Data Quality Punch List Item #45 Date: January 29, 2005 Description: Jadzia says" Since 2nd January Power Outage cft stereo data might be compromised. I am wandering if you noticed something that might point to that. I would be eager to go through the plots myself if you would point me to the data. Idealy I would like to see number of cft axial and cft stereo hits on tracks. As you probably know D0 is now reluctant to paricipate in available accesses because of fear of ramping of solenoid. Therefore I need to have evidence of the data being compromised before requesting access to swap bad module." Marj says: "I don't have hits on tracks, just hot or dead fibers. I didn't notice a change, but I can look more closely. Wouldn't the vertex examine find a serious problem with the tracking? Marj" Action: Point to RecoCert directory. Submitted more. Forwarded to Marj. Resolution: Item #44 Date: December 14, 2004 Description: The lumy operations report says run 201585 had some problems. Well, it's 223.59 nb-1 "lost in daq". I looked and found 458095 events written to tape in this run. Action: Email to lumy group. Resolution: Lumy group fixed it. Tim Andeen, I think. Item #43 Date: July 19, 2004 Description: From K. Harder: Last week we had a special run for muon and SMT, and I am trying to figure out whether the track reconstruction looked normal. The captain was unable to start the vertex examine, and it seems that all the GM code was not running either. Are there RecoCert plots or anything else available for this run? The run number is 195223, a special muon runs with new PDT DSP code. Action: Standard run quality evaluation. Resolution: SMT Reasonable. Item #42 Date: June 29,2004 Description: Please recocert 194556 and 194557 TV 12.37 and 13.02 or 3. Action: HTD. Done. Resolution: See July 9 ADM on differences in RecoCert plots for TV 12 vs 13. Item #41 Date: April 25, 2004 Description: Weber: "run 192308 is bad for SMT. Crate x60 was not reading out. Crate monitoring was unfortunately not running in crate 60 for that run... No entries in SMT examines, zero occupancy from crate 60 and also the 'cft only' track vs phi show a pronounced enhancement (vertex examine). It might be useful to look at recocert for 192307 and 192309 (one run before and one run after) to make sure things are fine there. Action: HTD ran RecoCert for 192307, 8, 9 on April 30. Resolution: SMT reasonable on 192307 and bad on 192308 and 9. Item #40 Date: April 22, 2004 Description: Tom Nunneman ran RecoCert and noted "Run 192079, which has been selected as the new reference run, shows many significant differences to the old reference. There is a very prominent problem in the calorimeter at -1mumu has poorer width (ratio is 14/9 fwhm) and more muon tracks with 0 smt hits. The tracking eff'y is 1 sigma lower than preshutdown data. The CFT shows normal number of hits for postshudown data (not quite as many as before). The postshutdown fraction is 50% tracks with no SMT hits. Preshutdown was 17%, if I understood clearly. Action: Ginther, Choi, Nomerotski, Diehl, Sanders, Nurse, & Weber met to think. We think the width is worse because of worse momentum resolution with no smt hits. We don't think Emily has focussed on runs where we know of SMT problems or muon problems, such as SMT off or toroid open. Some postshutdown Recocert plots show number of smt hits looks much more like normal SMT tracking. We don't yet know what the problem is. Emily is to check which runs have problems. We'll focus on that data. - HTD 3/3/4. This turned out to be a code bug. George put us onto it in his rough cross section calculation. Resolved: March 4, 2004. Item #36 Date: March 1, 2004 Description: Run 189818 and runs since 18:30ish Feb 29 through 16:00ish March 1 have some kind of noise in NEC and in N bottom PDT's. Are these runs any good. Please run recocert when on, say, 189818 when it's available. HTD Action: HTD ran RecoCert April 30, 2004. Resolved: Reasonable runs. Item #35 Date: February 13, 2004 Description: Run 188957. Dean S. writes: "However, in this case I believe that the problem was sever enough that had anyone actually looked at one event on the event display monitor with the standard zero bias trigger one would have known something was wrong. As Robert has recently found, the "bad BLS" is most likely a failed capacitor. This has happened before, but that time the BLS crate tripped, and we thought we had a failed BLS supply. This time it was not a "good" a short, so only appears to add AC ripple to