Jul - Dec 2001
ANDREW
Wed, 20 Jun 2001
We are pleased to announce the
"Run II Tevatron Diffractive Monte Carlo Workshop''
This small WORKshop will be July 18 - 20 at Fermilab with a goal of
developing the set of Monte Carlo generators for use at the Tevatron Run
II.
The aim will be to investigate each possible diffractive channel, for
example single diffractive dijets and double pomeron to dijets, W
production, Drell-Yan, and diffractive and double pomeron to Higgs, and
assess the different MC's capabilities and predictions. We will aim to match
the needs of the experiments with the capabilities of the Monte Carlos and
theoretical models. The current list of diffractive Monte Carlo generators
is
Pomwig
SCI/GAL
Pompyt
Phojet
Rapgap
The workshop will primarily focus on discussions of Pomwig and SCI/GAL as
they apply to the Tevatron. We will also discuss the status and future of
Pompyt. Phojet while useful for soft processes currently does not model
hard processes effectively and will not be one of the initial MC's used for
hard diffraction, although we are interested in possible upgrades. Rapgap
does not currently simulate proton-antiproton collisions, although this
upgrade is certainly of interest as well.
In addition to Monte Carlo experts, there will be representatives from
CDF+D0 and HERA, and others with an interest in the diffractive
measurement program at the Tevatron.
The program will be
Day 1 presentations on MC's and Tevatron experiments
discussions of improvements that are requested and
can be made, discussion of differences between MC's
and predictions
Day 2 working day to implement changes, resolve differences, train
experimentalists
Day 3 discussion of results and future plans
By scheduling discussion and work time as well as talks, we hope this
workshop will allow us to map out a plan of experimental measurements, Monte
Carlo and Theory Development which will allow us to make the most of the
Tevatron at Run II.
There will be no registration fee, no proceedings other than a web page
where talks can be posted, and no financial support available. Further
details about the program will be sent in two weeks. We apologize about the
lateness of this announcement and request that you forward it to others who
might be interested, with a CC to us so we can update the mailing list for
future announcements. Please confirm your plans to attend and if you would
like to give a talk and the topic (Note: we do plan to keep the talks to one
day and on subject).
Diffractive MC workshop Bulletin #2
Here is the draft agenda for the workshop:
(*=unconfirmed)
+++++++++++++++++++++++++++++++
Weds. July 18
9:00 1 East Wilson Hall 1st floor
CDF Run I Diffractive Results 40 min Kenichi Hatakeyama
D0 Run I Diffractive Results 30 min Andrew Brandt
Discussion: 10 min
CDF Run IIa Diffractive Plans 25 min Koji Terashi
Break 15 min
11:00 Curia 2 Wilson Hall 2nd floor Special Seminar
Rapidity gaps @ Tevatron: Soft Color Interactions, BFKL, Higgs
60 min Rikard Enberg
12:00 lunch
1:30 1 East
D0 Run II plans 30 min Michael Strang
CDF Run IIb plans/Higgs* 25 min Mike Albrow
Diffractive Higgs 30 min Christophe Royon
Higgs Discussion 20 min
Break 15 min
15:30 1 East
POMWIG MC 45 min Brian Cox
Status of POMPYT* 10 min Rikard Enberg
Beyond the MC 20 min Dino Goulianos
Break 15 min
17:00 Discussion of plans for Thursday Work
+++++++++++++++++++++++++++++++++++++++++++++++
Thurs. July 19
Working day
18:00-19:00 Discussions in 1 East
More work
20:00 Conference Dinner (Chez Leon or Meson Sabika)
++++++++++++++++++++++++++
Friday July 20
9:00 1 East
Discussions of Thursday work
Gaps between jets talk from H1?
Discussion of Gaps between jets
11:00 more work
13:00 lunch
14:30 final discussions 1East
16:00 Workshop adjourns
19:00 Red Sox vs. White Sox (baseball game)
++++++++++++++++++++++++++++++++++++++++++++++
Let me know of anything that is left out, or not given
enough time. Let me know if you plan to attend the dinner
and/or the baseball game. If you need any information
about hotels or other logistics let me know that too!
________________________________________________________________________________
CARLOS
Sun, 17 Jun 2001
Two files with plots obtained from the scope. channels 1 and 2 in both files
correspond to one per revolution clock and our in time clock, respectively.
http://www-d0.fnal.gov/fpd/FPD_News/p1dp2d.eps.
http://www-d0.fnal.gov/fpd/FPD_News/a1ua2d.eps.
File p1dp2d:
===========
channel 3 is P1D raw signal compared to our in time clock
channel 4 is P2D raw signal compared to our in time clock.
as one can see the rate we observe in P2D is mainly from Protons.
Rate is P1D has a greater contribution from antiprotons than for protons.
File a1ua2d:
============
channel 3: is A1U raw signal
channel 4: A2D signal
If one observes carefully the plot one finds that
A1U rate has a greater contribution from protons
A2U has a greater contribution from antiprotons.
The plots were taken during the store last friday.
################
Sun, 17 Jun 2001
I used store 538 on sunday to look at our single rates.
using a discriminator at 30 mv.
what I find is the following:
HV P1U P1D P2U P2D A1U A2U A2D
HZ HZ HZ HZ HZ HZ HZ
=== ==== ===== ===== ====== ===== ==== ====
50% 6 4 0 47 0 1 0
60% 385 315 0 402 11 32 15
70% 2535 3126 1 3012 119 183 97
80% 9030 11573 99 7947 291 883 536
90% 21326 26230 978 12376 989 2316 1447
100% 33626 40740 4117 14870 2463 4110 3006
A plot with those numbers is at:
http://www-d0.fnal.gov/fpd/FPD_News/singles.ps.
If one looks at the plot one sees that rates are increasing
with HV withouth much evidence of a plateau. This is quite strange
because not even at -1700 V we are finding a plateau.
If plateaus were measured at lab 6 starting at about -1400 volts,
we might expect higher voltage needed at our SCR due to the attenuation
of the cables from the tunnel.
I will keep investigating, ideas are wellcome. I will try to do a finer
scan in HV.
Finally, I am concerned about beam stability if we need to run at High
voltages. On friday I tried to look at rates but the rates were
quite high at P2D and jumping too much, they went from 200HZ up to 90 kHZ
in a period of 20 minutes and the HV in the counter was set only at 50%.
It will be good to figure a way to have a plot of rates versus time.
################
Sat, 23 Jun 2001
Here is a ps plot with rates we took when we moved the pots 20mm in. as Mike
said, we didnot observe any change in phalo or pbar halo, while the rates in
A2U increased by about 1.4 kHZ.
http://www-d0.fnal.gov/fpd/FPD_News/fpdrates.ps.
################
Tue, 26 Jun 2001
I took some data with detectors at home position, L0 pmts at 1600 V
and MAPMTS at 700 V.
The data is located at
/d0olcluster/scratch/D0fpd/data/
there are two runs for now:
run000001.dat --> trigger was P1U in time
run000002.dat --> trigger was P1U.P2U in time.
#################
Tue, 03 Jul 2001
You find tdc plots and some adc plots from the data we took yesterday with
an elastic trigger:
http://www-d0.fnal.gov/fpd/FPD_News/tdc2jul01.ps.
http://www-d0.fnal.gov/fpd/FPD_News/adc2jul01.ps.
The conditions were the following:
TRIGGER = A1D.A2D.P1U all signals in time to the clock.
L0 counters at -1600V, discriminator thresholds at -50mv
MAPMT at 900 V.
A1D and A2D were at 30 mm
P1U at 25 mm from home position.
TDC PLOTS
=========
The tdc plots for the counters that were part of the trigger have a
strange structure, peaking up towards late pulses. I believe those
late pulses migh be due to some hits with low amplitudes(just passing
thresholds) on the L0 counters. However, the total width of the TDC
distributions is about 40 channels = 10 nsec.
The widhts of the tdc distributions for the counters that were not
part of the trigger look gaussian with an approximate width of 50
channels (~12 ns). P2U and P2D counters had HV at 0 volts.
ADC PLOTS
=========
I just printed out some of the ADC plots for some channels of P1U
in two different planes. (the name os the planes migh not be correct)
the important point is that we have many channels which show to be
efficient for signal. Since we have 900 volts on the multianodes, we can
take next time a run at 950 volts to have a better gap between
signal to pedestal.
UNPACKING PROGRAM
=================
The unpacking program is working for IRIX machines as d0mino, but not
working for linux machines as the d0ol13. The problem is linking to
the cern libraries. I am investigating the difference. I created
a directory /home/d0fpd/unpack at d0ol13 with the unpacking subroutines,
everybody is wellcome to use them as a starting point to look at the data
and make suggestion for triggers and to help understand the detectors.
###########
Wed, 27 Jun 2001
We moved A2U 30 mm today, at the last hour of the store. I am attaching two
ps plots that show that there was no effect on the other rates, while the
rate in A2U increased by about 2.5 kHZ. The L0 pmts were run at -1600 Volts,
and discriminator thresholds were at -30 mv.
http://www-d0.fnal.gov/fpd/FPD_News/rates27jun01.ps.
http://www-d0.fnal.gov/fpd/FPD_News/otherrates27jun01.ps.
We will try to move other P1U and P2U at the end of next store.
I am proposing also that once we check the motion in each pot, we should
go to an elastic trigger mode and take data with that trigger. That would
need A1D and A2D and P1U and P2U to be moved. I would say, to move those 4
pots by 20 mm. We should see clean elastics that the tracking program
should be able to recognize.
###########
Thu, 05 Jul 2001
If the beam that we have been using had an emmitance of 20 pi mm-mrad the
8*sigma location from beam would be:
8*sigma
pot mm from beam
========== ================
A1U,A1D 8.47 mm
A2U,A2D 6.59 mm
P1U,P1D 15.18 mm
P2U,P2D 12.68 mm
We have been able to move P1U and P1D to 15 mm from beam with just a small
fraction of losses( this is 30mm from home position). So we might have moved
the P1 pots already to 8*sigma from beam.
However the A pots have been moved only to 15 mm from beam withouth
observing much increase in our single rates and not increase at all in halo
rates, the reason is beacuse we moved them only to about 17 to 18 sigma from
beam. we still need to go other 7 to 8 mm further in.
SO I propose to change the software limit of 30 mm from home we have right
now to the values that correspond to about 6*sigma from the beam at an
emmitance of 20 pi, these values, measured from home position are:
SOFTWARE LIMIT FOR P1 POTS MUST BE 33 mm from home.
SOFTWARE LIMIT FOR P2 POTS MUST BE 35 mm from home.
SOFTWARE LIMIT FOR A1 POTS MUST BE 38 mm from home.
