceeded the summed single-signal D/U ratios by at least 4 dB.  
Thus, data could be missing because the IM3 effects are small, because the interference effects of the 
corresponding single-channel undesired signals are high, or because some single-channel measurements 
on which the summed D/U’s are based are at the measurement limit of the test setup.  This makes it 
harder (and in many cases, impossible) to determine worst-case, second-worst, and median values from 
the existing measurements without introducing biases. 
 
We carefully examined both the paired-signal D/U measurements and the corresponding single-channel 
D/U measurements including combinations in which IP3 is not reported due to the paired-signal D/U ratio 
not meeting the threshold requirement described above.  This evaluation was used to determine, where 
possible, the values for median, second-worst, and worst (lowest) IP3 values among the channel pairs.  
The results are shown in Table 9-5 for both channels 30 and 51.  The final three columns combine the 
data from the two desired channels—by averaging when measurements exist on both channels—to 
provide a larger view of the IP3 variation with channel-pair spacing. 
 
 
Table 9-5.  IP3/SNR
R
1/2
 Statistics 
 IP3 / SNR
R
1/2
 (dB) 
 Channel 30 Channel 51 Combined 
 Worst 
2nd 
Worst Median Worst 
2nd 
Worst Median Worst 
2nd 
Worst Median
N-5/N-10 -12.2 -5.5 -4.5 -12.2 -5.5 -4.5
N-4/N-8 -9.4 -7.1 -5.0 -9.4 -7.1 -5.0
N-3/N-6     
N-2/N-4 
N-1/N-2     
N+1/N+2 
N+2/N+4     
N+3/N+6 -17.1 -7.1 -2.7 -17.8 -3.6 -2.0 -17.5 -5.3 -2.4
N+4/N+8 -4.8 -4.3 2.2 1.7 -4.8 -4.3 1.9
N+5/N+10 -1.3 2.0 4.4 0.4 2.7 4.7 -0.5 2.3 4.5
N+6/N+12   3.1 4.0 6.1 3.1 4.0 6.1
N+7/N+14   
N+8/N+16   4.1 5.0 9.2 4.1 5.0 9.2
 
 
The results can be used to compute threshold values of undesired signal power U and D/U ratio when the 
two undesired signals have equal power: 
 
U|
dB
 = (1/3) [2 (IP3 / SNR
R
1/2
)|
dB
 
 + (D – D
MIN
)|
dB
] 
 
where (D – D
MIN
)|
dB
 is computed by converting the desired power levels to linear power units (e.g., mW), 
performing the subtraction, and then converting back to dB. 
 
9-13 
 
The calculation has been performed for desired signal levels of -68 dBm and D
MIN
 + 3 dB, under the 
assumption that D
MIN
 = -84 dBm.  The results are shown in Tables 9-6 and 9-7. 
 
We note that the absence of data for channel pairs N-3/N-6, N-2/N-4, N+2/N+4 and N+7/N+14 in these 
tables does not indicate that IM3 was immeasurable on all of the receivers.  It means, rather, that 
conclusions regarding the worst, second-worst, and median values could not be reached from the 
receivers that were measurable.  On the other hand, the absence of data in the tables for the first-adjacent 
pairs (N-1/N-2 and N+1/N+2) was due to the fact that IM3 effects did not exceed single-channel effects 
on any of the receivers by a sufficient amount to support measurement of IM3 effects. 
 
 
Table 9-6.  Computed Threshold U Due to IM3 for Equal Paired Signals 
 Threshold U for Paired Signals With U
N+K
 = U
N+2K
 = U (dB) 
 D = -68 dBm D = D
MIN
 + 3 dB D = D
MIN
 + 1 dB 
  Worst 
2nd 
Worst Median Worst
2nd 
Worst Median Worst 
2nd 
Worst Median
N-5/N-10 -30.8 -26.3 -25.7 -36.1 -31.7 -31.0 -38.1 -33.6 -33.0
N-4/N-8 -29.0 -27.4 -26.0 -34.3 -32.7 -31.4 -36.2 -34.7 -33.3
N-3/N-6            
N-2/N-4            
N-1/N-2            
N+1/N+2            
N+2/N+4            
N+3/N+6 -34.3 -26.2 -24.3 -39.6 -31.6 -29.6 -41.6 -33.5 -31.5
N+4/N+8 -25.9 -25.6 -21.4 -31.2 -30.9 -26.7 -33.1 -32.8 -28.7
N+5/N+10 -23.0 -21.1 -19.6 -28.3 -26.4 -25.0 -30.3 -28.4 -26.9
N+6/N+12 -20.6 -20.0 -18.6 -25.9 -25.3 -23.9 -27.9 -27.3 -25.9
N+7/N+14           
N+8/N+16 -20.0 -19.3 -16.6 -25.3 -24.7 -21.9 -27.2 -26.6 -23.8
   Computation assumes D
MIN
 = -84 dBm 
 
 
 
9-14 
Table 9-7.  Computed Threshold D/U Due to IM3 for Equal Paired Signals 
 Paired-Signal D/U for U
N+K
 = U
N+2K
 = U (dB) 
 D = -68 dBm D = D
MIN
 + 3 dB D = D
MIN
 + 1 dB 
 Worst 
2nd 
Worst Median Worst 
2nd 
Worst Median Worst 
2nd 
Worst Median
N-5/N-10 -37.2 -41.7 -42.3 -44.9 -49.3 -50.0
-44.9 -49.4 -50.0
N-4/N-8 -39.1 -40.6 -42.0 -46.7 -48.3 -49.6
-46.8 -48.3 -49.7
N-3/N-6       
  
N-2/N-4 
  
N-1/N-2       
  
N+1/N+2 
  
N+2/N+4       
  
N+3/N+6 -33.7 -41.8 -43.7 -41.4 -49.4 -51.4
-41.4 -49.5 -51.5
N+4/N+8 -42.1 -42.4 -46.6 -49.8 -50.1 -54.3
-49.9 -50.2 -54.3
N+5/N+10 -45.0 -46.9 -48.4 -52.7 -54.6 -56.0
-52.7 -54.6 -56.1
N+6/N+12 -47.4 -48.0 -49.4 -55.1 -55.7 -57.1
-55.1 -55.7 -57.1
N+7/N+14       
  
N+8/N+16 -48.1 -48.7 -51.4 -55.7 -56.3 -59.1
-55.8 -56.4 -59.2
    Computation assumes D
MIN
 = -84 dBm 
 
 
 
 
9-15 
 
-50
-40
-30
-20
-10
0
TV Channel Number
Normaliz
ed Spect
rum Level
 (dB
)
30 34 38 4232 3628 40 4431 35 39 4333 3729 41
Shoulders caused by IM3
8-VSB Signal
on Chan 34
 
(a) Single 8-VSB  Signal 
 
-50
-40
-30
-20
-10
0
TV Channel Number
Nor
m
al
iz
ed Spec
tru
m L
e
vel (d
B)
30 34 38 4232 3628 40 4431 35 39 4333 3729 41
Original
8-VSB Signal
on chan 34
Add a 2nd signal
on chan 38
(WGN, but 8-VSB width)
New shoulders
are higher
New spectral bumps appear
at chan's 30 and 42
 
(b) Pair of Equal-Amplitude Signals 
 
Figure 9-1.  Third-Order  Intermodulation Distortion Spectra of Single and Paired Signals 
 
9-16 
-50
-40
-30
-20
-10
0
TV Channel Number
N
o
r
m
a
l
iz
ed S
p
ec
tr
um
 
Le
ve
l (dB
)
30 34 38 4232 3628 40 4431 35 39 4333 3729 41
15 dB
15 dB
5 dB5 dB
Drop both
signals
by 5 dB
Bumps & shoulders
drop by 15 dB
 
(a) Effect of Changing Amplitude of Both Signals 
 
-50
-40
-30
-20
-10
0
TV Channel Number
Normaliz
ed Spec
trum Level 
(dB)
30 34 38 4232 3628 40 4431 35 39 4333 3729 41
10 dB
5 dB
5 dB
Drop ONE
signal
by 5 dB
Closer bump
drops 10 dB
Farther bump
drops 5 dB
 
(b) Effect of Changing Amplitude of One Signal 
 
Figure 9-2.  Third-Order Intermodulation Distortion Spectra of Paired Signals Versus Amplitude 
 
9-17 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channels
Susceptib
ility to
 Inter
f
er
ence 
[D
/U
 at Threshold 
(dB)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
G4
A/74 Max of Pair
 
Figure 9-3.  Paired-Signal D/U of 8 receivers at D = -68 dBm on Channel 30 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
Suscept
i
bi
lit
y t
o
 I
n
t
e
r
f
er
e
n
c
e
 
[D
/U
 at 
Threshold (
d
B)]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 9-4.  Paired-Signal D/U Statistics of 8 receivers at D = -68 dBm on Channel 30  
 
9-18 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channels
Susceptib
ility to
 Inter
f
er
ence 
[D
/U
 at Threshold 
(dB)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
G4
A/74 Max of Pair
 
Figure 9-5.  Paired-Signal D/U of 8 receivers at D = -53 dBm on Channel 30 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
Suscept
i
bi
lit
y t
o
 I
n
t
e
r
f
er
e
n
c
e
 
[D
/U
 at 
Threshold (
d
B)]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 9-6.  Paired-Signal D/U Statistics of 8 receivers at D = -53 dBm on Channel 30 
 
9-19 
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channels
Susceptib
ility to
 Inter
f
er
ence 
[D
/U
 at Threshold 
(dB)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
G4
A/74 Max of Pair
 
Figure 9-7.  Paired-Signal D/U of 8 receivers at D = -28 dBm on Channel 30—1st-Adjacent Only 
 
 
Susceptibility to Interference (D/U at Threshold [dB)]  
N-1/N-2 N+1/N+2 
Worst -18.1 -16.4 
2
nd
 Worst -19.3 -20.1 
Median < -20.7 < -20.2 
Best < -20.7 < -20.2 
 
Figure 9-8.  Paired-Signal D/U of 8 receivers at D = -28 dBm on Channel 30—1st-Adjacent Only 
 
 
9-20 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
S
u
sce
ptibi
l
i
t
y
 to 
Inte
r
f
e
r
ence
 
[D/
U
 a
t
 T
h
resh
old (
d
B
)
]
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Receiver A3
 
Figure 9-9.  Paired-Signal D/U of Receiver A3 on Channel 30 with Summed D/U’s as Reference 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
S
u
sce
ptibi
l
i
t
y
 to 
Inte
r
f
e
r
ence
 
[D/
U
 a
t
 T
h
resh
old (
d
B
)
]
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Receiver D3
 
Figure 9-10.  Paired-Signal D/U of Receiver D3 on Channel 30 with Summed D/U’s as Reference 
 
9-21 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
S
u
sce
ptibi
l
i
t
y
 to 
Inte
r
f
e
r
ence
 
[D/
U
 a
t
 T
h
resh
old (
d
B
)
]
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Receiver I1
 
Figure 9-11.  Paired-Signal D/U of Receiver I1 on Channel 30 with Summed D/U’s as Reference 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
S
u
sce
ptibi
l
i
t
y
 to 
Inte
r
f
e
r
ence
 
[D/
U
 a
t
 T
h
resh
old (
d
B
)
]
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Receiver J1
 
Figure 9-12.  Paired-Signal D/U of Receiver J1 on Channel 30 with Summed D/U’s as Reference 
 
9-22 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
S
u
sce
ptibi
l
i
t
y
 to 
Inte
r
f
e
r
ence
 
[D/
U
 a
t
 T
h
resh
old (
d
B
)
]
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Receiver M1
 
Figure 9-13.  Paired-Signal D/U of Receiver M1 on Channel 30 with Summed D/U’s as Reference 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
S
u
sce
ptibi
l
i
t
y
 to 
Inte
r
f
e
r
ence
 
[D/
U
 a
t
 T
h
resh
old (
d
B
)
]
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Receiver N1
 
Figure 9-14.  Paired-Signal D/U of Receiver N1 on Channel 30 with Summed D/U’s as Reference 
 
9-23 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
S
u
sce
ptibi
l
i
t
y
 to 
Inte
r
f
e
r
ence
 
[D/
U
 a
t
 T
h
resh
old (
d
B
)
]
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Receiver O1
 
Figure 9-15.  Paired-Signal D/U of Receiver O1 on Channel 30 with Summed D/U’s as Reference 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Interfering Channel
S
u
sce
ptibi
l
i
t
y
 to 
Inte
r
f
e
r
ence
 
[D/
U
 a
t
 T
h
resh
old (
d
B
)
]
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Receiver G4
 
Figure 9-16.  Paired-Signal D/U of Receiver G4 on Channel 30 with Summed D/U’s as Reference 
 
9-24 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channels
Susc
ep
tibi
lity
 to 
Inte
r
f
e
r
en
ce
 
[D/U a
t
 T
h
re
sho
l
d
 (dB)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
A/74 Limit
 
Figure 9-17.  Paired-Signal D/U of 7 receivers at D = -68 dBm on Channel 51 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channels
Suscepti
b
ili
ty to 
In
ter
f
er
en
ce 
[
D
/
U
 at
 Th
r
e
sh
ol
d (
d
B
)
]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 9-18.  Paired-Signal D/U Statistics of 7 receivers at D = -68 dBm on Channel 51  
 
9-25 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channels
Susceptibili
ty to
 Interference 
[D/U
 at 
Thres
hold (dB)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
A/74 Limit
 
Figure 9-19.  Paired-Signal D/U of 7 receivers at D = -53 dBm on Channel 51 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channels
Su
scepti
bi
li
ty to In
terferen
ce 
[D/U a
t
 
Thre
shold (d
B)]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 9-20.  Paired-Signal D/U Statistics of 6 receivers at D = -53 dBm on Channel 51 
 
9-26 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channels
Susceptibili
ty to
 Interference 
[D/U
 at 
Thres
hold (dB)]
Measurement Limit
A3
D3
I1
J1
M1
N1
O1
A/74 Limit
 
Figure 9-21.  Paired-Signal D/U of 7 receivers at D = -28 dBm on Channel 51 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channels
Susceptib
ilit
y to
 Inter
f
er
ence 
[D
/U
 at T
h
res
h
old 
(dB)]
Measurement Limit
ATSC A/74 Recommendation
Median (w/Best & Worst)
2nd Worst
 
Figure 9-22.  Paired-Signal D/U Statistics of 6 receivers at D = -28 dBm on Channel 51 
 
9-27 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channel(s)
S
u
scept
i
bi
l
i
t
y
 to I
n
t
e
rfere
n
ce 
[D
/U
 a
t
 Th
r
e
sh
o
l
d
 
(
d
B)
]
Paired-Signal D/U at D = -28 dBm
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -28 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Meas. Limit at D = -28 dBm
Receiver A3
 
Figure 9-23.  Paired-Signal D/U of Receiver A3 on Channel 51 with Summed D/U’s as Reference 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channel(s)
S
u
scept
i
bi
l
i
t
y
 to I
n
t
e
rfere
n
ce 
[D
/U
 a
t
 Th
r
e
sh
o
l
d
 
(
d
B)
]
Paired-Signal D/U at D = -28 dBm
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -28 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Meas. Limit at D = -28 dBm
Receiver D3
 
Figure 9-24.  Paired-Signal D/U of Receiver D3 on Channel 51 with Summed D/U’s as Reference 
 
9-28 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channel(s)
S
u
scept
i
bi
l
i
t
y
 to I
n
t
e
rfere
n
ce 
[D
/U
 a
t
 Th
r
e
sh
o
l
d
 
(
d
B)
]
Paired-Signal D/U at D = -28 dBm
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -28 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Meas. Limit at D = -28 dBm
Receiver I1
 
Figure 9-25.  Paired-Signal D/U of Receiver I1 on Channel 51 with Summed D/U’s as Reference 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channel(s)
S
u
scept
i
bi
l
i
t
y
 to I
n
t
e
rfere
n
ce 
[D
/U
 a
t
 Th
r
e
sh
o
l
d
 
(
d
B)
]
Paired-Signal D/U at D = -28 dBm
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -28 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Meas. Limit at D = -28 dBm
Receiver J1
 
Figure 9-26.  Paired-Signal D/U of Receiver J1 on Channel 51 with Summed D/U’s as Reference 
 
9-29 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channel(s)
S
u
scept
i
bi
l
i
t
y
 to I
n
t
e
rfere
n
ce 
[D
/U
 a
t
 Th
r
e
sh
o
l
d
 
(
d
B)
]
Paired-Signal D/U at D = -28 dBm
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -28 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Meas. Limit at D = -28 dBm
Receiver M1
 
Figure 9-27.  Paired-Signal D/U of Receiver M1 on Channel 51 with Summed D/U’s as Reference 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channel(s)
S
u
scept
i
bi
l
i
t
y
 to I
n
t
e
rfere
n
ce 
[D
/U
 a
t
 Th
r
e
sh
o
l
d
 
(
d
B)
]
Paired-Signal D/U at D = -28 dBm
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -28 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Meas. Limit at D = -28 dBm
Receiver N1
 
Figure 9-28.  Paired-Signal D/U of Receiver N1 on Channel 51 with Summed D/U’s as Reference 
 
9-30 
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Interfering Channel(s)
S
u
scept
i
bi
l
i
t
y
 to I
n
t
e
rfere
n
ce 
[D
/U
 a
t
 Th
r
e
sh
o
l
d
 
(
d
B)
]
Paired-Signal D/U at D = -28 dBm
Paired-Signal D/U at D = -53 dBm
Paired-Signal D/U at D = -68 dBm
Summed D/U's at D = -28 dBm
Summed D/U's at D = -53 dBm
Summed D/U's at D = -68 dBm
Meas. Limit at D = -28 dBm
Receiver O1
 
Figure 9-29.  Paired-Signal D/U of Receiver O1 on Channel 51 with Summed D/U’s as Reference 
 
9-31 
-30
-25
-20
-15
-10
-5
0
5
10
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Undesired Channel Pair
A3
D3
I1
J1
M1
N1
O1
G4
IP
3
/
SNR
R
1/2
(d
B)
D = -68 dBm
 
Figure 9-30.  Third-Order Intercept Point Parameter with Desired Signal = -68 dBm on Channel 30 
-25
-20
-15
-10
-5
0
5
10
15
N-5/N-10 N-4/N-8 N-3/N-6 N-2/N-4 N-1/N-2 N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10
Undesired Channel Pair
A3
D3
I1
J1
M1
N1
O1
G4
IP3
/
SNR
R
1/2
(dB)
D = -53 dBm
 
Figure 9-31.  Third-Order Intercept Point Parameter with Desired Signal = -53 dBm on Channel 30 
 
9-32 
-30
-25
-20
-15
-10
-5
0
5
10
15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Undesired Channel Pair
A3
D3
I1
J1
M1
N1
O1
IP3
/
SNR
R
1/2
(dB)
D = -68 dBm
 
Figure 9-32.  Third-Order Intercept Point Parameter with Desired Signal = -68 dBm on Channel 51 
-25
-20
-15
-10
-5
0
5
10
15
N+1/N+2 N+2/N+4 N+3/N+6 N+4/N+8 N+5/N+10 N+6/N+12 N+7/N+14 N+8/N+16
Undesired Channel Pair
A3
D3
I1
J1
M1
IP3
/
SNR
R
1/2
(dB)
D = -53 dBm
 
Figure 9-33.  Third-Order Intercept Point Parameter with Desired Signal = -53 dBm on Channel 51 
 
9-33 
-30
-25
-20
-15
-10
-5
0
5
10
15
20
N-
5/
N-1
0
N-
4/
N
-8
N-
3
/
N
-
6
N
-2
/
N
-
4
N-
1/
N
-
2
N+
1
/
N
+
2
N+
2
/
N
+
4
N+
3
/N
+
6
N+
4
/N
+
8
N
+5
/N
+
1
0
N+
6
/
N+
1
2
N+
7
/
N
+
1
4
N+
8
/
N
+
16
Undesired Channel Pair
A3 / 30 A3 / 51
D3 / 30 D3 / 51
I1 / 30 I1 / 51
J1 / 30 J1 / 51
M1 / 30 M1 / 51
N1 / 30 N1 / 51
O1 / 30 O1 / 51
G4 / 30
IP3
/
SN
R
R
1/2
(dB)
D = -68 dBm
 
Figure 9-34.  Third-Order Intercept Point Parameter—Channel 30/51 Comparison 
 
10-1 
CHAPTER 10 
IM3 WITH PAIRED SIGNALS OF UNEQUAL AMPLITUDES 
The previous chapter showed the results of D/U measurements on eight receivers for pairs of undesired 
signals that were equal in power.  In this chapter, we select two receivers and one channel offset pair for 
each for additional measurements with unequal undesired input signals.  A model of the results is 
presented to allow measurements from the previous chapter to be to be extended to unequal signal levels 
on the other receivers.  It will be seen that the model performs best at low desired signal levels. 
MEASUREMENTS ON TWO RECEIVERS 
The tests for this chapter were performed on receiver G4 with an undesired signal pair at N+2/N+4 and 
receiver M1 with an undesired signal pair at N+3/N+6.  The selections of these receivers and channel 
offsets were based on their relatively high (though not the highest) differences between paired-signal D/U 
and summed D/U measurements, as presented in Chapter 9.  This selection criterion ensured that the tests 
would be performed on receivers that exhibited easily measurable IM3 effects—at least for paired signals 
of equal levels.  Plots of paired-signal D/U’s and summed D/U’s for these receivers were shown in 
Figures 9-16 and 9-13, respectively, in Chapter 9. 
 
Figure 10-1 shows measurements of threshold signal levels for a pair of undesired signals on N+2 and 
N+4 for receiver G4.  The three curves correspond to three desired signal levels:  -53 dBm, -68 dBm, and 
D
MIN
 + 3 dB.  The measurements were performed by attenuating one undesired signal with respect to the 
other, and then adjusting an attenuator that affected both undesired signal levels until TOV for the DTV 
receiver was found.
*
  The X-axis shows the signal level of the undesired signal on channel N+2.  The Y-
axis shows the level of the undesired signal on channel N+4.  The dashed lines represent a model, to be 
discussed in the next section. 
 
If one moves down the chart toward very low signal levels on N+4, each curve asymptotically approaches 
the respective threshold level for N+2 alone.  Similarly, moving leftward on the chart toward very low 
levels on N+2 causes at least one curve (that for D
MIN
 + 3 dB) to asymptotically approach the threshold 
value for N+4 alone.
†
  In between these two conditions is the region where IM3 between the pair of 
signals is the dominant interference mechanism. 
 
Figure 10-2 shows the same type of measurement performed on receiver M1.  This receiver exhibited 
very odd behavior when the undesired signal power on N+3 (U
N+3
) was between -27 and -21 dBm.  As 
U
N+3
 is increased in this range, the upper two curves exhibit first a dip in U
N+6
 relative to the previous 
trajectory, then a sharp increase, and finally a return back down to the previous trajectory.  The lower 
curve exhibits the dip, but the threshold value of U
N+3
 is reached before subsequent behavior had an 
opportunity to occur.   
 
Figure 10-3 isolates the curve corresponding to a desired signal power of -68 dBm and adds a shading 
effect to illustrate the bizarreness of the behavior.  Consider moving along the solid diagonal line, 
representing equal values of undesired signals, while the desired signal level remains at -68 dBm.  
Beginning at the lower left and moving rightward, DTV reception is visually flawless.  When each of the 
                                                      
*
 Referring to the block diagram in Figure 4-1, the output level of one of the two “AWGN” sources was lowered 
with respect to the other.  Step Attenuator-U, which operates on the sum of the two sources, was then adjusted to 
find the TOV point. 
†
 We note that this is the receiver that exhibited intermittent changes in performance of up to 6 dB, as noted in 
Chapter 7.  Consequently, there is approximately a 6-dB mismatch between three of the threshold shown here and 
equivalent measurements presented elsewhere in this report. 
 
10-2 
undesired signals reaches a level of -26.6 dBm, the TV picture begins to exhibit visual errors.  As the 
level of the undesired signal pair increases about another dB, the picture is lost entirely.  However, a 
further increase causes the picture to come back, and, with a little additional increase, to again become 
error free.  But the new error-free condition exists only over a very narrow range of undesired signal 
levels before picture errors—followed by complete picture failure—occur again as the undesired signal 
levels rise further. 
 
This behavior—of first losing the picture, then regaining it, and then losing it again with increases in 
undesired signal levels—was observed at D = -68 dBm for increases in both undesired signals at equal 
levels, as well as for both D = -68 dBm and D = -53 dBm when U
N+6
 was fixed in power (at an 
appropriate level) and U
N+3
 was varied.  We have no explanation for the observed behavior, but we 
suspect that AGC operation—driven by the signal level on N+3—is involved. 
 
MODELING AND EXTRAPOLATION 
Appendix B demonstrated that a pair of undesired signals at N+K and N+2K can create an IM3-based 
interference effect that is proportional to U
N+K
2
 and to U
N+2K
.  The resulting relationship between the 
threshold values of D, U
N+K
, and U
N+2K
 is given in Chapter 8 as 
 
D = (SNR
R
 / IP3
2
) U
N+K
2
U
N+2K
 + D
MIN
. 
 
where SNR
R
 / IP3
2
 quantifies the IM3 characteristics of the receiver for the specified channel offsets 
(N+K and N+2K) and AGC gain state.  It is clear from the equation that, for a fixed desired signal power, 
increasing the power of one of the undesired signals must result in a reduction of threshold power of the 
other if AGC gain remains constant with the variations.  Furthermore, because the U
N+K
 term is squared 
and the U
N+2K
 term is not, a change in U
N+K
 must be countered with a larger, but opposite, change in 
U
N+2K
, in order to remain at threshold.  For example, a 3 dB increase in U
N+K
 results in the threshold value 
of U
N+2K
 dropping by 6 dB. 
 
Chapter 8 shows that one can compute the threshold of one undesired signal given the level of the other 
by means of the following equations: 
 
U
N+K
|
dB
 = (IP3 / SNR
R
1/2
)|
dB
 + (D|
dB
 - U
N+2K
|
dB
)/2 
 
U
N+2K
|
dB
 = 2(IP3 / SNR
R
1/2
)|
dB
 + D|
dB
 - 2 U
N+K
|
dB
  
 
If there are no AGC-induced changes in tuner gain prior to the point of the nonlinearity that causes IM3 in 
a given situation, then the value of (IP3 / SNR
R
1/2
)|
dB
 is a constant, which can be computed from a 
measurement of threshold made with equal levels of undesired signals on the two channels (i.e., U = U
N+K
 
= U
N+2K
). 
 
 (IP3 / SNR
R
1/2
)|
dB
 = (3 U|
dB
 - D|
dB
 )/2 
 
Values of IP3 / SNR
R
1/2
 computed in this way for eight receivers were shown in Tables 9-3 through 9-5 
and Figures 9-30 through 9-34 of Chapter 9.  (Note that for values of D approaching D
MIN
, the quantity 
D|
dB
 should be replaced by (D – D
MIN
)|
dB
 where the subtraction is performed in linear power units, before 
conversion to dB.) 
 
Figures 10-1 and 10-2 included dashed line representing modeled performance in three regions: 
• A horizontal line represents the threshold value of U
N+2K
 in the absence of an undesired signal at 
N+K; this portion of the model is based on direct measurement of the threshold under that condition; 
 
10-3 
• A vertical line represents the threshold value of U
N+K
 in the absence of an undesired signal at N+2K; 
this portion of the model is based on direct measurement of the threshold under that condition; 
• A diagonal line with a slope of -2 dB/dB represents the IM3 contribution to interference based on the 
formulas above; the required value of IP3 / SNR
R
1/2
 was computed by measurement of threshold with 
U
N+K
 = U
N+2K
. 
 
Thus, three measurements (indicated by the black circles) were required to create each modeled 
performance curve.  Given that those same three measurements exist for each of the eight receivers at 
various channel offsets, we can apply the modeling technique to the other receivers and other channel 
offsets.  We choose to do this only at the lower signal levels (D = -68 dBm and below) since the model 
appears to be a better fit to measurements at such levels.  (At higher levels, AGC is more likely to 
influence the results.) 
 
We note that, even if the AGC did not engage to reduce gain prior to the IM3-generating nonlinearity 
during measurement of the threshold for equal-powered undesired signals, it is still possible that it might 
engage under some non-equal undesired signal conditions.  This is more likely to occur on the basis of 
U
N+K
, since it is likely to be subjected to less attenuation by the tuner’s tracking filter than is U
N+2K
.  If the 
AGC does engage on U
N+K
, the IM3 segment of the model curve will switch at that point from being a 
diagonal line to a horizontal line extending rightward from the AGC engagement point (per Appendix B).  
Since we haven’t attempted to determine AGC thresholds for each case, the reader should recognize 
that the curves shown in the models presented in this chapter will be invalid to the right of such an 
AGC engagement point, if one occurs.  In Chapter 14, we identify one such case. 
 
Models for D = -68 dBm 
Figures 10-4 through 10-10 show paired-signal IM3 models for the eight DTV receivers at a desired 
signal level of -68 dBm.  Each graph represents one channel-offset pair—e.g., N-5/N-10 for the first.  
Each graph contains one curve for each DTV receiver for which an IP3 / SNR
R
1/2
 was determined in 
Table 9-3. 
 
No plots are shown for the first-adjacent channel cases (N-1/N-2 and N+1/N+2) because measurements 
presented in Chapter 9 were not adequate to ensure that a paired-signal IM3 effect was measured for those 
channels (as opposed to individual-channel effects).  Chapter 11 will show, in detail, that the paired-signal 
IM3 on N+1/N+2 is limited to certain regions of the amplitude range for receiver D3 (at least for equal 
amplitude undesired signals). 
 