the power rails. This causes noise in both the L1 CAl trigger, and in the precision readout system for all BLS cards in that crate (16 trigger towers, 64 readout towers). So the "fix" applied last night to suppress the one really bad BLS card, was not enough. And the present attempt to kill that one trigger tower in L1 is very unlikely to work. If that does not work (which I suspect), the only "reasonable" thing to do, is power down that BLS crate, and wait for an access to replace the BLS card. Marking all the data taken this store as "BAD CAL". Action: Please recocert. Resolution: Done. Item #34 Date: February 2, 2004 Description: Silke N., on shift, notes that after a pedestal run, zero bias run 188917 shows little evidence of a warm region that occurred for 1/2 the previous store. The last run of the previous store was 188911. Now, reco uses pedestals taken 1 or 2 years ago. Of course, those are barely relevant. Zero bias, new peds, MET components are 0.8 and 0.4 Gev. The question is, what is the value Reco gets on previous run? Have we trashed the data? The GM plot for 188911 shows warm jets in eta 0.3 to 1.0 at phi=6. Silke says this is part of 2 ADCs. Action: Please run Recocert on 188911 when it is reco'd. Lima reco'd it Feb 5th. Metx mean is -1.37. Mety is at 2.16. Overall, worse. Jet cone NC comparison plots failed. but the warm region stands out in eta and phi. Also in good jets. Cal all cells shows the warm region. All cc cells is even more clear. The jets are rotten-looking. I'd like to see what happens if better peds are used. I asked SN to do so. April 30: Silke has made a couple of tries and she is checking thta she has repedestaled the correct cells. Resolution: Done. It helped. We didn't write a D0Note, though I wanted to. Silke wrote a talk for the Calgo meeting. Item #33 Date: January 30,2004 Description: Top group throws out runs including 180301-180305. There's more MET_x. It's peaked ~-4 GeV. Met_y is similar, but not as far, offset. There's extra jets at ieta = -8 at phi = 6.3 and 1.5. The logbook for Aug. 20, 2003 indicates there are no hot cells in this area but that there is a "warm region in crate 9". It seems to simply go away without solution and have appeared without cause. What happened? (Above info is form recocert. Call list: Carried plots to an expert. Action: Silke writes: "The files are not (anymore) on the CAL disk, so I could not actually look to see a hot tower. I then started looking at the log book and see: (ieta,iphi)=(18,1) Layer FH2 (12) (crate 7 adc 10 bls 6 tower 3 depth 8) E/event: 5.63 RMS: 0.38 occupancy: 1.00, the pT is about 2GeV so I decided not to kill it." Details on the pedestals were provided. "So there was a hot cell, that didn't quite make the cut to be killed online, but it was suppressed in the next run as the pedestal/sigma was too high. So there was a hot tower in addition to the one cell that did not make the "online-suppression"? Laurent also thinks to have seen a warm region, but with these plots, it's hard to tell. So this here is a case where you have to decide whether to kill the cell in the run before the pedestals killed it also, or not. But old/new pedestals would not have cured it. But then, it was not a hot tower. If there was one and you have the coordinates (any: ieta. iphi is ok), let me know and I'll look. About the warm region, I forgot again that there is actually still code missing that would need to be written. I'll see if I find the time to do that today, but this will definitely cause some delay. I'll let you know when I've some further." Resolved Date and Comment: April 30, 2004. I think the warm regions we experienced around this time may have been precursors of the "torrid torroid" noise. (Not all agree). -HTD Item #32 Date: January 28, 2004 Description: It's unclear whether central muon OK in 188032. This applies to all runs Jan 11 to Jan 13. There are NO notes in the GM log. Call List: HTD Description: I submitted RecoCert on it 01/28/04. 207 BC stubs in 100,000 events, 7000 unmatched muons. 