SOFTWARE LIMIT FOR A2 POTS MUST BE 40 mm form home.
Let me know if you agree with those limits, so Lionel can modify
the software limits for next time we can move pots at end of store.
############
Tue, 10 Jul 2001
There is cross talk due to patch pannel and to ribbon cable itself.
We did some tests today with Petra to understand things better.
CROSS TALK IN THE RIBBON CABLE
==============================
We put a signal to pin 8 in a ribbon cable in the SCR and checked
adjacent pins at other end of ribbon cable, what we find is:
pin5 pin6 pin7 pin8 pin9 pin10 pin11
signal amplitud 3.4mv 4.9mv 30mv 4.5V 30mv 4mv 2.7mv
3.0mv 3.0mv 13mv 2.0V 14mv 4mv 3.0mv
3.0mv 3.0mv 8.4mv 1.0V 8mv 3mv 3mv
====
^
CROSS TALK THROUGH PATCH PANNELS
================================
We put a signal at bnc 10 in patch pannel at collision hall and check
outputs of patch pannel at SCR. We observe two effects from collison hall,
one is the amplitude of signal due to cross talk, the other is a dc offset
that we checked that it is independent of the pulser signal.
pin7 pin8 pin9 pin10 pin11 pin13 pin14 pin16
D2 A1D P1D A1U P2D A2U A2D
signal 7mv 9mv 19mv 1.5V 15mv 14mv 7mv 6mv
DC_offset 30mv 15mv 8mv 10mv 8mv 8mv 8mv
The conclusion is:
1) We should use every other two channels in the ribbon cable to
send L0 signals, that means use two ribbon cables to send the 10 L0
signals.
2) Have two patch pannels in the collision hall, one for each ribbon
cable.
3) Modify the patch pannel in the SCR to get signals from the two ribbon
cables.
4) ground the channels in the ribbon cable we are not using. This should
be done at both sides: the SCR and the collision hall.
Jorge Molina also Talked to Mike Utes, I think he agrees with our approach
mentioned above, Jorge will write and independent mail with Mr Ute's
suggestions.
Any comment or new idea is welcome,
VICTOR BODYAGIN
Fri, 13 Jul 2001
Let me inform you were we are with VME daq.
After Dave Slimmer installed link on the backplane of the A2 controller it
started to recieve triggers. Using branch terminator did not help. Read
Event finction still sees only 45 scalers out of 48 and 15 TDCs out of 16.
After the discussion with Dave I decided to try the other mode of redout
scalers and TDCs. Probably it's a feature of this type of controller and
there is nothing wrong with. I had a talk with Petr Neustroev. He is the guy
who designed and built probably one half of all A2 controllers in the world.
He came to SCR and told that with such length of a cable we can work without
a branch terminator. So, I'm in a preparation of a new event readout. I'll
try a new version no later then tomorrow morning or probably this night.
################
Sun, 15 Jul 2001
I believe it makes sense to try it on different trigger rates and after
that on real data. I tested full scale version for few hours and it works
smoothly, i.e.:
client rundaq - VME running process does its job OK.
logdaq - child process shows its HISTOSCOPE pictures.
server daq - accpets data from rundaq and controls the other
processes as I expected.
Which is also good is that Pots Moving program was running in the
same time with rundaq. No interference was observed exepting usage
of CPU time.
Two concerns:
1. Histoscope completes its job after getting End of Run message not
in the way I would like it. Once, Carlos told me that this small
bug is also a feature of the good old daq. I'm going to look into
this problem later.
2. If FERA ADCs too fast for A2 we can get wrong data.
So far I did not try to write to disk, though I'm positive it will
be OK.
Further plans:
1. Revise PEDS for new version.
2. Coordinate further activity with Mike Martens on the subject of
sending our scalers to MCR.
________________________________________________________________________________
MIKE STRANG
Thu, 28 Jun 2001
The HV controls have settled into the final configuration. The HV Control
Sheet in the FPD CR has been changed to reflect this.
The biggest change is that you now use:
setup -t onl_apps
/online/config/fpd/hv/fpd.hv &
to turn on the HV control program. When it is turned on, you will see a
sheet with two tabs. One tab is labeled L0 and the other is labeled MAPMT.
You can choose the tab that reflects which tubes you are going to turn on.
It is very important to NEVER turn on the MAPMTs unless you have verified
the corresponding rates at the L0 tube.
Also, for now never turn on the tubes above 90% unless we are testing the
tubes.
Attached is the control sheet that is also in the control room.
http://www-d0.fnal.gov/fpd/FPD_News/Standalone HV Control.doc.
################
Sun, 08 Jul 2001
Here are the access plans with (hopefully) the last modifications.
Collision Hall
--------------
Replace patch panel near PW08 (1 hr, Michael Strang):
Take patch panel and signal generator into collision hall along
with plug map (remember A1I is A1D). Plug ribbon cable into flat
cable and send signal through BNC connector and verify that it is
received in the SCR. If everything works, plug in L0 cables from
tunnel into correct BNC location.
Tunnel (in order of importance)
-------------------------------
Before entering the tunnel, those going into the tunnel and those
remaining in the control room need to read the HV safety
documentation. A note must be placed at the computer in the SCR to
not turn on HV to any castle unless notified by those in the tunnel.
Copies of the HV safety documentation will be taken down into the
tunnel and placed near each castle.
Items needed before access:
We need to secure the Gauss Meters (Prep? D0?), find which cables
can carry the signals, and get the 16x2 pin extenders. We need to
take screw driver, multi-meter (to test HV splitter integrity),
tape to mark cables, marker. Need to make copies of the HV safety
instructions.
Also, all L0 HV cables will be marked with tape to distinguish from
MAPMT HV cables. Whenever a pot is connected to the HV and signals
it must be verified by at least two people.
If HV is to be turned on for a test, no other work can be performed on
other castles until the HV has once again been disabled.
Fix P2U (30 min if simple, 3 hrs min if complicated, Mike S. et al):
See which HV cable is plugged into the PMTs at P2U. If it is
wrong, put the correct cable. If the correct cable is connected,
the cartridge top will need to be opened to verify the HV splitter
is intact. If there is a problem with the HV splitter, the
cartridge top will need to be removed from the tunnel so that it
can be repaired. If the HV splitter is intact then verify
connections at the amplifier. If everything looks ok, I'm not
sure what else can be done.
Using 16x2 pin extender (I still don't know what it is called),
add these to the new pin tubes to help increase the connectivity
between signal cables and tubes.
Once the cartridge top is reconnected, an initial test is to turn
on the MAPMT HV and verify that current is being drawn by the
tubes. Then the LMB can be used to strobe the MAPMTs to verify
connections.
Fix P1U (30 min if simple, 3 hours min if complicated, Mike S. et al):
Only fix P1U if P2U has been fixed and tested.
Open cartridge top and use 16x2 extender to make better signal to
cable connections. Look at tube 5 to see if reason for no signals
can be determined. If HV cable has come loose, reattach if able
If pin is broken, cartridge top will need to be removed from the
tunnel for repairs. If HV is connected properly verify connection
cables at the amplifier.
Once the cartridge top is reconnected, an initial test is to turn
on the MAPMT HV and verify that current is being drawn by the
tubes. Then the LMB can be used to strobe the MAPMTs to verify
connections.
Measure Magnetic Field (1-2 hours, Mike S et al.):
Verify which cables in the tunnel can carry signal from a Gauss
Meter to the SCR. Position meters near castles and connect to
proper cables. Take a magnet of known strength and test that the
signal received in the SCR is of the expected value. We will take
measurements over the next few stores to get a feeling for the
field.
We can also get an idea of the effect by measuring the MAPMT
signal using the LMB both with and without magnetic field. The
field signal can be measured at any time there is beam. The
without field signal can be measured during the access or whenever
there is dead time.
Clear Camera View (30 Min, Mike S. et al):
Move some cables so the camera has a clear view of the motors.
Verify with the SCR that the view is cleared. Use a marker to put
markings on the motors to make it clear when they are moving.
Test new 5V circuit (4 hrs?, Lionel et al):
Reconfigure two castles to use the new 5V circuit and verify pot
motion.
Replace RM at dipoles (2 hrs, Lionel et al):
Place RM at dipole location and verify that the dipoles can be
moved using the software.
Calibrate LVDTs (1 hr, Lionel et al):
From the SCR move the castles through the entire range of motion
to properly calibrate the LVDTs and the values that they pass in
to ACNET
################
Fri, 13 Jul 2001
We seem to be making some progress in the CAMAC to VME rundaq program.
Lionel is also working on the new version of the pot motion software.
We have reinstalled the RMI at the dipoles and they will be activated
for the next time we have beam.
We have moved the in and out pots manually to make sure they are not
stuck.
We are able to see LMB signal from P1U but the new HV scheme appears to
have introduced some additional noise at the ~20mV level. Pierrick and I
will investigate this tomorrow.
P2U will not accept voltage above 400V. I went into the tunnel and
plugged P2I into the HV so that I was using a different Slot and Channel
from the HV crate and observed the same problem. We pulled the pot from
the tunnel but could not see anything obviously wrong. There were a few
places that weren't shrink-wrapped like P1 so we wrapped them in
electrical tape and will test again tomorrow morning. We swapped tube #4
(channels 8-16 are noisy even though the connection at the tube appears
good) to slot #2 (mixed plane) so that the problem would have minimal
impact.
We installed the Gaussmeter at P1 using the L0 P1I cable to carry the
signal out. While in the tunnel, carlos moved the meter around various
places and was getting a few gauss even with the magnet off. This
appeared to be perpendicular fields to the tubes however.
So, tomorrow we will keep working on P1U and P2U. If we can't get the
problem solved with P2U we will go back to the old way of soldering
everything together since it has been shown to work so that we can
design something better to split the voltages. All of the tubes have
been moved and there are no more spares we shouldn't have to swap any
more tubes until October anyway.
We will also try to find available machines to modify the cartridge top
for one pseudo-detector to accept the "official" L0 PMT magnetic shield.
If we can't access the equipment we will then wrap one tube with
multiple layers of the mu-metal foil (this will be A1U whichever
solution.) We will then wrap A2U L0 tube with a single thickness of
mu-metal foil and we will wrap P1U with what is left (at least 1
thickness, probably more).
################
Sun, 15 Jul 2001
Today we just verified everything in the tunnel. We used a pulser at
each L0 signal cable and verified that it arrived in the tunnel. We
verified that all L0 tubes draw current and recorded the current being
drawn for each location at -1800V. We observed that A1 appears to draw
about 0.4 mA more current than all other locations and that when the
tubes are first turned on, the current shoots past the steady state
valued and takes about a minute to settle to the steady current. We also
recorded the current being drawn by the MAPMTs at both positions
at -900V.