Each curve shows that the presence of one undesired signal can affect the TV receiver’s susceptibility to 
the other.  As an example, we examine the curve corresponding to receiver D3 in Figure 10-5, where K is 
-4 (i.e., the channel pair is N-4/N-8).  Starting at the top left end of that curve, we see that the TV is 
susceptible to interference from an undesired signal at -13 dBm on channel N+2K (y-axis).  Moving to 
the right, we see that, if an undesired signal is also present on channel N+K at a level exceeding -37 dBm, 
the susceptibility of the receiver to interference on channel N+2K will increase (i.e., the receiver will be 
affected by smaller undesired signals on that channel).  As the level of the undesired signal on channel 
N+K increases, the signal level on channel N+2K necessary to cause interference drops by 2 dB for each 
1-dB increase in power on channel N+K—eventually reaching -45 dBm when the undesired signal on 
channel N+K reaches -21 dB.  Similarly, we can view the undesired signal on channel N+2K as causing 
the TV to be more susceptible to interference from channel N+K. 
 
This example could be applied to the case in which a desired signal—broadcast from a DTV station on 
channel N—is received at a level of -68 dBm at the input to a DTV receiver and another DTV broadcast 
on channel N-4 is the first undesired signal.  The analysis described above could be used to predict the 
vulnerability of the DTV receiver to emissions from a white-space device or another DTV station 
 
10-4 
operating on channel N-8, as a function of the undesired signal level on channel N-4 at the input to the 
receiver. 
 
Table 10-1 summarizes the information in the model plots.  Note that the statistics provided here apply to 
the subset of combinations of channel offsets and TVs for which a measurement of IP3 / SNR
R
1/2
 was 
obtained.  In general, the greatest susceptibility to interference is predicted to be on channel N+2K when a 
large signal is present on N+K.  If no undesired signal is present on channel N+K, the receivers can 
tolerate undesired signal levels as high as -27 to -1 dBm on N+2K.  With an undesired signal on N+2K, 
the receivers are predicted to be susceptible to undesired signal powers as low as -79 to -28 dBm—an 
increase in susceptibility ranging from 23 to 63 dBm.  The signal level on N+K necessary to cause such 
an increase in interference susceptibility can range from -28 to -3 dBm, but the susceptibility increase 
begins at levels of -45 to -24 dBm on channel N+2K. 
 
If one were interested in determining what undesired signal level could cause interference to DTV 
reception, one might consider two different approaches with respect to IM3 from paired signals: 
(1) Identify a level that could cause interference if a similar signal level happens to occur at another 
channel offset that would place IM3 products in the desired channel;  
(2) Identify levels that will cause interference given a specific knowledge of signal levels that already 
exist on other channels that might combine with the signal of interest to cause IM3. 
 
In case (1) analysis could be performed based on the equal-power paired-signal test results from 
Chapter 9 (or summary charts presented in Chapter 15 of this report).  In case (2), the modeled results in 
this chapter could provide a basis for analysis. 
 
 
Table 10-1.  Range of Impact of IM3 from Pairs of Undesired Signals When D = -68 dBm 
Statistics of Undesired Signal Levels 
(dBm) 
 
Min. Median Mean Max. 
Standard 
Deviation
Susceptibility increase on N+K due to N+2K:   
Susceptibility to N+K begins increasing at U
N+2K
= -79.4 -46.6 -47.7 -28.0 11.9
Susceptibility to N+K reaches max. at U
N+2K
= -27.0 -5.2 -6.9 -1.4 5.4
U
N+K
 threshold before increase in susceptibility -28.2 -12.5 -13.3 -2.6 5.7
U
N+K
 threshold after increase in susceptibility -45.2 -34.2 -33.7 -24.2 5.4
Net increase in susceptibility caused by U
N+2K
 11.7 19.4 20.4 31.6 5.1
Susceptibility increase on N+2K due to N+K:   
Susceptibility to N+2K begins increasing at U
N+K
= -45.2 -34.2 -33.7 -24.2 5.4
Susceptibility to N+2K reaches max. at U
N+K
= -28.2 -12.5 -13.3 -2.6 5.7
U
N+2K
 threshold before increase in susceptibility -27.0 -5.2 -6.9 -1.4 5.4
U
N+2K
 threshold after increase in susceptibility -79.4 -46.6 -47.7 -28.0 11.9
Net increase in susceptibility caused by U
N+K
 23.5 38.8 40.8 63.1 10.1
 
Models for D = D
MIN
 + 3 dB 
Figures 10-11 through 10-17 show plots of models corresponding to a desired signal level that is 3 dB 
above the D
MIN
 value that was measured for each receiver.  Since we made no measurement of paired-
signal thresholds at D
MIN
 + 3 dB (except for those shown in Figures 10-1 and 10-2, which were not used 
in Figures 10-11 through 10-17), the plots are based on an assumption that  IP3 / SNR
R
1/2
 is the same at 
this low desired signal level as it was at D = -68 dBm.  To help ensure the validity of this assumption, 
models are shown only for DTVs and channel offsets for which Table 9-3 indicates that -68 dBm was 
likely to be low enough to avoid AGC operation (based on the change in  IP3 / SNR
R
1/2
 as desired the 
 
10-5 
signal decreased from -53 dBm to -68 dBm).  Nonetheless, this extrapolation to lower signal levels 
constitutes an additional potential source of error in these plots that was not present in the Figures 10-4 to 
10-10. 
 
Comparing the modeled curve for receiver M1 in Figure 10-15 to the measurements for that receiver at 
D = D
MIN
 + 3 dB in Figure 10-2, we see that the modeled threshold along the equal-power line is about 
-32 dBm, whereas the measured threshold is about -30 dBm.  This 2 dB error can be attributed to the 
anomalous behavior of the receiver for D = -68 dBm in the vicinity of the equal power line (Figure 10-2).  
In fact, if the model had been computed from the IP3 / SNR
R
1/2
 value for D = -53 dBm, it would match 
the measurement within 0.3 dB. 
 
Table 10-2 summarizes the information in the model plots.  Note that the statistics provided here apply to 
the subset of combinations of channel offsets and TVs for which a measurement of IP3 / SNR
R
1/2
 was 
obtained and for which AGC was judged to be inactive at -68 dBm (for tuner stages prior to the 
nonlinearity that causes the observed IM3) so that IP3 could be assumed constant below that level.  In 
most cases, the greatest susceptibility to interference is predicted to occur on channel N+2K, where the 
susceptibility may increase by amounts ranging from 6 to 59 dB when a large signal is present on N+K. 
 
 
Table 10-2.  Range of Impact of IM3 from Pairs of Undesired Signals When D =D
MIN
 + 3 dB 
Statistics of Undesired Signal Levels (dBm) 
Min. Median Mean Max. 
Standard 
Deviation 
Susceptibility increase on N+K due to N+2K:   
Susceptibility to N+K begins increasing at U
N+2K
= -68.7 -39.5 -41.3 -12.4 14.4
Susceptibility to N+K reaches max. at U
N+2K
= -32.3 -9.0 -11.1 -2.9 6.7
U
N+K
 threshold before increase in susceptibility -38.5 -23.9 -23.7 -12.3 6.4
U
N+K
 threshold after increase in susceptibility -50.0 -39.4 -38.9 -30.7 5.2
Net increase in susceptibility caused by U
N+2K
 3.2 15.4 15.1 29.6 7.2
Susceptibility increase on N+2K due to N+K:   
Susceptibility to N+2K begins increasing at U
N+K
= -50.0 -39.4 -38.9 -30.7 5.2
Susceptibility to N+2K reaches max. at U
N+K
= -38.5 -23.9 -23.7 -12.3 6.4
U
N+2K
 threshold before increase in susceptibility -32.3 -9.0 -11.1 -2.9 6.7
U
N+2K
 threshold after increase in susceptibility -68.7 -39.5 -41.3 -12.4 14.4
Net increase in susceptibility caused by U
N+K
 6.4 30.8 30.2 59.3 14.3
 
 
10-6 
-70
-60
-50
-40
-30
-20
-10
0
-80 -70 -60 -50 -40 -30 -20 -10 0
Undesired Signal Level on N+2 (dBm)
U
n
d
esi
r
e
d
 S
i
g
n
a
l
 
L
evel
 o
n
 N
+
4 
(
dB
m
)
D = -53 dBm
D = -68 dBm
D = Dmin + 3 dB
Line of Equal U Levels
Model
Model Basis
 
Figure 10-1.  Threshold U for Paired, Unequal Undesired Signals on Receiver G4 
-70
-60
-50
-40
-30
-20
-10
0
-70-60-50-40-30-20-10 0
Undesired Signal Level on N+3 (dBm)
U
n
d
esi
r
e
d
 S
i
g
n
a
l
 L
evel
 
o
n
 N
+
6 
(
dB
m
)
D = -53 dBm
D = -68 dBm
D = Dmin + 3 dB
Line of Equal U Levels
Model
Model Basis
 
Figure 10-2.  Threshold U for Paired, Unequal Undesired Signals on Receiver M1 
 
10-7 
-60
-50
-40
-30
-20
-10
-60 -50 -40 -30 -20 -10 0
Undesired Signal Level on N+3 (dBm)
U
n
d
esi
r
e
d
 S
i
g
n
a
l
 L
evel
 
o
n
 N
+
6 
(
dB
m
)
D = -68 dBm
Line of Equal U Levels
"PERFECT" DTV RECEPTION
(free of visual errors)
PICTURE ERRORS
OR NO PICTURE
 
Figure 10-3.  Threshold U for Paired, Unequal Undesired Signals on Receiver M1 at D = -68 dBm 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
de
s
i
r
e
d S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
D3
I1
J1
M1
N1
O1
G4
Equal Power
D = -68 dBm
Channel Pair = N-5/N-10
 
Figure 10-4.  Modeled Thresholds for N-5/N-10 with D = -68 dBm 
 
10-8 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
d
esi
r
e
d
 S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
D3
M1
O1
G4
Equal Power
D = -68 dBm
Channel Pair = N-4/N-8
 
Figure 10-5.  Modeled Thresholds for N-4/N-8 with D = -68 dBm 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
de
s
i
r
e
d S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
D3
M1
O1
G4
Equal Power
D = -68 dBm
Channel Pair = N-3/N-6
 
Figure 10-6.  Modeled Thresholds for N-3/N-6 with D = -68 dBm 
 
10-9 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
d
esi
r
e
d
 S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
D3
M1
G4
Equal Power
D = -68 dBm
Channel Pair = N+2/N+4
 
Figure 10-7.  Modeled Thresholds for N+2/N+4 with D = -68 dBm 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
de
s
i
r
e
d S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
A3
D3
J1
M1
G4
Equal Power
D = -68 dBm
Channel Pair = N+3/N+6
 
Figure 10-8.  Modeled Thresholds for N+3/N+6 with D = -68 dBm 
 
10-10 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
d
esi
r
e
d
 S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
A3
I1
J1
M1
N1
O1
G4
Equal Power
D = -68 dBm
Channel Pair = N+4/N+8
 
Figure 10-9.  Modeled Thresholds for N+4/N+8 with D = -68 dBm 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
de
s
i
r
e
d S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
A3
I1
J1
M1
O1
G4
Equal Power
D = -68 dBm
Channel Pair = N+5/N+10
 
Figure 10-10.  Modeled Thresholds for N+5/N+10 with D = -68 dBm 
 
10-11 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
d
esi
r
e
d
 S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
I1
J1
M1
N1
O1
G4
Equal Power
D = Dmin + 3 dB
Channel Pair = N-5/N-10
 
Figure 10-11.  Modeled Thresholds for N-5/N-10 with D = D
MIN
 + 3 dB 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
de
s
i
r
e
d S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
M1
O1
G4
Equal Power
D = Dmin + 3 dB
Channel Pair = N-4/N-8
 
Figure 10-12.  Modeled Thresholds for N-4/N-8 with D = D
MIN
 + 3 dB 
 
10-12 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
d
esi
r
e
d
 S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
D3
M1
Equal Power
D = Dmin + 3 dB
Channel Pair = N-3/N-6
 
Figure 10-13.  Modeled Thresholds for N-3/N-6 with D = D
MIN
 + 3 dB 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
de
s
i
r
e
d S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
D3
M1
G4
Equal Power
D = Dmin + 3 dB
Channel Pair = N+2/N+4
 
Figure 10-14.  Modeled Thresholds for N+2/N+4 with D = D
MIN
 + 3 dB 
 
10-13 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
d
esi
r
e
d
 S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
A3
J1
M1
G4
Equal Power
D = Dmin + 3 dB
Channel Pair = N+3/N+6
 
Figure 10-15.  Modeled Thresholds for N+3/N+6 with D = D
MIN
 + 3 dB 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
de
s
i
r
e
d S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
A3
I1
J1
M1
N1
O1
G4
Equal Power
D = Dmin + 3 dB
Channel Pair = N+4/N+8
 
Figure 10-16.  Modeled Thresholds for N+4/N+8 with D = D
MIN
 + 3 dB 
 
10-14 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
n
d
esi
r
e
d
 S
i
g
na
l
 Le
v
e
l
 on
 
N
+
2
K
A3
I1
J1
M1
O1
G4
Equal Power
D = Dmin + 3 dB
Channel Pair = N+5/N+10
 
Figure 10-17.  Modeled Thresholds for N+5/N+10 with D = D
MIN
 + 3 dB 
 
 
11-1 
CHAPTER 11 
SINGLE AND PAIRED REJECTION RATIOS VERSUS DESIRED 
SIGNAL LEVEL—A DETAILED EXAMPLE 
This chapter presents measurements of D/U ratio and of U at threshold as functions of desired signal level 
D based on a detailed set of measurements for one DTV receiver, D3.  The tests included single 
interferers as well as paired interferers spaced to place IM3 products in the desired signal channel.  The 
amount of effort involved in making these measurements precluded such an evaluation of other receivers; 
however, the results from this one set of measurements provide some insight into the behavior of 
interference susceptibilities that will aid in developing an understanding of test results for other DTV 
receivers. 
 
Figure 11-1 shows measurements of D/U versus D for receiver D3—measured for the desired channel N 
= 51.  Chronologically, this was the first rejection performance work done under this test program apart 
from a crude set of measurements used to select the TV for this test.  The test was performed to provide 
some insight into the variation of certain interference effects with desired signal level prior to testing at 
fixed desired signal levels.
*
 
 
The dashed curves on the plot represent measurements using pairs of equal-level undesired signals at 
N+1/N+2, N+2/N+4, N+3/N+6, and N+4/N+8.  The solid lines represent measurements with single 
undesired signals on channels N+1 through N+7, N+14, and N+15.  Curves labeled “N+1 (No Filter)” and 
“N+1/N+2 (No Filter)” are measurements performed with the filter removed from the test setup; this 
allowed the test setup to create higher interfering signal levels.
†
 
 
Though the number of curves on each plot makes it difficult to identify individual results, the curves are 
all combined on one graph to illustrate both the diversity and the commonality in the behavior of the 
various interference phenomena.  The data will be dissected into separate charts later in the chapter. 
  
The first point on each curve is a measurement with the desired signal at approximately 1 dB above the 
minimum signal D
MIN
 for the TV.  (D
MIN
 is the desired signal level corresponding to the threshold of 
visibility (TOV) of picture degradation in the absence of interference.)  The second point is 3 dB above 
D
MIN
.  The third point is at D
MIN
 + 5 dB, and all subsequent points are at 5-dB intervals. 
 
Note the following reference levels on the graph. 
• Diamonds ( ) on the X-axis mark the following desired signal levels. 
                                                      
*
 Our expectation at the onset of this project was that the primary vulnerabilities that would need to be investigated 
would be the first-adjacent channel response, the mixer image response, and possibly the effects of third-order 
intermodulation (IM3) distortion for signal pairs.  In order to gain an understanding of the nature of these 
interference vulnerabilities, we decided to conduct detailed measurements on a TV receiver that exhibited high 
enough interference susceptibilities for those cases to permit measurement over a wide range of signal levels.  Prior 
to procurement of custom filters for this project, measurements on first-adjacent channels were not feasible with the 
signal generators available at the FCC Laboratory; however, crude measurements could be made at other channel 
spacings.  Such measurements were performed on eight TVs (including three that were used in tests reported in this 
document) with the goal of identifying a TV that exhibited relatively high mixer image and third-order 
intermodulation (IM3) effects.  The mixer image measurements were performed using a single undesired signal at 
N+15 and a desired signal level of -68 dBm; IM3 measurements were performed for undesired signal pairs at 
N+1/N+2 and at N+2/N+4 with each undesired signal set for -14 dBm.  (It was thought, incorrectly, that such high 
signal levels might be necessary to enable observation of IM3 effects.)   The selected receiver (D3) had the highest 
IM3 effects and the second highest mixer image response among the three tested receivers at these levels. 
†
 Removal of the filter was possible because the high D/U ratios (above -20 dB) for those few measurements 
reduced the otherwise stringent requirement on splatter of the undesired signal into the desired channel. 
 
11-2 
? -84 dBm is the received level of a DTV signal at the edge of coverage of a broadcast station (per 
OET-69); this is also marked by a black vertical dashed line. 
? -68 dBm, -53 dBm, and -28 dBm are the desired signal levels that the ATSC designated as 
“weak”, “moderate”, and “strong”, respectively; most D/U measurements shown elsewhere in this 
report were made at these three levels. 
? Not shown (off scale to the right) is -8 dBm, the largest expected DTV signal. 
• A diagonal dashed line corresponds to an undesired signal level of -8 dBm—equal to the maximum 
DTV signal that is expected at the input to a TV receiver. 
 
Figure 11-2 shows the same measurement data as Figure 11-1, but it is presented as undesired signal level 
at threshold rather than as D/U ratios.  Thus, in Figure 11-2, high points on the graph represent high 
rejection performance rather than high susceptibility to interference. 
 
Figures 11-3 and 11-4 are plots of slope of log-D versus log-U and of log-U versus log-D, respectively.  
The slope information will aid in evaluating order of the interference mechanisms represented by the 
curves.  The slope of log-D versus log-U directly indicates order of the interference mechanism (except 
where AGC affects the results and at desired signal levels near D
MIN
); thus, an interference mechanism 
that is third-order in terms of the undesired signal level should have a slope of 3 in Figure 11-3.  Figure 
11-4 is included because the interference effect of certain interference mechanisms such as cross-
modulation is directly proportional to desired signal as well as being dependent on the undesired signal 
level; the result is predicted to be an infinite slope in Figure 11-3 or a zero slope in Figure 11-4.  Slopes 
were computed from adjacent pairs of measurement points from Figures 11-1 and 11-2 and are plotted at 
the midpoint of the pair.  Thus, for example, the slope that was computed based on measurements at 
desired signal levels of -68 dBm and -63-dBm was plotted at -65.5 dBm. 
 
GENERAL OBSERVATIONS 
Returning to Figure 11-1, we make some general observations regarding the interference rejection results. 
 
First, in examining the left portion of the graph—for desired signals less than about -60 dBm—we note 
the following. 
• Some of the curves are—for the most part—horizontal lines, indicating constant D/U ratio as desired 
signal is varied.  This result is true for interference at N+14 and N+15 because interference effects on 
these channels result from the receiver’s mixer image—a linear interference mechanism.  Less 
expected is that the curves for N+1 and N+2 are flat, a topic that will be discussed later.  Totally 
unexpected is the fact that interference from a pair of undesired signals at N+1/N+2 also exhibits a 
flat D/U and that the D/U is nearly identical to that for N+1 alone, over part of the curve.  N+1/N+2 is 
expected to generate IM3—a third order nonlinear process (third order when the amplitudes of both 
undesired signals are adjusted together). 
• Some of the curves exhibit upward slopes of log(D/U) versus log-D.  Many of the slopes appear to be 
identical, but others—such as those for N+3 and for N+7—are steeper. 
• While most of the curves appear to be nearly straight lines, all bend upward at their left-hand ends, 
where the desired signal level approaches D
MIN
. 
 
Next, looking at the middle to right hand portions of the graph, we note the following. 
• Most of the curves that are upward-sloped on the left exhibit an abrupt bend—becoming 
approximately horizontal.  The D/U’s, which were increasing as straight lines (on the dB versus dB 
scale),
*
 dip down gradually after the bend, but then eventually begin to increase again.  The bend is 
believed to be associated with AGC action in the receiver for reasons discussed in Chapter 8. 
                                                      
*
 Note that linear appearance on a log-log plot (such as these plots in decibels) does not necessarily indicate a linear 
function but rather a fixed-order function such as X
M
 where M is a fixed exponent.  The slope of the plot of D versus 
 
11-3 
• The abrupt bend in most of the curves occurs at the same desired signal level of approximately 
-53 dBm.  We believe that the receiver AGC begins reducing RF gain at this desired signal level. 
• One curve, that for interference at N+2/N+4, exhibits its bend at a lower desired signal level—
approximately -58 dBm.  In this case, it is considered likely that AGC action is initiated—not based 
on the desired signal level alone, but rather based on the total power reaching an AGC sensing point 
in the receiver.  In this case, the approximately -37 dBm undesired power on channel N+2 (see 
Figure 11-2) is sufficient that—after filtering within the tuner—it combines with the desired signal 
power to engage the AGC.   
• The curves for N+1 and for the N+1/N+2 pair gradually shift from their horizontal slope on the left of 
the graph to upward slopes on the right half of the plot.  The flat portions at the left side of both 
curves is believed to be due to AGC action similar to that discussed in the previous bullet; in this 
case, AGC action based on the level of the undesired signal at N+1 exceeding a threshold on the order 
of -40 dB or lower results in the curves being flat in the left hand region.  In the upward-sloped 
region, the D/U of the N+1/N+2 curve tracks that of the N+1 curve, but at a D/U ratio that is about 
3-dB higher than that for N+1.  This portion of both of the curves is likely to be the result of IM3 
occurring either at an early RF amplifier stage that is not controlled by the AGC, or at the same point 
as the other IM3 examples (probably the mixer), but after the AGC’s gain reduction capability for 
those stages has reached its limit. 
 
IM3 WITH PAIRED UNDESIRED SIGNALS 
We start by assessing the paired-signal results, because those results enable us to understand the N+1 
results.  To provide a simpler view, Figure 11-5 was created to show D/U for the four tested pairs (the 
dashed lines) and for the single interferers associated with them (solid lines).  All other data has been 
removed from the plot. 
 
N+2/N+4, N+3/N+6, and N+4/N+8 
We first examine the paired signal test results that did not involve the first adjacent channel.  We note that 
the D/U curves for N+2/N+4, N+3/N+6, and N+4/N+8 (Figure 11-5) have two distinctive regions:  an up-
sloped region on the left and more horizontal, but dish-shaped region to the right.  The dish-shaped region 
slopes downward initially, but then begins to trend upward (on two of the three curves) on the far right.  
The two regions are separated by an abrupt bend in each curve, occurring at a desired signal level of 
either -53 dBm or -58 dBm. 
 
Looking at the sloped portions of the curves to the left of D = -58 dBm, we see that the D/U ratios (and 
thus the interference effects) of the paired signals (dashed curves) are higher than those of the 
corresponding single signals.
*
  For example, the D/U ratios for N+2/N+4 (the top curve) are significantly 
higher than those for N+2 and N+4 individually.  This implies that the pair of undesired signals combines 
synergistically so as to create an interference effect larger than the sum of the interference effects of the 
individual signals.  This synergism is the result of third-order intermodulation (IM3) distortion occurring 
somewhere in the DTV receiver—most likely in the mixer. 
 
Looking at the slope of log-D versus log-U (Figure 11-3), we see that the slope measurements of the 
paired signal curves in this region fall between 3 and 4 dB/dB.  The expected slope for an IM3 process 
with both undesired signals varied in amplitude together is 3 dB/dB.  We take this as further confirmation 
of IM3 as the primary cause of the paired signal interference. 
                                                                                                                                                                           
U on a log-log scale provides the value of the exponent that describes the order of the interference mechanism.  A 
linear process (M=1) will have a slope of one on a log-log plot. 
*
 Measurements of D/U for N+8 were not performed in this test; however, measurements presented in Chapter 6 
show that D/U for N+8 is quite small. 
 
11-4 
 
While we had expected the paired-signal IM3 to be significant only at high signal levels, it is interesting 
to note in Figure 11-1 that IM3 for the N+2/N+4 pair causes higher interference susceptibility than any 
single-channel interferer—including the first-adjacent channel—all the way down to a desired signal level 
only 1 dB above the lowest signal level at which the receiver can operate (D
MIN
).  Similarly, susceptibility 
to interference from an undesired signal pair at N+3/N+6 is greater—even at signal levels down to 
D
MIN
 + 1 dB—than from any single channel except the first adjacent channel, and it exceeds even that 
except when the desired signal drops to D
MIN
 + 3 dB or below.  Even on N+4/N+8, interference 
susceptibility exceeds that of any single channel, except the first adjacent one, down to 2 dB above D
MIN
; 
mixer image interference at N+15 barely surpasses N+4/N+8 interference effects only at D
MIN
 + 1 dB.   
 
We also note in Figure 11-5 that D/U ratios of the N+K/N+2K pairs fall gradually as K increases.  The 
fact that IM3 is observed even for N+4/N+8 suggests that it occurs at a place in tuner at which undesired 
signals all the way out to N+8 (for the N+4/N+8 pair) are still present.  The gradual fall with increasing K 
suggests that the nonlinearity causing these IM3 effects is after the tracking filter in the RF section.  We 
surmise that the IM3 effects observed on this part of each of the three paired-signal curves being 
discussed in this section is caused by nonlinearity at the mixer. 
 
We now note the abrupt bend in each of the three curves (for N+2/N+4, N+3/N+6, and N+4/N+8), 
representing a change in character of the interference effect.  Based on discussions in the Chapter 8, we 
attribute that bend and subsequent leveling of the D/U plots to AGC action in the tuner.  Specifically, we 
conclude that, beginning at this bend, the AGC adjusts gain downward with further increases in signal 
level and that this gain adjustment occurs in the RF amplifier stage since the nonlinearity that is affected 
is probably at the mixer. 
 
We further note that the curves for N+3/N+6 and N+4/N+8 exhibit this AGC bend at identical levels of 
desired signal, D = -53 dBm.  Four of the single-interferer curves in Figure 11-1 exhibit this same bend 
point.  We conclude that the RF AGC engages whenever the desired signal exceeds -53 dBm. 
N+2/N+4—A Different AGC Point 
On the other hand, the N+2/N+4 curve exhibits its AGC bend when D = -58 dBm.  This suggests an AGC 
action that is influenced by the undesired signal level.  The bend occurs when the undesired signal level 
applied to both N+2 and N+4 reaches -36.5 dBm (Figure 11-2).  The desired-signal-driven AGC 
thresholds for the other curves occur when undesired signals on channel N+3 or beyond are at signal 
levels ranging from -31 to -17 dBm (Figure 11-2); these levels apparently do not result in AGC action 
since all bends occurred at the same desired signal level.  In particular, the bend in the N+4 curve at D = 
-53 dBm and U = -25 dBm suggests that the early bend in the N+2/N+4 curve is entirely due to N+2 as 
opposed to N+4. 
 
Hence, it is clear that the RF gain reduction in at least this one case—and maybe in all—is triggered by 
total signal level (desired plus undesired) at a point in the tuner where total power from each channel is 
influenced by filtering.  Assuming both AGC mechanisms are the same, the fact that AGC action begins 
at a desired signal level of -53 dBm for most cases, but begins at a desired signal of -58 dBm in the case 
at hand implies that the undesired signal is making up the difference—at total of about -54.7 dBm referred 
to the input
*
 to get to the apparent AGC threshold of -53 dBm referred to the input.  (This suggests that, 
had the desired signal level been much smaller, the threshold at which N+2 would engage the AGC would 
be 1.7 dB higher than that observed—i.e., at U = -36.5 dBm + 1.7 dB = -34.8 dBm).  Since the undesired 
input level on channel N+2 at the point of RF AGC action occurred at an input level of -36.5 dBm, it 
appears that the total power in the N+2 channel is attenuated by filtering by about 18.2 dB (i.e., 
-36.5 dBm - [-54.7 dBm]) relative to the desired channel.    
                                                      
*
 This value was obtained by converting -53 dBm and -58 dBm to milliwatts, subtracting the two, and then 
converting back to dBm. 
 
11-5 
 
We examine the N+2 curve (Figure 11-2) to see whether it exhibits evidence of RF AGC action beginning 
at an undesired signal level in the region of -34.8 to -36.5 dBm.  There are no obvious AGC bends in the 
curve; however, the undesired signal level for points on the curve ranges from -34.1 to -11.7 dBm.  Thus 
it is likely that the AGC reductions of RF gain began just prior to the left-most point on the N+2 curve 
and continued throughout the curve. 
Deviation From Straight Line Near Receiver Threshold 
The portion of each of the three D/U curves (N+2/N+4, N+3/N+6, and N+4/N+8 in Figure 11-5) to the 
left of the AGC bend exhibits a primarily straight-line appearance with a slope consistent with a third-
order interference process; however, there is an upward deviation from the straight line at the left end of 
each curve.  Chapter 8 predicted that such deviations will occur as the desired signal level approaches 
D
MIN
 for the receiver.  Specifically, Table 11-2 predicts that, for a third-order interference process, the 
undesired signal at threshold will deviate from a straight-line projection by 1.0 dB when D = D
MIN
 + 3 dB 
and by 2.3 dB when D = D
MIN
 + 1 dB.   
 
By using these predicted deviations together with the theoretical slope for a third-order process 
(Table 11-1) and a single measurement at D = -68 dBm (15 dB above D
MIN
), one can predict the actual 
measured threshold values of undesired signal at D
MIN
 + 3 dB and  D
MIN
 + 1 dB with errors not exceeding 
0.6 dB for the former prediction and 1.2 dB for the latter in this particular case. 
 