187734 202 in 60,000 events and 4600 unmatched. I don't see how I can declare the runs bad. It turns out to be a shifgter screw up. Linda Stutte showed me how to look at the muon rootuples and they are OK. Resolved: 01/29/04. HTD Item #31 Date: January 28, 2004 Description: It's unclear whether MDT SBWT is OK in 187714 Call List: HTD Description: I submitted RecoCert on it 01/28/04. Resolved: 01/29/04 Seems OK. HTD Item #30 Date: January 24, 2004 Description: Marco reports checkerboard pattern in calorimeter runs. EM, in particular, one has to look at a lot of runs to see it. Call List: HTD Description: I'd like a run number - something - to go on. I've never seen it in the runs I've looked at. MV says one needs a lot of EM data to see it. Resolved: On or about end of winter shutdown this was determined to be due to the cal gain switching errors due to timing problems in electronics. Item #29 Date: January 15, 2004 Description: Sanders reports: "In run 188156, there is a spike around phi=3.2 in the global track distribution, as seen by PhysEx. Do we see this offline?" Call List: HTD or offline shifter. Action: HTD submitted 01/24/04. I looked at the recocert plots. There is what I'd call a nubbin of tracks at phi = 3.2 but not a spike. There is a problem with one of the smt layers (4) at that phi. I asked M.S. if the plots I sent him showed the same thing as the online. Item #28 Date: January 11, 2004 Description: HTD reports: "I'm looking at run 187230 using Recocert. The plots are in /rooms/greatlakes/michigan/dataquality/jobs /run187230-p140600... I'm comparing to run180951-p140302-DST/run180951-p140302-DST.root The tracking plots seem poor in the newer data. Specifically, "Tracks All" shows: many fewer tracks in "ntrks" otherwise same? "Hits All" seems normal, except seems to be fewer with 16 CFT hits. I can't see where they went. "Isolated Tracking Eff'y" has lower efficiencies in low trk-multiplicity events. Tracking in Jets shows overall lower tracking eff'y. Electrons, matched muons about same, but they have tracking eff'y normalized out. There's more unmatched muons. Jets are similar except the new run has nothing with F90 > 0.8. (I don't think that's related to this problem). "Detector Occupancy" says There's way fewer CFT clusters, cal cells, and cps clusters. This seems to me to be at the heart of the problem. But "Cal All cells" seems same and SMT cluster plots seems same. There's more dead CFT channels, but I expected that due to post-shutdown problems. CFT occupancy looks about the same. There's fewer muon local tracks but I recall normalization problems in this figure of old. The Reco timing is significantly faster as if there's fewer clusters. Call List: HTD sent email to Melanson and Suyong. Action: HTD ran recocert on more runs. 187800 has similar number of tracks but the eff'y in isolated tracks doesn't look terrible at low-pT. Other distributions look similar. Note 187800 is p14.06.00. 187230 is same. Drew Alton asked about recent RecoCert stuff so I steered him at this run. He reports: "...page 26. The number of cps clusters, the number of strips on those clusters, the phi distribution all look very bad. I don't think it leads to the other problem [cft clusters - htd]. I haven't checked if it's present in other runs, but I figured out most of the bug that caused it...it's database constants. As you know the CFT (and cps) map changed. When Fred first told us about this he said crate 51 VRB 15 moved to VRB 19, but in fact crate 52 VRB 15 moved to VRB 19. When I prepared my mip calibration numbers I used the swap for crate 51. The CPS channels in crate 51 VRB 15 are now coming out without MIP calibrations. (I don't understand this, because I thought the DB would use the calibration from the previous calibration but it smells too much like the problem). The thing that confuses me more is that the actual channels that moved: crate 52 VRB 15 ->19 all seem to be fine. I guess the DB doesn't work the way I thought it should. I will get together with Brent and fix this today." Drew further replies: "I guess all runs processed since the shutdown are screwed up." Brigitte Vachon says (well she didn't actually say it). "Hey, don't try to pin this on the CTT". Actually she asked if the CFT calibration was correct. HTD forwarded this to Fred, Jadwiga and Stefan. Jadwiga says: "No, there is no run with bad calibration constants. We look online at occupancy plots after calibration is applied for all runs. All runs have expected occupancies. The calibration run number for this run is 187158. I checked that online database has data for this run. I have log files from transfer to offline databaes, this log file does not show anything unusual. May be Taka would be willing to look at offline data base and compare, for example, with the next calibration run 187661." Taka says: "I find nothing unusual about the pedestals in the offline database for the calibration runs 187158. 