We installed plane mirrors at P1 and A1 that allow us to vaguely see the
down motor. We will have to wait until we try to move the pots to see if
it actually works. We will most likely not be able to keep the mirrors
in the current locations once we install the in and out pots. We were
unable to effective used the convex mirror and were unable to see the
racks. In order to do something like this, and to better focus on the
mirrors we need to optimize the positions of the cameras in the October
shutdown.
Lionel said that the new pot motion software is working good and Victor
is continuing to work on the VME DAQ and making good progress.
Carlos is going to gather data from the MAPMTs in no magnet conditions
and we will record the data about the PW08 rack tomorrow morning before
they close the collision hall.
________________________________________________________________________________
PETRA
Wed, 27 Jun 2001
After the access (when two MAPMTs at P1U position were replaced) I checked
the response of MAPMTs at P1U and P2U using the LMB.
I turned on the HV of MAPMTs at P1U and then checked the signals channel
by channel.
Results:
http://www-d0.fnal.gov/~krivkova/fpd/mapmt0627.html.
It looks that one MAPMT is not still working and the half of the another one
has a bad contact. (I turned off HV at MAP1U).
Then I started to check the P2U. I turned on the HV at P2U. At -700V I still
saw no pulses when I was doing the random check. (I check P2U again
tomorrow, maybe I did some stupid mistake).
################
Thu, 28 Jun 2001
Today we (Mike S. and me) went to SCR to check MAPMTS again. First we
checked few channels of P1, it had the same behavior as yesterday. Then we
turned on the MA HV at P2U and we saw that no current was drawn (we compared
it with P1U). My guess is that the HV cable is not connected.
I also wrote the LMB manual:
http://www-d0.fnal.gov/~krivkova/fpd/lmb_manual.html.
Please send me your comments.
################
Tue, 10 Jul 2001
n the fpd disk at d0mino:
/prj_root/810/fp_1/
I created the directory:
FPD_data
it has two subdirectories:
commisioning_run
engineering_run
I gave the write permition for this two directories:
FPD_data and FPD_data/commisioning_run
to the fp_write group.
All data files, I copied to the directory FPD_data/commisioning_run, are
read-only.
So anyone, who is going to copy our data from the online machine, should use
this directory (/prj_root/810/fp_1/FPD_data/commisioning_run/) to store them
and make the data read-only.
If you think that I should change anything, please, let me know.
################
Wed, 11 Jul 2001
On Friday I did the following test:
Tevatron magnets were on (later when I took next data they were off) and
there was no beam.
I generated the SLP pulse and the trigger using the pulse generator. I
turned on the P1U MA HV.
I ran ./peds to load pedestals. Then I ran the 'pedestal run' using ./rundaq
(at this time I had the DC offset of the SLP signal set to zero). Then I ran
the 'calibration run' using ./rundaq (at this time the DCoffset was
nonzero). I repeated this for the various values of HV, DCoffset and with
and without the Tevatron magnetic field.
Unfortunately, later on when I analyzed the offline data, I found out that
the daq was not working properly. We (me and Carlos) tested the daq again
later on Friday and yesterday, but we don't know was is wrong - we noticed
that Victor was running logdaq_v1 on d0ol13 since Sunday, but I don't know
if he was running it also on Friday ?
Anyway the result is that some ntuples are filled only with zeros and also
some pedestal data contain only zeros.
Anyway I have few (I hope good) data files - just to demonstrate you now
the idea of the test:
run8 - pedestal run (no DCoffset), HV=-900V, magnets on
run9 - calibration run (DCoffset=+7.7V), HV=-900V, magnets on
in the attachment you can find the ps-file: run8_9.ps
it contains the adc values (in CAMAC only cables of P1U were connected
in the following order:
slot 1 2 3 4 5 6 7
ribbon RC-32-7 RC-32-6 RC-32-5 RC-32-4 RC-32-3 RC-32-2 RC-32-1
n-tuple fera(1, fera(2, fera(3, fera(4, fera(5, fera(6, fera(7,
variable 1-16) 1-16) 1-16) 1-16) 1-16) 1-16) 1-16)
on the plots they are the
http://www-d0.fnal.gov/fpd/FPD_News/run8_9.ps.
run8 data are drawn as the 'striped' histograms
run9 data are drawn as the 'blank' histograms
You can see the clear separation of the signal and the residual pedestal
in some channel. In the other channel, unfortunately, The pedestal overlap
with the data.
The idea of this test is to take the data for the same HV and same DC offset
with and without the magnetic field and compare the amplitudes of the
signals.
Anyway this test can be also used to study the ADC response.
________________________________________________________________________________
CHRISTOPHE
Thu, 21 Jun 2001
D0 Forward Proton Detector (FPD) call-in list.
----------------------------------------------
=================================
| FPD PAGER : (630) 266 2635 |
=================================
Andrew Brandt UTA 817 272-2706
Home (TX) 817 496-9053
Fermilab 630 840-2778
Cell phone 817 723-7268
Mike Martens Fermilab 630 840-8241
Pager 630 314-1734
Home 630 983-8694
Sergio Novaes Cell phone 630 272-1819
Fermilab 630 840-2454
Christophe Royon Fermilab 630 840-8598/6851
+Lionel Peyrichoux Home 630 753-9182
Michael Strang Fermilab 630 840-2691
Home 630 375-0366
Petra Krivkova Fermilab 630 840-8786
Home 630 375-0076
Jorge Molina Fermilab 630 840-6625
Home 630 840-3545
Carlos Avila Fermilab 630 840-5199
Home 630 840-4873
Victor Bodyagin Fermilab 630 840-6833
Lab6 630 840-3377
Tunnel Phone at D1 (P1 pot) 630 840-2334
Tunnel phone at C4 (A1 pot) 630 840-2335
Tunnel phone at C48 (Dip pots) 630 840-3517
D0 control room 630 840-8800
Beams Division control room 630 840-3721
################
Tue, 10 Jul 2001
WORK TO BE DONE DURING THE SHUTDOWN BY ORDER OF PRIORITY:
==========================================================
Up to now, the access in the tunnel is until Friday evening, and into the
collsion hall until Monday. The access into the tunnel might be extended until
Sunday evening as well.
1) Python software: (mainly Lionel)
====================
TESTS BETWEEN WEDNESDAY AND MONDAY
The final version of the code (realased in cvs) should be frozen on Saturday
evening sothat further tests can be done Sunday/Monday
test the new software which should include:
- the treatment of false smoke alarms
- the database (being done now)
- trigger rate from EPICS variables (Victor should give the information for
Lionel)
The tests include the motion of all pots.
Lionel should also write down a mail with what is new in the program and what is
still missing.
The LVDT calibration should also be performed for all pots while moving them.
2) Connections in tunnel:
==========================
Wednesday morning, Lionel is going into the tunnel to connect all the
disconnected pots to be able to move them. The tests of the program start
in the afternoon.
At the same time, Lionel will draw a scheme for the connections of each rack
which will be the reference of our electronic connections in tunnel.
3) Rack monitor status:
========================
Mike will contact PREP to know the status of the RM on Wednesday. If it is not
yet ready, he will try to get a new one on Wednesday from Fritz.
Thursday morning, the new RM can be put back (Lionel, Mike), and the dipoles
can move starting Thursday afternoon.
4) 5 V connections:
====================
Thursday, Lionel and Victor build the 5V cables and connection pannels for P1
and P2, and install it in the tunnel. They can test extensively after this.
5) Changes of PM at P1-P2:
============================
Wednesday morning, Molina and Mike go into the tunnel to remove P1-P2, and
change the bad PMTs. Thursday, the can be reinstalled back in the tunnel and
tested.
6) Test SLP control for LMBs:
==============================
Petra
7) Patch pannel in collsion hall:
=================================
Molina is finishing the patch pannel on Tuesday. It can be installed and tested
over the week-end.
8) Mirrors:
=============
Mike will meet the camera guys.
9) Further tasks if the acees in tunnel is also on Saturday/Sunday:
=====================================================================
Further tests of the IIB problems can be performed in tunnel (Lionel, Mike,
Molina)
Intensive tests of python program / LVDT calibration (Lionel)
Installation of 5V control line on A side (Molina, Mike, Lionel)
________________________________________________________________________________
LIONEL
Mon, 25 Jun 2001
This is the list of our ACNET variables (Pot position in mm):
- C:FPA1UL (A1 UP)
- C:FPA1DL (A1 DOWN)
- C:FPA1IL (A1 IN)
- C:FPA1OL (A1 OUT)
- C:FPA2UL (A2 UP)
- C:FPA2DL (A2 DOWN)
- C:FPA2IL (A2 IN)
- C:FPA2OL (A2 OUT)
- C:FPP1UL (P1 UP)
- C:FPP1DL (P1 DOWN)
- C:FPP1IL (P1 IN)
- C:FPP1OL (P1 OUT)
- C:FPP2UL (P2 UP)
- C:FPP2DL (P2 DOWN)
- C:FPP2IL (P2 IN)
- C:FPP2OL (P2 OUT)
- C:FPD1IL (D1)
- C:FPD2IL (D2)
################
Fri, 29 Jun 2001
You can find in my home page:
http://www-d0.fnal.gov/~peyricli/home.html.
links for Epics and ACNET variables.
In Acnet Variables, you'll find a short description how to access
to the value of the variables.
________________________________________________________________________________
MARIO
Tue, 03 Jul 2001
I added some information to my FPD page at
http://d0br1.lafex.cbpf.br/~mario/fpgg.htm.
One is a mechanical drawing for the AFE to ribbon cable interface at:
http://d0br1.lafex.cbpf.br/~mario/FPD/fpdflexmech.JPG.
Based in this view I am making the layout of the FPD flex cable. The idea is
to bring the flat ribbon cables from the top of rack PW08 into A&S
backplanes. Under these backplanes we have the FPD flex cables that go to
AFE boards.
One flex cable join one Cin::Apse connector to four 16 x 2 pin header
connector with a flat ribbon cable in the other side as in the drawing at
http://d0br1.lafex.cbpf.br/%7Emario/fpdflex.htm.
There are 6 backplanes with 21 header connectors with 16 x 2 pins, plus two
backplane pieces with just one header connector each, that makes the whole
set fit well in the PW08 rack.
Please send me your comments.
MIKE MARTENS
Thu, 05 Jul 2001
Some notes on Flying wires.
Fly times:
During a store the flying wires are flown once an hour. The wires are
triggered to fly on TCLK event $CF which is triggered 4 seconds after TCLK
event $DB. The $DB event is set to automatically trigger every hour. You can
determine the time since the last $DB event was issued by using the D33
ACNET console application.