N+1/N+2 
Following the trend that was observed in the sub-section titled, “N+2/N+4, N+3/N+6, and N+4/N+8”, one 
might expect that the curve for N+1/N+2 would look just like those for the other three paired signals 
except that it would be positioned above those curves on the D/U chart (Figure 11-5)—indicating that the 
TV is more susceptible to interference from the adjacent-channel pair than from the pairs that are further 
away from the desired channel.  The measurements dash this expectation.  Not only is the N+1/N+2 curve 
below the other three for most of its trajectory, but its shape is completely different—flat where the others 
are sloped and sloped where the others are essentially flat. 
 
Also unexpected is that the N+1 curve overlays the N+1/N+2 curve for desired signals ranging from 
-73 dBm to -68 dBm.  Apparently in this region, adding an undesired signal at N+2 to an existing one at 
N+1 causes no additional interference effect.  Clearly, in this region, the IM3 contribution from the pair is 
negligible compared to the interference effect of N+1 alone, and the overall interference effect is less than 
that of the other pairs. 
 
We can take a clue as to the reason for the unexpected behavior from the observations about AGC in the 
previous section.  We note that a signal level exceeding -34.8 dBm on channel N+2 is able to cause the 
AGC to reduce the RF amplifier.  (The observed threshold of -36.5 dBm was slightly less because the 
desired signal was also contributing to the energy seen by the AGC sampling point.)  Given that RF AGC 
operation begins at a level of -53 dBm on channel N, at -34.8 dBm on channel N+2, and a higher, but 
unknown level on channels N+3 and beyond, we would expect that an undesired signal on channel N+1 
would activate the RF AGC at a level somewhere between the channel N and channel N+1 values—i.e., 
between -53 and -34.8 dBm.  Furthermore, a typical bandpass filter at channel N would take a much 
bigger bite out of a signal at N+2 than out of a signal at N+1, so we might expect the AGC threshold for 
N+1 to be closer to that at N than to that at N+2.  This suggests that the AGC threshold for N+2 is 
somewhere between that for N and the midpoint between those for N and N+2; hence, we would expect a 
threshold between -44 and -53 dBm.  Looking at the U versus D plot in Figure 11-2, we see that only the 
left-most point on the N+1/N+2 curve has a U value within this range; all other points are above the 
range—suggesting that the RF AGC is active throughout the plotted curve, with the possible exception of 
the left-most point. 
 
 
11-6 
RF AGC operation throughout the curve provides a reason for the different behavior of N+1/N+2 
interference relative to the other pairs.  At all measured threshold points for N+1/N+2 (except possibly 
D
MIN
 + 1 dB) the RF gain is reduced, reducing the IM3 at the mixer—apparently to a level below the N+1 
interference mechanism that is at work. 
 
The N+1/N+2 D/U ratio matches that for N+1 virtually exactly for desired signal levels of -73 to -68 dBm 
or so.  Outside of this range, the paired signal shows more interference effect than N+1 alone.  Since N+2 
has a relatively low D/U ratio, the fact that interference susceptibility for the pair significantly exceeds 
that of N+1 indicates that the IM3 of the pair becomes significant outside of this narrow range.  Why? 
Desired Signal Levels Above -58 dBm  
At desired signal levels above -58 dBm, both the N+1/N+2 and N+1 plots of D/U (Figure 11-5) become 
upward sloped.  In this region it is assumed that the interference mechanism for both single adjacent-
channel case (N+1) and the paired signal case (N+1/N+2) is IM3.  (Recall that IM3 creates shoulders 
around a single undesired signal that spill into the adjacent channel.  Thus, for the first-adjacent-channel 
case, a pair of signals is not needed to create IM3 in the desired channel.)  The slope of log-D versus log-
U (Figure 11-3) goes to about three—consistent with IM3.  The D/U of N+1/N+2 exceeds that of N+1 
alone by about 3 dB—consistent with the expectation that more IM3 will be generated in the paired-signal 
case than in the single-signal case. 
 
The fact that more IM3 occurs with N+1/N+2 than with N+1 in the right-hand part of the curves suggests 
that the interference is being generated by a nonlinearity occurring before the IF filter; so, again, it 
appears that the relevant nonlinearity is likely to be at the mixer or an earlier point in the tuner.  The 
upturn of the D/U curves at a desired signal level around -63 dBm indicates that either the interference is 
being generated by a nonlinearity at an RF amplifier stage that is not controlled by the AGC or the 
interference is being generated by nonlinearity at the mixer and the AGC range for the RF amplifier has 
run out—i.e., the RF amplifier is at its minimum gain.  As a result, the D/U ratio rises with further 
increases in signal level.  In Figure 11-5 the beginnings of a similar trend can be seen at the right-hand 
end of the curve for N+2/N+4, where D/U begins rising, but does so at a level below the D/U of 
N+1/N+2, indicating that N+2/N+4 generates less IM3 than does N+1/N+2 due to rolloff of the tracking 
filter in the RF stage. 
Desired Signal Levels Below -73 dBm 
The upturn of D/U for N+1/N+2 relative to N+1 below -73 dBm was unexpected.  Measurements were 
repeated to confirm the behavior; the result was the same. 
 
In general all of the D/U curves exhibit an upturn from their straight-line projections as one moves 
leftward on the plot—toward D
MIN
.  Such an upturn is predicted in Chapter 8 (Table 8-2).  The increase in 
D/U for N+1/N+2 above that of N+1, while D/U for N+2 remains low, suggests that IM3 for the signal 
pair becomes dominant again at low signal levels near the receiver’s threshold.  In fact, by the time 
desired signal drops to -82 dBm, D/U at N+1/N+2 has returned to its “rightful place” above the D/U for 
the other signal pairs.  Thus, the adjacent undesired signal pair (N+1/N+2) finally exhibits a greater 
interference effect than that of the more distant signal pairs, an expectation which was foiled by AGC 
gain reductions in the RF stage at most other points along the curve.  Apparently, the AGC-induced RF-
gain reductions are gone or nearly gone by this point on the graph. 
SINGLE UNDESIRED SIGNALS 
To facilitate the discussions below, Figure 11-6 and 11-7 were created to show the D/U ratio and 
threshold U, respectively, for single undesired signals (excluding the paired signal data that are included 
in Figures 11-1 and 11-2).  For the most part the plots exhibit a near-straight-line character (in the log-log 
plots due to units of dB) in a region from about D = -78 dBm to D =  -58 or -53 dBm.   
 
11-7 
 
To the left of this region, all D/U curves exhibit upturns from their straight line projections.  Those 
upturns are attributed to the effect of receiver noise as the minimum signal threshold for the DTV receiver 
is approached. 
 
Moving to the right, five of the plots (like three of the paired-signal plots) exhibit an abrupt downward 
bend in D/U to a relatively constant D/U ratio beginning at what we have termed the “AGC bend” at D = 
-53 dBm.  The bend is caused by a stabilization of nonlinear interference mechanisms by AGC operation. 
N+1 
The case of N+1 interference is discussed extensively in the previous section.  The observed behavior 
differs among three regions of the curve. 
 
In the middle section, from D = -78 dBm to D = -58 dBm, the D/U is constant, as would be expected for a 
linear interference process.  However, we showed, in the “N+1/N+2” section of this chapter, that the 
AGC is actively adjusting RF gain throughout this section, so the constant D/U could be due to a linear 
interference mechanism or a non-linear one whose effect has been stabilized by the AGC.  The 
“N+1/N+2” section also argued that the interference in this region occurs at a point beyond the first IF 
filter stage.  It could be due to a nonlinearity in an IF amplifier stage or due to a linear process, such as 
inadequate IF selectivity allowing the edge of the N+1 signal to leak into the demodulator. 
 
In the right-hand section, beyond D = -53 dBm, interference was shown to be the likely result of IM3 in 
the mixer after the AGC has reached the limit of its ability to reduce gain of the RF amplifier. 
 
In the left-hand section, to the left of D = -73 dBm, there is a rise in the left-most two points on the curve 
that is attributable to the influence of receiver noise as the receiver threshold is approached.  The fact that 
the rise is somewhat smaller than expected, coupled with a possible slight dip in D/U to the left of D = 
-73 dBm may suggest a diminishment of the interference mechanism that was dominant in the middle 
section of the curve.  Such a diminishment could occur if that interference mechanism were nonlinear and 
if the AGC action ends as one moves leftward on the curve through that region.  An alternative hypothesis 
is that the AGC-induced gain reductions might have increased the noise figure of the tuner and that the 
additional noise diminishes as one moves to the left on the curve and the AGC causes the RF gain to 
increase.  (Cowley and Hanrahan state that noise figure “increases at 1 dB per dB of gain backoff…with 
some AGC architectures.”
*
) 
N+2 
The D/U ratio for N+2 is shown in the bottom-most horizontal curve in Figure 11-6. 
 
In the “N+2/N+4” section of this chapter, we showed that the AGC begins to reduce RF gain when the 
undesired signal level on N+2 exceeds a threshold of -34.8 dBm (or lower if the desired signal level 
approaches -53 dBm).  Figure 11-7 shows that the undesired signal level throughout the N+2 curve 
exceeds this AGC threshold; thus, it appears that the AGC was controlling the RF gain throughout this 
curve.   
 
The constant D/U ratio versus D throughout most of the plot is consistent with either a linear interference 
mechanism or an AGC-stabilized nonlinear one.  The upturn at the left of the D/U plot (Figure 11-6) as D 
approaches D
MIN
 is caused by the effect of receiver noise. 
N+3 
The D/U curve for N+3 exhibits a steeper slope than the IM3-driven curves (Figure 1-1).  On the plot of 
threshold U versus D, one can see that, for desired signal levels between -73 dBm and -53 dBm, the 
                                                      
*
 Cowley and Hanrahan, 2006. 
 
11-8 
threshold U is almost constant—increasing only by only 1.1 dB in a span of 20 dB change in desired 
signal level.  A constant value of U would be expected for interference caused by a cross-modulation 
mechanism.  The slope of log-U versus log-D for such a mechanism would be nominally zero, a value 
approached in this case, as can be seen in Figure 11-4. 
 
We suspect cross modulation in the mixer (a likely candidate for N+2 interference as well).  We attribute 
the abrupt bend in U (downward bend in D/U) toward a more constant D/U ratio at D = -53 dBm to AGC 
gain reductions in the RF amplifier driven by the desired signal level. 
N+4 
For desired signal levels below D = -63 dBm, the log-D versus log-U curve for N+4 approaches 3—
consistent with third-order interference mechanism (Figure 11-3).  We are unable to propose such a 
mechanism in this case.  The so-called the “half IF” taboo channel for analog TV was based on the second 
harmonic of an undesired signal beating with the second harmonic of the local oscillator frequency.  Such 
an effect would create a susceptibility to interference that is centered 22 MHz (3 
2
/
3
 channels) above the 
desired channel—placing it predominantly at N+4.  However, we anticipate that such an effect would be 
second-order in terms of the undesired signal level, and thus would exhibit a log-D versus log-U slope of 
two. 
 
Between D = -63 dBm and D = -53 dBm, the slope increases significantly—possibly indicating a shift 
toward cross-modulation as the dominant interference mechanism. 
 
As with many of the other curves, the D/U for N+4 flattens when D exceeds -53 dBm, due to AGC action 
in the RF amplifier. 
N+5 
The slope of the N+5 curve to the left of the AGC bend at D = -53 dBm appears to be consistent with a 
third-order process.  We are unable to propose a likely mechanism. 
N+6 
Oddly, the slope of log-D versus log-U for N+6 appears to linearly increase with desired signal level until 
the AGC bend point, as can be seen in Figure 11-3.  Mathematically, this would suggest that log-D is 
linearly related to U.  Indeed, when desired signal level in dB is plotted against the threshold undesired 
signal level in linear power units (Figure 11-8), the result is strikingly linear to the left of the AGC bend.  
Whether there is a physical reason for this behavior or it is just a fluke is not known; however, we note 
that N+6 is one of the smaller interference vulnerabilities for this receiver, so no additional assessment 
has been performed. 
N+7 
As was noted in Chapter 5 of this report, most of the DTV receivers tested for this program exhibited an 
increased susceptibility to interference at N+7 relative to the surrounding channels.  The interference 
threshold at N+7 was found to be nearly constant in terms of undesired signal level as desired signal level 
was varied over a wide range up to -53 dBm.  (See for example even numbered Figures 5-12 through 5-
16.)  The behavior can be seen here in Figure 11-7. 
 
Initially, we attributed the interference phenomenon to the “IF beat” mechanism, which was one reason 
for the N+7 analog taboo.
*
  “IF beat” interference is caused by an undesired signal on channel N+7 or N-7 
beating with the desired signal—creating interference that can pass through the IF filter of the receiver.  
The lack of a corresponding susceptibility at N-7 (based on tests presented in Chapter 5) suggests that 
another mechanism may be at work.  Furthermore, tests with a narrower-bandwidth interferer, presented 
                                                      
*
 The other reason was the potential for direct radiation of one TV’s local oscillator to cause interference to another 
nearby receiver tuned seven channels higher. 
 
11-9 
in Chapter 7, suggest that the interference occurs only when the undesired signal spectrum overlaps the 
local oscillator frequency of the receiver.  (Located 44 MHz above the center of the desired channel, the 
local oscillator falls within channel N+7.)  This suggests some sort of direct interaction between the 
undesired signal and the local oscillator, although we have not identified a specific mechanism. 
 
N+14 and N+15 
The mixer image band for a single-conversion TV tuner with a 44 MHz IF overlaps parts of channels 
N+14 and N+15.  In Figure 11-6, the D/U ratio on these channels is seen to be essentially constant except 
when D approaches D
MIN
.   This is consistent with the fact that the mixer image interference mechanism is 
linear. 
 
SUMMARY 
The detailed assessment of one receiver’s out-of-channel interference rejection performance provides a 
basis for understanding the less detailed measurements for the other receivers.  In particular, we note the 
following. 
 
• Paired-signal IM3 can be one of the more dominant interference mechanisms, even at low signal 
levels. 
• The expected increase in D/U ratio with increasing signal levels for nonlinear interference 
mechanisms is flattened above levels at which the AGC engages to stabilize the nonlinearity. 
• AGC flattening of D/U ratios may begin at lower signal levels for close-in interference (e.g., N+1) 
than for interferers spaced further from the desired channel.  This can cause the unexpected result 
that, at some signal levels, a TV may be more tolerant of interference on the first-adjacent channels 
than on some other channels. 
• As signal levels continue to increase, the flatting of D/U ratios by AGC may end, allowing D/U ratios 
to, again, increase with increasing signal levels.  
 
 
11-10 
-60
-50
-40
-30
-20
-10
-90 -80 -70 -60 -50 -40 -30 -20
Power of Desired Signal (dBm)
D
esi
red
 /
 
U
n
desi
red
 R
a
t
i
o (dB)
N+1
N+2
N+3
N+4
N+5
N+6
N+7
N+14
N+15
N+1/N+2
N+2/N+4
N+3/N+6
N+4/N+8
N+1 (No Filter)
N+1/N+2 (No Filter)
DTV Signal Range
Sig Lvl Names
MODERATEWEAK
DT
V Rec
e
iver
 
S
ensit
ivit
y (O
ET
 69)
STRONG
 
Figure 11-1.  D/U Versus D for Receiver D3 (Desired Signal on Channel 51) 
-50
-40
-30
-20
-10
0
-90 -80 -70 -60 -50 -40 -30 -20
Power of Desired Signal (dBm)
P
o
wer
 
of E
a
ch U
n
desi
r
e
d 
S
i
g
n
al (dB
m
)
N+1
N+2
N+3
N+4
N+5
N+6
N+7
N+14
N+15
N+1/N+2
N+2/N+4
N+3/N+6
N+4/N+8
N+1 (No Filter)
N+1/N+2 (No Filter)
DTV Signal Range
Sig Lvl Names
MODERATEWEAK STRONG
D
T
V
 R
e
c
e
iv
er
 
S
e
n
s
i
t
iv
it
y
 (
O
E
T
 
6
9
)
U = -8 dBm
 
Figure 11-2.  Threshold U Versus D for Receiver D3 (Desired Signal on Channel 51) 
 
11-11 
0
1
2
3
4
5
6
7
8
-90-80-70-60-50-40-30-20
Power of Desired Signal (dBm)
Del
t
a 
D / D
e
lt
a
 U 
(
d
B/
d
B
)
N+1 N+2
N+3 N+4
N+5 N+6
N+7 N+14
N+15 N+1/N+2
N+2/N+4 N+3/N+6
N+4/N+8 N+1 (No Filter)
N+1/N+2 (No Filter) Sig Names
DTV Signal Levels
MODERATEWEAK STRONG
D
T
V
 R
e
c
e
iv
er
 
S
e
n
s
i
t
iv
it
y
 (
O
E
T
 
6
9
)
Cross-Modulation or IF Beat
3rd Order
AGC-Stabilized
Linear or
 
Figure 11-3.  Slope of Log-D versus Log-U for Receiver D3 (Desired Signal on Channel 51) 
-0.33
0.00
0.33
0.67
1.00
1.33
1.67
-90 -80 -70 -60 -50 -40 -30 -20
Power of Desired Signal (dBm)
De
lt
a
 U /
 De
lt
a 
D (
d
B
/
d
B
)
N+1
N+2
N+3
N+4
N+5
N+6
N+7
N+14
N+15
N+1/N+2
N+2/N+4
N+3/N+6
N+4/N+8
Sig Names
Sig Lvls
MODERATEWEAK STRONG
DT
V
 Re
c
e
i
v
er
 Sensit
i
v
i
t
y
 
(
O
ET
 
69)
Cross-Modulation or IF Beat
3rd Order
Linear or
AGC-Stabilized
 
Figure 11-4.  Slope of Log-U versus Log-D for Receiver D3 (Desired Signal on Channel 51) 
 
11-12 
-60
-50
-40
-30
-20
-10
-90 -80 -70 -60 -50 -40 -30 -20
Power of Desired Signal (dBm)
D
esi
red
 /
 
U
n
desi
red
 R
a
t
i
o (dB)
N+1
N+2
N+3
N+4
N+6
N+1/N+2
N+2/N+4
N+3/N+6
N+4/N+8
N+1 (No Filter)
N+1/N+2 (No Filter)
Sig Lvl Names
MODERATEWEAK
DT
V Rec
e
iver
 
S
ensit
ivit
y (O
ET
 69)
STRONG
 
Figure 11-5.  D/U Versus D for Receiver D3—Paired Signals And Their Constituents 
-60
-50
-40
-30
-20
-10
-90 -80 -70 -60 -50 -40 -30 -20
Power of Desired Signal (dBm)
De
sir
e
d / Unde
s
i
r
e
d Ra
tio 
(dB)
N+1
N+2
N+3
N+4
N+5
N+6
N+7
N+14
N+15
N+1 (No Filter)
N+1/N+2 (No Filter)
DTV Signal Range
Sig Lvl Names
MODERATEWEAKDT
V Rec
e
iver 
S
ens
i
t
ivity
 (
O
E
T
 69)
STRONG
 
Figure 11-6.  D/U Versus D for Receiver D3—Single Undesired Signals 
 
11-13 
-50
-40
-30
-20
-10
0
-90 -80 -70 -60 -50 -40 -30 -20
Power of Desired Signal (dBm)
P
o
w
e
r
 of
 E
ach U
ndesi
r
e
d S
i
gn
al
 (d
B
m
)
N+1
N+2
N+3
N+4
N+5
N+6
N+7
N+14
N+15
N+1 (No Filter)
N+1/N+2 (No Filter)
DTV Signal Range
Sig Lvl Names
MODERATEWEAK STRONG
DT
V
 Receiver
 
Sensitiv
i
t
y (O
E
T
 69)
U = -8 dBm
 
Figure 11-7.  Threshold U Versus D for Receiver D3—Single Undesired Signals 
0
5
10
15
20
25
30
-90 -80 -70 -60 -50 -40 -30 -20
Power of Desired Signal (dBm)
P
o
w
e
r 
Un
d
e
sired
 Sig
n
al (m
icrow
a
tt
s)
MODERATEWEAK STRONG
DT
V R
e
ceiver
 Sensitivit
y 
(O
ET
 
69)
 
Figure 11-8.  Linear Undesired Power Versus Log-Desired Signal Power for N+6 
 
12-1 
CHAPTER 12 
EXTENDING THE RESULTS TO LOWER SIGNAL LEVELS 
 
This chapter extrapolates the channel-30, single-channel interference rejection measurements from 
Chapter 5 of this report to a lower desired signal level, D
MIN
 + 1 dB.  It also employs measured data from 
Chapter 11 to evaluate the extrapolation method.  The motivation for the extrapolation was explained in 
Chapter 2. 
ORDER OF INTERFERENCE PROCESSES AND EFFECT OF AGC 
Chapter 8 presented a theoretical framework for understanding DTV interference susceptibilities.  
Table 12-1 summarizes the effects of the order of an interference mechanism and the state of the tuner’s 
AGC on plots of threshold undesired versus desired signal power at the input to a TV according to that 
framework. 
 
 
Table 12-1.  Characteristics of Log-Log Plots of Undesired Versus Desired Signal Power At TOV 
 Characteristics of Log-Log Plot of U Versus D 
Tuner AGC StateÆ 
and 
Input Condition? 
Fixed gain (or AGC operating to 
adjust gain of a tuner stage after the 
relevant nonlinearity
1
) 
AGC operating to adjust gain of a 
tuner stage prior to relevant 
nonlinearity
1 
D >> D
MIN
 
Straight line with slope determined by 
the order of the interference 
mechanism.  Slope is unity for linear 
interference. 
Straight line with unity slope—
matching that of a linear interference 
mechanism 
D approaching D
MIN
 
Deviation from straight-line 
determined by order of interference 
mechanism 
Deviation from straight-line: 
• matches linear process if AGC is 
driven by U; 
• determined by order of 
interference mechanism if AGC is 
driven by D 
1
 Relevant nonlinearity refers to the nonlinearity responsible for a given observed interference effect. 
Slopes are listed in Table 8-1 and deviations from straight line are listed in Table 8-2 of Chapter 8. 
 
 
Behavior for Desired Signal Levels From -68 dBm to -53 dBm 
Over the desired signal range from -68 dBm to -53 dBm, the desired signal power is far enough above 
D
MIN
 that the model predicts essentially straight-line behavior for log-U versus log-D, assuming that the 
interference mechanisms and AGC state remain the same throughout the region. 
 
Figure 12-1 shows the slopes of threshold U vs D in dB units over this signal range for each of the 
receivers and channel offsets for which threshold measurements were successful (i.e., the rejection 
performance was not beyond the measurement limit imposed by the test setup) at N = channel 30.  The 
graph also includes horizontal reference lines corresponding to the slopes for the various interference 
mechanisms discussed in this report.   
 
We note that some points clearly fit the expectations for a category of interference.  For example, most of 
the receivers exhibit linear-like interference behavior when the undesired signal is on channels N-4 
through N+2.  It is considered likely that the actual interference mechanisms in those cases are nonlinear 
 
12-2 
and that the linear-like behavior is caused by AGC operation.  Interference to receiver D3 from undesired 
signals on channels N-15 through N-4 clearly fits the pattern of cross-modulation—a fact that was also 
discussed in Chapter 5 in the section entitled, “Taboo Effects and Other Observations”, based on other 
evidence.  The susceptibility of most receivers at N+7 also matches the slope expected for cross-
modulation, although we suspect that the actual mechanism is different from those that have been 
postulated here based on the discussion in Chapter 7.   
 
Some of the points fall in-between categories, possibly as a result of changes in the dominant interference 
mechanisms over the 15 dB range of desired signal amplitudes, or due to changes in AGC operation over 
that range. 
 
Behavior for Desired Signal Levels From D
MIN
 + 3 dB to -68 dBm 
As the desired signal drops to a point 3-dB above D
MIN
, the deviation from straight-line behavior is 
expected to become significant—ranging from 1 to 3 dB (Table 8-2 of Chapter 8). 
 
Because of this deviation, we have chosen to estimate the slope of the straight-line portion of the log-
undesired versus log-desired signal curves in the region between D
MIN
 + 3 dB and -68 dBm by what we 
will call the adjusted slope.  The slope will be estimated by first shifting the left-hand point of the range 
(D = D
MIN
 + 3 dB) by 3 dB to the left (to D = D
MIN
).  Figures 8-1 through 8-4 (Chapter 8) showed that, in 
cases that are not stabilized by AGC operation, such a shift returns that point to the straight-line whose 
slope we are trying to predict, in all cases except the cross-modulation case.  With AGC operation driven 
by the undesired signal level, the same will be true.  Slope estimated in this way will deviate from that of 
the straight-line portion we would like to estimate in cases involving cross-modulation or involving AGC 
operation driven by the desired signal level.  Neglecting these cases, the expected slopes of log-U versus 
log-D for linear (or AGC-stabilized nonlinear), second order, and third order interference processes are 1, 
1/2, and 1/3 dB/dB, respectively, and the expected slopes of log-D versus log-U are 1, 2, and 3 dB/dB. 
 
For the cross-modulation case, Figure 8-4 showed that the straight-line projection should have a slope of 
zero for log-U versus log-D or an infinite slope for log-D versus log-U.  The shift of the left hand point of 
the range by 3 dB (for a log-U versus log-D plot) does not return that point to the straight line in this case.  
Rather, in the nominal case, where D
MIN
 = -84 dBm, we are measuring the slope of a line connecting the 
points (D = -68 dBm, U = U
T
) and (D = -84 dBm, U = U
T
 - 1.5 dB), where U
T
 is the constant threshold 
value of U along the straight line.  The expected slope, then, is 0.09 for log-U versus log-D or 10.7 for 
log-D versus log-U, though small measurement errors could cause the latter number to vary widely. 
 
Similarly, for the case of nonlinear interference stabilized by AGC operation driven by desired signal 
level, the adjusted slope of log-U versus log-D will be somewhat less than the unity slope of the straight-
line portion of the curve, and the slope of the log-D versus log-U curve will exceed unity. 
 
Figure 12-2 shows the “adjusted slope” of log-log curves of U versus D computed by the above method 
for D from D
MIN
 + 3 dB to -68 dBm.  Figure 12-3 shows the adjusted slope of log-log curves of D versus 
U—the reciprocal of the slopes shown in Figure 12-2.  For reasons discussed above, the slopes of the 
cross-modulation case generally fall above zero on the first plot and below infinity for the second.  
Similarly, some of the AGC-stabilized cases fall below unity slope of the first plot and above unity on the 
second. 
 
EXTRAPOLATION TO D = D
MIN
 + 1 DB 
The extrapolations to D = D
MIN
 + 1 dB will be based on measurements at D = D
MIN
 + 3 dB—an 
extrapolation distance (in terms of desired signal level) of only 2 dB.  Though the extrapolation distance 
is short, it should be recognized that threshold undesired signal levels are expected to change rapidly as 
 
12-3 
D
MIN
 is approached, as was illustrated in Figures 8-1 through 8-4.  Consequently, both measurements and 
extrapolations can be subject to greater errors as D approaches D
MIN
, due to high sensitivity to the D/D
MIN
 
ratio.  As an example, when D/D
MIN
 is expected to be 1 dB, misjudging the D/D
MIN
 ratio by 0.25 dB could 
cause the threshold undesired signal level to change by amounts ranging from 0.3 to 1.1 dB, depending on 
the direction of the error and the order of the interference process. 
 
The extrapolation will consist of two parts:   
• a straight-line projection part based on the slopes identified in Table 8-1 for each of the interference 
categories; 
• an estimate of the deviation from straight-line projection based on the deviations listed in Table 8-2; 
more specifically, we will use the difference between the straight line deviation for D/D
MIN
 = 1 dB 
and that for D/D
MIN
 = 3 dB. 
 
The first part will require that we categorize each interference case (each channel offset for each receiver) 
into one of four categories:  (1) linear or AGC-stabilized; (2) second order; (3) third order; (4) cross-
modulation.  The second part will require additional categorization within the linear or AGC-stabilized 
category. 
 
The category of each interference process will be estimated from the adjusted slope of the log-D versus 
log-U data from D = -68 dBm to D = D
MIN
 + 3 dB, computed as described in the previous section.  The 
boundary between third-order interference and cross-modulation will be set at an adjusted slope of 5.  The 
categories will serve as the basis for defining the slope of the straight-line portion of the extrapolation 
(Table 8-1), as well as for estimating the deviation from the straight-line projection.  The measurement 
data at D = D
MIN
 + 3 dB will serve as the anchor point for the extrapolation. 
 
The deviation from straight-line behavior will create a need to adjust the threshold undesired signal 
downward from the straight-line projection by an amount equal to the difference between the deviation 
for D
MIN
 + 1 dB and that for D
MIN
 + 3 dB, as determined by Table 8-2 (Chapter 8).  Those differences 
range from -1.3 to -3.8 dB
*
—a span of 2.5 dB. 
 
But selection of the correct values from Table 8-2 will, in some cases, require more knowledge than we 
have.  For cases that appear to be nonlinear (order higher than 1) based on the above, the appropriate 
values from Table 8-2 can be used directly; however, for channel offsets that appear to exhibit linear 
behavior based on the above, Table 8-2 shows that the expected offset depends on whether the underlying 
interference mechanism was truly linear or was made to appear linear by AGC action.  We will assume 
that the underlying mechanism is linear only for N+14 and N+15, the mixer image frequencies.  For the 
other cases, the correction depends both on the order of the nonlinearity and on whether the AGC action 
was driven primarily by desired signal level or primarily by the undesired signal level.  In the former case 
the deviation matches that for the underlying nonlinear process, which we don’t know; in the latter, the 
deviation will be the same as for a linear process. 
 