96% of channels have values between 50 and 80. For the calibration run 187661, this value is 95%. Do I need to look these in more details?" Further: "97.7% of channels in the thresholds table have values between 50 and 100 for run 187158. 97.6% for run 187661. HTD: I now believe most of these problems are due to axes differences in p14.06.00 vs p14.03.02 RecoCert. I still don't understand why 187230 has low eff'y in the low-pT end of pT vs trk multiplicity on page 4. Jan-29: Start keeping stats on Runs 187734 trkmulty not ok 187714 trkmulty not ok 188032 trkmulty ok Item #27 Date: January 6, 2004 Description: Horst Wahl notices phi dip in Vertex Examine. It seems to be a persistent thing. He writes: "VertexExamine shows dips in the L3 track phi distribution at phi = 1.5, 4.5; (also seen in hit distributions) -- all due to mapping and old algorithm? many CFT only tracks, with peak at phi=2, dips at phi=1.5, 4.5." We are in runs 187804-812, for instance. Call List: HTD sent email to Sanders. Action: Sanders reports this is fixed up Jan 15th: "Vertex examine will have similar tracking as L3 later today, to "fix" dead cft fibres, as is done in L3 itself. Resolved: Jan 15th, 2004. Item #26 Date: December 19, 2003 Description: Jadwiga and Mike H. note the post-shutdown CFT dead channels. Not understood. Perhaps associated with warm-ups. Call List: Detector people are on it. Resolved: Over the next few days. 1% CFT channels didn't come back from the shutdown. Item #25 Date: December 4, 2003 Description: HTD ran recocert on 185751 and compared it with run 180951 and noticed something interesting in the SMT occupancy histograms. It's SMT clusters in barrel 4 plots. In "phi layer 4", around phi=1 there's occupancy in the comparison run but not in 185751. Conversely, in "phi layer 5" it's the other way around at phi = 4.5. Call List: Michael Weber and Sergey Burdin. Action: Michele W. says:"Thanks for looking at the plots. If you go through all the plots you will find ca. 35 only black ("appears new") and ca. 9 only red ("disappears"). This corresponds roughly to number of devices we enabled (repaired, 40) odd were newly lost (13) since the shutdown. I'm in the process of going through our lists and comparing it with the recocert plots. He reports he doesn't understand the shape (tilt) in the occupancy plots as you go across the face of the device. He thought it should be flat. It was expressed at the 12/15/03 Verification ADM that maybe this is a mechanical problem that causes the HDI's to be flaky. Recall, they seem to recover if you wiggle the connectors. An autopsy might be in order. Resolved: Jan. 19, 2004. Sergey Burdin: "Drew told us about this about 2 weeks ago. I looked offline on run 187910 and it looks like run 187169 in December (the distributions were affected by the CFT map though). The event display does not show anything unusual also. There was some discussion at the beginning of December about the region phi=1 (please see below). Is it the same problem? Can you compare your vertex examine with the run taken at the beginning of December?" Sanders: "Rick noticed a spike at phi=1 for CFT only tracks in vertex examine. This means that the L3 tracker didn't pick up any SMT hits in that region. Does this imply that the SMT is "dead" in one small phi-wedge? I think offline (physics examine) things look quite ok. Item #24 Date: December 2, 2003 Description: HTD noticed out-of-time hits (~-50 ns) in central muon C-layer scintillator plots from all recent data since startup after shutdown. Previously such hits indicated a bad or flaky SFE. Call List: Al Ito, Steve Doulas, Anatoli Evdokimov. Action: Al Ito and Robert Harrington replaced the SFE on 12/3 after- noon. Robert notes: SFE $20 in CES, which is for PDT counters 200, 210, 220). Please let me know if any of you see any more hits < -20 ns. Resolved Date and Comments: Done. See above. Item #23 Date: October 16, 2003 Description: HTD ran recocert on 175645 for the purpose of studing RecoCert and noticed (p13.06.01) that there was terrific calorimeter noise. Call List: Dean and Robert. Action: They identified that the problem was an analog driver (ADC) for a specific eta-phi cell set (16 eta-phi units). 