>From D33: On the left hand side near the bottom interrupt on the green
"Normal" and select "Time since event". Then interupt on the DB in the array
of clock events. The time since the last $DB event was issued will be
displayed.
On Emittances:
The flying wire emittances for each proton and each pbar bunch are saved in
ACNET array devices. The devices are T:FWVEMI[0-72] for vertical emittance
and T:FWHEMI[0-72] for horizontal emittance. The array elements with indices
0-35 are for proton bunches 1 through 36 and array elements with indices
36-71 are for pbar bunches 1 through 36.
As far as I can tell, the Tev group is not yet data logging these devices so
I don't have a history of wire emittances. However, a spot check this
afternoon shows that the proton 95% normalized vertical emittances are
between 38 and 43 pi mm-mrad and the pbar 95% normalized emittances are
between 29 and 36 pi mm-mrad depending on which bunch. This was about 44
hours into the store. (If you are interested in how long the store has
lasted you can look at ACNET device T:STRDUR.) I also believe that the pbars
are bigger at the start of a store and may be as large as 50 pi mm-mrad.
The horizontal emittances that are reported are probably bullshit since they
read between 0 and 12 pi mm-mrad. This is due to the fact that the
horizontal beam sigma is a function of both transverse emittance and
momentum spread and small errors in beam sigma measurements in the wires can
translate into big errors in emittance and momentum spread. The best
estimate for horizontal emittance is probably the same as the vertical
emittance since it is believed that coupling effects in the Tevatron results
in equal vertical and horizontal emittances.
Since these devices haven't been datalogged I tried to add them to the
datalogger but for some reason I cannot log all of the bunches. Instead I
have added the first 12 proton bunch vertical emittances and the first 12
pbar vertical emittances. This should at least be a start for logging these
devices.
CARLOS
Mon, 16 Jul 2001
We need to change the software limits so we can experiment how close we
can get the pots to the beam.
Since the beam emmitance is about 30 pi, right now, we want to have a pot
position limit at 6*sigma from beam, the numbers are shown below. I
propose to set up the software limits shown in the last column, so next
week once we get beam we can try to see how far in we can move the pots.
rounded off
software limit
From home 6*sigma from beam from home
=========== ================= ==============
D1 38.862 mm 6.138 mm 39 mm
D2 39.432 mm 5.568 mm 39 mm
A1 36.792 mm 8.208 mm 37 mm
A2 38.626 mm 6.374 mm 39 mm
P1 30.306 mm 14.694 mm 30 mm
P2 32.718 mm 12.282 mm 33 mm
My proposal is that we move the pots in as long as single rates are
lower than 50 kHZ and the antiproton and proton halo rates are lower
than a factor of 1.5 from the value of those rates at home position.
Once we reach either high single rates or 50% losses we retarct the pots.
Objections and suggestions are welcome,
To get sigma of beam from emmitance one can use the following equation:
sigma^2 = emmitance*beta/(2*pi*p*ln(1-f))
beta= beta value at detector location for normal run operation.
f= fraction of normalizaed emmitance at 95% = 0.95
p=beam momentum=900 GeV
Tue 17/7/2001
Let me explain how I got them:
I used the collider run II lattice parameters that Mike Martens have
in his home page:
http://www-ap.fnal.gov/~martens/mad/tevatron/run2/lattice.html
I used the last entrance that is for beta=0.35 and beam colliding at B0
and D0.
from that tables I could extract the beta values for each detector
position and used the following equation that relates emmitance
and beam width:
sigma^2 = emmitance*beta/(2*pi*p*ln(1-f))
beta= beta value at detector location for normal run operation.
f= fraction of normalizaed emmitance at 95% = 0.95
p=beam momentum=1000 GeV
The sigmas that Mike have in another file for fpd tracking are for 20 pi
emmittance.
The emmittance for the runs we have taken is closer to 30 pi and this
is the one we want to use for the limits.
Going through the numbers again you made me realise that I used 900 GeV
from beam momentun. I should have used 1000 GeV.
So I send now the values corrected which will produce a small change in
the limits proposed before:
rounded off
software limit
From home 6*sigma from beam from home
=========== ================= ==============
D1 39.174 mm 5.826 mm 39 mm
D2 39.714 mm 5.286 mm 40 mm
A1U,A1D 37.218 mm 7.782 mm 37 mm
A2U,A2D 38.946 mm 6.054 mm 39 mm
P1U,P1D 31.062 mm 13.938 mm 31 mm
P2U,P2D 33.348 mm 11.652 mm 33 mm
If I compare the 6*sigma values I got in this table with the ones you sent
me they differ in average by about 1.22 which is equal to
sqrt(30*pi emmit/20*pi emmit) and is what we expect, the sigma of the
beams change as square root of emmittance.
Please let me know if you agree with those last numbers, so we
ask Lionel to modify the software limits.
VICTOR BODYAGIN
Tue 17/7/2001
This night Lionel and me tested simultaneous running of our tasks on
the subject of CPU time usage. In VME crate usage of CPU was around 80%.
d0ol13 CPU usage time varied depending on the value of the
"nanosleep" delay I introduced into camac function polling trigger.
I introduced values: 0 msec
1/4 msec
1 msec
Here are the average values of CPU usage:
at 1/4 msec : daq_v01 ~ 1%
logdaq_v01 at peak ~ 5%
histo ~ 1%
Python ~ 30%
at 1 msec it was too dull to watch daq jobs
Python ~ 30%
at 0 msec daq_vo1 ~ 1.5%
logdaq_v01 less than 6%
histo ~ 1.5%
Python 5 - 18% (we got this snake in the
crate!)
Note, while running physics data collection we can expect
logdaq_v01 and histo will go higher. Nevertheless I do not expect total
time usage by daq processes greater than 20-25%.
So far, I decided to set 1/4 msec sleep.
Now, a problem: when trigger cable was not plugged in properly rundaq
running at 0 msec sleep did not get triggers and was in a state of
permanent poll of the trigger latch, at that time Python job crashed few
times. I quess when it can not get some "EPICS variables" it does not
use some default set in order send a message about it but sirvive!
LIONEL
Wed 18/7/2001
The new version is now available on d0fpd account. The old version
is on python/OLD directory and can be used.
To start the new version, go to the 'python' directory and type
'./FPDGui.py'.
In the menubar on the Main window, 1 button 'Init' has been
added. All pots will go Home if you click on it. Also, they'll go home
automatically when the soft will start.
I've changed the Expert mode. You need to unlock the pot that you
want to move. Then select the speed and the position. Click on Run, and
OK. The button will change to 'CHECK', to 'MOTION' to 'CHECK' to
'Lock'. When Lock appears, then you can choose a new position. But, with
the 'MOTION' button, yiu can also 'STOP' the motor. You need to Change the
button to 'STOP' then click on 'STOP' button.
MIKE STRANG
Thu 19/7/2001
Jorge and I were using your new pot motion software to test A2U after we
replaced the coupler and it doesn't appear to be very stable. Here is
what we observed:
When we first turned it on, it told us that P2U was not in the home
position even though the LVDT reported a position of 0.0. It then tried,
unsuccessfully to send the pot home.
We turned the new code off and went to the old code. It also said that
P2U was not in the home position but as soon as we sent the command to
go home in expert mode, it switched to being in the home position
without seeming to actually use the motor. Maybe if a pot is almost home
but not quite, the method you are using to send it home does not work
effectively? I'm not sure but you might want to look at this.
We were successfully able to move A2U, send the stop command and send it
back home using the old program.
We then tried the new program again to see if the previous problem
reappeared. It didn't come back, but a new one did appear. A2D started
to move itself in without our having even opened the expert mode. I
opened the expert mode and issued a stop command, but it was not
accepted. I had to use the old software to stop the pot and send it back
home. When it got back home, the LVDT said it was at 0.6 even though the
home bit was turned on. I think we need to better calibrate the Lads.
Over lunch, Jorge and I moved all pots to the home position and left the
new software running to see what would happen. When we returned, A2D was
at 10.5mm without having expert mode opened. Also, A2U is now showing a
mechanical failure even though it was working before.
I notice that the new software does not seem to accept the home command.
When I unlock a pot that is inserted and issue the home command it does
not confirm that I have sent the command and the pot does not go home.
Finally, I've turned on the old program and I'm going to leave it
running for a while to see if the same phantom motions of the pots occur
to see if there is an IIB problem.
CHRISTOPHE
Fri 20/7/2001
Mike and I went into the tunnel, and we noticed that A2 up was completely
stuck in the IN position. The problem is that it seems that the pot went
there by itself (it should not go completely IN by the software).
we had to unplug the motor, move it a bit by hand, and then move the pot
out using the switch on the IB. We also switched on and off the RM to see
if it could be a RM problem. We left the motors on, and the pots at HOME
to see if they are moving by themselves again.
Mike Martens
Thursday, July 26, 2001
Pot survey data
These are the surveyed positions of the four quadrupole spectrometers
and the two dipole pots.
This is from the V-star data of the pots while the pots were at the IN
position after the re-work. The pots were aligned in the tunnel to place
the Tev center line vertically in the center of the UP and DOWN
detectors when they are in their IN position, and to place the Tev
center line horizontally in the center of the IN and OUT detectors when
they are in their IN position.
Positive values for the y coordinate are vertically upward, and positive
values for the x coordinate are toward the inside (i.e. toward the
tunnel aisle.) The coordinates are at the inside surface of the pot tips
and at the center of the pot tips. This does not include the thickness
of the thin metal foil. The edge of the detectors will be offset from
these coordinates by approximately the thickness of the thin foil
(200um)
assuming that the detector is flush against the thin foil.
The numbers that I trust the most are the distances between opposing
pots. The lateral position of the pot has a larger uncertainty due to
the limitations of the V-star system and the geometry of pot wells. I do
not have any error estimates for these numbers. The tunnel alignment is
accurate to no better than 10 mils (0.25 mm).
MATINADA, D11 upstream, P1 pot
X Y
P1U +0.39 mm +3.71 mm
P1D -0.77 mm -3.71 mm
P1I +4.03 mm -0.18 mm
PIO -4.03 mm +0.75 mm
LARANJEIRAS, D11 downstream, P2 pot
X Y
P2U +0.39 mm +3.14 mm
P2D -0.14 mm -3.14 mm
P2I +4.55 mm -0.47 mm
P2O -4.55 mm -2.21 mm
PACOVAL, C49 downstream, A1 pot
X Y
A1U -0.41 mm +4.10 mm
A1D -1.15 mm -4.10 mm
A1I +3.70 mm +1.92 mm
A1O -3.70 mm +0.52 mm
GENTIL, C49 upstream, A2 pot
X Y
A2U +0.41 mm +4.36 mm
A2D -0.45 mm -4.36 mm
A2I +3.45 mm +1.12 mm
A2O -3.45 mm +0.59 mm
GUAJARA, C48 upstream, D2 pot
X Y
D2I +3.00 mm +0.00 mm
MACACAO, C48 downstream, D1 pot
X Y
D1I +3.00 mm +0.00 mm
Lionel
Friday, July 27, 2001
New Software
The new pot motion software is now available on d0fpd account. To
start it, go to python directory and type ./FPDGui.py. Also, a new small
soft ./FPDStop.py will run all the time. This soft has 2 buttons
('EMERGENCY BUTTON' and 'Exit'). It sends the command 0 (no motion,
position 0) to the output of the Rack monitor when the EMERGENCY BUTTON is
activated. It can be used if the pot is moving by itself or doesn't stop.