Though we have enough information to resolve some of these questions for receiver D3, on which 
detailed tests were performed, the limited measurements performed on the other receivers are inadequate 
for such resolution.  Consequently, we take the following approach to computing the deviation from a 
straight-line projection for all receivers: 
• If the interference behavior appears nonlinear, use the values from Table 8-2 for the estimated order 
of the interference; 
• For N+14 and N+15, use the values from Table 8-2 for linear processes; 
• For all other cases that appear linear, select a deviation from straight line behavior as the midpoint 
between the two extremes that could be occur—i.e., nonlinear with U driving the AGC (-3.8 dB 
                                                      
*
 The differences of the rounded numbers is the table, -6.9 - (-3.0) is -3.9, however, if the calculation is performed 
before rounding, and the answer is then rounded, the result is -3.8 dB. 
 
12-4 
adjustment) and third-order with D driving the AGC (-1.3 dB); we will accept the error of up to 1.3 
dB in either direction that could result from adjustments based on the midpoint (-2.6 dB). 
 
Table 12-2 summarizes the entire extrapolation process. 
 
 
Table 12-2.  Process for Extrapolation from D = D
MIN
+3dB to D
MIN
+1dB 
Extrapolation of Threshold U From 
D = D
MIN
+3dB to D
MIN
+1dB (dB) 
Undesired 
Channel 
Adjusted 
Slope of 
Log-D Versus 
Log-U 
(dB/dB) 
Assumed 
Interference 
Mechanism 
Straight-Line 
Projection 
Deviation 
From Straight 
Line Total 
N+14 or N+15  Linear -2.0 -3.8 -5.8 
All others < 1.5 
AGC Stabilized 
Nonlinear 
-2.0 -2.6 -4.6 
All others 1.5 to 2.5 2
nd
 Order -1.0 -1.9 -2.9 
All others 2.5 to 5 3
rd
 Order -0.7 -1.3 -1.9 
All others Magnitude >5 
Cross 
Modulation 
0 -1.9 -1.9 
 
 
Extrapolation Test 
In Table 12-3 we have applied this extrapolation process to data from the detailed measurements that 
were made on receiver D3 on channel 51 (Chapter 11).  Values of threshold U at D = D
MIN
 + 1 dB 
extrapolated from measurements at D = D
MIN
 + 3 dB and -68 dBm are compared to measurements at D = 
D
MIN
 + 1 dB.  The extrapolation errors were less than 1 dB in each case. 
 
Extrapolation of Channel-30 Measurements on All Eight Receivers 
Extrapolations of threshold undesired signal level to a desired signal level of D
MIN
 + 1 dB were performed 
for all channel-30 measurements of single-channel rejection performance for the eight fifth-generation 
DTV receivers.   
 
Figure 12-4 shows D/U ratios for the eight receivers based on the extrapolation.  Note that the plot 
includes data for a given receiver at a given channel offset only if a valid measurement was obtained at 
D
MIN
 + 3 dB, from which to extrapolate, and if the slope could be estimated.  Slope estimation requires a 
valid measurement at D = -68 dBm, except in the cases of N+14 and N+15, where the interference 
process was assumed to be linear.   
 
The plot also includes the ATSC performance guidelines corresponding to a desired signal level of -68 
dBm.  (ATSC does not specify rejection performance at a lower desired signal level.) 
 
The extrapolated data are combined with measured data in graphs in Chapter 13 and tabulations in 
Appendix A. 
 
 
12-5 
 Table 12-3.  Error Test for Extrapolation From D = D
MIN
 + 3 dB to D
MIN
 + 1 dB 
Change in U as D goes from D
MIN
+3dB to 
D
MIN
+1dB (dB) 
Extrapolated Change In U 
 
Adjusted 
Slope of 
Log-D 
Versus 
Log-U 
(dB/dB) 
Modeled 
Interference 
Mechanism 
Straight-
Line 
Portion of 
Projection 
Deviation 
From 
Straight 
Line Total 
Measured 
Change in 
U 
Extrapolation 
Error 
(dB) 
N+1 1.3 AGC-
Stabilized 
Nonlinear 
-2.0 -2.6 -4.6 -4.2 0.4
N+2 1.2 AGC-
Stabilized 
Nonlinear 
-2.0 -2.6 -4.6 -4.7 -0.2
N+3 5.5 Cross-
Modulation 
0.0 -1.9 -1.9 -1.7 0.2
N+4 3.0 3
rd
 Order -0.7 -1.3 -1.9 -1.8 0.1
N+5 3.2 3
rd
 Order -0.7 -1.3 -1.9 -1.6 0.3
N+6 4.5 3
rd
 Order -0.7 -1.3 -1.9 -1.2 0.8
N+7 6.3 Cross-
Modulation 
0.0 -1.9 -1.9 -2.4 -0.4
N+14 1.1 Linear -2.0 -3.8 -5.8 -5.0 0.8
N+15 1.1 Linear -2.0 -3.8 -5.8 -5.6 0.3
N+1/ 
N+2 
1.1 AGC-
Stabilized 
Nonlinear 
-2.0 -2.6 -4.6 -5.2 -0.6
N+2/ 
N+4 
3.4 3
rd
 Order -0.7 -1.3 -1.9 -1.4 0.6
N+3/ 
N+6 
3.3 3
rd
 Order -0.7 -1.3 -1.9 -2.1 -0.1
N+4/ 
N+8 
3.4 3
rd
 Order -0.7 -1.3 -1.9 -2.4 -0.5
Notes 
1
 For cases shown in bold italics, the apparent linearity is assumed to be a possible result of AGC action 
and deviation from straight line is calculated as described in text. 
Rounding to 0.1 dB after calculations are performed may cause some apparent discrepancies of up to 
0.1 dB. 
 
 
12-6 
0
5
10
15
20
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N-
8
N-
7
N-
6
N-
5
N-
4
N-
3
N-
2
N-
1
N+
1
N+
2
N+
3
N+
4
N+
5
N+
6
N+
7
N+
8
N+
9
N+
1
0
N+
1
1
N+
1
2
N+
1
3
N+
1
4
N+
1
5
N+
1
6
Interfering Channel
C
h
an
ge i
n
 Th
re
sho
l
d
 U
 fo
r 15
-d
B
 I
n
cr
eas
e in
 
D
(
dB
)
0.00
0.33
0.67
1.00
1.33
Slop
e of
 Th
resho
l
d
 
U
 
vs D (d
B/dB
)
A3
D3
I1
J1
M1
N1
O1
G4
Fi d
Linear Process or
AGC-Stabilized Process
3rd Order Process
Cross Mod or IF Beat
2nd Order Process
 
Figure 12-1.  Slope of Threshold U Versus D from D = -68 dBm to -53 dBm 
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N-
9
N-
8
N-
7
N
-
6
N-
5
N-
4
N-
3
N-
2
N-
1
N+
1
N+
2
N+
3
N+
4
N+
5
N
+
6
N+
7
N+
8
N+
9
N+
1
0
N
+
1
1
N
+
1
2
N+
1
3
N+
1
4
N+
1
5
N+
1
6
Interfering Channel
S
l
o
p
e of
 Thr
esh
ol
d 
U
 vs D
 (
d
B
/
d
B
)
A3
D3
I1
J1
M1
N1
O1
G4
Fi d
 
Figure 12-2.  Slope of Threshold U Versus D from D =D
MIN
+3dB to -68 dBm 
 
12-7 
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N-
9
N-
8
N-
7
N-
6
N-
5
N-
4
N
-
3
N-
2
N
-
1
N+
1
N+
2
N+
3
N+
4
N+
5
N+
6
N+
7
N
+
8
N+
9
N+
1
0
N+
1
1
N+
1
2
N
+
1
3
N+
1
4
N
+
1
5
N+
1
6
Interfering Channel
S
l
o
p
e of
 Thr
esh
ol
d 
U
 vs D
 (
d
B
/
d
B
)
A3
D3
I1
J1
M1
N1
O1
G4
 
Figure 12-3.  Slope of Threshold D Versus U from D =D
MIN
+3dB to -68 dBm 
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
S
u
scep
tibility to 
Inter
f
er
ence 
[D/U at
 T
h
resho
l
d
 (d
B
)
]
A3
D3
I1
J1
M1
N1
O1
G4
A/74 Limit
 
Figure 12-4.  D/U of 8 Receivers at D = D
MIN
 + 1 dB on Channel 30 (Extrapolation) 
 
13-1 
CHAPTER 13 
COMBINING MEASURED AND EXTRAPOLATED RESULTS 
This chapter shows single-channel rejection performance for eight DTV receivers for a desired signal on 
channel 30.  The results shown combine the measurements of Chapter 5 (D = -28 dBm, -53 dBm, 
-68 dBm, and D
MIN
 + 3 dB) with the extrapolations of Chapter 12 (D = D
MIN
 + 1 dB). 
 
Figure 13-1 shows D/U ratios for receiver A3.  Figure 13-2 shows the same information as Figure 13-1, 
but shows it as the threshold value for undesired signal power.  The first graph will be useful for those 
who prefer to work in terms of D/U ratios and for identifying channel offsets that behave in a linear 
manner (constant D/U ratio as D changes) either because the interference mechanism is linear (N+14 and 
N+15) or due to AGC action.  The second is useful for identifying the absolute signal levels that cause 
interference effects and for identifying channel offsets where thresholds tend to be constant in terms of 
absolute power of the undesired signal (e.g., N+7). 
 
Figures 13-3 to 13-16 are the same pair of plot formats for each of the remaining seven receivers. 
 
We note that the D/U plots (odd-numbered Figures 13-1 through 13-15) show four measurement limit 
curves.  From top to bottom, these correspond to limits at D = -28 dBm, -53 dBm, -68 dBm, and 
D
MIN
 + 3 dB, respectively.  In the top two curves, all points are determined by the maximum undesired 
signal power that the test setup could deliver to the input of a TV receiver.  The third curve, for D = -68 
dBm, has two sources of measurement limits:  at N-1 and N+1, the measurement limitation is based on 
leakage of the undesired signal into the desired channel; at all other channel offsets, the measurements are 
limited by maximum undesired signal level.  For D = D
MIN
 + 3 dB, the measurement limitation (shown by 
the bottom, shaded region of the graph) is caused by leakage of the undesired signal into the desired 
channel; since this is a soft limit, values below the limit are shown, but their accuracies are influenced by 
the leakage.  In the case of data extrapolated to D
MIN
 + 1 dB, data points are shown only if the 
measurements on which they were based were not subject to measurement limits. 
 
The undesired signal threshold plots (even-numbered Figures 13-2 through 13-16) show only one 
measurement limit curve—the curve associated with the maximum undesired signal power that the test 
setup could inject into the receiver.  The N-1 and N+1 offsets for D = -68 dBm and all of the offsets for D 
= D
MIN
 + 3 dB are subject to an additional limitation, shown only in the D/U plots, based on leakage of 
the undesired signal into the desired channel. 
 
We note in particular the case of receiver D3 in Figures 13-3 and 13-4.  In Chapter 5 we stated that the 
smooth rise in D/U (or corresponding smooth fall in threshold U) as one moves from N-15 to N-4 is 
suggestive of a particular broadband interference mechanism—cross-modulation.  Chapter 8 showed that 
cross-modulation is expected to exhibit a constant threshold U with changes in desired signal level except 
as the desired signal level approaches D
MIN
 or if the AGC begins to reduce gain prior to the tuner 
nonlinearity at which the cross-modulation is occurring.  We see in Figure 13-4 that the curves 
corresponding to D = -53 dBm and D = -68 dBm are essentially identical from N-15 to N-4 (except for a 
small bump associated with a single-channel interference susceptibility at N-11.  Interference 
susceptibility increases by a few dB as one moves past the curve for D = D
MIN
 + 3 dB to the curve 
corresponding to D = D
MIN
 + 1 dB; the increased susceptibility is an expected result of receiver noise 
becoming significant at lower signal levels.  All four of the curves appear to be the result of cross-
modulation.  
  
At N-3 things change.  The D/U ratio, which had been smoothly increasing as the undesired channel 
moved toward the desired channel, takes an abrupt drop—indicating that the receiver’s AGC has acted to 
decrease the RF gain prior to the mixer—the likely point of the nonlinearity that caused the cross-
 
13-2 
modulation.  Since each curve corresponds to fixed desired signal level, it is clear that the AGC must have 
been engaged by the undesired signal, which is likely to be larger in amplitude at the AGC sampling point 
when the undesired signal is on N-3 than when it was on N-4 and beyond because of the tuner’s RF 
tracking filter response.  The smaller change in the D
MIN
 + 3 dB curve as compared to the -68 dBm curve 
suggests that the AGC gain reduction was relatively small in the former case (where U is about -22 dBm).  
Thus, it is clear that the AGC engages on an undesired signal level somewhat lower than -22 dBm on 
N+3—a factor that will become important in some analysis in the next chapter. 
 
Chapter 15 includes composite charts for median, second-worst, and worst performance among the 
receivers. 
 
13-3 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
ceptibility to 
Inter
f
e
r
ence
 
[D
/U
 at T
h
reshold (dB
)
] Meas. Limit--
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
(Extrap)
Meas. Limits
Receiver A3
 
Figure 13-1.  D/U of Receiver A3 at Five Desired Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Dmin+1dB (Extrap)
Receiver A3
 
Figure 13-2.  Threshold U of Receiver A3 at Five Desired Signal Levels on Channel 30 
 
13-4 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
ceptibility to 
Inter
f
e
r
ence
 
[D
/U
 at T
h
reshold (dB
)
]
Meas. Limit--
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
(Extrap)
Meas. Limits
Receiver D3
 
Figure 13-3.  D/U of Receiver D3 at Five Desired Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Dmin+1dB (Extrap)
Receiver D3
 
Figure 13-4.  Threshold U of Receiver D3 at Five Desired Signal Levels on Channel 30 
 
13-5 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
ceptibility to 
Inter
f
e
r
ence
 
[D
/U
 at T
h
reshold (dB
)
] Meas. Limit--
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
(Extrap)
Meas. Limits
Receiver I1
 
Figure 13-5.  D/U of Receiver I1 at Five Desired Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Dmin+1dB (Extrap)
Receiver I1
 
Figure 13-6.  Threshold U of Receiver I1 at Five Desired Signal Levels on Channel 30 
 
13-6 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
ceptibility to 
Inter
f
e
r
ence
 
[D
/U
 at T
h
reshold (dB
)
]
Meas. Limit--
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
(Extrap)
Meas. Limits
Receiver J1
 
Figure 13-7.  D/U of Receiver J1 at Five Desired Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
U
n
desir
e
d Sig
n
al Level 
(U
) at 
Thr
esh
o
ld 
(dB
m
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Dmin+1dB (Extrap)
Receiver J1
 
Figure 13-8.  Threshold U of Receiver J1 at Five Desired Signal Levels on Channel 30 
 
13-7 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
ceptibility to 
Inter
f
e
r
ence
 
[D
/U
 at T
h
reshold (dB
)
]
Meas. Limit--
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
(Extrap)
Meas. Limits
Receiver M1
 
Figure 13-9.  D/U of Receiver M1 at Five Desired Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Dmin+1dB (Extrap)
Receiver M1
 
Figure 13-10.  Threshold U of Receiver M1 at Five Desired Signal Levels on Channel 30 
 
13-8 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
ceptibility to 
Inter
f
e
r
ence
 
[D
/U
 at T
h
reshold (dB
)
] Meas. Limit--
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
(Extrap)
Meas. Limits
Receiver N1
 
Figure 13-11.  D/U of Receiver N1 at Five Desired Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Dmin+1dB (Extrap)
Receiver N1
 
Figure 13-12.  Threshold U of Receiver N1 at Five Desired Signal Levels on Channel 30 
 
13-9 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
ceptibility to 
Inter
f
e
r
ence
 
[D
/U
 at T
h
reshold (dB
)
] Meas. Limit--
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
(Extrap)
Meas. Limits
Receiver O1
 
Figure 13-13.  D/U of Receiver O1 at Five Desired Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Dmin+1dB (Extrap)
Receiver O1
 
Figure 13-14.  Threshold U of Receiver O1 at Five Desired Signal Levels on Channel 30 
 
13-10 
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
ceptibility to 
Inter
f
e
r
ence
 
[D
/U
 at T
h
reshold (dB
)
] Meas. Limit--
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
(Extrap)
Meas. Limits
Receiver G4
 
Figure 13-15.  D/U of Receiver G4 at Five Desired Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
Dmin+1dB (Extrap)
Receiver G4
 
Figure 13-16.  Threshold U of Receiver G4 at Five Desired Signal Levels on Channel 30
 
14-1 
CHAPTER 14 
VERIFICATION TESTS USING ALTERNATIVE METHODS 
This chapter presents the results of tests and analyses performed to validate the test methodology and test 
setup used for the measurements in this report.  Readers primarily interested in DTV receiver test results 
may prefer to skip ahead to Chapter 15 or to the “Summary” section of this chapter. 
 
Most of tests described in this section of the report were intended to eliminate various artifacts that might 
be associated with the primary test setup used for measurements throughout this report, and then to retest 
a TV to see whether the interference rejection results change.  One test was designed to directly measure 
the effect of one such artifact—the noise plateau around the desired channel caused by using a band-reject 
filter to “clean up” the output spectrum of the undesired signal. 
 
The other reason for the validation tests was to confirm unexpected results of the interference rejection 
tests—most notably, the peak in interference susceptibility that was observed to varying degrees at 
channel N+7 for nine out of ten receivers that was tested. 
 
The verification tests were performed only for channel 30. 
 
ALTERNATIVE TEST SETUP WITH BANDPASS FILTER ON 
UNDESIRED SIGNAL 
The primary test setup for measurements presented in this report used a band-reject filter to limit leakage 
of the undesired signal into the desired channel, as described in Chapter 4.  This differs from the more 
conventional approach of placing a band-pass filter around the undesired signal (or signals).  The band-
reject approach leaves a plateau in the out-of-band spectrum of the undesired signals—a plateau which is 
filtered out only in the desired channel and its immediate vicinity.  (See, for example, Figures 4-1 and 4-2 
of Chapter 4.)  
 
For the measurements described in this section of the report, the test approach was changed to a more 
conventional one.  Instead of subjecting the undesired signal to a band-reject filter designed to limit 
leakage into the desired channel (channel 30), a bandpass filter was employed to limit leakage into any 
channel outside of the undesired channel.  The bandpass filter used for this test was an existing seven-
section cavity filter (Micro Communications, Inc, Type "N" IDBP Filter, Part # 220035 C/N) for TV 
broadcast channel 29.  Filter response was down 0.7 to 0.8 dB at the channel-29 band edges (0.4 dB at the 
DTV pilot frequency) and more than 73 dB beyond the first-adjacent channels.  Measured frequency 
response is shown in Figure 14-1. 
 
The use of the fixed bandpass filter for the undesired signal as opposed to a band-reject filter at the 
desired channel required a change in test approach.  The undesired signal was fixed at channel 29 and the 
desired channel, along with the TV tuner, were switched among channels 14, 15, 21 to 28, and 30 to 37—
creating test cases corresponding to N+15, N+14, N+8 through N+1, and N-1 through N-8.  (Note that we 
were changing N rather than the undesired channel to achieve these offsets.) 
 
Though the main purpose of testing in this way was to eliminate the noise shoulders that appeared around 
the desired channel in the primary test setup, several other changes were made in the test setup in order to 
rule out the influence of other possible artifacts.  Specifically, the test setup (Figure 4-1 of Chapter 4) was 
modified as described below and as shown in Figure 14-2. 
 
14-2 
• The Rohde and Schwarz SFU 8-VSB signal generator was used as the undesired signal source for all 
of these measurements.  This was intended to eliminate the effects of any generator-specific artifacts 
that might be produced by the Agilent E4437B signal generator that was configured to produce 
bandlimited Gaussian noise as the undesired signal source for much of the testing described in this 
report. 
• All amplifiers and step attenuators, as well as some fixed attenuator pads, were eliminated to simplify 
the test setup and to reduce the maximum signal levels that existed within the test setup in order to 
reduce the possibility of unintended coupling of signals (e.g., by radiation). 
 
Rejection ratio measurements were performed on one DTV (receiver J1) using the alternative test setup.  
Figure 14-3 compares the results with the previous measurements using the primary test setup.  All 
measurements matched within 1.5 dB.  The measurements differed on N-1 and N+1 by only 0.4 and 
0.8 dB, respectively.   On the other channels, the D/U ratios with the alternative test setup averaged 
0.7 dB lower than those with the primary test setup.  Since the alternative test setup used an 8-VSB signal 
rather than bandlimited Gaussian noise as the interferer,
*
 the alternative test setup was expected to exhibit 
lower D/U ratios than the primary one by about 1.2 dB based on the difference in signal type (Chapter 7).   
 
The match between test results is quite close, especially given that the different test approaches forced 
testing to be performed on different TV channels (mostly channels 21 through 37 for the alternative test 
setup and channel 30 for the primary one).   
 
BROADBAND NOTCHED NOISE 
Since the primary test setup left a plateau of noise surrounding the desired channel, a decision was made 
to test each TV receiver to determine the effect of that plateau by itself (i.e., without the undesired signal).  
The signal generator used as an undesired signal source for non-adjacent tests on channel 30 had a noise 
floor that was 56 to 59 dB below the undesired signal power when measured in a 6-MHz bandwidth 
beyond the first adjacent channels.  The primary test setup subjected the undesired signal to a band-reject 
filter that further attenuated this noise floor within the desired channel.  The result was a signal spectrum 
that included the main undesired signal spectrum, plus a noise plateau 56 to 59 dB below it, but with a 
valley cut into the noise plateau at channel N, as shown in Figure 4-2. 
Measurements 
The “broadband notched noise” tests were performed by replacing the undesired signal source with a 
white Gaussian noise generator having a spectrum that extended from about 40 to 850 MHz—i.e., 
covering the entire broadcast television spectrum and beyond.  The noise signal spectrum was “notched” 
by the same band-reject filter that had been used to filter the undesired signal in all of the non-adjacent 
tests at channel 30 (i.e., all tests except at N-1 and N+1).  Figure 14-4 shows a portion of the resulting 
spectrum.  Essentially, the undesired signal for these tests consisted of white noise filling the entire TV 
spectrum and beyond, but with a 43-dB deep hole cut in it at channel 30 (N), and with some filter rolloff 
through the first-adjacent channels.
†
 
 
The interference rejection performance of the eight DTV receivers were then measured using this 
broadband notched noise source.  Tests were performed for desired signal levels of -68 and -53 dBm.  
D/U ratios were computed with U being the undesired power per TV channel averaged across channels N-
4 through N-2 and N+2 through N+4.   
 
                                                      
*
 The undesired signal for all channel offsets except N-1 and N+1 was Gaussian noise in measurements made with 
the primary test setup. 
†
 Total noise power in channel N was 42.8 dB below the average power per channel on channels N-4, N-3, N-2, 
N+2, N+3, and N+4. 
 
14-3 
The results are shown in the first two data columns of Table 14-1.  At a desired signal power of -68 dBm, 
the interference thresholds occurred at a D/U ratio ranging from -27.1 to -23.0 dB, where U is measured 
on a per TV channel basis, as described above.  The corresponding D/U range when D = -53 dBm was 
slightly higher at -25.9 to -20.6 dB. 
 
 
Table 14-1.  Broadband Notched Noise D/U Ratios at Threshold 
 
D/(U/channel) For Notched 
Noise Only (dB) 
D/U Caused by Noise 
Plateau If Single-Channel U 
Were 56 dB Above the 
Noise Plateau (dB) 
Receiver D = -68 dBm D = -53 dBm D = -68 dBm D = -53 dBm 
A3 -26.9 -25.9 -82.9 -81.9 
D3 -25.5 -22.9 -81.5 -78.9 
I1 -25.4 -23.6 -81.4 -79.6 
J1 -23.0 -20.6 -79.0 -76.6 
M1 -27.1 -25.6 -83.1 -81.6 
M2 -23.1 -22.0 -79.1 -78.0 
N1 -26.1 -22.2 -82.1 -78.2 
G4 -23.4 -23.7 -79.4 -79.7 
Max -23.0 -20.6 -79.0 -76.6 
Min -27.1 -25.9 -83.1 -81.9 
 
 
Thus, all of the TVs could operate with the noise plateau 23 dB or more above the desired channel, on a 
noise-power-per-channel basis when D = -68 dBm and 20.6 dB or more above the desired channel when 
D = -53 dBm.  Since the actual undesired noise source power used in the tests was 56 dB above its noise 
plateau, one would expect that the effect of the noise plateau by itself would limit D/U measurements to 
about -79 dB at D = -68 dBm and about -77 dB at D = -53 dBm.  If we want to maintain at least 10 dB 
margin between measured D/U’s and the failure point due to the plateau, these numbers increase to -69 
and -67 dB, respectively.   
 
In reality, D/U ratio measurements were limited by the maximum undesired signal level that the test setup 
could generate to values of about -52 dB at D = -53 dBm and -66 dB at D = -68 dBm; consequently, 
plateau noise alone would not have affected the measurements, even for the most vulnerable TV. 
 
IM3 Analysis 
We saw in Chapter 10 that a high level signal on a channel N+K can cause a receiver to be susceptible to 
interference from very low level signals on channel N+2K.  We could consider the intended undesired 
signal to be on N+K and the plateau noise to contain a signal at N+2K at a level 56 dB below the intended 
undesired signal.  (Note that, for N+/-2 and beyond, the plateau noise ranges from 56 to 59 dB below the 
undesired signal power, so this is a worst-case assumption.)  The same could be considered with the roles 
of N+K and N+2K reversed. 
 
To test whether IM3 between the undesired signal and the plateau could have been responsible for any of 
the measured single-channel D/U ratios, lines corresponding to the cases of U
N+K
/U
N+2K
 = 56 dB 
U
N+K
/U
N+2K
 = -56 dB were overlaid on the modeled IM3 charts from Figures 10-4 to 10-17.  The 
intersections of those overlaid lines with the modeled curve for each channel offset, each TV, and each of 
two desired signal levels (-68 dBm and D
MIN
 + 3 dB) were then examined. 
 
 
14-4 
In all of the 67 cases available on the plots, the U
N+K
/U
N+2K
 = -56 dB line crossed the horizontal segment 
of the model plot, indicating that the measured U
N+2K
 threshold was reached before the plateau-based IM3 
threshold would have been reached. In 63 out of the 67 cases available on the plots, the U
N+K
/U
N+2K
 = 
56 dB line crossed the vertical segment of the model plot, indicating that the measured U
N+K
 threshold 
was reached before the plateau-based IM3 threshold would have been reached.  These results indicate that 
IM3 between the undesired signal and the plateau was not a factor in these cases. 
 
In three cases, the U
N+K
/U
N+2K
 = 56 dB line crossed model plot near the intersection of the vertical 
segment and the diagonal IM3 segment.  This suggests that, in those cases, the “single-channel” threshold 
that was measured and reported in Chapter 5 may have actually been caused by IM3 between the 
undesired signal and the plateau.  The model plots for those cases are shown in Figures 14-5 to 14-7.  The 
three cases are summarized in the following bullets. 
• Receiver D3 at D = -68 dBm, N+2 threshold occurred at U = -20.5 dBm, D/U = -47.5 dB. 
• Receiver G4 at D = -68 dBm, N-5 threshold occurred at U = -5.4 dBm, D/U = -62.7 dB; 
• Receiver G4 at D = D
MIN
 + 3 dB, N+4 threshold occurred at , D/U = -13.5 dB,U = -69.6 dBm; 
The two cases involving receiver G4 correspond to D/U ratios below (more negative than) -60 dB.  If IM3 
between the undesired signal and the plateau caused measurement limitations at those levels, the 
measurement limitations occurred beyond the planned measurement range of the test setup and in a region 
that represents very good rejection performance by the DTV receivers; consequently, the cases are not 
considered important.  The remaining case, involving receiver D3, corresponds to the second best 
rejection ratio among the eight receivers at N+2 for a desired signal level of -68 dBm.  If the results were 
influenced by IM3, it would mean that the actual performance was even better and falls only 0.8 dB short 
of the best performance among the eight receivers. 
 
In one other case (Figure 14-8), the U
N+K
/U
N+2K
 = 56 dB line crossed a model plot on the diagonal IM3 
segment at a U
N+K
 value about 5 dB below the single-channel threshold that had been measured for that 
case (receiver D3 at D = -68 dBm, N-3 threshold occurring at U = -8.2 dBm, D/U = -59.8 dB).  This is 
expected to be an impossible situation in that the IM3 effect is predicted to occur at a lower desired signal 
level than the actual -8.2 dBm threshold that was measured for the single-channel interferer.  
Theoretically, IM3 should have caused an artificially low reading of about -13 dBm at threshold.   
 
Our use of worst-case assumptions regarding the undesired signal-to-plateau ratio (56 dB, whereas the 
range was 56 to 59 dB) is not enough to account for the difference; however, Figures 13-3 and 13-4 and 
the related discussion in Chapter 13 offer some insight into what is happening.  The text in Chapter 13 
argues that the receiver’s AGC activates to reduce the tuner gain prior to the mixer when the undesired 
signal level on channel N-3 exceeds a level somewhat below -22 dBm.  Referring to Figure 14-8, we see 
that the AGC would engage beginning at a point above and to the left of U
N+K
 = -22 dBm and U
N+2K
 = -52 
dB.  The Appendix B shows that, when AGC engages on U
N+K
, the threshold value of U
N+2K
 becomes 
constant with further increases in U
N+K
.  This means that the sloped IM3 portion of the model for receiver 
D3 in Figure 14-8 should end its downward trajectory and switch to a horizontal trajectory when the AGC 
engages.  The horizontal trajectory would continue until the curve intersects the vertical segment 
corresponding to the measured single-channel threshold for U
N+K
.  An examination of the plot shows that, 
with this new trajectory, the dashed line representing U
N+K
/U
N+2K
 = 56 dB would not intersect the IM3 
segment of the line.  The apparent contradiction was actually caused by a failure to show the AGC 
engagement effect in the model for that receiver.  Chapter 10, the origin of the modeled results cautions 
that the model becomes invalid beyond the signal levels at which the AGC engages to reduce gain prior to 
the nonlinearity that causes the IM3 (usually the mixer). 
 