1 layer (3rd em) all with same energy. It's a little curious becasue only 14 cells show hot. But that is not unheard-of. The ADC card is in 8 pairs so if the first pair has a problem the rest can get stuck umm, high. It was noticed by the shift and fixed by Parua the same day. It will be interesting to see what p14.03.02 does with this. Item #22 Date: October 3rd, 2003. Description: HTD ran recocert on run 180911 for the purpose of studying one thing (cal. noise, see item #19) and noticed something else. The CFT adc distribution, stereo view, has an excess over run 178722 at low adc counts. I guess this is indicative of hot channels in the adc readout but wanted to check with you on that. The cft occpuancies do seem slightly higher and u and v, all layers, but in such a way that doesn't point me to a specifically hot area. It changed between 180796 and 180828. Call List: Jadzia W. and Mike Friedman. Action: Requested advice as to what this is and how serious. Jadzia wrote: "I looked in CFT eLog for entries for run 178722 and 180911. I do not see anything abnormal there and that is all I know about these runs. It is disturbing that such change occured and we are not aware of that. For the future I will try the get in our cft_examine the plots similar to the ones you sent me." HTD: It changed between 180796 and 180828. 180796 is on 9/1/3. There were CFT calib runs ~180801-180805 in between these global runs. One of those wasn't cft calib. Well, logbook notes problems with the calibration download, but expert thought it worked. The thing is those calibrations occur every day. These were for crate 50. The next crate 50 calib was 9/5 run 180921. But global runs after that still show the problem. Mike Hildreth showed plots in the 12/19 ADM and said it was OK now. I should compare these with those. Resolved Date and Comments: April 24, 2004. This has not re-occurred as of April 2004, so far as we know. Item #21 Date: September 16th, 2003 Description: Ricardo P. compared run 180881 to 180789. He writes: "Normalization when comparing to previous run is screwed up on some plots: all tracks, electrons, muons. Seems to be a software problem on the verifyRecoCert code." Follow-up indicates that comparison with 178722 doesn't show a serious normalization problem. Call List: HTD. Action: Follow-up indicates that comparison with 178722 doesn't show a serious normalization problem with 180881. It seems 180789 is the outlier. I sent this off to Harry M. Harry suggests the tracks are the problem in 180789, not the normalization. Maybe. I am sticking with normalization for the later muon plots. Also, Harry wonders about the instantaneous luminosity for the later run. It was first of store, started at 40e30 and fell over time to 30e30. Run 180881 is 19-15e30. Harry's guess may be borne out. Item #20 Date: September 13th, 2003 Description: Raimund S. saw run 180748 had no entries in the SMT Cluster Bar2 in phi layer 2 in range 0.5-1. (p20 bottom left). Call List: TBD Action: TBD Resolved Date and Comments: Item #19 Date: September 9th, 2003 Description: Shift Captain complained that calorimeter jet trigger rates are 50% to 1000% higher than normal in Run 180911. We are curious as to whether something was going on. Incidentally, Captain logbook notes: "Just before 6am there was a glitch in the luminosity and the background rates. About 15 minutes after that, the L1Cal trigger lost power due to what appears to be a faulty interlock readout box. It was replaced with a 5volt supply, and we began running again around 7:20....but with the jet rates returned to normal." Curious. More logbook notes: "MCR called to say that CDF called them to say that at ~5-6am CDF noticed shift of ~5cm in beam position. Our vertex examine shows a shift of ~2.5cm. MCR is aware of that and working on it. Mike Martins is coming in to work on it in ~20min. They will call before they make any changes." Call List: Any offline shifter, Schamberger and Parua to be informed. Action: Submitted 09/30/03 HTD. VerifyRecoCert against run 178722 on Oct. 3rd. Vertices distribution close, 180911 has one peak in x and one peak in y, though the y-peak is 1/20th cm higher. There is a slight excess on the +z side, maybe centered around 5 cm, in 180911. Very small, though. Electrons(all) identical. Muon matching eff't is a little better in 178722. MET is close. There's a little more MET(y) in the downward direction (1.5 GeV?). This caused me to wonder if a net met (or a shift in it) can be a vertex position effect. Well, obviously it can, if you think about it in the limit, but how much for a small shift? -1.5 GeV for +1/20 cm shift? Jets(coneall) has a softer Et spectrum than 178722. It also shows an excess at eta=1, phi=2. This is not reflected in the cal all cells distribution, though. CFT adc summary has excess at low adc counts not seen in the comparison run. I don't