Victor Bodyagin
Friday, July 27, 2001
EPICS and ACNET
Here is our list of names for our EPICS and ACNET variables:
E P I C S A C N E T
FPD_CAMAC_0308_SCR/D1IS C:FPD1IS
FPD_CAMAC_0309_SCR/D2IS C:FPD2IS
FPD_CAMAC_0306_SCR/A2US C:FPA2US
FPD_CAMAC_0307_SCR/A2DS C:FPA2DS
FPD_CAMAC_0600_SCR/A2IS C:FPA2IS *
FPD_CAMAC_0601_SCR/A2OS C:FPA2OS *
FPD_CAMAC_0304_SCR/A1US C:FPA1US
FPD_CAMAC_0305_SCR/A1DS C:FPA1DS
FPD_CAMAC_0602_SCR/A1IS C:FPA1IS *
FPD_CAMAC_0603_SCR/A1OS C:FPA1OS *
FPD_CAMAC_0300_SCR/P1US C:FPP1US
FPD_CAMAC_0301_SCR/P1DS C:FPP1DS
FPD_CAMAC_0604_SCR/P1IS C:FPP1IS *
FPD_CAMAC_0605_SCR/P1OS C:FPP1OS *
FPD_CAMAC_0302_SCR/P2US C:FPP2US
FPD_CAMAC_0303_SCR/P2DS C:FPP2DS
FPD_CAMAC_0606_SCR/P2IS C:FPP2IS *
FPD_CAMAC_0607_SCR/P2OS C:FPP2OS *
* NOTE: variables marked with "*" do not exist physically at present,
since the detectors will be installed in October.
Pierrick Hanlet Saturday, July 28, 2001 Amp/Shaper to AFE interface I've been looking into how we will connect to the AFE boards. The present plan is to mount the AFE cassettes upside down and feed the cables from the top of the rack to a transition patch panel. The transition patch panel would accept the present connectors and feed through to modified flex cables on the other side, which would then connect to the AFE boards. After having spent a couple of hours with George Ginther, I've come to several conclusions and recommendations: 1) I don't believe that we can safely mount the cassettes upside down. The devices that are used for the Synaps connectors and the rails for mounting the cards are spring loaded, with the springs working in conjuction with gravity. It might be possible to mount them upside down, but I suspect that we would find ourselves modifying and tweeking to get them to work. I propose that we use the system as designed. The backplanes are only 1/2 the width of the rack, so we can mount the AFE cards/backplane on the RHS of the rack and run the cables from the top of the rack to the bottom on the LHS. We would then make a transition patch panel that would mount underneath the cassettes. 2) There remains some questions as to how we connect to the flex cables and what flex cables we will use. We want to keep the Synaps connectors at the AFE end of the cable, since we don't want to modify the boards. That leaves us with the transition patch panel end of the cable. The present proposal is to split a flex cable into 4 16-pin cables onto which we can put a standard connector and then plug them into bulkhead connectors on the patch panel. We need 32 of these 64-conductor flex cables. If we use existing flex cables, we can get some of George's rejects. These pass continuity tests, but when visually inspected with a magnifying glass, they have some imperfections such as a damaged trace or a scratch. Should we use these, we would not easily modify these cables to split into 4 16-conductor parts. Hence, the patch panel would require a PCB which could accept the Synaps connector pad structures, allow us to mount the connectors on the board, and fan out the traces to our standard connectors. The nice thing about this board is that we could put whatever passive components we need to make our signals match the required inputs to the SIFT chips; i.e. <= 600fC in a 90 ns window. Another possibility is to purchase new flex cables that are not low mass, and would also allow us to split the cable into the 4 16-conductor cables for our patch panel. The problems here are: we don't know whether or not we can get such cables, and we won't have as convenient a means of adding capicitors to our cables. The first of these problems concerns me the most as I don't see how one can make a cable that on one end is thin enough to use the Synaps connector and on the other hand is thick enough to use the relatively bulky Amphenol connectors; I would be concerned about manhandling these and loosing the connections at the AFE boards. 3) For either of the scenarios, I am concerned about noise pickup. The tracking system has part of their flex cables in a cryostat and the rest behind some shielding which serves as a Faraday cage. We will not have this, and should we use the existing flex cables, we have ~2 feet of cables exposed on the platform. Mike Matulik assumes that we will purchase some shorter flex cables which should help. I propose that we plan to put a screen, or some sort of Faraday cage from below the transition patch panel and around the incoming cables up to the cassettes where we can ground it. __________________________________________________________________________ Carlos Avila Monday, July 30, 2001 Status of L0 trigger counters We had beam this morning. Checking the L0 pmts I found the following things: 1) The gauss meter reading is 0.83 gauss. So magnetic field at the detector location seems to be rather low (at least in the direction we oriented the gauss meter). 2) A2U seems to be connected backwards. I dont know how this could happen we checked with Mike and Jorge the connections by sending a pulse from the tunnel, early at the access. We sould check anyway if it is at the patch panel in the SCR. 3) A2D: signal cable might be disconnected at some place, the scaler rate is always 0 but current drawn is about the same at all other counters. I think is also at the patch pannel in the SCR. 4) D1 and D2 are always giving the exact same scaler reading, like if they were connected together. I believe this also is happenning at the patch panel in SCR. Jorge could you check A2U, A2D, D1 and D2 connections in the patch pannel? 5) A1U and P1D are picking up noise with an amplitude of 50 mv. We have to setup the discriminator threshold to 55 mv. 6) the rates for proton halo and pbar-halo when I was looking at rates were of the order of 4kHZ and 500 HZ. and the rates, in HZ, in our counters were: HV P1U P1D P2U P2D A1U A1D A2U A2D D1 D2 1200 310 0 0 1496 0 0 2 0 0 0 1300 622 2 87 2675 0 0 4 0 0 0 1400 1088 5 418 3650 0 0 7 0 0 0 1500 1679 9 1172 4683 0 0 12 0 0 0 1600 2155 23 2296 5456 1 1 21 0 0 0 1700 2876 62 3323 6082 5 16 48 0 0 0 1800 3092 90 4437 6739 24 33 45 0 0 0 1900 3622 388 5277 7750 95 137 280 0 52 52 2000 4543 683 6049 8955 129 632 851 0 452 452 p-halo(HZ) = 4 kHZ a-halo(HZ) = 0.5 kHZ Raes at the antiproton side were very low because the antipronton intensities were very low. Proton rates seem to be smaller compared to data taken previous to the access. The conclusion is that we do not have a plateaus yet. __________________________________________________________________________ Mike Martens Tuesday, July 31, 2001 BD Shutdown A message from the John Marriner about the October shutdown: As you probably already know, I have asked Bob Mau to coordinate the fall accelerator shutdown. We are planning to start the shutdown on October 8 with a duration that has not yet been determined but is "about" 5 weeks. Our strategy on the shutdown is that we will hold maintenance to a minimum and do limited work on MiniBoone and SY120 (both of which are expecting beam in FY02). The bulk of the resources will go to the Recycler. It appears that we can finish most of the Recycler work provided that we can find 38 techs in addition to those in the Mechanical Support Department. As a consequence: 1. Projects should expect that Mechanical Support techs will be unavailable starting October 8 until the end of the shut down. I expect that support will be available until October 8, but I can't guarantee that there won't be interuptions before then to prepare for the shut down. I appreciate that this policy is very harmful to progress on our projects, but there is no reasonable alternative. 2. Bob Mau will be contacting department heads to discuss whether some of their "captive" technicians could work on the Recycler. I request that department heads cooperate with Bob so that we can understand the shut down work load and develop a plan that utilizes the existing manpower on the highest priority jobs. 3. We have requested additional manpower from other divisions in the lab and from outside users. We believe we will get some help but so far we don't know how much. If you have shutdown jobs that require help from Mechanical Support or require significant coordination, please make sure that Bob Mau knows about them and that they get on Dave Augustine's list. Regards, John __________________________________________________________________________ Christophe Monday, August 06, 2001 Friday summary We moved the pots we already moved before last Friday to measure the rates. A1 UP does not show any rate at all. We also observe huge differences in the rates between two consecutive pots (see the logbook in the SCR for more details) which seems strange. They are probably due to differences in the L0 PMTs / pseudo-detectors. The pots moved without any problems at all. One pot was moved at a time, left at 25 mm, and then the next one was moved. Molina is now checking the patch pannel, and repairing it during the access to the collsion hall. Unfortunaltely no access is foreseen to the tunnel yet... Some problems occured last week and provoked a crash of the power PC. (It happened almost all nights). The reason for this is still unclear. It might be due to problems of compatibility between Victor's program and the python program. To test this further, we should get the python program running tonight without Victor's program and see what happens... __________________________________________________________________________ Christophe Tuesday, August 07, 2001 Power PC crash Once again the power PC crashed this night, and the only program running was the python software. We clearly have a problem with the python code now, which is not clear at present. It might be due to the fact that there are now too many communications between the 1553s and the power PC according to Victor. This is clearly a high priority for today activities at the same time as the tunnel checks of the L0 PMTs. Lionel and Victor will investigate this problem further in the SCR today and should have the priority to use d0ol13. Our system is not reliable now. One test could be to increase by hand the communications and check if the crash happens more often... Please let me know in case of problems. __________________________________________________________________________ Victor Bodyagin Tuesday, August 07, 2001 Power PC crash Let me deliver you a few comments on the crashes of our MVME-162 (Power PC). I was watching these crashes last week. I think it is not the problem of PC itself, i.e. CPU and its memory. The way I could restart all system: PC, MIL1553 controller and CBD8210 after such a crash makes me think so. When I was debugging my DAQ code I crashed PC for a few times. In these cases to restart everything it was enough to push "reset" button on the front panel of the MVME-162. Last week it was different. To reset PC was not enough. The only thing which helped was the cycling of the VME crate power. It means that either MIL1553 or CBD8210 or both of them were preventing PC from coming back. At the moment of crash CBD8210 was idling, though it could cause some spurious interrupt (I do not know which and why) which could not be handled by vworks properly. Nevertheless I think it is unlikely. It is more likely that it was some status of MIL1553 which caused the crash. So far I do not know exactly how vxworks kernel is being downloaded while resetting. But if it is being done through MIL1553 then it can explain a lot. In this case being in an abused state MIL1553 cannot download vxworks kernel properly and PC can not come back. How MIL1553 can get into such a state and why seing that PC dies instead of proper handling of such a situation? The only thing we know about it that at this time it was communicating with the rack monitors in the tunnel. As for handling of such a situation I think we have to talk with Sten Hansen (he knows MIL1553 very well) and Fritz Bartlett (he knows evrything about vxworks, downloading and probably handling). __________________________________________________________________________ From: Dmitri_Denisov Tuesday, August 07, 2001 Long term Tevatron plans Dear Colleagues! Below are long term Tevatron plans as discussed during today meeting between Beams Division, D0, and CDF: 1. During August 2001 Tevatron will have 6 days of colliding beams and one day of accelerator studies per week. By the end of August 2001 typical Tevatron luminosity is expected to be about 1.E31. Total expected integrated luminosity before October shutdown is ~12pb-1. 2. October shutdown plans: a) shutdown starts on October 8th; b) shutdown duration is 6 weeks driven by recycler work. There are no more planned shutdowns before July 1st 2002. There will be no 2001 Christmas shutdown at Fermilab. 