Thus we conclude that the only cases in which we have identified a potential for IM3 between the 
undesired signal and the plateau to have influenced the test results are cases in which measured rejection 
performance of a receiver was very good.  If the unintended IM3 did actually influence those 
measurements, then the actual receiver performance was even better than the measurement indicated—a 
result that would not change any conclusions of this report. 
 
14-5 
 
SCREEN-ROOM TEST 
We are unaware of other reports of DTV receivers exhibiting an N+7 interference vulnerability, such as 
was observed on nine of ten receivers tested for this report.  In an effort to rule out test environment 
anomalies as the cause, receiver D3 was placed in a closed screen room.  No equipment inside the screen 
room, other than the TV, was powered on.  The rejection performance test equipment was placed outside 
of the screen room about 18 feet from the TV.  Rejection performance was measured with a desired signal 
of -68 dBm on channel 30 and the white Gaussian noise source (bandlimited to match the 3-dB width of 
an 8-VSB signal) was placed on channel N+7 and was adjusted to determine TOV.  The resulting D/U 
ratio was -42.1 dB.  The measurement one month earlier using the primary test setup without the screen 
room was -41.6 dB.  The difference of only 0.5 dB provides further evidence that the observed N+7 
sensitivity is not an anomaly of the test environment.   
 
SUMMARY 
A single screen-room measurement confirmed that the N+7 susceptibility is not caused by direct pickup 
involving the TV receiver.   
 
The noise plateau discussed in this report was about 56 dB or more below the undesired signal power 
when measured on a per-channel basis.  Tests were performed using an alternative test setup that (1) 
employed a bandpass filter approach to attenuate the noise plateau globally rather than just within the 
desired channel and (2) eliminated amplifiers and attenuators to create a simpler (and more limited) 
configuration that maintained lower signal levels in the test setup to minimize potential for unintended 
signal coupling.  The test results closely matched those obtained with the primary test setup used in this 
report. 
 
Tests of each of the eight receivers against notched plateau noise demonstrated that the rejection 
performance tests were not influenced by the presence of the noise plateau by itself. 
 
Analysis was performed to determine whether IM3 effects between the undesired signal and a portion of 
the undesired signal plateau 56 or more dB lower could have affected the measured results for single-
channel rejection performance.  Since IM3 properties vary among the individual receivers as well as with 
channel offset and desired signal amplitude, a global answer is not feasible.  Measurements of IM3 effects 
performed with equal-powered paired signals (Chapter 9) were used to develop models for paired-signal 
IM3 for 67 combinations of TV receivers, channel offsets, and desired power levels (Chapter 10).  Each 
of these 67 was examined to determine whether IM3 could have influenced the single channel 
measurements in either of two cases (undesired signal at N+K and undesired signal at N+2K)—a total of 
134 cases of single-channel rejection measurements.  The evaluation suggested that three of the 134 
measurements might have been affected by IM3 with the noise plateau.  All three involved very good 
rejection performance (two with D/U’s beyond -60 dB and one only -48 dB, but the second best 
performance among eight receivers); if IM3 influenced the measurements, the actual rejection 
performance was even better than that measured—a result that would not change any conclusions of this 
report. 
 
14-6 
-80
-70
-60
-50
-40
-30
-20
-10
0
-9 -6 -3 0 3 6 9
Frequency Offset From Channel Center (MHz)
Filter R
esponse 
(dB
)
 
Figure 14-1.  Frequency Response of Channel-29 Bandpass Filter 
 
50-75 ohm
min loss pad
25-FT
Low-Loss
Coax
DTV Rcvr
6-dB
PAD
SPECTRUM
ANALYZER
MLP#13
P#M
Chan 29 
Band-
Pass 
Filter
ATSC997
@ Variable 
Output 
Level
75-50 ohm
min loss pad
6-dB
PAD
6-dB
PAD
COMBINER SPLITTER
8-VSB Source A
3-dB
PAD
3-dB
PAD#1
3-dB
PAD#2
SFU 8-VSB 
GENERATOR
@ Variable Output 
Level
8-VSB Source C
UNDES
I
RE
D
SI
GN
A
L
DE
SI
RED
SIGNA
L
25-FT
Low-Loss
Coax
UNDES
I
RE
D
SI
GN
A
L
DE
SI
RED
SIGNA
L
 
Figure 14-2.  Alternative Test Configuration With Bandpass Filter on Undesired Signal 
 
14-7 
-60
-55
-50
-45
-40
-35
-30
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
1
4
N
+
1
5
Interfering Channel
D/
U R
a
tio
 
(d
B
)
Measurements w/Primary Test Setup
Measurement w/Alternative Test Setup
Meas. Limit w/Alternative Test Setup
 
Figure 14-3.  D/U Measurements on Receiver J1 with Primary Versus Alternative Test Setup 
-60
-50
-40
-30
-20
-10
0
10
542 548 554 560 566 572 578 584 590 596
Frequency (MHz)
N
o
rmal
iz
ed Spectrum 
L
evel (dB 
in 30-kHz
 R
B
W)
N+2 N+4N-2N-4 N
 
Figure 14-4.  Spectrum of Undesired Signal for Broadband Notched Noise Test 
 
14-8 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
ndesir
e
d
 S
i
gnal
 L
e
vel
 
on N+
2K
A3
D3
I1
J1
M1
N1
O1
G4
N+K to N+2K Power Ratio = 56 dB
D = -68 dBm
Channel Pair = N-5/N-10
 
Figure 14-5.  Model Plot for “Plateau IM3” On Receiver G4 at D = -68 dBm with N-5/N-10 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
ndesir
e
d
 S
i
gnal
 L
e
vel
 
on N+
2K
A3
D3
I1
J1
M1
N1
O1
G4
N+K to N+2K Power Ratio = 56 dB
D = -68 dBm
Channel Pair = N+2/N+4
 
Figure 14-6.  Model Plot for “Plateau IM3” On Receiver D3 at D = -68 dBm with N+2/N+4 
 
14-9 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
ndesir
e
d
 S
i
gnal
 L
e
vel
 
on N+
2K
A3
D3
I1
J1
M1
N1
O1
G4
N+K to N+2K Power Ratio = 56 dB
D = Dmin + 3 dB
Channel Pair = N+4/N+8
 
Figure 14-7.  Model Plot for “Plateau IM3” On Receiver G4 at D = D
MIN
+3dB with N+4/N+8 
-80
-70
-60
-50
-40
-30
-20
-10
0
-
8
0
-
7
0
-
6
0
-
5
0
-
4
0
-
3
0
-
2
0
-
1
0 0
Undesired Signal on N+K
U
ndesir
e
d
 S
i
gnal
 L
e
vel
 
on N+
2K
A3
D3
I1
J1
M1
N1
O1
G4
N+K to N+2K Power Ratio = 56 dB
D = -68 dBm
Channel Pair = N-3/N-6
 
Figure 14-8.  Model Plot for “Plateau IM3” On Receiver G4 at D =-68 dBm with N-3/N-6 
 
 15-1
CHAPTER 15 
SUMMARY AND CONCLUSIONS 
We now summarize the main findings of the measurement program. 
 
This report presented the following: 
• Tests of 30 consumer DTV receivers to determine tuner type (e.g., single conversion or double 
conversion); 
• Measurements of out-of-channel interference rejection performance of eight DTV receivers tuned to 
channel 30; 
• Measurements of out-of-channel interference rejection performance of seven DTV receivers (a subset 
of the eight) tuned to channel 51; 
• A theoretical framework for understanding the results; 
• Extrapolation of the measurement results to a desired signal level 1 dB above the minimum signal 
threshold for each receiver. 
 
The out-of-channel interference tests included single-channel interferers and pairs of interferers at channel 
spacings that could place third-order intermodulation distortion products in the desired TV channel.  Most 
of the tests used undesired (i.e., interfering) signals that occupied most of the 6-MHz width of a DTV 
channel assignment.  An 8-VSB DTV signal was used as the undesired signal for all tests on first-adjacent 
channels.  For other channel spacings, the source was either an 8-VSB source (for the channel-51 tests) or 
a Gaussian noise source bandlimited to match the 3-dB width of an 8-VSB signal (for the channel-30 
tests).  For paired-signal tests, the second undesired signal was always a Gaussian source.  A limited 
number of tests were performed to measure the relative interference effects of 8-VSB, Gaussian noise, 
and OFDM signals.  One TV was tested to determine the effects of a narrower-band undesired signal—a 
Gaussian noise source bandlimited to a 3-dB width of 1 MHz. 
 
The out-of-channel interference rejection performance measurements are presented in this report in terms 
of desired and undesired signal power levels at the RF input (i.e., antenna terminal) of each DTV receiver.  
While the test results are intended to be useful in assessing interference potential and developing 
protection criteria to prevent interference, the results do not translate directly into such criteria.  Assessing 
existing protection criteria or developing new criteria will require steps that are beyond the scope of this 
report, such as definition of the scenarios on which protection is to be based, modeling of propagation 
from the undesired signal source to the DTV antenna, modeling of antenna gain and of losses from the 
TV antenna to the DTV receiver, and policy decisions regarding the DTV receiver performance and 
signal margins to be assumed in developing the protection criteria.   
 
We have attempted to provide receiver interference rejection performance data over a parameter range 
sufficient to support the broad needs of the Commission and the technical community.   While we note 
that the rejection performance measurements in this report were performed on only a small sample of 
consumer DTV receivers and thus do not provide a robust statistical basis for identifying the overall range 
of performance of consumer DTV receivers, they do provide a more representative sample of the 
performance of currently available products than did the earlier tests of the prototype receiver. 
 
In forming conclusions from the measurements, we focus primarily on the measurement results at channel 
30, since that channel is more central to the UHF band and thus is likely to be more representative of 
performance across the band.  Results from channel 51 are used for comparative purposes, and to fill in 
gaps in the measurements at channel 30. 
 
 
 15-2
DTV TUNER TYPE 
Tuner type is of significance because the “taboos” that limited local analog TV channel allotments at 
certain channel spacings were based primarily on interference susceptibilities that were unique to single-
conversion tuners having a 44-MHz intermediate frequency (IF).  The reduced vulnerability of the 
prototype ATSC DTV receiver—relative to analog TV—was achieved by a combination of the more 
robust ATSC digital transmission system that is inherently less susceptible to noise and interference than 
the NTSC analog system and the use of a double-conversion tuner in the prototype receiver.   
 
Tests of 30 consumer DTV receivers demonstrated that all 30 have single-conversion tuners with 44-MHz 
IF.  (28 were identified as single conversion by means of a small but detectable leakage of the tuner’s 
local oscillator frequency from its antenna port.  The remaining two were identified by means of 
interference vulnerabilities that we expect to be unique to single-conversion tuners with 44-MHz IF.) 
 
EFFECT OF UNDESIRED SIGNAL TYPE 
Tests were performed to determine the relative interference effects of the following types of signals: 
• White Gaussian noise bandlimited to the 3-dB width of an 8-VSB signals; 
• 8-VSB DTV signal; 
• DVB-H—an orthogonal frequency division multiplexing (OFDM) signal—set for a 5-MHz channel 
width; and 
• White Gaussian noise bandlimited to a 3-dB width of 1 MHz. 
 
The first three signal types were tested on eight DTV receivers at the five non-first-adjacent channel 
offsets that exhibited the most interference potential among the receivers at low signal levels (N+2, N-2, 
N-3, N-4, and N-6).  (Channels N-1 and N+1 were not tested because the Gaussian source did not have 
adequate band-edge rolloff to permit testing on first-adjacent channels.) 
 
The interference effects of the OFDM signal and of the Gaussian noise signal averaged 1.0 and 1.2 dB, 
respectively, greater than the interference effects of the 8-VSB signal.  That is, the signal level of an 
OFDM signal or of a Gaussian noise signal that is required to interfere with DTV reception is about 1 dB 
lower than the level of an 8-VSB signal that would cause interference at the same channel offset. 
 
Tests with the reduced-bandwidth interferer (1 MHz wide) were performed on only one receiver.  The 
results generally tracked those of the wider Gaussian noise interferer except where narrowband 
interference susceptibilities existed, such as at N+7, where reception was susceptible to interference when 
the undesired signal spectrum overlapped the TV’s local oscillator frequency. 
 
CLIFF EFFECT 
The ATSC digital television broadcast system can achieve flawless picture reception under interference 
conditions that would produce an unusable picture for analog broadcast TV;
*
 however, once an undesired 
signal reaches a level at which picture impairments become visible on a DTV receiver, the picture 
degrades extremely rapidly with further increases in undesired signal level.  The rapid degradation from 
flawless picture to no picture at all is known as the cliff effect. 
                                                      
*
 The co-channel D/U ratio for DTV into DTV interference is 15.3 dB.  Analog TV reaches the point of unusability 
at a D/U ratio of 16.1 dB for interference from DTV and requires a significantly higher D/U ratio to achieve a high 
quality picture.  Results are from: 
Wayne Bretl and Gary Sgrignoli, “Summary of the Grand Alliance VSB Transmission System Laboratory Tests”, 
IEEE Transactions on Consumer Electronics, Vol 42, No. 3, June 1996, sections 3.7.1 and 3.8.1. 
 
 15-3
 
The interference rejection test results in this report are presented as signal levels at the threshold of 
visibility (TOV) of picture degradation—i.e., the point at which picture degradation (in the form of 
pixilation, image freezes, or dropouts) becomes visible.  Increases in interfering signal level above this 
point result in further degradation—and ultimately complete loss—of the TV picture. 
 
Though the “hardness” of the thresholds was not one of the measurement parameters for this study, a 
strong “cliff effect” was observed during the tests.  For example, in most cases, increasing interference 
level about 1 dB above TOV caused complete loss of picture.  In some cases picture loss didn’t occur 
until the undesired signal level rose as much as much as 3 dB and in one case, 5 dB (though picture errors 
occurred continuously in that case after only a 1.5 dB increase).  In a few cases picture loss occurred 
concurrently with appearance of errors or with only an additional 0.1 dB increase in interference—an 
extremely abrupt cliff!  By contrast, interference to analog TV occurs much more gradually.  An 8-dB 
increase in signal level of an interferer from the TOV level for analog TV may cause the interference 
effect to grow to the “slightly-annoying” level, from the TV viewer’s point of view.
*
  A total increase of 
20 to 30 dB may be required to make the analog picture unusable.
†
 
INTERFERENCE FROM A SINGLE UNDESIRED SIGNAL 
The interference tests with a desired signal on channel 30 were performed using an 8-VSB undesired 
signal for first-adjacent channels (N-1 and N+1) and a white Gaussian noise signal bandlimited to match 
the 3-dB width of an 8-VSB signal for all other channels. 
 
Measurements of interference rejection performance are specified in terms of desired signal power D and 
undesired (i.e., interfering) signal power U at the RF input of the DTV receiver.  We have chosen to 
present the interference rejection performance measurements in two ways:  as D/U ratios and as threshold 
values for the undesired signal level U.  Each has its own application.  D/U ratios might be preferable for 
all analytical work if they were constant with desired signal power; however, nonlinearity of interference 
mechanisms and the effects of receiver noise at low desired signal levels cause D/U ratios to be variable.  
Nonetheless, D/U ratios can be convenient to use in applications like DTV-into-DTV interference 
assessment because estimation of D/U ratios may be easier and more accurate that estimation of absolute 
levels where long-distance propagation is involved—especially if the broadcast stations are co-sited.  Use 
of absolute signal level thresholds may be more appropriate for assessing shorter distance interference 
from low-power devices because the effects of TV antenna height and placement on the undesired signal 
are likely to be very different from their effects on the desired signal. 
 
No receiver appeared to fully achieve the ATSC recommended guidelines for interference rejection 
performance.  After taking into account differences between the Gaussian-noise interferer used for most 
of the tests and the 8-VSB interferer specified by the ATSC, the best-performing receiver appears to fail 
the guidelines at only one channel offset, and there by only 1 dB.  A second receiver failed to meet the 
voluntary guidelines by 1 to 2 dB at two channel offsets.  The remaining five receivers failed to meet the 
guidelines at two to 16 channel offsets; the worst failure for each of those receivers ranged from 8 to 
24 dB. 
 
The single-channel rejection performance measurements performed for this report are best summarized by 
Figures 15-1 through 15-6.  The first graph presents the rejection performance in terms of D/U ratios and 
                                                      
*
 Tests of interference rejection thresholds for DTV into analog NTSC TV for 8 taboo channels show that D/U ratios 
differed by an average of 8.1 dB between TOV and CCIR Grade 3 (“Slightly Annoying”) at a weak signal level.  
These results were obtained by averaging data from table in Bretl and Sgrignoli, 1996, section 3.8.3. 
†
 Tests of lower-adjacent interference rejection thresholds for DTV into analog NTSC TV showed a difference of 
20.2 dB between D/U ratio at TOV and D/U ratio at the point of unusability at a weak signal level.  Similar tests for 
co-channel interference showed a difference of 32 dB between TOV and the point of unusability (Bretl and 
Sgrignoli, 1996, sections 3.8.1 and 3.8.2.1). 
 
 15-4
the second in terms of the undesired signal power at TOV.  The first pair of graphs presents the median 
rejection performance across the eight DTV receivers at each of the tested channel offsets (from N-16 to 
N+16—omitting the co-channel case).  The second pair presents the second-worst performance across the 
eight receivers.  The third pair presents the worst performance. 
 
Each graph shows the interference thresholds at five desired signal levels:  -28 dBm, -53 dBm, -68 dBm, 
D
MIN
 + 3 dB, and D
MIN
 + 1 dB, where D
MIN
 is the desired signal level corresponding to the TOV for the 
receiver in the absence of interference.  For measurements at D
MIN
 + 3 dB, each TV was tested by first 
measuring its minimum signal threshold D
MIN
, then setting the desired signal power 3 dB higher that 
threshold for the interference rejection tests.  Thus the desired signal level was different for each DTV 
receiver.  The results at D
MIN
 + 1 dB were extrapolated from the measured values at D
MIN
 + 3 dB by 
means of a model developed in Chapters 8 and 12. 
 
Each graph shows measurement limitations imposed by the test setup—in the form of solid black lines 
and a shaded region.  The D/U plots (Figures 15-1, 15-3, and 15-5) show four measurement limit curves 
that correspond, from top to bottom, to limits at D = -28 dBm, -53 dBm, -68 dBm, and D
MIN
 + 3 dB, 
respectively.  In the case of data extrapolated to D
MIN
 + 1 dB, data points are shown only if the 
measurements on which they were based were not subject to measurement limits.  The undesired signal 
threshold plots (even-numbered Figures 15-2, 15-4, and 15-6) show only one measurement limit curve—
the curve associated with the maximum undesired signal power that the test setup could inject into the 
receiver; the N-1 and N+1 offsets for D = -68 dBm and all of the offsets for D = D
MIN
 + 3 dB are subject 
to an additional limitation—shown only in the D/U plots—that is caused by leakage of the undesired 
signal into the desired channel. 
 
We make the following observations regarding the results. 
• In terms of absolute signal levels that can cause interference, the TVs are at their most vulnerable 
when operating at low desired signal levels. 
• At low desired signal levels the TV receivers are as susceptible to interference from the second-
adjacent channels (N-2 and N+2) as from first-adjacent channels (N-1 and N+1) in terms of median 
performance of the receivers.  In terms of worst and second-worst performance, the receivers are 
more susceptible to interference from second-adjacent channels than from first-adjacent channels.  
(This contradicts the assumptions of OET-69 and the receiver performance guidelines of ATSC 
Document A/74.) 
• The receivers tend to be more susceptible to interference from N+2, N+1, N-1, N-2, N-3, N-4, and 
sometimes N-6 than from the mixer image channel offsets of N+14 and N+15. 
• At moderate desired signal levels, the receivers exhibit relatively high susceptibility to interference 
from channel N+7.  This interference threshold is nearly constant in terms of absolute power of the 
undesired signal necessary to cause interference at different levels of desired signals.  At lower 
desired signal levels, other channel offsets become more vulnerable. 
 
INTERFERENCE FROM IM3-GENERATING PAIRS OF UNDESIRED 
SIGNALS 
Pairs of undesired signals placed on channels N+K and N+2K, where K is a positive or negative integer, 
create an opportunity for third-order intermodulation (IM3) occurring in the DTV tuner to create spectral 
products that fall in the desired channel N.  We had anticipated paired-signal IM3 effects would be 
significant only at high signal levels; however, detailed measurements on one DTV receiver (Chapter 11) 
demonstrated that such effects can constitute a dominant interference susceptibility even at desired signal 
levels very near the minimum signal threshold for the TV, when such signal pairs exist.  
 
 
 15-5
Measurements of interference thresholds were performed with equal-powered undesired signals on 
N+K/N+2K combinations for eight DTVs on channel 30 and seven of those DTVs on channel 51.  Tests 
were performed for K = -5 to 5 when N was 30 and for K = 1 to 8 when N was 51.  In both cases, desired 
signal levels were set to -68 dBm and -53 dBm.  Not all measured cases produced interference effects that 
were sufficiently higher than the single-channel interference effects to allow measurement of the IM3 
effects.  For those that were, a third-order intercept point (IP3) was computed.   
 
The data was carefully examined to identify cases for which median, second-worst, and worst IP3 across 
the eight receivers could be determined.  In cases in which corresponding values (e.g., “second-worst”) 
were obtained from both the channel 30 measurements and the channel 51 measurements, the results were 
averaged across the two channels; in cases where only one channel yielded a value, that value was used.  
IP3’s computed from that data were used to extend the results, by calculation, to desired power levels of 
D
MIN
 + 3 dB and D
MIN
 + 1 dB.  For simplicity, that calculation assumed that D
MIN
 = -84 dBm for each 
receiver. 
 
Figure 15-7 shows the median D/U ratios across the eight tested receivers for a desired signal level of 
-68 dBm.  The “Single Signal” curve is a duplicate of the corresponding curve from Figure 15-1.  Paired 
signal combinations for which IM3 effects were measurable across a sufficient set of receivers to allow 
the median IP3 to be computed are shown as pairs of large red dots connected by straight lines.  The 
horizontal pairs represent rejection thresholds for equal-level undesired signals.  For example, the 
connected pair of dots at N+3 and N+6 indicates that the median D/U ratio for a signal pair at N+3 and 
N+6 is –43.7 dB when the desired signal level is -68 dBm.  (For equal-powered undesired signals, the 
“U” in the D/U ratio is U = U
N+K
 = U
N+2K
.) 
 
The right-most signal pair (N+8/N+16) was plotted in three ways—as an equal-power signal pair 
(horizontal line) and as two sets of unequal pairs.  Unequal pairs were created by raising and lowering the 
N+8 point by 10 dB, which results in a +/-20 dB change in the threshold for N+16.  The example 
illustrates behavior that would be exhibited by any of the signal pairs for unequal signals with such a 
signal level deviation.  That is, if U
N+K
 were to change by X dB from the equal-level threshold value, the 
U
N+2K
 threshold would change by -2X dB.  Thus, the presence of a signal stronger than the equal-power 
threshold on one channel in a pair, makes the other channel susceptible to weaker undesired signals.  In 
the example, increasing the undesired signal power at channel N+8 by 10 dB from the equal-power 
threshold (resulting in a 10 dB decrease in D/U at N+8) causes the threshold D/U ratio at N+16 to 
increase from its equal power value of -51.4 dB to -31.4 dBm.  The result is that the susceptibility of the 
TV to interference on channel N+16 is now greater than its susceptibility to interference on the first 
adjacent channels (N-1 or N+1).  Charts and tables in Chapter 10 illustrate the range of thresholds that can 
result. 
 
We note that the behavior of unequal signal pairs as described in the previous paragraph and quantified in 
the charts and tables of Chapter 10 is valid only over signal level regions for which the receiver’s 
automatic gain control (AGC) does not act to reduce tuner gain prior to the nonlinearity that causes the 
IM3 effects (usually in the mixer).  In some cases (including one identified in Chapter 14), a large 
increase in U
N+K
 for small K values may cause such AGC gain reductions.  When that happens, the model 
predicts that the sensitivity of the receiver to interference on N+2K will freeze—exhibiting no further 
increases in interference susceptibility with further increases in U
N+K
. 
 
Regarding signal pairs that are missing from Figure 15-7 (and from the subsequent plots that will be 
introduced), we note the following. 
• In the case of the first-adjacent pairs (N-1/N-2 and N+1/N+2) the values are missing because the 
measurements of paired-signal thresholds for those channel pairs did not exceed single-channel 
effects on any of the receivers by a sufficient amount to support measurement of IM3 effects.  (We 
note that equal undesired signal powers levels are not necessarily optimal for detecting IP3 effects, so 
 
 15-6
it is not known whether IM3 estimates could have been obtained from measurements at unequal 
levels.) 
• In the case of channel pairs N-3/N-6, N-2/N-4, N+2/N+4 and N+7/N+14, IM3 effects were 
successfully measured for some of the receivers, but the receivers that did not support such 
measurement created uncertainty in trying to determine worst, second-worst, and median values of 
the IP3 parameter; consequently, those values are missing from the graphs. 
• No signal pair measurements were performed beyond N-5/N-10 in on the left side of the plots. 
 
It can be seen on the right side of the plot, where measurements are more plentiful, that the paired-signal 
interference effects gradually decrease with separation from the desired channel. 
 
If one were interested in determining general undesired signal levels at which IM3 can interfere with TV 
reception, the “Equal Signal Pairs” plots in Figure 15-7 (and in subsequent charts to be described next) 
could be used to identify a level that could cause interference if a similar signal level happens to occur at 
another channel offset that would place IM3 products in the desired channel.  If, on the other hand, one 
wanted to determine case-specific interference vulnerabilities that take into account existing undesired 
sources (e.g., a nearby DTV broadcast station when the receiver is tuned to a more distant station), one 
could use the data from Chapter 10 to determine the signal level associated, for example, with a new non-
TV service that would cause interference under specific reception conditions.  Such analysis using signal 
levels from existing channel allotments could reveal greater interference susceptibilities than those based 
on equal signal pairs. 
 
Figure 15-8 presents the same data as Figure 15-7, but shows it as the threshold value of undesired signal 
power U rather than as a D/U ratio. 
 
The previous two plots corresponded to median receiver performance with a desired signal level of 
-68 dBm.  The subsequent two pairs of charts show similar data for the second-worst and worst 
performing receivers.  That six-chart sequence is followed by six charts corresponding to a desired signal 
level of D
MIN
 + 3 dB and six more charts corresponding to a desired signal of D
MIN
 + 1 dB. 
 
The plots show that IM3 between paired signals can be the dominant source of interference at many 
channel offsets, when undesired signals exist at IM3-generating spacings (N+K/N+2K).  IM3 interference 
effects from paired signals on the first-adjacent channel pair (N+1/N+2 or N-1/N-2) appear to be less 
important than the single-channel interference effects for first adjacent channels; the same is sometimes, 
but not always, true for the second adjacent pair (N+2/N+4 or N-2/N-4).  At other channel spacings the 
paired signal IM3 effects dominate when the desired signal level is -68 dBm. 
 
As desired signal level drops, IM3 effects diminish more rapidly than first- and second-adjacent single-
channel interference effects.  This can be seen by comparing, for example, Figures 15-7 and 15-19.  
Nonetheless, paired-signal IM3 appears to be the dominant interference vulnerability for channel offsets 
from about N+4 to N+16 (with the exception of the mixer image at N+14 and N+15) and from about N-5 
to N-10, even at desired signal levels near D
MIN
 and even if the paired signals are assumed to be equal in 
level.  No paired signal measurements were performed beyond N+16 and N-10, so it is not know how far 
out the effect continues; however, the effect is seen to diminish with increasing channel offset from the 
desired channel. 
 
Paired signals at IM3-generating spacings have the potential to create even greater interference 
susceptibilities if an existing undesired signal on one of the IM3-generating channels (e.g., a nearby DTV 
broadcast station when the receiver is tuned to a more distant station) exceeds the measured equal-power-
level threshold for paired signals.  In such a case, the presence of that signal can greatly increase 
susceptibility to interference on the other channel of the IM3-generating pair.  This situation generally 
creates the greatest vulnerabilities when the stronger undesired signal is on channel N+K and it exceeds 
 
 15-7
the equal-power paired-signal threshold; in that case, the receiver susceptibility to interference on the 
N+2K channel increases by twice the N+K signal excess above the equal power threshold. 
 
The ATSC Receiver Guidelines document (A/74) provides no recommended performance levels for 
rejection of paired-signal interference. 
 