usually study this, though and don't know what it means. Probably a hot channel. I sent a distribution to Mike Freidman (Marge C.) and to Jadzia. This spawned punch list item #22. Time in C-layer FAMUS (scintillator) shows an excess in the late peak (~+20 ns) but A and B times are normal. I don't know which end the excess is in. Resolved Date and Comments: I (HTD) was really interested in seeing if there was any effect as bad as ring-of-welding. Nothing like that here. Item #18 Date: September 4th, 2003 Description: Daria Z. wrote: "I'm looking at recocert plots for run 180206 (p14.03.02). /rooms/greatlakes/michigan/dataquality/jobs/ run180206-p140302-DST/run180206-p140302-DST.ps on clued0. The phi distribution of all SMT barrel clusers (p27 in the ps file) has a decreasing trend. In my experience it happens when there is a cutoff on the max total number of objects in the reconstruction, and the reconstruction goes over sectors increasing in phi. I think the fact that the max is saturated at low phi can lead to an overall inefficiency in cluster finding, and hence in tracking." The inst. lum'y fell from 25 to 21 e30 during the run. Action: Forwarded to Melanson and Slava Kulik. Slava replied (9/4): "As far as I know, there is a cut on SVX occupancy (offline), so per chip, and maybe now on HDI (online), but not on total number of SMT hits. At least not in reco." I replied, asking him what this means. 10/23/03 (HTD) Inactive item. Item #17 Date: August 21, 2003 Description: Please run RecoCert on run 180301. This run was taken after a BLS supply was changed, but no pedestal run followed it. Runs after 180305 have the pedestals. Call List: Any Offline Shifter. Action: Submitted by HTD 8/21/2003. Aug. 24: No Reco'd data in SAM. Submitted by HTD 9/3/2003. Sucessful submit. Comparison plots were made against run 178722. MET looks a few GeV higher. Cal CC cells doesn't look too good. We should compare against a run immediately after 180305. I compared against 180329, taken after a new pedestal run and download was performed. There is a significant improvement in jets, met, and cal cells in the latter run. Resolved Date and Comment: September 9, 2003. I will stress this point in an ADM. Item #16 Date: August 18, 2003 Description: Please run recocert on run 179614 and see if electrons look OK. From M.Sanders: "I was paged by the GM shifter earlier today who told me that he didn't see any L2 EM global objects in run 179614. These plots get filled by l2gbl_analyze into the L2gblEM block in the trigsim_analyze roottuple. I checked the roottuple and it did not have anything filled in this block. However, the L2gblJet stuff did appear in the roottuple and thus the histograms." Call List: Any Offline Shifter. Action: Submitted by HTD 8/18/2003. Aug. 19: (M. Sanders and L. Duflot). At some luminosity L2EM doesn't run (much) and therefore the online plots have no information. Run 179614 was one such run. Aug. 25: (HTD) Electrons look OK in this run. Jets are also OK, but the are some hot spots in the jet distribution at about eta=-1 phi=4.5 Resolved Date and Comment: August 25, 2003. Item #15 Date: August 18, 2003 Description: Please run recocert on run 180130 or 180131 when they become available. This has "significant fraction of events with large calorimeter coherent noise, mainly at boundaries of the central and end cryostats. No obvious correlation with any DAB work has been found." - Shift Captain Summary. Please compare to run 179334. Call List: Any Offline Shifter. Action: Submitted by HTD 9/4/3. I looked at the output of RecoCert in comparison with run 178722. Run 180130 looked fine. There was no notable discrepency between it and the comparison. The muon plots still have the normalization difference because we aren't using an update yet. There are tiny little excesses in the CC cells that are in the same place as tiny little excess in the much older comparison run. This does not look so bad as "Ring-of-Welding". - HTD 9/8/2003. Resolved Date: September 8, 2003. Item #14 Date: August 4, 2003 Description: Thomas Nunneman notes that comparisons for all three "muon local tracks" plots isn't working right. It might be a normalization error. See, for instance, run 179236. Call List: Melanson, Daria Z., Mulders. Action: Item #13 Date: July 29, 2003 Description: Please run recocert on run 179332 when it becomes available. This has "Ring-offire". Please compare to run 179334, maybe. Call List: Any Offline Shifter. Action: Diehl did it. Resolved Date and Comment: August 14, 