3. Starting early 2002 typical Tevatron luminosity is planned to be in 5.E31 range. Total integrated luminosity delivered by the end of June 2002 is estimated at 200pb-1. 4. 132ns bunch spacing operation: a) Beams Division is not planning to actively commission 132ns bunch spacing operation untill luminosity of 1.E32 is reached; b) according to estimates Beams Division will start 132ns commissioning (including crossing angle operation) by late summer 2002 with one 132ns store every 1-2 weeks; c) starting early 2003 Tevatron is planned to run full time in 132ns bunch spacing mode. Dmitri. __________________________________________________________________________ Pierrick Hanlet Wednesday, August 08, 2001 Yesterday's activities Yesterday, George and I worked on fixing the bad trigger channels in A1U and A1D. To do so, we went into the tunnel. George quickly found the problem with A1D; due to a short A1D cable, we have been using the A1I to carry the signal to the SCR, and this had been swapped with the A1D cable. He fixed that and now it works well. At -2kV, the tube shows a very large dark noise signal, tens of mV. It is also a very noisy tube. Note that this is NOT pickup, as we looked at the signal at the PMT itself. For A1U, we still have problems. After a couple of trips into the tunnel, we found that we have a quiet PMT when looking at it directly; the dark signal is ~6mV. A problem that we have is that the holes in the cover to the castle do not line up well. When we force it enough, i.e. twist it, to line up the holes, we get a very large pickup as seen on the scope. We were able to reproduce this with the cover off simply by pushing the L-bracket on which the BNC connector is mounted. Being limited in time as set by the accelerator, we reinstalled the cover and managed to untwist the BNC connector enough to not cause the noise; however, I don't consider this fixed for there is a possibility that one bumping against it might recreate the noise. We need to fix this in the next access. We had also checked the signals for both A1U and A1D at the patch panel in the collision hall. There we saw both signals. However, after George and I came out of the late afternoon tunnel access, I checked the signals in the SCR and do not see the A1U signal. Hence, we have a problem at the CH patch panel. A1D is fine, though still noisy. Since we have had so much trouble with the patch panel, I propose that we make up a group of 10 RG58-U cables of 48m, label them, bundle them, and test them. The bundle can then be put into the cable tray during the next access. My understanding is that the LMR cables from the tunnel are 75 ohm; so we can put 50 ohm terminators on the patch panel. Though the 48m of RG58-U may degrade the signals a bit, I do believe that reliable signals are more important than pristine signals. Petra told George and I that the RG58-U cable is slower than the present ribbon cable that we are using and that The only question is can we afford the additional delay in the trigger. From the numbers that she gave us, the RG58-U signal would be delayed by ~10ns if we cut them to 48m. Presently, the cables are 97m, and introduce an 85ns delay. Another problem with the present patch panel is that we have a source for noise pick-up in the collision hall. Making an all coax run, should help us here. We have had several problems with the patch panel. I think that it would behoove us to have reliable signals into the SCR. Pulling out the existing cables, cutting them, putting on new connectors, labeling, testing, rebundling, and reinstalling them would be less than 1 day's work.
Pierrick Thursday, August 16, 2001 Summary of Amp/Shaper to AFE connection I realize that I've been giving piecemeal information regarding making the connection from our detectors via the Amp/Shaper cards to the AFEs. Andrew has asked me to give some dimensions; so I've expanded on this idea to give what I think is the complete picture to making the link. I've attached a short document: http://www-d0.fnal.gov/fpd/FPD_News/afemap.ps. with some rough sketches from which our engineers might work. We need to settle these issues soon, so as not to have this as a bottleneck come October. __________________________________________________________________________ Andrew Thursday, August 16, 2001 Pot motion Christophe tells me that one pot moved out without being commanded to near the end of the run. This was not detected by Lionel's program (or anyone in the control room). We need for it to be detected before we run again, as it could just as easily have gone in. We will be discussing with Lionel why this wasn't detected and how to make sure it is in the future. There was supposed to be a check of LVDT value--perhaps this was not active in the current version-we are investigating. Solutions: 1) more debugging of boards and RM in SCR to fully address the cause of problem (timescale uncertain--Lionel only has less than two weeks left) Since the solution is still a ways off, we have to be absolutely bullet proof in the reaction to the problem. a) need a window to pop up whenever a pot is in motion--since we only move pots one at a time for now, a second window popping up would certainly be noticed by shifter running the program this could be activated by checking the Run bit. Even better would be alarm that said run bit is set but noone asked for it to be. b) there is a software limit that doesn't allow you to choose an LVDT value that is too close, but in this case that doesn't help since noone chose the value. Need an alarm and a stop command in the case where a pot moves in past this value. c) need an alarm if the LVDT value is different from the selected value--we thought we had this already, needs to be tested when there is no beam, by turning off LVDT for example d) need rates of all in pots to be displayed on screen so that a change in these can be notified--also could have this change color if a rate >50khz is observed and another color if >100khz e) need one shifter to be fully focused on pot motion and rates while 5V is active Other suggestions welcome. __________________________________________________________________________ Jorge Molina Thursday, August 16, 2001 We moved the pots today to study the rates and elastic hits. We moved all the A side pots and only P1U and P2U on the P side. We saw that A1U rates didn't change when we put inside the beam, and also when we reach 27.5 mm it retracted by itself to home position. Then we proceed to move the remaining A pots until 30mm and we observed that the rates in the respectives pots changed without affect the beam rates given by the luminosity monitor. Then we moved P1U until 25mm (with all the A pots at 30mm inside) without problem, but when we were moving the P2U (from 20 to 25mm) the A2U pot suddenly moved by itself going to home position, without obeying the stop command. Then we decided to continue to measure elastic rates with the PU + AD pots. We did with the A1D and A2D at 30mm and P1U and P2U at 25mm without any new problem. After that we decided to finish without move the PD pots, all the pots were retracted to home position and we turn off the HV and the 5V line as well. __________________________________________________________________________ Lionel Thursday, August 16, 2001 Just for information: - A1 UP was going home because the direction was wrong. We asked it to go in IN direction (27.5 mm) and the direction was OUT. So the command 'Go Home' was activated. This is not a software problem but a IIB problem. - the A2 UP motion was maybe produced by a fake smoke alarm. When a smoke is detected in one rack, the command 'Go Home' is sent to the pots. We didn't see the Smoke Alarm Window because this window is destroyed if the alarm disappeares, and no message in the logbook (I'll modify the software). We saw immediatly becasue the color of the scaler became green (Motion) when the pot was moving (blue when the pot is stopped) and the LVDT was decreasing. It moved at low speed until it reached the Home switch. I've tried to stop it but it was not possible because the motion was coming from the software (if the problem was coming from the fake smoke alarm) and it was the 'Go Home' command (but I can modify this). Friday, August 17, 2001 The smoke alarm sends, only the pots in the same rack, at home position. For instance, only the pots at A2 will be send at home when the smoke alarm at A2 will be activated. I checked and A2 Down was also In when A2 Up was coming back. So, it seems that it was not a fake smoke alarm. I saw also that A2 UP started to run when P1 UP stopped, but I think it's only a coincidence. __________________________________________________________________________ Victor Bodyagin Friday, August 17, 2001 MIL1553 test is over Roman Pots program was running successfully for more than 24 hours without crashing MVME-162 and MIL1553. A few minutes ago I had a talk with Vladimir. We agreed to consider the test to be over. It is very likely that the crashes were caused by some hardware bug in that particular MIL1553 controller we used before. So, I'm going to proceed with DAQ. __________________________________________________________________________ Michael Strang Friday, August 17, 2001 Run results After a phone conversation with Andrew this morning, Lionel and myself inserted A1U into the beam at the end of the store to see how close we could get before we begin to impact halo losses. On the P side, we impact pbar losses, on the A side we would expect to impact p losses. We have inserted to 30mm on the A side with no problems and from Carlos prediction, for A1 we would expect the 6 sigma value to be: ~ 36.5 mm. At 12:15 we began to insert A1U. We were going to insert A1D, but this is the hot tube and was already showing large (20 kHz rates). Also, A1U didn't show any change in rates during the move yesterday and I wanted to see if the rate would start to change beyond the previous 30mm limit. At 12:15 we began to insert the pots after notifying the shift captain who notified MCR and CDF. Victor's rundaq was running on the PC at this time. We were able to get rates from the program in the pot motion program (therefore, something is making it through to EPICS) but the values were strange. Victor has an idea as to why this might be. After moving the pots to 20mm, we performed this test and then the rundaq program was terminated. Somewhere between 34 and 35 mm we begin to affect the p losses at D0 as can be seen in the attached plot. We tried to move to 36 mm, we observed a large spike in the losses that came back down while the pot was at 36mm. Lionel feels that the effects of the losses are different when a pot is moving compared to when the pot is in position. I'm not sure that this spike is correlated with the pot motion. It appeared to momentarily effect CDF as well but when we impact rates, in the past it has been a steady increase and not a sudden spike. We retracted the pot after this since we were not positive what happened and the store was soon to be stopped. We didn't notice any dramatic increase in the rates from A1U so I feel that both tubes at A1 should be replaced with tubes that have been tested at Lab 6 during the next tunnel access. The pot was removed shortly after 12:40 and everyone notified. We did not notice any other strange behavior with other pots moving. __________________________________________________________________________ Mike Martens Friday, August 17, 2001 I took a look at the figure Mike provided and looked at the datalogger of this same data on a finer scale. My opinion is that we found the halo when C:FPA1UL reached 34.5 mm. At this point the D0 proton rates clearly increased from 23.8 to 25.7 kHz. During the next pot motion step, when C:FPA1UL goes from 34.5 to 36.1 mm, there is a noticeable increase in both the D0 and B0 proton halo rates. This suggests that we cut into the halo and the rates increased for a short time until we scraped away that portion of the halo. A 1.5 mm step is a fairly large one (although not unreasonable this far from the beam) and I believe we were scraping away halo for a while. After the pot was in at 36.1 mm for a while the D0 proton halo rates were around 33kHZ. Finally when the pots were withdrawn again the D0 proton halo rates dropped back to 23.8kHz. As we get nearer the beam I think we will want to make smaller steps (like 0.5mm or even 0.1mm) so that we can keep an eye on halo rates. I checked at the flying wires were flown every hour at 48 minutes after the hour so the spike in losses were not related to flying wires. The behavior of losses increasing suddenly and then decaying away while moving objects near the beam is something we see when moving collimators. This was a good experiment and I think it is worth doing this again with all of the pots so that we can find the edge of the beam.