 15-8
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
c
e
p
tibility
 to Inte
r
f
er
e
n
ce
 
[
D
/U 
at Th
r
e
shold (
d
B)
]
Meas. Limit at
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
Meas. Limits
MLiit
Median of 8 DTV Receivers
 
Figure 15-1.  Median D/U of 8 Receivers at Five Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
Median of 8 DTV Receivers
 
Figure 15-2.  Median Threshold U of 8 Receivers at Five Signal Levels on Channel 30 
 
 15-9
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
c
e
p
tibility
 to Inte
r
f
er
e
n
ce
 
[
D
/U 
at Th
r
e
shold (
d
B)
]
Meas. Limit at
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
Meas. Limits
MLiit
2nd Worst of 8 DTV Receivers
 
Figure 15-3.  2
nd
 Worst D/U of 8 Receivers at Five Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
2nd Worst of 8 DTV Receivers
 
Figure 15-4.  2
nd
 Worst Threshold U of 8 Receivers at Five Signal Levels on Channel 30 
 
 15-10
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Sus
c
e
p
tibility
 to Inte
r
f
er
e
n
ce
 
[
D
/U 
at Th
r
e
shold (
d
B)
]
Meas. Limit at
Dmin+3dB
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
Meas. Limits
MLiit
Worst of 8 DTV Receivers
 
Figure 15-5.  Worst D/U of 8 Receivers at Five Signal Levels on Channel 30 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Undesired Signal Level (U) 
at Thresho
l
d (dBm
)
Measurement Limit
D = -28 dBm
D = -53 dBm
D = -68 dBm
D = Dmin+3dB
D = Dmin+1dB
Worst of 8 DTV Receivers
 
Figure 15-6.  Worst Threshold U of 8 Receivers at Five Signal Levels on Channel 30 
 
 15-11
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susc
eptibility
 to Inte
r
f
erence
 
[
D
/U a
t
 T
h
re
s
h
ol
d 
(dB)]
Single Signal
Meas. Limit---68dBm
Equal Signal Pairs
Unequal Signal Pairs
Median of 8 DTV Receivers
D = -68 dBm
 
Figure 15-7.  Median D/U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
U
n
desir
e
d Sig
n
al Level 
(U
) at 
Thr
esh
o
ld 
(dB
m
)
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
Median of 8 DTV Receivers
D = -68 dBm
 
Figure 15-8.  Median Threshold U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs 
 
 15-12
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susc
eptibility
 to Inte
r
f
erence
 
[
D
/U a
t
 T
h
re
s
h
ol
d 
(dB)]
Single Signal
Meas. Limit---68dBm
Equal Signal Pairs
Unequal Signal Pairs
2nd Worst of 8 DTV Receivers
D = -68 dBm
 
Figure 15-9.  2
nd
 Worst D/U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
U
n
desir
e
d Sig
n
al Level 
(U
) at 
Thr
esh
o
ld 
(dB
m
)
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
2nd Worst of 8 DTV Receivers
D = -68 dBm
 
Figure 15-10.  2
nd
 Worst Threshold U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs 
 
 15-13
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susc
eptibility
 to Inte
r
f
erence
 
[
D
/U a
t
 T
h
re
s
h
ol
d 
(dB)]
Single Signal
Meas. Limit---68dBm
Equal Signal Pairs
Unequal Signal Pairs
Worst of 8 DTV Receivers
D = -68 dBm
 
Figure 15-11.  Worst D/U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
U
n
desir
e
d Sig
n
al Level 
(U
) at 
Thr
esh
o
ld 
(dB
m
)
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
Worst of 8 DTV Receivers
D = -68 dBm
 
Figure 15-12.  Worst Threshold U of 8 Receivers at D = -68 dBm With IM3 Signal Pairs 
 
 15-14
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susc
eptibility
 to Inte
r
f
erence
 
[
D
/U a
t
 T
h
re
s
h
ol
d 
(dB)]
Meas. Limit at Dmin+3dB
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
Median of 8 DTV Receivers
D = Dmin+3dB
 
Figure 15-13.  Median D/U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Un
de
si
re
d Si
gn
a
l
 L
e
v
e
l
 (U) 
a
t
 Th
re
sh
ol
d
 (dBm
)
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
Median of 8 DTV Receivers
D = Dmin+3dB
 
Figure 15-14.  Median Threshold U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs 
 
 15-15
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susc
eptibility
 to Inte
r
f
erence
 
[
D
/U a
t
 T
h
re
s
h
ol
d 
(dB)]
Meas. Limit at Dmin+3dB
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
2nd Worst of 8 DTV Receivers
D = Dmin+3dB
 
Figure 15-15.  2
nd
 Worst D/U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
U
n
desir
e
d Sig
n
al Level 
(U
) at 
Thr
esh
o
ld 
(dB
m
)
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
2nd Worst of 8 DTV Receivers
D = Dmin+3dB
 
Figure 15-16.  2
nd
 Worst Threshold U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs 
 
 15-16
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susc
eptibility
 to Inte
r
f
erence
 
[
D
/U a
t
 T
h
re
s
h
ol
d 
(dB)]
Meas. Limit at Dmin+3dB
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
Worst of 8 DTV Receivers
D = Dmin+3dB
 
Figure 15-17.  Worst D/U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
U
n
desir
e
d Sig
n
al Level 
(U
) at 
Thr
esh
o
ld 
(dB
m
)
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
Worst of 8 DTV Receivers
D = Dmin+3dB
 
Figure 15-18.  Worst Threshold U of 8 Receivers at D = D
MIN
+3dB With IM3 Signal Pairs 
 
 15-17
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Suscepti
b
ili
ty to 
In
ter
f
er
en
ce 
[
D
/
U
 at
 Th
r
e
sh
ol
d (
d
B
)
]
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
Median of 8 DTV Receivers
D = Dmin+1dB
 
Figure 15-19.  Median D/U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
U
ndesi
r
e
d 
Sign
al 
Level (U
) at T
h
r
e
shol
d (d
B
m
)
D = Dmin+1dB
Equal Signal Pairs
Unequal Signal Pairs
Median of 8 DTV Receivers
D = Dmin+1dB
 
Figure 15-20.  Median Threshold U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs 
 
 15-18
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susc
eptibility
 to Inte
r
f
erence
 
[
D
/U a
t
 T
h
re
s
h
ol
d 
(dB)]
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
2nd Worst of 8 DTV Receivers
D = Dmin+1dB
 
Figure 15-21.  2
nd
 Worst D/U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
U
n
desir
e
d Sig
n
al Level 
(U
) at 
Thr
esh
o
ld 
(dB
m
)
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
2nd Worst of 8 DTV Receivers
D = Dmin+1dB
 
Figure 15-22.  2
nd
 Worst Threshold U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs 
 
 15-19
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
Susc
eptibility
 to Inte
r
f
erence
 
[
D
/U a
t
 T
h
re
s
h
ol
d 
(dB)]
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
Worst of 8 DTV Receivers
D = Dmin+1dB
 
Figure 15-23.  Worst D/U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs 
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
N
-
1
6
N
-
1
5
N
-
1
4
N
-
1
3
N
-
1
2
N
-
1
1
N
-
1
0
N
-
9
N
-
8
N
-
7
N
-
6
N
-
5
N
-
4
N
-
3
N
-
2
N
-
1
N
+
1
N
+
2
N
+
3
N
+
4
N
+
5
N
+
6
N
+
7
N
+
8
N
+
9
N
+
1
0
N
+
1
1
N
+
1
2
N
+
1
3
N
+
1
4
N
+
1
5
N
+
1
6
Interfering Channel
U
n
desir
e
d Sig
n
al Level 
(U
) at 
Thr
esh
o
ld 
(dB
m
)
Single Signal
Equal Signal Pairs
Unequal Signal Pairs
Worst of 8 DTV Receivers
D = Dmin+1dB
 
Figure 15-24.  Worst Threshold U of 8 Receivers at D = D
MIN
+1dB With IM3 Signal Pairs 
 
A-1 
APPENDIX A 
TABULATED REJECTION PERFORMANCE RESULTS 
This appendix includes tabulations of statistical data from the interference rejection performance 
measurements of eight DTV receivers on tuned to channel 30.  The data are presented as D/U ratios and 
as threshold levels of the undesired signal U in separate major sections of this appendix. 
 
Within each major section, a separate table is provided for rejection performance data at each of five 
desired signal levels: 
• D
MIN
 + 1 dB 
• D
MIN
 + 3 dB 
• -68 dBm 
• -53 dBm 
• -28 dBm 
 
The rejection performance results for the four higher levels were obtained by direct measurement.  For 
that data, “>” or “<” symbols appear when the statistical parameter shown depends on a value that was at 
the measurement limit of the test setup.  For example, if the “Best U” value is listed as >-6.6 dBm, that 
means that a valid measurement of rejection performance of the best performing receiver was not 
obtained due to measurement system limitations.  Specifically, at that channel offset, the maximum 
undesired signal power that the test setup was capable of supplying to the DTV input was -6.6 dBm and 
the TV picture was still flawless at that level; thus we know only that the threshold undesired signal 
power of the best receiver was greater than -6.6 dBm.  If only the best receiver measurement was so 
limited, values will be listed for the median, 2
nd
 worst, and worst performance, but not for the mean or 
standard deviation, because those values can be computed only if all of the measurement values were 
valid.   
 
The rejection results for D
MIN
 + 1 dB were extrapolated from measurements at D = -68 dBm and D = 
D
MIN
 + 3 dB, but only if valid measurements were available at both of those levels.  Otherwise the 
corresponding cell in the table is left blank.   
 
In the case of the measurements at D = D
MIN
 + 3 dB or extrapolations for D = D
MIN
 + 3 dB, the desired 
signal level was different for each receiver, based on separately measured D
MIN
 values for the receivers.  
Table A-1 shows statistics for the D
MIN
 values. 
 
 
Table A-1.  Statistics of D
MIN
  for 8 Receivers on Channel 30 
  D
MIN
 D
MIN
+3dB 
Minimum (dBm) -86.2 -83.2
Median (dBm) -84.0 -81.0
Mean (dBm) -83.9 -80.9
Maximum (dBm) -81.8 -78.8
Standard 
Deviation (dB) 1.3 1.3
 
 
A-2 
THRESHOLD D/U STATISTICS FOR 8 FIFTH-GENERATION DTV 
RECEIVERS 
Desired Signal = DMIN + 1 dB (Extrapolated from Measurements at DMIN + 3 dB) 
 
Table A-2.  D/U Statistics for 8 Receivers at D = D
MIN
 + 1 dB on Channel 30 
Undesired 
Channel 
Best 
D/U 
(dB) 
Median 
D/U 
(dB) 
Mean 
D/U 
(dB) 
2nd 
Worst 
D/U 
(dB) 
Worst 
D/U 
(dB) 
Standard 
Deviation 
(dB) 
N-16         
N-15         
N-14         
N-13  -73.0 -64.4 -64.2  
N-12 -72.3 -62.7 -60.9
N-11  -70.6 -61.0 -59.0  
N-10 -68.2 -59.3 -57.0
N-9  -66.8 -56.5 -55.0  
N-8 -72.6 -62.7 -60.3 -52.5 -45.1 8.9 
N-7 -71.7 -62.1 -58.7 -50.0 -44.6 8.8 
N-6 -72.9 -53.0 -51.6 -43.2 -26.6 13.9 
N-5 -71.0 -53.3 -54.0 -47.2 -40.6 8.9 
N-4 -65.6 -45.4 -47.9 -41.8 -36.2 9.5 
N-3 -57.5 -45.5 -46.5 -40.4 -36.9 7.1 
N-2 -49.1 -36.0 -36.5 -29.3 -23.2 8.4 
N-1 -38.2 -35.4 -35.4 -33.0 -32.8 2.0 
N+1 -38.9 -36.7 -36.2 -33.9 -32.2 2.3 
N+2 -50.7 -37.5 -37.8 -30.3 -23.5 9.2 
N+3 -61.9 -50.2 -50.6 -44.3 -43.5 5.8 
N+4 -69.8 -55.3 -56.1 -50.9 -47.7 6.7 
N+5 -67.3 -61.6 -61.0 -59.6 -50.2 5.0 
N+6         
N+7 -72.1 -64.4 -64.0 -57.8 -52.8 7.1 
N+8         
N+9         
N+10         
N+11         
N+12         
N+13         
N+14   -54.4  -48.1 -47.7   
N+15    -48.8  -44.9 -42.9   
N+16         
 
 
A-3 
Desired Signal = D
MIN
 + 3 dB 
 
Table A-3.  D/U Statistics for 8 Receivers at D = D
MIN
 + 3 dB on Channel 30 
Undesired 
Channel 
Best 
D/U 
(dB) 
Median 
D/U 
(dB) 
Mean 
D/U (dB)
2nd 
Worst 
D/U 
(dB) 
Worst 
D/U (dB)
Standard 
Deviation 
(dB) 
N-16 < -76.0 -73.2 < -72.3 -67.8 -67.8 > 3.4 
N-15 < -76.0 -71.4 < -72.1 -69.2 -69.0 > 2.8 
N-14 < -75.3 -72.7 < -71.8 -67.6 -67.1 > 3.4 
N-13 < -74.8 -72.9 < -71.4 -66.7 -65.3 > 3.9 
N-12 < -74.8 -72.3 < -70.7 -65.2 -63.5 > 4.5 
N-11 < -74.6 -70.5 < -68.9 -63.5 -61.6 > 4.8 
N-10 < -74.7 -68.7 < -67.8 -61.8 -59.6 > 5.0 
N-9 < -74.1 -67.2 < -65.3 -59.1 -57.5 > 5.9 
N-8 -72.5 -64.1 -61.8 -55.1 -47.7 7.9 
N-7 -71.7 -63.0 -60.2 -52.6 -47.1 7.9 
N-6 -72.9 -55.5 -53.5 -45.7 -29.2 13.1 
N-5 -71.0 -55.9 -56.0 -49.7 -43.2 8.0 
N-4 -66.6 -47.9 -50.0 -44.3 -38.8 8.8 
N-3 -60.1 -48.1 -49.1 -42.9 -39.5 7.1 
N-2 -51.7 -38.5 -39.1 -31.9 -25.7 8.4 
N-1 -40.7 -38.0 -38.0 -35.6 -35.3 2.0 
N+1 -41.5 -39.3 -38.7 -36.4 -34.7 2.3 
N+2 -51.5 -40.1 -40.1 -32.9 -26.1 8.9 
N+3 -62.8 -52.8 -52.6 -46.8 -46.0 5.5 
N+4 -69.6 -57.1 -57.6 -53.5 -47.6 6.4 
N+5 -67.2 -63.2 -62.0 -61.8 -50.1 5.1 
N+6 -73.7 -72.1 -70.0 -66.7 -61.3 4.2 
N+7 -72.0 -64.4 -63.9 -57.8 -52.7 7.1 
N+8 < -77.1 <-74.6 < -74.5 -73.2 -72.1 > 1.5 
N+9 < -78.4 <-75.2 < -75.5 <-74.8 -73.3 > 1.5 
N+10 < -77.9 <-74.7 < -75.0 -73.1 -72.8 > 1.9 
N+11 < -78.3 <-75.8 < -75.8 <-75.0 -73.2 > 1.5 
N+12 < -79.3 <-75.8 < -76.3 <-75.1 -73.1 > 2.0 
N+13 < -79.1 <-76.0 < -76.4 <-75.0 -73.2 > 1.9 
N+14 < -78.8 -58.2 < -61.5 -51.9 -51.6 > 9.9 
N+15 < -78.3 -52.6 < -56.5 -48.7 -46.8 > 10.2 
N+16 < -77.7 <-75.7 < -75.8 <-74.9 -73.2 > 1.4 
 
A-4 
Desired Signal = -68 dBm 
Table A-4.  D/U Statistics for 8 Receivers at D = -68 dBm on Channel 30 
Undesired 
Channel 
Best 
D/U 
(dB) 
Median 
D/U 
(dB) 
Mean 
D/U (dB)
2nd 
Worst 
D/U 
(dB) 
Worst 
D/U (dB)
Standard 
Deviation 
(dB) 
ATSC 
Performance 
Guideline 
(dB) 
N-16 < -66.9 -64.2 < -63.9 -62.2 -59.2 > 2.7  
N-15 < -67.0 -64.1 < -64.0 -61.8 -59.7 > 2.7 -50.0
N-14 < -67.2 -64.1 < -63.8 -61.6 -59.3 > 2.8 -50.0
N-13 -67.0 -64.1 -63.6 -61.4 -58.8 2.8 -57.0
N-12 -66.1 -62.7 -63.0 -61.3 -58.3 2.7 -57.0
N-11 -65.9 -62.0 -62.1 -60.8 -56.7 2.8 -57.0
N-10 -66.0 -63.5 -62.4 -59.6 -57.1 3.0 -57.0
N-9 -65.0 -60.9 -60.9 -58.1 -56.0 3.1 -57.0
N-8 -64.2 -58.0 -58.3 -55.0 -50.7 4.5 -57.0
N-7 -64.2 -58.2 -57.2 -53.5 -48.4 5.1 -57.0
N-6 -62.5 -52.9 -52.0 -48.3 -31.8 9.7 -57.0
N-5 -62.6 -55.5 -54.3 -49.7 -44.5 5.7 -56.0
N-4 -61.9 -47.4 -49.3 -45.8 -40.7 6.3 -52.0
N-3 -59.8 -49.6 -50.1 -44.3 -41.9 6.2 -48.0
N-2 -49.7 -40.9 -40.1 -32.0 -28.1 7.4 -44.0
N-1 -40.1 -39.3 -39.1 -38.0 -37.9 0.8 -33.0
N+1 -42.1 -39.7 -39.7 -38.3 -37.9 1.4 -33.0
N+2 -48.3 -42.3 -41.1 -34.3 -29.8 6.9 -44.0
N+3 -57.0 -54.6 -51.9 -48.3 -39.8 5.8 -48.0
N+4 -60.0 -56.6 -54.9 -54.4 -41.1 5.9 -52.0
N+5 -65.4 -58.4 -56.9 -53.6 -42.6 7.1 -56.0
N+6 < -66.3 -63.1 < -61.6 -56.3 -52.9 > 4.6 -57.0
N+7 -60.3 -53.3 -51.9 -45.0 -41.6 6.8 -57.0
N+8 < -66.6 -65.0 < -64.5 -63.1 -60.9 > 2.0 -57.0
N+9 < -66.6 -65.9 < -65.2 -63.4 -62.1 > 1.6 -57.0
N+10 < -66.5 -65.0 < -65.0 -63.1 -62.9 > 1.4 -57.0
N+11 < -66.2 <-65.9 < -65.2 -63.2 -63.2 > 1.3 -57.0
N+12 < -66.0 <-65.7 < -65.0 -62.9 -62.9 > 1.3 -57.0
N+13 < -65.9 <-65.5 < -64.9 -63.3 -63.2 > 1.1 -57.0
N+14 < -65.3 -60.3 < -59.2 -52.9 -51.2 > 5.3 -50.0
N+15 < -64.8 -55.0 < -56.0 -50.1 -48.7 > 5.7 -50.0
N+16 < -64.5 <-64.2 < -64.1 -63.6 -63.4 > 0.4  
N-5/N-10 -49.8 -42.3 -43.6 -41.7 -37.2 4.2 
N-4/N-8 -46.3 -41.9 -42.4 -39.1 -38.2 3.1 
N-3/N-6 -45.2 -41.7 -40.6 -36.0 -33.0 4.1 
N-2/N-4 -51.4 -40.7 -39.3 -30.7 -29.5 7.0 
N-1/N-2 -39.2 -35.7 -34.8 -29.9 -27.3 4.1 
N+1/N+2 -39.7 -36.5 -36.1 -34.9 -29.0 3.3 
N+2/N+4 -42.3 -38.0 -37.2 -31.0 -28.9 5.1 
N+3/N+6 -51.1 -43.5 -44.0 -40.6 -33.9 5.8 
N+4/N+8 -50.8 -46.8 -46.0 -42.1 -39.6 4.2 
N+5/N+10 -52.5 -48.3 -47.9 -44.4 -42.5 3.3 
 
 
A-5 
Desired Signal = -53 dBm 
 
Table A-5.  D/U Statistics for 8 Receivers at D = -53 dBm on Channel 30 
Undesired 
Channel 
Best 
D/U 
(dB) 
Median 
D/U 
(dB) 
Mean 
D/U (dB)
2nd 
Worst 
D/U 
(dB) 
Worst 
D/U (dB)
Standard 
Deviation 
(dB) 
ATSC 
Performance 
Guideline 
(dB) 
N-16 < -52.2 -51.4 < -50.5 -49.2 -45.7 > 2.2  
N-15 < -52.4 <-52.0 < -50.6 -49.0 -45.1 > 2.5 -45.0
N-14 < -52.6 <-52.2 < -50.6 -48.8 -44.5 > 2.8 -45.0
N-13 < -52.8 -52.0 < -50.5 -48.8 -44.3 > 2.9 -45.0
N-12 < -52.9 -50.8 < -50.1 -48.6 -43.7 > 3.1 -45.0
N-11 < -52.9 -50.9 < -49.8 -46.8 -42.9 > 3.5 -45.0
N-10 < -52.9 -52.3 < -50.5 -48.2 -42.3 > 3.6 -45.0
N-9 < -52.7 -52.2 < -50.2 -48.1 -41.4 > 3.9 -45.0
N-8 < -52.6 -51.4 < -49.5 -47.1 -40.2 > 4.3 -45.0
N-7 < -52.3 -50.5 < -48.9 -45.7 -38.8 > 4.7 -45.0
N-6 < -52.1 -47.7 < -45.7 -36.9 -32.4 > 7.3 -45.0
N-5 < -51.9 -51.6 < -48.4 -43.8 -34.7 > 6.2 -42.0
N-4 < -51.9 -47.0 < -45.7 -39.9 -32.3 > 6.6 -40.0
N-3 < -51.8 -49.1 < -48.2 -45.2 -41.9 > 3.3 -40.0
N-2 -49.0 -41.5 -39.8 -31.8 -27.4 7.1 -40.0
N-1 -40.0 -39.0 -38.5 -37.7 -34.4 1.8 -33.0
N+1 -41.9 -39.4 -39.4 -38.3 -36.0 1.8 -33.0
N+2 -46.8 -42.2 -40.6 -34.5 -30.3 6.3 -40.0
N+3 < -50.9 -49.8 < -46.8 -48.1 -25.1 > 8.8 -40.0
N+4 < -51.0 -50.1 < -46.6 -45.1 -29.1 > 7.5 -40.0
N+5 < -51.2 -49.0 < -46.7 -44.5 -31.9 > 6.5 -42.0
N+6 < -51.4 -50.8 < -48.9 -46.3 -39.8 > 4.1 -45.0
N+7 -51.2 -38.7 -38.8 -32.6 -26.9 8.0 -45.0
N+8 < -51.7 <-51.4 < -50.9 -49.6 -48.7 > 1.1 -45.0
N+9 < -51.6 <-51.3 < -51.1 <-51.2 -49.2 > 0.8 -45.0
N+10 < -51.5 <-51.2 < -51.1 <-51.1 -49.8 > 0.5 -45.0
N+11 < -51.2 <-51.0 < -51.0 <-50.9 -50.4 > 0.3 -45.0
N+12 < -51.2 <-50.8 < -50.8 <-50.7 -50.4 > 0.3 -45.0
N+13 < -50.9 <-50.6 < -50.6 <-50.4 <-50.4 > 0.2 -45.0
N+14 < -50.5 <-50.1 < -50.2 <-50.0 -50.0 > 0.2 -45.0
N+15 < -50.0 <-49.7 < -49.5 <-49.6 -47.6 > 0.8 -45.0
N+16 < -49.5 <-49.3 < -49.3 <-49.2 <-49.2 > 0.1  
N-5/N-10 -39.0 -33.6 -34.3 -32.2 -30.6 2.8 
N-4/N-8 -38.4 -34.6 -34.2 -32.5 -28.2 3.1 
N-3/N-6 -43.6 -37.4 -36.4 -32.5 -25.4 5.8 
N-2/N-4 -41.9 -38.2 -36.4 -28.6 -28.6 5.2 
N-1/N-2 -39.0 -36.0 -34.4 -27.4 -26.3 4.9 
N+1/N+2 -38.7 -35.4 -35.2 -33.3 -30.1 2.6 
N+2/N+4 -38.9 -35.1 -33.4 -31.4 -20.8 5.7 
N+3/N+6 -41.6 -34.1 -34.4 -29.9 -22.9 6.1 
N+4/N+8 -41.2 -35.8 -36.0 -33.6 -28.4 4.2 
N+5/N+10 -41.6 -38.2 -37.7 -34.3 -32.2 3.1 
 
A-6 
Desired Signal = -28 dBm 
 
Table A-6.  D/U Statistics for 8 Receivers at D = -28 dBm on Channel 30 
Undesired 
Channel 
Best 
D/U 
(dB) 
Median 
D/U 
(dB) 
Mean 
D/U (dB)
2nd 
Worst 
D/U 
(dB) 
Worst 
D/U (dB)
Standard 
Deviation 
(dB) 
ATSC 
Performance 
Guideline 
(dB) 
N-16 < -27.3 <-27.2 < -27.1 <-26.9 <-26.8 > 0.2  
N-15 < -27.4 <-27.3 < -27.2 <-27.0 <-26.9 > 0.2 -20.0
N-14 < -27.6 <-27.5 < -27.4 <-27.2 <-27.1 > 0.2 -20.0
N-13 < -27.8 <-27.7 < -27.6 <-27.3 <-27.3 > 0.2 -20.0
N-12 < -27.9 <-27.8 < -27.7 <-27.5 <-27.4 > 0.2 -20.0
N-11 < -28.0 <-27.8 < -27.7 <-27.5 <-27.4 > 0.2 -20.0
N-10 < -27.9 <-27.8 < -27.7 <-27.5 <-27.4 > 0.2 -20.0
N-9 < -27.8 <-27.7 < -27.6 <-27.3 <-27.3 > 0.2 -20.0
N-8 < -27.6 <-27.5 < -27.4 <-27.2 <-27.1 > 0.2 -20.0
N-7 < -27.4 <-27.3 < -27.2 <-27.1 <-27.0 > 0.1 -20.0
N-6 < -27.2 <-27.1 < -27.0 <-26.9 <-26.7 > 0.2 -20.0
N-5 < -27.0 <-26.9 < -26.9 <-26.9 <-26.7 > 0.1 -20.0
N-4 < -26.9 <-26.9 < -26.9 <-26.8 <-26.7 > 0.1 -20.0
N-3 < -26.9 <-26.8 < -26.8 <-26.7 <-26.6 > 0.1 -20.0
N-2 < -26.6 <-26.5 < -26.5 <-26.5 <-26.4 > 0.1 -20.0
N-1 < -20.9 <-20.8 < -20.8 <-20.7 -20.7 > 0.1 -20.0
N+1 < -20.3 <-20.2 < -20.2 <-20.2 -19.8 > 0.1 -20.0
N+2 < -26.2 <-26.0 < -26.0 <-26.0 <-26.0 > 0.1 -20.0
N+3 < -26.0 <-25.8 < -25.8 <-25.8 -25.3 > 0.2 -20.0
N+4 < -26.0 <-25.8 < -25.8 <-25.8 <-25.7 > 0.1 -20.0
N+5 < -26.2 <-26.0 < -26.0 <-26.0 <-25.9 > 0.1 -20.0
N+6 < -26.5 <-26.3 < -26.3 <-26.2 <-26.1 > 0.1 -20.0
N+7 < -26.6 <-26.4 < -25.3 -23.9 -20.0 > 2.3 -20.0
N+8 < -26.7 <-26.4 < -26.4 <-26.3 <-26.2 > 0.1 -20.0
N+9 < -26.6 <-26.3 < -26.3 <-26.2 <-26.1 > 0.2 -20.0
N+10 < -26.5 <-26.2 < -26.2 <-26.1 <-26.0 > 0.2 -20.0
N+11 < -26.3 <-26.0 < -26.0 <-25.9 <-25.8 > 0.2 -20.0
N+12 < -26.1 <-25.8 < -25.8 <-25.7 <-25.6 > 0.2 -20.0
N+13 < -25.9 <-25.6 < -25.6 <-25.4 <-25.4 > 0.2 -20.0
N+14 < -25.5 <-25.2 < -25.2 <-25.1 <-25.0 > 0.2 -20.0
N+15 < -25.1 <-24.8 < -24.8 <-24.7 <-24.6 > 0.2 -20.0
N+16 < -24.5 <-24.3 < -24.3 <-24.2 <-24.1 > 0.2  
N-5/N-10 NA NA NA NA NA NA 
N-4/N-8 NA NA NA NA NA NA 
N-3/N-6 NA NA NA NA NA NA 
N-2/N-4 NA NA NA NA NA NA 
N-1/N-2 < -20.9 <-20.8 < -20.3 -19.3 -18.1 > 1.0 
N+1/N+2 < -20.3 <-20.2 < -19.7 -20.1 -16.4 > 1.4 
N+2/N+4 NA NA NA NA NA NA 
N+3/N+6 NA NA NA NA NA NA 
N+4/N+8 NA NA NA NA NA NA 
N+5/N+10 NA NA NA NA NA NA 
 
A-7 
 
THRESHOLD U STATISTICS FOR 8 FIFTH-GENERATION DTV 
RECEIVERS 
Desired Signal = D
MIN
 + 1 dB (Extrapolated from Measurements at D
MIN
 + 3 dB) 
 
Table A-7.  Threshold U Statistics for 8 Receivers at D = D
MIN
 + 1 dB on Channel 30 
Undesired 
Channel 
Best U 
(dBm) 
Median U 
(dBm) 
Mean U 
(dBm) 
2nd 
Worst 
U 
(dBm) 
Worst 
U 
(dBm) 
Standard 
Deviation 
(dB) 
N-16      
5  
N-14      
N-13  -11.0 -18.8 -19.3  
N-12 -11.4 -20.3 -22.8  
N-11  -12.5 -22.0 -24.7  
N-10 -14.4 -23.7 -26.7  
N-9  -15.8 -26.5 -28.8  
N-8 -12.7 -18.8 -22.5 -31.2 -37.9 8.8 
N-7 -13.5 -20.6 -24.2 -33.7 -38.5 8.8 
N-6 -12.4 -29.6 -31.3 -39.2 -56.4 13.5 
N-5 -14.3 -29.4 -28.9 -35.8 -43.1 8.6 
N-4 -19.6 -37.3 -34.9 -41.3 -47.5 9.1 
N-3 -26.7 -37.1 -36.4 -42.6 -46.8 7.1 
N-2 -36.2 -47.0 -46.4 -53.7 -57.6 7.7 
N-1 -44.8 -47.0 -47.4
-49.9
-50.9 2.2 
N+1 -43.2 -46.0 -46.6
-49.1
-51.5 2.7 
N+2 -34.6 -45.9 -45.1 -52.7 -57.3 8.6 
N+3 -23.3 -32.5 -32.3 -38.8 -40.2 5.5 
N+4 -15.5 -27.7 -26.8 -32.2 -34.3 6.3 
N+5 -17.2 -21.1 -21.9 -23.5 -31.8 4.6 
N+6        
N+7 -12.4 -17.1 -18.9 -25.3 -29.2 6.4 
N+8        
N+9  
N+10        
N+11  
N+12        
N+13  
N+14    28.3 33.9 36.0   
N+15   33.3 37.1 40.8   
N+16       
 