2003. Item #12 Date: July 17, 2003 Description: Run 178737. "Jets Cone R=0.7 No Cuts" shows larger horns at ~|eta|=1.5 than the reference. There is some strong phi structure not seen in the reference. Coarse hadronic fraction is wider at the low end. Perhaps all of this reflects the hot areas visible in the Jet phi vs eta plot? Question for experts: why does the EMF distribution contain little between 0.8 < EMF < 1 in comparison with the reference below? This was also found by R.I. 7/17/2003 in run 178765. And Sabine in run 178763. Call List: 7/17 MC Gao is Cal on-call. Also, sent to Wobisch 7/18. Also sent to Bernardi, Groer, Parua on 7/20. Action: No, there was no change in the jet selection criteria. Maybe a trigger bias (less electrons triggered) or a higher rate of jets due to the noisier ch, which masks the peak at high EM fraction, or a combination of the two. July 23: HTD determined this change may have occurred on Trigger version 11.04 to 12.10 boundary. That occurred 178721 to 178722. Plots exist. Email to Bernardi, Wobisch, Parua, Groer. July 28: The trigger information from Lum'y guys is on a crapped-out disk and this has been causing some delay. Resolved Date and Comment: August 4, 2003. Attributed to tightening of L3 EM filters at change in trigger version. Item #11 Date: July 17, 2003 Description: Run 178737. "Tracks All" has a minor spike in the Z at DCA plot at Z=0. This looks like it might be axial-only tracks. It's also reflected in the "Vertices Primary" plots. I thought we would not be seeing this kind of thing very frequently any more. Is it still expected to happen from time-to-time? Call List: Drew Alton is CFT on-call. -> Gordon Watts. Action: Drew commented and went on to summarize: "In fact I dont think this is axial only tracks. Since everything you point to is a vertex quantity...I'd ask a vertex expert. If you want to give me more histograms I'd be glad to look at them." In short, maybe these are zero-bias or minbias or muon events without a reco'd vertex. Drew followed up with: "page 2 of your plots shows its not axial-only, because there aren't >100 events with 8 cft hits. There are only 50." July 29, 2003. Gordon W. thinks it might be "hot cal triggers" making events without vertices. Resolved Date and Comment: 10/23/03 (HTD) Declared obselete. Item #10 Date: May 20, 2003 Description: Several events in zero-bias run 176055 have non-correspondance between L1 and L2 (which don't agree) and L3 masks (which agrees with L2). This was identified by Begel. Call List: Jon Hayes (script-runner) and Gerry Gugliemo (data-logger). Action: Gerry is working to determine if the problem is upstream of the data collector. He found out that there are "pathological" features of this run which have not been observed in physics data-taking. It's a problem upstream of datalogger/collector router. Some of these are "known" occasional blow-ups or stuck bits. Resolved Date and Comment: Never resolved. Obsoleted July 17, 2003. Item #: 9 Date: May 15, 2003 Description: A dip was seen around phi=1 and phi=5 in the cft track distribution from physics examine, at the beginning of run 176929. The problem seemed to disappear during the run. Similar dips are seen in run 176880. Could be bad CFT or bad SMT, no dips seen in global tracks, nothing seen in vertex examine, L3 tracks or cft examine. Call List: Drew Alton Action: Resolved Date and Comment: Never resolved. Obsoleted July 17, 2003. Item #: 8 Date: April 26, 2003 Description: A dip is seen around phi=2 in both the L3 tracks plot in trigger examine, and in vertex examine. From vertex examine, it is clear that a small region is missing in CFT axial layer 2. However, CFT examine doesn't show a problem. Call List: Jadzia Warchol, Fred Borcherding, Michiel Sanders Action: This was first linked to a "new" CFT calibration, whereas vertex examine is using an "old" calibration. However, the real problem was in one SVX chip, SVX0 on slot 9A5. It was not seen in the examine because: "The confusion was the consequence of the way our hit maps are filled: discriminator hits without SVX hits are assigned ADC=0 and put into SVX hit maps. This an old issue we argue about since somtime but we never reached a consensus. Luckily cft_examine delivers pulse hight distribution for chips specified on rcp file. Pulse height of the bad SVX shows a lot of entries with ADC=0, meaning entries to SVX hit map come mainly from events where discriminator fired but SVX did not." Resolved Date and Comment: April 29, 2003 Bad AFE