Lionel Peyrichoux
Tuesday, August 21, 2001
This is the result of the tests during the last shutdown (Friday,
August 17t).
TEST #1 : Pots motion
-------
Move all pots one by one : home --> 10mm --> 25mm --> 32.5mm --> IN -->
27mm --> home
Speed : High
Pots : D1, D2, A1U, A1D, A2U, A2D, P1U, P1D, P2U and P2D
D1 : In position at 39.66 mm, motion OK
D2 : In position at 36.2 mm, motion OK
A1U : In position at 40.11 mm, motion OK
A1D : In position at 39.36 mm, motion OK, mechanical failure close to
In position (39.2 mm)
A2U : In position at 37.6 mm, motion OK
A2D : In position at 37.45 mm, motion OK
P1U : In position at 39.77 mm, motion OK
P1D : In position at 40.30 mm, motion OK
P2U : In position at 39.40 mm, motion OK
P2D : In position at 37.77 mm, motion OK
The 'In position' is the LVDT value.
TEST #2: LVDT Calibration
-------
All pots at the same time for LVDT calibration : home --> 10mm -->
20mm --> 30mm --> home.
Speed : High
Pots : D1, D2, A1U, A1D, A2U, A2D, P1U, P1D, P2U and P2D
I stopped the motors by using the individual STOP command, and general
STOP command. All pots stopped correctly.
No problem during the motions.
TEST #3: fake alarms
--------
All pots at 10mm
I tried to generate some fake alarms by increasing the CPU of the power
PC. Only one fake alarm was detected (Smoke detector) at P1 (P1U and P1D
went back at home).
TEST #4: Emergency Line
--------
D1, A1U, A2U, P1U and P2U at 5 mm
Turn ON the emergency line. The software detected the motion and
then popped up the window 'BAD STATUS - MOTION DETECTED'. This is OK
because the software didn't know when the Emergency line is turned ON.
Tuesday, August 21, 2001
We (Sergio, Victor and me) move A1 Down : home until 36 mm. The
trigger rate was 13kHz at home position and goes up to 37kHz at 36mm. The
P and PBar Halo didn't change during the motion (29 KHz for Phalo, 1.2 kHz
for Pbar Halo).
Tuesday, August 21, 2001
A2 Up motion was a result of bad communication with the rack
monitor at A2. But why only A2 Up moved and not A2 Down ? Because these 2
motors are not connected to the same rack monitor.
A2 UP is connected to the rack monitor #3, and A2 Down #4. The
smoke detector is connecting to #3 for the control, and #4 for the
status.
During a bad communication with the rack monitors, The bits of
Inputs go to 1 for a short time and then come back to the correct
status (0 or 1). For the software, 1 from the inputs of rack monitor
#4 means that : the smoke detector is in alarm at A2 and also the pot (A2
Down) is at home position. That's while the software didn't send the
command "GO HOME" at A2 Down because, for the software, this pot seemed to
be at home. A2 Up is connected to the rack monitor #3 and was not at home
position (no bad communication) so the software send it back to home.
__________________________________________________________________________
Molina
Thursday, August 23, 2001
L0 rates
Yesterday we made another measurement of the halo rates, and then we
tried to compare with previous measurements.
In this table you can see the variation of the Luminosity (L) and the
variation of the response of the detectors respect to the
measurement of the 8/9/01 (all the numbers are in %, and - means
decrease). All the values where made at HV 100%.
Day dL(%) P-halo a_halo P1U P1D P2U P2D A1U A1D D1 D2 A2U A2D
8/10 -21 3.1 -27 3.1 11 -8.8 -15 -17 -30 -32 -25 -33 -30
8/16 53 55 314 26 56 77 82 183 82 108 126 20 67
8/17 50 40 36 -2.8 -4.7 67 13 233 -14 16 32 -19 -16
8/22 78 75 20 48 45 66 79 61 59 102 89 59 57
May be somebody can see some pattern there, but I cannot see even an
agreement in the measurement of our detectors with the P-halo and
Pbar_Halo values.
Thursday, August 23, 2001
L0 rates
These are the values taken as reference of the 08/09/01 measurement (all
the numbers of counts were obtanied as mean value of three readings,
except the value of the luminosity)
Day L(e30) P-Halo Pbar_Halo P1U P1D P2U P2D A1U A1D D1
8/9 3.2 18980 1552 16588 13682 39833 63979 480 23898 4255
D2 A2U A2D
5410 25201 21080
In this table you can see the variation of the Luminosity (L) and the
variation of the response of the detectors respect to the measurement of
the 8/9/01 (all the numbers are in %, and - means decrease). All the
values where made at HV 100%.
Day dL(%) P-halo a_halo P1U P1D P2U P2D A1U A1D D1 D2 A2U A2D
8/10 -21 3.1 -27 3.1 11 -8.8 -15 -17 -30 -32 -25 -33 -30
8/16 53 55 314 26 56 77 82 183 82 108 126 20 67
8/17 50 40 36 -2.8 -4.7 67 13 233 -14 16 32 -19 -16
8/22 78 75 20 48 45 66 79 61 59 102 89 59 57
May be somebody can see some pattern there, but I cannot see even an
agreement in the measurement of our detectors with the P-halo and
Pbar_Halo values.
__________________________________________________________________________
Mike Martens
Thursday, August 23, 2001
L0 rates
It looks like these show a strong correlation between detectors in the
same spectrometer.
To me, this suggests that these rates are really related to beam halo.
If these rates were due to something else (noisy tubes, small light
leaks, etc.) I wouldn't expect such a strong correlation. The luminosity
may not be relevant since the halo can depend on many things. Also, the
proton and pbar halo rates are not neccessarily correlated either since
one species of beam can be "dirtier" than the other.
We (i.e. Tev people) have been doing some scraping studies to better
understand the collimators but we don't have any conclusions yet.
__________________________________________________________________________
Pierrick
Thursday, August 23, 2001
RG58 U
The idea is to cut the cables to 47m lengths and re-terminate them. This
would cause a delay of only 24ns.
Calculating the latency using known (what I've been told) cable lengths and
propagation into the SCR, we have (cables + Amp/Shapers):
D, A2, P2 MAPMTs - 661ns
A1, P1 MAPMTs - 452ns
We have 3 different cable lengths of LMR400 running from the detectors to
the CH. Adding those lengths plus different 47m cable types to run from the
CH to the SCR plus the ~25ns logic to make the coincidences, we have:
| Ribbon (present) RG58/U LMR400
-------+--------------------------------------------------
D PMTs | 676ns 704ns 651ns
P PMTs | 557ns 585ns 532ns
A PMTs | 521ns 549ns 496ns
At present, we only have tracking in the PU spectrometer. Assuming that we
want to make an ADC gate with the coincidence of 1 and 2 AND CLK, then we
can only look at the P2U MAPMT signals for any choice of cables. If we had
any of the A2 fiber detectors instrumented, then we could look at these with
the ADCs using RG58/U also.
In conclusion, since we can only look at the P2U (and A2 if we had them), I
think that we should add the RG58/U cables for the sake of reliability.
Pierrick
__________________________________________________________________________
Christophe
Shutdown task list
Draft of the list of activities for the shut down for comments.
LIST OF OCTOBER SHUTDOWN ACTIVITIES:
=======================================
I) HARDWARE AT LAB 6 / LAB 3:
==============================
1) Before shutdown:
====================
*** 5V scheme for control and emergency lines, to be discussed with Mario
(Molina, Christophe)- preparation of cables, modules
*** Test of AFE boards
*** Tests of MAPMTs
*** Tests of L0 PMTs
*** Detector building
*** Prepare screws for detectors (Mike M.)
*** Order small yellow screws for detectors (tbc)
*** Build circuit dividers for 10, 15 V
*** Build multiplexors (Brazil)
*** Modify IIB if needed (depends on Pierrick 's tests)
*** Modify the detectors to be able to calbrate L0 PMTs using LMBs
*** Build flex circuits (Mario)
*** Test amplifier boards
*** Prepare fibers for LMBs for L0 PMTs (Petra)
*** Prepare small circuit to amplify signal for LMBs if needed ?
*** Prepare cables between IIB and RM if needed (they are not shielded now)
*** Test IIBs to separate good ones from bad ones
*** Buy low voltage power supplies
*** Build small supports for ribbon cables in tunnel
*** Prepare patch pannel for LMR 400 cables
*** Prepare two additional small grey cables for LMBs (we need 5, we have 3)
*** Build AFE interface (Arlington)
*** Prepare LVDS cables (Fred?)
2) During shutdown:
===================
*** Prepare new couplers
II) HARDWARE IN TUNNEL:
=========================
*** 5V installation for control and emergency lines and tests
*** Multiplexor installation
*** Installation of detectors and pseudodetectors
*** electronic installation (IIBs, RM...)
*** Install new couplers
*** Circuit divider installation
*** Mirrors for down pots???
*** Install more LMBs
*** Check switches (the home position is not always detected by the program)
*** check pot motion with switches / IIBs
*** Install amplifier boards / low voltage power supplies
*** Install supports for ribbon cables / plug in ribbon cables / tests
*** HV connections and tests
*** FIRUS to be checked
III) HARWARE IN SCR / COLLISION HALL / MCH:
============================================
1) Before shutdown:
===================
*** Test IIB and understand what is wrong for all IN and OUT pots in SCR
using
a motor and an IIB (Pierrick)
*** Install two separate switches for 5V control line
2) During shutdown:
====================
*** Installation of AFE boards / plug ribbon cables / installation of flex
circuits
*** Installation of trigger modules
*** Pull last two green cables for LMBs in CH
*** Install patch pannel for LMR 400 cables
*** Install DFE boards
*** Lay cables connecting AFE to data streams (16 cables)
3) After shutdown:
===================
*** Installation of trigger modules (if not finished)
IV) SOFTWARE/ DAQ / TRIGGER / TESTS:
=====================================
1) Before shutdown:
====================
*** Trigger equations
*** Python program updates for multiplexors
*** Python program updates for 5V control
*** Python program update for database / upgrade of database
*** DAQ program (both PDAQ and SDAQ)
2) During shutdown:
====================
*** tests of pot motion
*** Python program updates for multiplexors
*** IIB tests
*** multiplexor tests
*** DAQ tests with AFE boards
*** Test full system with amplifier boards / ribbon cables connections
*** 5V control line / emergency line tests
*** HV tests for all detectors / pseudo-detectors
3) After shutdown:
===================
*** DAQ tests
*** get beam parameters / magnet currents in python program and database
General
Brief summary of accelerator PMG:
1. CDF has requested 4-7 shifts access. It is starting 8:00am on Monday,
September 10th. Exact duration of the access will be determined by Monday,
but it will not be shorter then 4 shifts.