 
A-8 
Desired Signal = D
MIN
 + 3 dB 
 
Table A-8.  Threshold U Statistics for 8 Receivers at D = D
MIN
 + 3 dB on Channel 30 
Undesired 
Channel 
Best U 
(dBm) 
Median U 
(dBm) 
Mean U 
(dBm) 
2nd 
Worst 
U 
(dBm) 
Worst 
U 
(dBm) 
Standard 
Deviation 
(dB) 
N-16 > -4.2 -8.2 > -8.6 -12.3 -13.9 > 3.5 
N-15 > -4.2 -9.7 > -8.8 -11.7 -12.8 > 3.2 
N-14 > -4.2 -8.6 > -9.1 -13.3 -14.7 > 3.7 
N-13 > -4.6 -9.1 > -9.5 -14.2 -16.4 > 4.2 
N-12 > -4.3 -9.4 > -10.2 -15.7 -18.3 > 4.8 
N-11 > -5.9 -10.5 > -12.0 -17.4 -20.1 > 5.0 
N-10 -7.1 -11.9 > -13.1 -19.2 -22.1 > 5.2 
N-9 -9.2 -13.4 > -15.6 -21.9 -24.2 > 6.0 
N-8 -10.8 -15.3 -19.1 -26.7 -33.4 7.8 
N-7 -11.6 -16.8 -20.7 -29.1 -33.9 7.8 
N-6 -10.4 -25.0 -27.4 -34.6 -51.8 12.7 
N-5 -12.3 -24.8 -25.0 -31.3 -38.5 7.8 
N-4 -16.7 -32.7 -30.9 -36.7 -43.0 8.3 
N-3 -22.1 -32.5 -31.8 -38.1 -42.2 7.1 
N-2 -31.6 -42.5 -41.8 -49.1 -53.1 7.7 
N-1 -40.2 -42.4 -42.8 -45.4 -46.3 2.2 
N+1 -38.6 -41.4 -42.1 -44.5 -46.9 2.7 
N+2 -31.6 -41.3 -40.8 -48.1 -52.7 8.4 
N+3 -20.4 -28.3 -28.3 -34.3 -35.7 5.3 
N+4 -13.5 -23.5 -23.3 -27.6 -32.3 5.9 
N+5 -15.2 -17.8 -18.9 -19.0 -29.9 4.6 
N+6 > -6.7 -8.9 -10.9 -14.4 -18.6 4.3 
N+7 -10.5 -15.2 -17.0 -23.4 -27.2 6.4 
N+8 > -3.2 >-6.4 > -6.4 -8.5 -9.4 > 2.3 
N+9 > -1.8 >-5.3 > -5.4 -8.4 -8.4 > 2.4 
N+10 > -2.9 >-6.4 > -5.9 -7.8 -8.7 > 2.1 
N+11 > -1.9 >-5.4 > -5.1 -7.8 -8.5 > 2.5 
N+12 > -1.8 >-4.3 > -4.6 -7.7 -8.6 > 2.8 
N+13 > -2.0 >-4.1 > -4.5 -7.4 -8.5 > 2.7 
N+14 > -2.3 -22.4 > -19.4 -28.1 -30.2 > 9.5 
N+15 > -2.8 -27.4 > -24.3 -31.3 -34.9 > 9.9 
N+16 > -3.3 >-4.6 > -5.1 -7.1 -8.5 > 2.0 
 
 
A-9 
Desired Signal = -68 dBm 
Table A-9.  Threshold U Statistics for 8 Receivers at D = -68 dBm on Channel 30 
Undesired 
Channel 
Best U 
(dBm) 
Median U 
(dBm) 
Mean U 
(dBm) 
2nd 
Worst 
U 
(dBm) 
Worst 
U 
(dBm) 
Standard 
Deviation 
(dB) 
N-16 > -1.1 -3.9 > -4.1 -5.9 -8.8 > 2.7 
N-15 > -1.0 -3.9 > -4.0 -6.3 -8.3 > 2.7 
N-14 > -0.8 -3.9 > -4.2 -6.4 -8.7 > 2.8 
N-13 -1.0 -3.9 -4.4 -6.7 -9.2 2.8 
N-12 -1.9 -5.3 -5.0 -6.7 -9.7 2.7 
N-11 -2.1 -6.1 -5.9 -7.2 -11.2 2.8 
N-10 -2.1 -4.5 -5.6 -8.4 -10.9 2.9 
N-9 -3.1 -7.1 -7.1 -9.9 -11.9 3.1 
N-8 -3.8 -10.0 -9.7 -13.0 -17.4 4.5 
N-7 -3.8 -9.8 -10.8 -14.5 -19.6 5.1 
N-6 -5.5 -15.1 -16.0 -19.7 -36.2 9.7 
N-5 -5.4 -12.5 -13.7 -18.3 -23.5 5.7 
N-4 -6.1 -20.6 -18.7 -22.2 -27.3 6.3 
N-3 -8.2 -18.4 -17.9 -23.7 -26.2 6.3 
N-2 -18.3 -27.2 -27.9 -36.0 -40.0 7.4 
N-1 -28.0 -28.7 -28.9 -30.0 -30.1 0.8 
N+1 -25.9 -28.3 -28.3 -29.6 -30.2 1.4 
N+2 -19.7 -25.7 -26.9 -33.7 -38.2 6.9 
N+3 -11.1 -13.3 -16.1 -19.7 -28.2 5.8 
N+4 -8.1 -11.4 -13.1 -13.6 -27.0 5.9 
N+5 -2.6 -9.6 -11.1 -14.5 -25.5 7.1 
N+6 > -1.7 -4.9 > -6.4 -11.8 -15.2 > 4.6 
N+7 -7.7 -14.8 -16.0 -23.0 -26.4 6.8 
N+8 > -1.4 -3.0 > -3.4 -4.8 -7.1 > 2.0 
N+9 > -1.4 -2.2 > -2.8 -4.5 -5.9 > 1.6 
N+10 > -1.5 -3.0 > -3.0 -4.9 -5.2 > 1.4 
N+11 > -1.8 >-2.2 > -2.8 -4.8 -4.9 > 1.3 
N+12 > -2.0 >-2.3 > -3.0 -5.0 -5.1 > 1.3 
N+13 > -2.1 >-2.5 > -3.0 -4.6 -4.8 > 1.1 
N+14 > -2.7 -7.7 > -8.8 -15.1 -16.8 > 5.2 
N+15 > -3.1 -12.9 > -12.0 -18.0 -19.3 > 5.7 
N+16 > -3.5 >-3.8 > -3.9 -4.5 -4.6 > 0.4 
N-5/N-10 -18.2 -25.7 -24.4 -26.3 -30.8 4.2 
N-4/N-8 -21.7 -26.1 -25.6 -28.9 -29.8 3.1 
N-3/N-6 -22.8 -26.3 -27.4 -32.0 -35.0 4.1 
N-2/N-4 -16.7 -27.3 -28.7 -37.3 -38.5 6.9 
N-1/N-2 -28.8 -32.3 -33.2 -38.1 -40.7 4.1 
N+1/N+2 -28.3 -31.5 -31.9 -33.1 -39.0 3.4 
N+2/N+4 -25.7 -30.0 -30.8 -37.1 -39.1 5.1 
N+3/N+6 -17.0 -24.5 -24.0 -27.4 -34.1 5.8 
N+4/N+8 -17.2 -21.2 -22.0 -25.8 -28.4 4.2 
N+5/N+10 -15.6 -19.7 -20.1 -23.6 -25.5 3.2 
 
A-10 
Desired Signal = -53 dBm 
Table A-10.  Threshold U Statistics for 8 Receivers at D = -53 dBm on Channel 30 
Undesired 
Channel 
Best U 
(dBm) 
Median U 
(dBm) 
Mean U 
(dBm) 
2nd 
Worst 
U 
(dBm) 
Worst 
U 
(dBm) 
Standard 
Deviation 
(dB) 
N-16 > -0.8 -1.6 > -2.5 -3.7 -7.3 > 2.2 
N-15 > -0.6 >-1.1 > -2.4 -4.0 -7.9 > 2.5 
N-14 > -0.3 >-0.9 > -2.4 -4.2 -8.5 > 2.8 
N-13 > -0.1 -1.1 > -2.5 -4.1 -8.8 > 2.9 
N-12 > -0.1 -2.2 > -2.9 -4.4 -9.4 > 3.1 
N-11 > -0.1 -2.0 > -3.2 -6.3 -10.1 > 3.5 
N-10 > -0.1 -0.7 > -2.5 -4.8 -10.7 > 3.7 
N-9 > -0.2 -0.8 > -2.8 -4.9 -11.6 > 3.9 
N-8 > -0.4 -1.6 > -3.5 -5.8 -12.8 > 4.3 
N-7 > -0.6 -2.5 > -4.1 -7.3 -14.2 > 4.7 
N-6 > -0.9 -5.3 > -7.3 -16.2 -20.5 > 7.3 
N-5 > -1.0 -1.4 > -4.6 -9.2 -18.3 > 6.2 
N-4 > -1.1 -6.0 > -7.3 -13.0 -20.8 > 6.6 
N-3 > -1.2 -3.8 > -4.9 -7.8 -11.1 > 3.3 
N-2 -4.0 -11.5 -13.2 -21.2 -25.7 7.1 
N-1 -13.1 -14.0 -14.6 -15.4 -18.6 1.8 
N+1 -11.0 -13.6 -13.6 -14.7 -17.1 1.9 
N+2 -6.2 -10.9 -12.4 -18.5 -22.7 6.3 
N+3 > -2.1 -3.2 > -6.2 -4.9 -27.9 > 8.8 
N+4 > -2.1 -2.9 > -6.4 -7.9 -23.9 > 7.4 
N+5 > -1.8 -4.0 > -6.3 -8.5 -21.1 > 6.5 
N+6 > -1.6 -2.2 > -4.1 -6.7 -13.2 > 4.1 
N+7 -1.8 -14.3 -14.2 -20.4 -26.1 8.0 
N+8 > -1.4 >-1.6 > -2.1 -3.4 -4.3 > 1.1 
N+9 > -1.5 >-1.7 > -1.9 >-1.8 -3.8 > 0.8 
N+10 > -1.6 >-1.8 > -1.9 >-1.9 -3.2 > 0.5 
N+11 > -1.8 >-2.0 > -2.1 >-2.1 -2.7 > 0.3 
N+12 > -1.8 >-2.2 > -2.2 >-2.3 -2.6 > 0.3 
N+13 > -2.2 >-2.5 > -2.4 >-2.6 >-2.6 > 0.2 
N+14 > -2.6 >-2.9 > -2.8 >-3.0 -3.1 > 0.2 
N+15 > -3.0 >-3.3 > -3.5 >-3.4 -5.4 > 0.8 
N+16 > -3.5 >-3.7 > -3.7 >-3.8 >-3.8 > 0.1 
N-5/N-10 -14.1 -19.4 -18.7 -20.8 -22.4 2.8 
N-4/N-8 -14.5 -18.5 -18.8 -20.5 -24.8 3.1 
N-3/N-6 -9.4 -15.6 -16.6 -20.5 -27.6 5.8 
N-2/N-4 -11.1 -14.8 -16.6 -24.4 -24.4 5.1 
N-1/N-2 -14.0 -17.1 -18.6 -25.6 -26.7 4.9 
N+1/N+2 -14.3 -17.6 -17.9 -19.7 -23.0 2.6 
N+2/N+4 -14.1 -17.8 -19.6 -21.6 -32.2 5.7 
N+3/N+6 -11.5 -18.8 -18.6 -23.1 -30.0 6.1 
N+4/N+8 -11.8 -17.2 -17.0 -19.4 -24.5 4.1 
N+5/N+10 -11.4 -14.8 -15.3 -18.7 -20.8 3.1 
 
 
A-11 
Desired Signal = -28 dBm 
Table A-11.  Threshold U Statistics for 8 Receivers at D = -28 dBm on Channel 30 
Undesired 
Channel 
Best U 
(dBm) 
Median U 
(dBm) 
Mean U 
(dBm) 
2nd 
Worst 
U 
(dBm) 
Worst 
U 
(dBm) 
Standard 
Deviation 
(dB) 
N-16 > -0.7 >-0.8 > -0.9 >-1.1 >-1.2 > 0.2 
N-15 > -0.6 >-0.7 > -0.8 >-1.0 >-1.1 > 0.2 
N-14 > -0.4 >-0.5 > -0.6 >-0.8 >-0.9 > 0.2 
N-13 > -0.2 >-0.2 > -0.4 >-0.6 >-0.6 > 0.2 
N-12 > -0.1 >-0.2 > -0.3 >-0.6 >-0.6 > 0.2 
N-11 > -0.1 >-0.2 > -0.3 >-0.5 >-0.6 > 0.2 
N-10 > -0.1 >-0.2 > -0.3 >-0.6 >-0.6 > 0.2 
N-9 > -0.3 >-0.3 > -0.4 >-0.7 >-0.7 > 0.2 
N-8 > -0.4 >-0.5 > -0.6 >-0.8 >-0.8 > 0.2 
N-7 > -0.6 >-0.7 > -0.8 >-0.9 >-1.0 > 0.1 
N-6 > -0.8 >-0.9 > -1.0 >-1.1 >-1.2 > 0.1 
N-5 > -0.9 >-1.1 > -1.1 >-1.2 >-1.3 > 0.1 
N-4 > -1.1 >-1.1 > -1.1 >-1.2 >-1.3 > 0.1 
N-3 > -1.2 >-1.2 > -1.2 >-1.3 >-1.4 > 0.1 
N-2 > -1.4 >-1.5 > -1.5 >-1.5 >-1.6 > 0.1 
N-1 > -7.2 >-7.2 > -7.2 >-7.3 -7.3 > 0.0 
N+1 > -7.7 >-7.8 > -7.8 >-7.8 -8.2 > 0.1 
N+2 > -1.9 >-2.0 > -2.0 >-2.0 >-2.0 > 0.1 
N+3 > -2.0 >-2.1 > -2.2 >-2.2 -2.7 > 0.2 
N+4 > -2.0 >-2.2 > -2.2 >-2.2 >-2.2 > 0.1 
N+5 > -1.8 >-2.0 > -2.0 >-2.0 >-2.1 > 0.1 
N+6 > -1.5 >-1.7 > -1.7 >-1.8 >-1.8 > 0.1 
N+7 > -1.5 >-1.6 > -2.7 -4.1 -8.0 > 2.3 
N+8 > -1.4 >-1.6 > -1.6 >-1.7 >-1.7 > 0.1 
N+9 > -1.4 >-1.7 > -1.7 >-1.8 >-1.9 > 0.2 
N+10 > -1.5 >-1.8 > -1.8 >-1.9 >-2.0 > 0.2 
N+11 > -1.7 >-2.1 > -2.0 >-2.1 >-2.2 > 0.2 
N+12 > -1.9 >-2.2 > -2.2 >-2.3 >-2.4 > 0.2 
N+13 > -2.1 >-2.5 > -2.4 >-2.6 >-2.6 > 0.2 
N+14 > -2.5 >-2.8 > -2.8 >-2.9 >-3.0 > 0.2 
N+15 > -2.9 >-3.3 > -3.2 >-3.4 >-3.4 > 0.2 
N+16 > -3.5 >-3.7 > -3.7 >-3.9 >-3.9 > 0.1 
N-5/N-10 NA NA NA NA NA NA 
N-4/N-8 NA NA NA NA NA NA 
N-3/N-6 NA NA NA NA NA NA 
N-2/N-4 NA NA NA NA NA NA 
N-1/N-2 > -7.1 >-7.2 > -7.7 -8.7 -9.9 > 1.0 
N+1/N+2 > -7.7 >-7.8 > -8.3 -7.9 -11.7 > 1.4 
N+2/N+4 NA NA NA NA NA NA 
N+3/N+6 NA NA NA NA NA NA 
N+4/N+8 NA NA NA NA NA NA 
N+5/N+10 NA NA NA NA NA NA 
 
 
B-1 
APPENDIX B 
THEORETICAL BASIS FOR OUT-OF-CHANNEL 
INTERFERENCE 
When a DTV receiver operates in the presence of white Gaussian co-channel interference, the threshold 
of visibility (TOV) of picture degradation occurs when the desired signal power D exceeds the co-channel 
interference by about 15 dB.
*
  This number may vary somewhat for noise having other statistical 
properties, and may be much lower if the noise is heavily concentrated at a band edge where filtering in 
the DTV provides additional rejection; nonetheless, one expects that, as signal power D varies, the 
undesired signal power at threshold will vary linearly with it—resulting a constant D/U ratio as D or U 
are varied.  This relationship holds whenever the co-channel interference is high enough that the effect of 
internal noise in the receiver becomes insignificant. 
 
For most out-of-channel interference mechanisms, the DTV receiver unintentionally converts a small 
portion of the out-of-channel power into co-channel power.  If one knows the amount of conversion into 
co-channel interference, one can treat the problem as a co-channel interference problem, which is 
relatively well understood, as described above.  In this formulation of the problem, measuring the desired 
signal power D at the TOV provides an indirect method of measuring the co-channel power created 
internal to the receiver, since we know that the co-channel power will be about 15 dB below the measured 
value of D. 
 
The conversion process by the DTV from out-of-channel interference to co-channel interference may be 
linear or nonlinear.  If it is linear, then the internally-created co-channel interference will vary linearly 
with the out-of-channel interference power U causing the value of the desired signal power D at threshold 
to vary linearly with U.  The result will be that threshold D/U ratio will be constant as D or U is varied.  If 
the conversion process is nonlinear, then the relationship between D and U will be nonlinear and the D/U 
ratio will vary with D and U. 
 
We will assume that the co-channel interference power created by the DTV receiver in response to an 
out-of-channel undesired signal power U will be proportional to D
L
U
M
, where L and M are integer 
constants that define the order of the interference mechanism.  For most interference mechanisms, L will 
be zero, so only the U
M
 term exists.  The following are among the interference mechanisms that can be 
modeled by this formulation. 
• Linear interference:  L=0, M=1.  Creates co-channel interference proportional to U.   
? Example:  mixer image.  The mixer in a TV receiver converts the spectrum of the intended 
channel of the received signal to an intermediate frequency (IF) where it can be filtered more 
precisely to pass the desired channel while rejecting the undesired frequencies.  Unfortunately, in 
single-conversion tuners a second a 6-MHz wide portion of the input spectrum centered 88 MHz 
above the desired channel is also converted to that same IF.  Filtering prior to the mixer strongly 
diminishes—but doesn’t fully extinguish—this unintended signal. 
? Example:  leakage of the adjacent channel signal through the channel selection filter of the DTV 
would also constitute a linear interference mechanism. 
• Second-order interference:  L=0, M=2.  Creates co-channel interference proportional to U
2
. 
? Example:  “half-IF” taboo.  The second harmonic of an undesired signal 22 MHz above the 
desired signal beats with the second harmonic of the receiver’s local oscillator, creating a 
difference frequency that falls within the IF band of the receiver. 
• Third-order interference:  L=0, M=3.  Creates co-channel interference proportional to U
3
. 
                                                      
*
 SHVERA Study results on 28 receivers showed that D must exceed U by amounts ranging from 14.9 to 15.8 dB, 
with a median value of 15.3 dB. 
 
B-2 
? Example:  third-order intermodulation (IM3) of a single, adjacent-channel undesired signal.  IM3 
creates spectral components that spill into each adjacent channel.  
? Example:  third-order intermodulation (IM3) of a pair of undesired channels placed at channels 
N+K and N+2K where N is the desired channel.  In this case, the interference power created in 
channel N is proportional to U
N+K
2
U
N+2K
.  The result is a process that is second-order in terms of 
U
N+K
 and linear in terms of U
N+2K
; however, if the two undesired signals are set to equal powers 
and varied in amplitude together, the resulting interference is third order. 
• Cross-modulation:  L=1, M=2.  Creates co-channel interference proportional to DU
2
. 
? Cross-modulation is essentially a third-order effect, but the co-channel interference created is 
proportional to D and to U
2
.  As a result, increasing the desired signal power does not improve the 
signal-to-interference ratio. 
 
We define the following: 
D = Power of desired signal on channel N at input to TV 
U
N+K
 = Power of interferer on channel N+K at input to TV 
U
N+2K
 = Power of interferer on channel N+2K at input to TV 
where D, U
N+K
, and U
N+2K
 refer to signal level combinations that place the TV at TOV 
 
R = Required SNR of the TV receiver at TOV 
D
MIN
 = D at TOV in absence of interference or external noise 
N
R
 = Receiver noise referred to the input of the TV 
 
Thus, 
R = D
MIN
/N
R
 
N
R
 = D
MIN
/R 
 
Consequently, N
R
 can be inferred from measurements of D
MIN
 and R 
 
Let  
 
P
CC
 = Total power of co-channel noise and interference affecting the demodulation of the DTV signal by 
the TV, referenced to the input.  P
CC
 includes co-channel interference created by non-linear effects in the 
TV. 
 
We will consider two cases.  That of a single interferer with power U, where  
 
P
CC
 =  N
R
 + c U
M
 D
L
   
 
And that of third-order intermodulation (IM) between a pair of signals U
N+K
, and U
N+2K
  
 
P
CC
 =  N
R
 + c
IM3
 U
K
2
 U
2K
  
 
The “c” terms are constants related to the nonlinear process in the receiver.  The 1st term in each equation 
is receiver noise.  The second is the interference term created by distortion in the TV tuner.  The terms M 
and L define the order of the nonlinear interference process with respect to the undesired and desired 
signals, respectively. 
 
We start with the case of a single interferer. 
SINGLE INTERFERER 
We have  
 
P
CC
 =  N
R
 + c U
M
 D
L
   
 
B-3 
 
We will generally be interested in three cases: 
• Linear interference mechanisms:  M = 1; L = 0 
• 2
nd
-order interference mechanisms: M = 2; L = 0 
• 3
rd
-order interference mechanisms:  M = 3; L = 0 
• Cross-modulation:  M = 2; L = 1 
 
At TOV the desired signal must exceed P
CC
 by a factor equal to the required signal-to-noise ratio R.  (We 
assume that the same value of R applies for both receiver noise and noise created by an undesired signal.)  
Thus 
 
D/P
CC
 = R, or, equivalently, 
 
D = R P
CC
  
 
Substituting, we have 
 
D = R (N
R
 + c U
M
 D
L
) 
 
Substituting N
R
 = D
MIN
/R, we have 
 
D = R c D
L
U
M
 + D
MIN
 
 
U
M
 = (D - D
MIN
) / (R c D
L
) 
 
And, finally, 
 
U = [(D - D
MIN
) / (R c D
L
)]
1/M
 
 
We will also find it useful to write this as 
 
U = D
(1-L)/M
 [(1 - D
MIN
/D) / (R c)]
1/M
 
 
High Signal Levels (D >> DMIN) 
When D >> D
MIN
, the equation simplifies to  
 
U ? [D
1-L
 / (R c)]
1/M
 
 
U ? D
(1-L) / M
 / (R c)
1/M
 
 
Similarly, D/U at threshold is given by 
 
D/U ? D/[ D
(1-L) / M
 / (R c)
1/M
] 
 
D/U ?  (R c)
1/M
 D
(M – 1 + L)/M
 
 
Now we wish to view U and D in log-based units, such as decibels. 
 
log(U) ? log[D
(1-L) / M
 / (R c)
1/M
] =  [(1 – L)/M] log(D) – (1/M) log(R c) 
 
log(D/U) ? log[(R c)
1/M
 D
(M – 1 + L)/M
] = [(M – 1 + L)/M] log(D) + (1/M) log(R c) 
 
 
B-4 
Thus a log-log plot of U versus D will be a straight line, with slope (1 – L) / M.  Similarly, the slope of 
log-D versus log-U will be M / (1 – L), and the slope of log(D/U) versus log-D is given by   
[(M – 1 + L)/M]. 
 
Table A-1 summarizes this slope information for the interference mechanisms of interest. 
 
 
Table B-1.  Slopes of Log-Log Plots of D, U, and D/U for Various Interference Mechanisms 
Interference Mechanism 
Slope of 
Log (D) 
Versus 
Log (U) 
in dB/dB 
Slope of 
Log (U) 
Versus 
Log (D) 
in dB/dB 
Slope of 
Log (D/U) 
Versus 
Log (D)  
in dB/dB 
Linear (M = 1, L = 0) 1 1 0 
Second order (M = 2, L = 0) 2 0.5 0.5 
Third order (including third-order 
intermodulation of a pair of equal-
power interferers) (M = 3, L = 0) 
3 0.333 0.667 
Cross modulation (M = 2, L = 1) Infinite 0 1 
 
 
Low Signal Levels 
The interference mechanisms described above are expected to result in linear relationships between log-U 
and log-D at threshold when the desired signal level is high enough that receiver noise is insignificant.  
Now we consider the case of smaller signal levels.  Recall that 
 
U = [(D - D
MIN
) / (R c D
L
)]
1/M
 
 
Note that the presence of receiver noise (causing D
MIN
 to be non-zero and the log-U versus log-D 
relationship to deviate from a straight line) results in U changing by a factor of 
 
[(D - D
MIN
) / D]
1/M
 
 
Consider the case where D is X dB above D
MIN
.  Then U is Y dB above the value it would have had based 
on a straight-line log-log projection from the results at a high desired signal level. 
 
X = 10 log(D / D
MIN
) 
 
Y = 10 log{(D - D
MIN
) / D]
1/M
 }  
 
Y = (1/M) 10 log[1 – 10
-X/10
] dB 
 
Table B-2 summarizes these results for three values of X. 
 
 
 
B-5 
Table B-2.  Deviation in Threshold U from Straight-Line Projection as D Approaches D
MIN
 
 
Deviation in Threshold U 
from Straight-Line Projection (dB) 
Interference Mechanism 
D/D
MIN
1
  
= 16 dBm 
D/D
MIN
 
= 3 dB 
D/D
MIN
 
= 1 dB 
D/D
MIN
 
= 0 dB 
Linear (M = 1) -0.11 -3.02 -6.87 Infinite 
Second order (M = 2) -0.06 -1.51 -3.43 Infinite 
Third order (including third-order 
intermodulation of a pair of equal-
power interferers) (M = 3) 
-0.04 -1.01 -2.29 Infinite 
Cross modulation (M = 2, L = 1) -0.06 -1.51 -3.43 Infinite 
Note: 
1
 For the nominal D
MIN
 value of -84 dBm, D/D
MIN
 = 16 dB when D = -68 dBm 
 
 
Effect of AGC 
The above relationships are expected to hold when automatic-gain-control (AGC) does not cause changes 
in gain between the TV antenna port and the point in the TV tuner at which the relevant nonlinearity 
occurs (i.e., the location of the nonlinearity that causes the observed interference, assuming it is caused by 
a nonlinearity). 
 
When either the desired signal power (D) or an undesired signal (U) rises sufficiently that AGC causes 
gain reductions prior to the point of a relevant nonlinearity, the relationships change.   
 
We define the following terms: 
 
G  = power gain from the antenna input terminal of the TV to the point of a relevant 
nonlinearity; 
G
MAX
  = the value of G when both D and U are low enough that AGC does not reduce the gain of 
any tuner stages prior to the nonlinearity; 
 
In a television, AGC operation may be invoked based on increasing levels of either the desired signal D 
on channel N or of some filtered combination of desired and undesired signals.
*
  In modeling AGC, we 
will assume that, if the AGC reduces gain of a tuner stage prior to the point of the relevant nonlinearity, it 
will do so in such a way as to achieve a constant power level with changes input signal level.  
Specifically, the power which is maintained constant by AGC action will be either the desired signal or 
the total power of some filtered combination of desired and undesired signals (as in the case of 
“broadband AGC”).  We will consider two bounding cases: 
• AGC driven by D.  AGC adjusts gain in such a way that the level of the desired signal at the point of 
the nonlinearity remains constant; 
• AGC driven by U
N+K
.  AGC is controlled by a filtered combination of desired and undesired signals, 
but with undesired signal at the AGC sensing point being much larger than the desired signal, so that 
the AGC adjusts gain to, in effect, maintain a constant undesired signal power at the point of the 
relevant nonlinearity. 
 
For each of these two cases, we define a signal level threshold above which the AGC reduces the gain of 
tuner stages prior to the point of the relevant nonlinearity: 
D
AGCthresh
   = the desired signal power at the TV input, above which AGC begins reducing G; 
U
N-K,AGCthresh
 = the undesired signal power on channel N+K, above which AGC begins reducing G. 
 
                                                      
*
 Bendov and Patel, 2005, p.38-39. 
 
B-6 
Thus, in the first case, we will assume that the AGC reduces gain by 1 dB for each 1-dB increase in D 
above D
AGCthresh
.  In the second case we assume that AGC reduces gain by 1 dB for each 1-dB increase in 
U beyond U
N-K,AGCthresh
. 
 