board replaced in a controlled access. A similar problem occured in July 2002, where CFT examine showed no problem but in reality one SVX was bad. This affects runs > 176175 and < 176346. Item #: 7 Date: April 20, 2003 Description: The GM shifter noticed a spike at eta=1.1, phi=5.7 in the tau trigger rate. This corresponds to a hot cell in calorimeter. This is run 175890 & 175891. Apparently this hot cell has been around a long time and is getting worse. Call List: Nirmalya Parua Action: No response. Obsoleted Jun 9th, 2003. Item #: 6 Date: April 14, 2003 Description: Run 174212 shows non-standard (fewer) hits in RecoCert tracks plots, lower efficiency, and two-peaked vertex distribution. Call List: Ron Lipton, Michiel Sanders Action: Ron found comments in the logbook. They said: "The vertex examine showed large losses in stereo tracks found look carefully at the fiber fired map and find a region that may be suppressed. Reset the cft_examine. Request that the daq shifter give the cft shifter a chance to init_VRBC when a crate goes missing rather than continuing via an SCL init." Sanders writes: To my knowledge, the beam should be ~x = -0.0193 cm, y = 0.0208 cm. The beam jumped by about 1 mm when the TeVatron magnet was replaced, early March. So, after March 12th, run ~ 174210, I expect a much larger x value. For run 174212, I expect something like x = 0.0760 cm, y = 0.0301 cm. (right now (current store) it is even further away from zero: x = +0.09672, y = +0.03515). So, the highest peak in the vertex plots is the correct one. I don't expect beam shifts during that run that would explain the second peak. Next guess would be a bias in the vertex algorithms for vertices with small number of tracks..... But, it is surprising that the second peak is exactly at zero. Note that the number of tracks attached to the vertex is very small. Also, lots of events with zero tracks. Probably a very bad run for tracking. Resolved Date and Comment: This seemed to be a one-time effect. Item #: 5 Date: April 11, 2003 Description: There is a dip in PHI at DCA on Tracks(All) RecoCert plot between 36-180 degrees. Call List: Drew Alton Action: Drew suggest this will go away with data after CFT timing move to occur April 11th. Resolved Date and Comment: It didn't. Now (June) it's thought to be related to the vertex position. Item #: 4 Date: April 11, 2003 Description: Readout Errors in SMT from end of Jan. (2003) shutdown to ~ Mar. 10th effected the reconstruction. Runs sould be graded as "Not perfect but analyzable". Can this be seen in RecoCert plots? Call List: HTD Action: Ron Lipton is still worried about this July 17, 2003. Resolved Date and Comment: Obsoleted 10/23/03 (HTD) Item #: 3 Date: April 3, 2003 Description: Muon Examine plots show central scintillator hits out of the expected time-gate. Call List: Al Ito, Steve Doulas, Dmitri Denisov Action: DD recommends to check SFE's. Resolved Date and Comment: Al Ito replaced SFE $2F in crate x50. Histos look good starting in run 175322. Item #: 2 Date: March 12, 2003 Description: Run 173528 and 173529 have many events with cal readout skipped? Call List: Groer, Bernardi, Duflot (calorimeter data quality), Sang-Joon Lee (Recocert), Parua (03/25/03 calorimeter data quality) Action: According to histograms produced by Reco-cert there are many events with no calorimeter cells. See Sang-Joon's histos at /work/yogsothoth-clued0/sangjoon/Reco_Mon_Plots/*/Cal_All_cells.ps where * is 173528_Plots or 173529_Plots. 4/4/2003 - Dean says that unless another cut on ET has been applied besides 2.5 sigma in E (not Et units) the probability for it to happen even once is vanishingly small. Resolved Date and Comment: April 4. Harry Melanson - "In order to get data from the calorimeter chunk, one needs to supply a vertex (???!!! Well, that's the way the code is written). I check whether there is a primary vertex found, and use it if so. I'm guessing the spike at zero is for events where no PV is available. I will fix that in an upcoming release...". A recocert-only problem. Item #: 1 Date: March 12, 2003 Description: Run 174207-174212 dimuon trigger problem. Drop in number of N dimuons should be fixed in store with run 174238. Call List: RobMcc (L1 musim), Sang-Joon Lee (recocert), Diehl (muon detector) Action: RobMCC discovered it and fixed it. Muon detector examine plots showed no discrepency. Reco-cert was busy with higher priority item Resolved Date and Comment: March 12, 2003. This was discovered ex-post-facto and was resolved on discovery.