In the single-interferer case, recall that the total co-channel noise plus interference that seen by the 
receiver, referenced to input levels, is 
 
 
P
CC
 =  N
R
 + c U
N+K
M
 D
L
 
 
We note that this formula applies when gain prior to the relevant point of nonlinearity is at its maximum 
(G = G
MAX
)—i.e., the AGC hasn’t caused any gain reductions.  Thus, the formula is valid only when D < 
D
AGCthresh
 and U
K
 < U
K-AGCthresh
.  If either AGC threshold is exceed, G is reduced and the formula is no 
longer valid.  We include the N+K subscript on U to emphasize that the AGC threshold will be different 
for different channel offsets because of filtering in the receiver. 
 
It should be recognized that the terms U
N+K
M
 and D
L
 describe nonlinear behavior at some point in the TV 
tuner—perhaps at the mixer, or at the output of the IF amplifier.  Thus, we could more correctly describe 
the nonlinearity in terms of signal levels at this point in the TV tuner.  If we use bold italicized terms to 
represent desired signal power, undesired signal power, and receiver noise referred to the point of 
nonlinearity in the tuner, we can rewrite the equation as follows: 
 
P
CC
 =  N
R
 + c
1
 U
N+K
M
 D
L
  when G = G
MAX
 
 
where 
P
CC
 = G P
CC
 
N
R
 = G N
R
 
c
1
 = a new constant describing the nonlinearity in terms of levels at the point of nonlinearity, instead of at 
referenced to the input 
U
N+K
 = G U
N+K
 
D = G D 
 
Performing substitutions, we have 
 
G P
CC
 = G N
R
 + c
1
 (G U
N+K
)
M
 (G D)
L
   
 
G P
CC
 = G N
R
 + c
1
 G
M+L
 U
N+K
M
 D
L
   
 
P
CC
 = N
R
 + c
1
 G
M+L-1
 U
N+K
M
 D
L
   
 
P
CC
 = N
R
 + c
1
 G
MAX
M+L-1
 (G/G
MAX
)
M+L-1
 U
N+K
M
 D
L
   
 
When G = G
MAX
, this formula must be equivalent to the previous version: 
 
P
CC
 =  N
R
 + c U
N+K
M
 D
L
  when G = G
MAX
 
 
Thus, it is clear that the relationship between the nonlinearity constant defined referenced to the input and 
that defined referenced to the point of the nonlinearity is 
 
c = c
1
 G
MAX
M+L-1
  
 
Consequently, we will rewrite the new formula as follows: 
 
 
B-7 
P
CC
 = N
R
 + c (G/G
MAX
)
M+L-1
 U
N+K
M
 D
L
   
 
At threshold, P
CC
 = D/R; also R N
R
 = D
MIN
.  So we have 
 
D = R [N
R
 + c (G/G
MAX
)
M+L-1
 U
N+K
M
 D
L
 
 
 
D = D
MIN
 + R c (G/G
MAX
)
M+L-1
 U
N+K
M
 D
L
 
 
 
AGC Driven By D 
We first consider the case in which the desired signal reaches a sufficient level to cause AGC gain 
reductions before the point of the relevant nonlinearity.  We assume that 
 
G/G
MAX
  = 1,  when D ? D
AGCthresh
, and 
  = D
AGCthresh
/D,  when D > D
AGCthresh
  
 
Thus, for the case of D > D
AGCthresh
, we have 
 
D = D
MIN
 + R c (D
AGCthresh
/D)
M+L-1
 U
M
 D
L
] 
 
D - D
MIN
 = R c D
AGCthresh
M+L-1
 U
M
 D
-M+1
] 
 
U
M
 = (D - D
MIN
) D
M-1
 /(R c D
AGCthresh
M+L-1
) 
 
U = [(D - D
MIN
) D
M-1
 /(R c D
AGCthresh
M+L-1
)]
1/M
 
 
U = [(D - D
MIN
) D
M
 /(R c D D
AGCthresh
M+L-1
)]
1/M
 
 
U = D [(D - D
MIN
) /(R c D D
AGCthresh
M+L-1
)]
1/M
 
 
U = D [(1 - D
MIN
/D) /(R c D
AGCthresh
M+L-1
)]
1/M
 
 
We recall and rearrange the original formula that applies when there are no AGC gain reductions, 
 
U = [(D - D
MIN
) / (R c D
L
)]
1/M
 
 
U = D
(1-L)/M
 [(1 - D
MIN
/D) / (R c)]
1/M
 
 
We now combine this with the case of no AGC gain changes. 
 
 
With AGC operation driven by desired signal level, 
 
U =  D [(1 - D
MIN
/D) /(R c D
AGCthresh
M+L-1
)]
1/M
,  when D > D
AGCthresh
 (i.e., AGC operating) 
  
 D
(1-L)/M
 [(1 - D
MIN
/D] / (R c)]
1/M
,   when D ? D
AGCthresh
 (i.e., no AGC operation) 
 
 
Consider the case when D >> D
MIN
.  The formula becomes, 
 
U ?  D [1 / (R c D
AGCthresh
M+L-1
)]
1/M
, when D > D
AGCthresh
 (i.e., AGC operating) 
 
B-8 
  
 D
(1-L)/M
 / (R c)
1/M
,  when D ? D
AGCthresh
 (i.e., no AGC operation) 
 
 
Notice that the first formula (for use when the AGC is operating) is linear in D.  I.e., the threshold 
undesired signal is directly proportional to the desired signal level.  This means that, at signal levels 
D >> D
MIN
, the interference behaves as if it is linear, even though the underlying mechanism is nonlinear.  
For D >> D
MIN
, once D exceeds the AGC threshold for gain adjustments prior to the point of the 
relevant nonlinearity, D/U ratio remains constant with further increases in D.  Beyond this AGC 
threshold, the interference behaves as if it derives from a linear mechanism, even though the actual 
interference mechanism may be nonlinear. 
 
The above statement applies to D >> D
MIN
.  We now examine further the case of small signal levels. 
 
U = D [(1 - D
MIN
/D) /(R c D
AGCthresh
M+L-1
)]
1/M
, when D > D
AGCthresh
 (i.e., AGC operating) 
 
Or 
 
U = D (1 - D
MIN
/D)
1/M
 / (R c D
AGCthresh
M+L-1
)
1/M
, when D > D
AGCthresh
 (i.e., AGC operating) 
 
 
We can see that the presence of receiver noise (causing non-zero D
MIN
) causes the undesired signal 
threshold U to change by a factor of 
 
(1 - D
MIN
/D)
1/M
, when D > D
AGCthresh
 (i.e., AGC operating) 
 
The effect of receiver noise on threshold U is the same result that was obtained when D was below the 
AGC threshold. 
 
AGC Driven by U 
 
Now consider the case in which the undesired signal reaches a sufficient level to cause AGC gain 
reductions before the point of the relevant nonlinearity.  We assume that 
 
G/G
MAX
 = 1,   when U
N+K
 ? U
N+K-AGCthresh
, and 
  = U
N+K,AGCthresh
/U
N+K
,  when U
N+K
 > U
N+K-AGCthresh
  
 
Thus, for the case U
N+K
 > U
N+K,AGCthresh
, we have 
 
D = D
MIN
 + R c (G/G
MAX
)
M+L-1
 U
N+K
M
 D
L
 
 
Substituting for G/G
MAX
,  
 
D = D
MIN
 + R c (U
N+K,AGCthresh
/U
N+K
)
M+L-1
 U
N+K
M
 D
L
 
 
Rearranging, 
 
D - D
MIN
 = R c U
N+K,AGCthresh
M+L-1
 U
N+K
1-L
 D
L
 
 
U
N+K
1-L
 = (D - D
MIN
) / [R c U
N+K,AGCthresh
M+L-1
 D
L
] 
 
There is no solution when L = 1, the cross-modulation case.  For other cases, where L = 0, 
 
 
B-9 
U
N+K
 = (D - D
MIN
) / [R c U
N+K,AGCthresh
M-1
] 
 
We now combine this with the case of no AGC gain changes. 
 
 
With AGC operation driven by the undesired signal level, 
 
U
N+K
 =  (D - D
MIN
) / (R c U
N+K,AGCthresh
M-1
),  when U
N+K
 > U
N+K,AGCthresh
, (i.e., AGC 
operating), L = 0; 
 
 D
(1-L)/M
 [(1 - D
MIN
/D] / (R c)]
1/M
,  when U
N+K
 ? U
N+K,AGCthresh
 (i.e., no AGC 
operation) 
 
 
By requiring that L = 0, we are excluding the case of cross-modulation from the solution when the AGC 
is operating on undesired signal level.  Recall that, for large desired signal levels well above D
MIN
, the 
solution to the no-AGC cross-modulation case is a fixed value of U, independent of D, because the co-
channel interference power created by the TV tuner is directly proportional to D; a 1-dB increase in D 
causes a 1-dB increase in co-channel interference power, so changing D doesn’t get you closer to, or take 
you further from, the TOV.  If we consider the case with AGC driven by undesired signal, our hypothesis 
is that the AGC acts to keep the power of the undesired signal at a fixed level at the point of the 
nonlinearity by driving down the gain as the undesired signal at the input increases.  The net effect, then, 
of a 1-dB increase in undesired signal at the input will be that the undesired signal power at the point of 
nonlinearity remains constant, but the power of the desired signal at that point decreases.  As we 
described above, such a change does not move the operating point either closer to, or further from, the 
TOV.  Rather, whether the TV operates error free will depend only on whether the AGC threshold is 
above or below the TOV threshold for U that results from the cross-modulation process. 
 
Note that, in the formula that applies when AGC is operating, we find that U is directly proportional to D 
if D >> D
MIN
.  Thus, the AGC operation causes the interference to act as if it were linear, even if the 
underlying mechanism is nonlinear. 
 
At low signal levels, the effect of receiver noise is identical to that for a linear process. 
 
Except in the case of cross-modulation, AGC operation that is driven by undesired signal level causes 
the interference to behave as if it were created by a linear process.  This conclusion applies both to the 
slope of log-U versus log-D at high signal levels, and to the deviation from that straight-line log-log 
curve at low signal levels. 
 
IM3 WITH PAIRED SIGNALS 
When a pair of undesired signals placed on channels N+K and N+2K, nonlinearities in the receiver can 
create third-order intermodulation products in the desired channel N. 
 
If the undesired signals are set to equal amplitudes (U = U
N+K
 = U
N+2K
), then the results are identical to the 
third-order interference case described above.  More generally, we substitute U
N+K
2
 U
N+2K
 for U
3
 in those 
formulas. 
 
D = R (N
R
 + c U
N+K
2
 U
N+2K
) 
 
D = D
MIN
 + R c U
N+K
2
 U
N+2K
 
 
 
B-10 
If the two undesired signals have equal power (U = U
N+K
 = U
N+2K
) and D >> D
MIN
 so that receiver noise is 
insignificant, the equation simplifies to 
 
D = R c U
3
 
 
Or, 
 
D/R = c U
3
 = IM3 power referred to the input of the TV receiver. 
 
Rhodes and Sgrignoli point out that IM3 is often computed in terms of the third-order intercept power 
(IP3) for an amplifier or receiver.
*
  In decibel units, this is written as 
 
IM3|
dB
 = 3 U|
dB
 – 2 IP3|
dB
 
 
In linear power units, the equation can be rewritten as 
 
IM3 = U
3
 / IP3
2
 
 
Using this in our equation,
†
  
 
D/R = c U
3
 = IM3 = U
3
 / IP3
2
 
 
In this form we see that our constant c is equal to 1/ IP3
2
 and our original equation (when receiver noise is 
insignificant) becomes 
 
D = (R / IP3
2
) U
N+K
2
 U
N+2K
 
 
Given measurements at threshold for D and U, along with knowledge of the required SNR of the DTV 
receiver (R), we could compute IP3 as follows (when the two undesired signals are equal): 
 
IP3 = (R U
3
 / D)
1/2
 
 
Rather than do this, we will group the IP3 and R terms. 
 
IP3 / R
1/2
 = (U
3
 / D)
1/2
 
 
 
IP3 / R
1/2
 = (U
3
 / D)
1/2
 
 
or, in decibel units,  
 
(IP3 / R
1/2
)|
dB
 = 1.5 U|
dB
 - 0.5 D|
dB
  
 
 
Once we know IP3 / R
1/2
, we can use it in our original, more general equation. 
 
D = D
MIN
 + R c U
N+K
2
 U
N+2K
 
                                                      
*
 For example, see Rhodes and Sgrignoli, 2005, p. 464. 
†
 We note that IP3 is typically defined in this way for narrowband signals.  Here we use a definition that, while 
similar to the narrowband case, is not the same because:  (1) we are using to model IM3 with broadband signals 
rather than sinusoids, and (2) we are interested only in the IM3 power than falls in TV channel N although the IM3 
signal also extends into channels N-1 and N+1. 
 
B-11 
 
D = D
MIN
 + U
N+K
2
 U
N+2K
 / (IP3 / R
1/2
)
2
 
 
If we know one of the two undesired signals, we can determine the threshold value of the other from: 
 
U
N+K
2
 U
N+2K
 = (IP3 / R
1/2
)
2
 (D - D
MIN
) 
 
 
U
N+K
 = (IP3 / R
1/2
) (D - D
MIN
)
1/2
 / U
N+2K
1/2
 
 
U
N+2K
 = (IP3 / R
1/2
)
2
 (D - D
MIN
) / U
N+K
2
 
 
 
If D >> D
MIN
, then the equations can be converted to dB as follows: 
 
 
U
N+K
|
dB
 = (IP3 / R
1/2
)|
dB
 + (D|
dB
 - U
N+2K
|
dB
) / 2 
 
U
N+2K
|
dB
 = 2 (IP3 / R
1/2
)|
dB
 
 + D|
dB
 - 2 U
N+K
|
dB
 
 
 
AGC With Paired-Signal IM3 
We consider the case of AGC operating in such a way as to maintain one of the two undesired signals at a 
constant power level at the point of the nonlinearity that causes the observed IP3. 
 
We begin with AGC operation based on the power of the first of the two undesired signals.  We assume 
that 
 
G/G
MAX
 = 1,   when U
N+K
 ? U
N+K-AGCthresh
, and 
  = U
N+K,AGCthresh
/U
N+K
,  when U
N+K
 > U
N+K-AGCthresh
  
 
Thus, for the case U
N+K
 > U
N+K,AGCthresh
, we have 
 
D = D
MIN
 + (G/G
MAX
)
2
 U
N+K
2
 U
N+2K
 / (IP3 / R
1/2
)
2
 
 
Substituting for G/G
MAX
,  
 
D = D
MIN
 + (U
N+K,AGCthresh
/U
N+K
)
2
 U
N+K
2
 U
N+2K
 / (IP3 / R
1/2
)
2
 
 
Rearranging, 
 
D - D
MIN
 = (U
N+K,AGCthresh
)
2
 U
N+2K
 / (IP3 / R
1/2
)
2
 
 
Thus when the AGC is driven by the power of U
N+K
, the desired signal power at threshold is linearly 
related to the power of the second undesired signal U
N+2K
 and independent of the power of first 
undesired signal power U
N+K
.  At a constant desired signal power, the threshold of U
N+2K
 is constant—
independent of U
N+K
. 
 
Now we consider the case of AGC operation based on the power of the second undesired signal.  We 
assume that 
 
 
B-12 
G/G
MAX
 = 1,   when U
N+2K
 ? U
N+2K-AGCthresh
, and 
  = U
N+2K,AGCthresh
/U
N+2K
,  when U
N+2K
 > U
N+2K-AGCthresh
  
 
Thus, for the case U
N+2K
 > U
N+2K,AGCthresh
, we have 
 
D = D
MIN
 + (G/G
MAX
)
2
 U
N+K
2
 U
N+2K
 / (IP3 / R
1/2
)
2
 
 
Substituting for G/G
MAX
,  
 
D = D
MIN
 + (U
N+2K,AGCthresh
/U
N+2K
)
2
 U
N+K
2
 U
N+2K
 / (IP3 / R
1/2
)
2
 
 
Rearranging, 
 
D - D
MIN
 = (U
N+2K,AGCthresh
)
2
 (U
N+K
2
 / U
N+2K
) / (IP3 / R
1/2
)
2
 
 
Recall that, for values below the AGC threshold, each 1-dB increase in the power of U
N+2K
 causes a 
0.5 dB decrease in the undesired signal power that can be tolerated on U
N+K
 for a given desired signal 
power.   When the AGC is driven by the power of U
N+2K
, this trend reverses above the AGC threshold.  
At constant a desired signal power, each 1 dB increase in U
N+2K
 above the AGC threshold causes a 
0.5-dB increase in the undesired signal power that can be tolerated on channel N+K. 
 
SUMMARY 
Single Undesired Signals 
Interference creating by linear or non-linear effects within a TV receiver acting on incoming signals has 
been modeled as a conversion of the incoming signals into co-channel interference with a power 
proportional to D
L
 U
M
, where D and U represent the desired and undesired signal powers, respectively, at 
the input to the TV receiver.  The model has been developed for the following types of interference 
mechanisms: 
• Linear (M = 1; L = 0) 
• Second-order (M = 2; L = 0) 
• Third-order (M = 3; L = 0) 
• Cross-modulation (M = 2; L = 1) 
 
The model includes the effects of receiver noise at low signal levels. 
 
The basic model applies to the case in which no AGC-induced gain changes occur between the input of 
the receiver and the point in the tuner at which the interference mechanism is created (usually a 
nonlinearity).  The model is then extended to include the changed behavior that occurs when AGC acts to 
reduce gain prior to the point at which the interference is created.  The AGC model assumes that, for 
signal levels above a certain threshold, the gain will be adjusted in such a way as to maintain a constant 
signal level at the point at which the interference is created.  In practice, that constant signal level 
assumption may apply to the desired signal power D or to a filtered sum of desired and undesired signal 
powers.  The AGC model considers two bounding cases of such operation: 
• AGC driven by desired signal power.  AGC adjusts gain in such a way that the level of the desired 
signal at the point of the nonlinearity remains constant; 
• AGC driven by undesired signal power.  Here we assumed that the AGC is controlled by a filtered 
combination of desired and undesired signals, but with undesired signal at the AGC-sensing point 
being much larger than the desired signal, so that the AGC adjusts gain to, in effect, maintain a 
constant undesired signal power at the point of the relevant nonlinearity. 
 
 
B-13 
Relevant formulas for undesired signal power at TOV are: 
 
U
N+K
 = D
(1-L)/M
 [(1 - D
MIN
/D) / (R c)]
1/M
, when D ? D
AGCthresh
 and U
N+K
 ? U
N+K,AGCthresh
 
(i.e., no AGC operation) 
 
U
N+K
 =  D [(1 - D
MIN
/D) /(R c D
AGCthresh
M+L-1
)]
1/M
, when D > D
AGCthresh
 (i.e., AGC operating to keep 
desired signal constant at the point in the 
receiver at which the interference is created) 
  
 
U
N+K
 =  (D - D
MIN
) / (R c U
N+K,AGCthresh
M-1
), when U
N+K
 > U
N+K,AGCthresh
 (i.e., AGC operating 
to keep the undesired signal power constant at 
the point in the receiver at which the 
interference is created) and L = 0 (i.e., the 
formula does not apply to cross-modulation). 
 
where, 
 
D = Power of desired signal on channel N at input to TV 
U = Power of the undesired, out-of-channel signal at channel N+K at input to TV 
(D and U refer to signal level combinations that place the TV at TOV) 
R = Required SNR by TV at TOV 
D
MIN
 = Desired signal at TOV in absence of interference or external noise 
c = a constant describing the interference mechanism 
 
When operating well above the minimum desired signal level that a TV can demodulate (in the absence of 
interference), the interference model predicts that a log-log plot of undesired signal power (U) versus 
desired signal power (D) or a log-log plot of D/U ratio versus desired signal power (D) (i.e., plots in units 
of decibels) will be linear, with a slope determined by the interference mechanism and the AGC 
operation.  The slopes are summarized in Table B-4. 
 
 
Table B-3.  Slopes of Log-Log Plots of D, U, and D/U for Various Interference Mechanisms 
Interference Mechanism 
Slope of 
Log (D) 
Versus 
Log (U) 
in dB/dB 
Slope of 
Log (U) 
Versus 
Log (D) 
in dB/dB 
Slope of 
Log (D/U) 
Versus 
Log (D)  
in dB/dB Characterization 
Linear (M = 1) 1 1 0 Constant D/U 
Second order (M = 2) 2 0.5 0.5  
Third order (including third-order 
intermodulation of a pair of equal-
power interferers) (M = 3) 
3 0.333 0.667  
Cross modulation (M = 2, L = 1) Infinite 0 1 Constant U 
AGC-Stabilized Nonlinear 1 1 0 Constant D/U 
 
 
As desired signal power approaches D
MIN
, the threshold of the receiver in the absence of interference, the 
undesired signal deviates from the log-log straight line by amounts shown in Table B-4.   
 
 
 
B-14 
Table B-4.  Deviation in Threshold U from Straight-Line Projection as D approaches D
MIN
 
 
Deviation in Threshold U 
from Straight-Line Projection (dB) 
Interference Mechanism 
D/D
MIN
1
  
= 16 dBm 
D/D
MIN
 
= 3 dB 
D/D
MIN
 
= 1 dB 
D/D
MIN
 
= 0 dB 
Linear (M = 1) -0.1 -3.0 -6.9 Infinite 
Second order (M = 2)  -1.5 -3.4 Infinite 
Third order (including third-order 
intermodulation of a pair of equal-
power interferers) (M = 3) 
 
-1.0 -2.3 Infinite 
Cross modulation (M = 2, L = 1)  -1.5 -3.4 Infinite 
AGC-Stabilized Nonlinear w/U 
driving AGC
2
 
-0.1 -3.0 -6.9 Infinite 
Note: 
1
 For the nominal D
MIN
 value of -84 dBm, D/D
MIN
 = 16 dB when D = -68 dBm 
2
 With desired signal driving AGC, deviation from straight-line projection matches that of the 
original nonlinear process, except in the case of cross-modulation, which is not addressed.
 
 
Third-Order Intermodulation With Paired Signals at N+K and N+2K 
Third-order intermodulation between paired signals at N+K and N+2K was modeled as follows. 
 
D = Power of desired signal on channel N at input to TV 
U
N+K
 = Power of interferer on channel N+K at input to TV 
U
N+2K
 = Power of interferer on channel N+2K at input to TV 
 
where D, U
N+K
, and U
N+2K
 refer to signal level combinations that place the TV at TOV. 
 
R = Required SNR of TV at TOV 
D
MIN
 = Desired signal at TOV in absence of interference or external noise 
IP3 = Third-order intercept point of the receiver under the current AGC conditions 
 
We define the interference performance in terms of a parameter that combines IP3 with the required SNR 
of the DTV receiver (nominally 15.3 dB, or 33.9).  The parameter is computed from measurements of 
threshold values of undesired and desired signals when the two undesired signals have equal power (U = 
U
N+K
 = U
N+2K
). 
 
 
IP3 / R
1/2
 = (U
3
 / D)
1/2
 
 
or, in decibel units,  
 
(IP3 / R
1/2
)|
dB
 = 1.5 U|
dB
 - 0.5 D|
dB
  
 
 
Once we know IP3 / R
1/2
, we can use it in one of the following equations to determine the threshold for 
one undesired signal in terms of the other undesired signal power. 
 
 
B-15 
 
U
N+K
 = (IP3 / R
1/2
) (D - D
MIN
)
1/2
 / U
N+2K
1/2
 
 
U
N+2K
 = (IP3 / R
1/2
)
2
 (D - D
MIN
) / U
N+K
2
 
 
 
If D >> D
MIN
, then the equations can be converted to dB as follows: 
 
 
U
N+K
|
dB
 = (IP3 / R
1/2
)|
dB
 + (D|
dB
 - U
N+2K
|
dB
) / 2 
 
U
N+2K
|
dB
 = 2 (IP3 / R
1/2
)|
dB
 
 + D|
dB
 - 2 U
N+K
|
dB
 
 
 
AGC With Paired-Signal IM3 
When gain is constant, each 1-dB increase in the power of U
N+K
 causes a 2-dB decrease in the undesired 
signal power that can be tolerated on U
N+2K
 at a constant desired signal power.  Conversely, each 1-dB 
increase in the power of U
N+2K
 causes a 0.5-dB decrease in the undesired signal power that can be 
tolerated on U
N+K
. 
 
If the AGC acts to keep the power of U
N+K
 constant at the point of the nonlinearity that creates the 
observed IM3, the threshold of U
N+2K
 becomes a linear function of desired signal power and is 
independent of U
N+K
.   
 
If the AGC acts to keep the power of U
N+2K
 constant at the point of the nonlinearity that creates the 
observed IM3, the fixed-gain trend reverses.  At constant a desired signal power, each 1 dB increase in 
U
N+2K
 above the AGC threshold causes a 0.5-dB increase in the undesired signal power that can be 
tolerated on channel N+K. 
 
 
 
BIBLIOGRAPHY 
 
____, Code of Federal Regulations (CFR) Title 47, Telecommunications, Section 73. 
____, “Longley-Rice Methodology for Evaluating TV Coverage and Interference”, Office of Engineering and 
Technology (OET) Bulletin No. 69, <OET-69>, Federal Communications Commission, 6 February 2004. 
Advanced Television Systems Committee (ATSC), “ATSC Recommended Practice:  Receiver Performance 
Guidelines”, <ATSC Receiver Guidelines>, ATSC Doc. A/74, 17 June 2004. 
Advanced Television Systems Committee (ATSC), “ATSC Standard:  Digital Television Standard (A/53), Revision 
C Including Amendment No. 1”, Doc. A/53C, 13 July 2004. 
Advanced Television Systems Committee, “Recommended Practice:  Guide to the Use of the ATSC Digital 
Television Standard”, ATSC Doc. A/54A, 4 December 2003. 
Advanced Television Systems Committee, “Transmission Measurement and Compliance for Digital Television”, 
ATSC Standard Doc. A/64 Rev A, 30 May 2000. 
Bendov, Oded and C. B. Patel, “Television Receiver Optimization in the Presence of Adjacent Channel 
Interference”, IEEE Transactions On Broadcasting, Vol. 51, No. 1, March 2005, p.38-39. 
Bendov, Oded and others, “Planning Factors for Fixed and Portable DTTV Reception”, IEEE Transactions On 
Broadcasting, Vol. 50, No. 3, September 2004. 
Bretl, Wayne and Gary Sgrignoli, “Summary of the Grand Alliance VSB Transmission System Laboratory Tests”, 
IEEE Transactions on Consumer Electronics, Vol 42, No. 3, June 1996. 
Bretl, Wayne and others, “VSB Modem Subsystem Design for Grand Alliance Digital Television Receivers”, IEEE 
Transactions on Consumer Electronics. Vol. 41, No. 3, August 1995, p.773. 
Chong, Young-Jun and others, “The Design And Implementation Of TV Tuner For Digital Terrestrial 
Broadcasting”, International Conference on Consumer Electronics (ICCE), 2000 Digest of Technical Papers, 
13-15 June 2000, p40 – 41. 
Cowley, Wayne and Robert Hanrahan, “ATSC Compliance and Tuner Design Implications”, Electronic Engineering 
Times, May 1, 2006. (http://www.eetasia.com/ART_8800416208_480700_f6d4765f200605.HTM) 
Egli, J., “Radiowave propagation above 40 Mc over irregular terrain”, Proceedings of the IRE, Vol. 45, Oct. 1957, 
pp.1383-1391. 
Eilers, Carl and Gary Sgrignoli, “Digital Television Transmission Parameters—Analysis and Discussion”, IEEE 
Transactions on Broadcasting, Vol. 45, No. 4, Dec 1999, p.368. 
Federal Communications Commission Advisory Committee on Advanced Television Service, “Final Technical 
Report”, Oct 31, 1995, p.19. 
Henderson, John and others, “ATSC DTV Receiver Implementation”, Proceedings of the IEEE, Vol. 94, No. 1, 
January 2006, p.125. 
Martin, Stephen R., “Tests of ATSC 8-VSB Reception Performance of Consumer Digital Television Receivers 
Available in 2005”, Report FCC/OET TR 05-1017, <SHVERA Study>, November 2, 2005.  
(www.fcc.gov/oet/info/documents/reports/TR-05-1017-ATSC-reception-testing.pdf) 
Rhodes, Charles W. and Gary J. Sgrignoli, “Interference Mitigation for Improved DTV Reception”, IEEE 
Transactions on Consumer Electronics, Vol. 51, No. 2, May 2005, p. 463-470. 
Rhodes, Charles W., “DTV interference could be mitigated by receivers,” TV Technology Magazine, Vol. 22, No. 
17, p.21-23, Aug. 18, 2004. 
Rhodes, Charles W., “Interference Between Television Signals due to Intermodulation in Receiver Front-Ends”, 
IEEE Transactions On Broadcasting, Vol. 51, No. 1, March 2005, p.36. 
Rhodes, Charles W., “The Challenge of Channel Election, TV Technology.com, January 19, 2005.  
(http://www.tvtechnology.com/features/digital_tv/Features_Rhodes-01.19.05.shtml) 
Scheinberg, N. and others, “A GaAs Up Converter Integrated Circuit for a Double Conversion Cable TV ‘Set-Top’ 
Tuner”, IEEE Journal of Solid-State Circuits, Vol. 29, No. 6, June 1994, p.688. 
Wu, Yiyan, “Performance Comparison of ATSC 8-VSB and DVB-T COFDM Transmission Systems for Digital 
Television Terrestrial Broadcasting”, IEEE Transactions on Consumer Electronics, Vol. 45, No.3, August 1999, 
p. 922.