We should try to make as much use of this access as possible. We will be
in Supervised Access mode. Captains will handle the keys.
2. Before the October shutdown Tevatron plan is:
a) have 3-4 stores (~6 days) with 1 day of Tevatron studies per week;
b) deliver about 1.2pb-1 per week;
c) have peak luminosity about 1.E31cm-2sec-1
3. It is confirmed that October shutdown is starting on October 8th with
duration of 6 weeks. Beams Division is working on option to reduce this
shutdown down to 5 weeks. Both D0 and CDF confirmed their plans to
accomplish shutdown work within 5 weeks.
*** Confirmed the dates for upcoming shutdowns. ***
There will be a 7-shift shutdown starting 8:00am Monday, September 10.
There will be no other scheduled shutdown before October 8.
The "October" shutdown starts October 8 and ends November 16. The Main
Injector will be turned on November 17, and we expect to have pbarp
collisions in the Tevatron by November 24.
__________________________________________________________________________
Wagner Carvalho
Monday, August 27, 2001
L1 Firmware
I'm sending below a short document which is a kind of
compilation of what Jorge and I have learned lately about the L1 firmware.
This document is evolving and there are, of course, many questions to be
addressed/clarified yet. Last meeting we had on this subject we reported
that we had to use doublewide DFE, as Manoel Martins had convinced us
about that. We are not so convinced anymore and tomorrow will try to talk
to him again to ask this and a few more questions. Also tomorrow, after
our 1:00 meeting, M. Martens, Jorge and myself plan to meet for discussing
the L1 equations, in case somebody else wants to attend.
The AFE channels mapping is not according to the Pierrick drawing
found in his recent note about the patch panel (it is based on Mario's
former drawing). That is because at this point the precise mapping seems
not be particularly important for what follows. Actually, we even have to
consider a scheme with fewer AFE's, in case we can't get the four we expect.
There is still a lot of work ahead. Jorge is working mainly on
trying to reduce/simplify the equations. We are trying also to figure out
what is the best design option: 2 x 3 DFE motherboards; singlewide x
doublewide doutherboard, etc.
Comments/suggestions are welcome.
L1 Trigger Equations Implementation
===================================
General Records
--------------
Number of spectrometers = 9 (PU, PD, PI, PO, AU, AD, AI, AO, D)
Number of channels per spectrometer = 224
Total number of channels = 2016
Tentative Designs
----------------~
Four AFE Boards + Two DFE MotherBoards
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The 4 AFE boards handle the 2016 channels sent through flat
cables from the amplifier shapers:
512 @ AFE1 ( PU1 + PU2 + PD1 + PD2 + PO1[1:64] )
496 @ AFE2 ( PO1[65:112] + PO2 + PI1 + PI2 +D1 )
496 @ AFE3 ( AO1[65:112] + AO2 + AI1 + AI2 +D2 )
512 @ AFE4 ( AU1 + AU2 + AD1 + AD2 + AO1[1:64] )
The digitized signals from AFE's are sent through LVDS (low
voltage differential signal) cables to the DFE boards. There are 4
such cables coming out from each AFE. Data is transmited through these
cables in 28-bit long words about every 19ns (~53MHz ). All data must
be transfered from AFE board to DFE board in no more than 132ns. Thus,
seven 28-bit words from a given event can be sent through a single LVDS
cable (7*28=196 bits). Actually, not all the 28 bits are available for
data, some are for control purposes (it looks like 24 bits are for data
carrying).
Each DFE motherboard (MB) receives data from 10 LVDS cables. Assuming
two DFE's, a possible distribution of spectrometers signals to be handled by
each board is the following:
DFE1:
Spectrometers: PU, PD, PI, PO
Number of channels = 896
Number of LVDS cables = 7
DFE2:
Spectrometers: AU, AD, AI, AO, D
Number of channels = 1120
Number of LVDS cables = 9
Also shown are the number of LVDS cables allocated to carry spectrometers
data from AFE to DFE.
There are two DFE daughterboard (DB) types: singlewide and doublewide.
The DB fits onto MB and is a card where the programmable devices (FPGA's)
are
assembled. One DFE MB fits two singlewide or one doublewide DB. The
singlewide
board accepts up to 5 FPGA's while doublewide accepts up to 3 devices only.
On
the other hand, more powerful FPGA's can be used with doublewide boards.
If we intend not to have FPGA's sharing equations from different
spectrometers (what seems to be the case due to the huge number of equations
for each spectrometer) then we need to go for the singlewide DB in this
scheme
with only two DFE MB.
Four AFE Boards + Three DFE MotherBoards
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Using three DFE MB instead of two would allow us to make use of the
doublewide DB, as would have just enough FPGA's to proccess data from all 9
spectrometers. In such a scheme data coming from a single LVDS cable would
have to be split among two different DFE DB (this is something we don't know
yet if is possible or not). A foreseen distribution of spectrometers signals
to be handled by each board is the following:
DFE1:
Spectrometers: PU, PD, PI
Number of channels = 672
Number of LVDS cables = 5 + 1 shared
DFE2:
Spectrometers: PO, D, AO
Number of channels = 672
Number of LVDS cables = 4 + 2 shared
DFE3:
Spectrometers: AU, AD, AI
Number of channels = 672
Number of LVDS cables = 5 + 1 shared
Alternative Design
------------------
The aforementioned schemes assume that the L1 track pattern finding
will be made in one step only. Digitized signals from one spectrometer are
input to an FPGA, pre-programmed with the suitable equations, which makes
all the processing and produces the trigger output for that spectrometer.
An alternative approach is trying to split the trigger processing
in more than one step. For instance, one FPGA might try to find pixels at
detector 1 while another (or the same) FPGA would do the same for detector 2
and then their inputs would be sent to a third (or second) one which task
would be try to match pixels found in both detectors. There is at least one
drawback with this scheme: signals of one detector would need to wait finish
the processing of another detector. There are also some issues that would
need
to be addressed:
- More than 3000 pixels can be formed by using information from the
six planes of one detector (we are considereing using this fine
segmentation for detector 1 and a coarser one for detector 2,
ignoring X and X' signals). Then we would have to come up with a
protocol to send the pixel information from "pixel finding" FPGA's
to "pixel matching" FPGA, as the number of possible pixels exceed
by far the number of physical FPGA connections.
- The "pixel matching" FPGA's would have to be downloaded with the
matching equations for more than one spectrometers. We would need
to figure how large these equations would be. _
______________________
Tuesday, September 04, 2001
End of run
We (Jorge, Wagner, Eugene & Pierrick) started at 15:00 and finished
the run at ~18:00.
Luminosity was 3.27E30. Halo rates were 22kHz and 1.9kHz for p and
pbar, respectivelly.
After turning HV on we recorded the detectors rates at home (out)
position, namely:
-----------------------------
P1U P2U A1D A2D
15 35 18 14 (kHz)
-----------------------------
Pots were then inserted to the planned positions (24mm for P1U and
P2U and 34mm for A1D and A2D) with no noticeble changes in halo rates.
After inserting P1U and P2U, Pierrick intended to make some
coincidence tests using the MAPMT signals but then we experienced what
seems to have been a syncronization problem. Suddenly we started getting
nearly random readings from scalers. After some time trying to find what
was causing the problem we started suspecting of the pulse generator and
then Pierrick tried using an external generator for create the gate. It
worked and then we could read rates again. Thinking about this latter, I
remembered that the first time I went to SCR something similar may have
hapened and I remember Petra making some fine tuning to the pulse
generator in order to get rates readings stable again.
We then took a measurement (over a 10 seconds time window) of
detectors, spectrometers and elastics rates (all in kHz):
------------------------------------------------------------------
P1U P2U A1D A2D P1U*P2U A1D*A2D Elastics
33 49 39 21 7.7 3.5 0.723 kHz
------------------------------------------------------------------
After recording these numbers we started making studies on the
early signals. We've found that
EA1D*EA2D rate = 0.30 kHz
P1U*P2U*NOT(EA1D) = P1U*P2U*NOT(EA2D) = P1U*P2U (exact),
what means AD halo vetoes do not impact PU rates. We were not sure if a
complete agreement (zero impact) was to be expected in this case and tried
to figure out if we're doing something wrong. After quite some time we
could not see something that was obviously wrong.
We decided to finish the run as Jorge could not stay very longer
and then there would be not enough time neither for plateauing the tubes
nor to make the same study we had just made for the other elastic
configuration (P1D, P2D, A1U, A1D).
We turned HV off, retracted the pots and turned control lines off.
We have to think about the pulse generator setting for next time
we run. We need also some feedback on the early vetoes (NO)impact on the
PU rates, is it reasonable?
________________________
Thursday, September 06, 2001
Moving pots
We (Jorge, Mike, Eugene, Andrew, Wagner) finished the run at around
8pm.
Pots were retracted to home position, control lines and PMT HV turned off
and shift
captain informed.
We spent a lot of time just trying to set the delay for the early
signals properly, as Phillips 794 Gate/Delay generator seems to have
the strange feature that to set delay time you have to adjust the width
screw and to set the width you need to play with the delay screw. It
looks crazy, but that's the way we did it work (I guess...).
At the end we managed to have an EA1D signal arriving a little
before the P1U*P2U coincidence to use it as a veto. We set the EA1D signal
delay to 150 ns and its width to 80 ns (latter on to 100 ns) and checked
with the osciloscope they were in time (actually, EA1D was set to arrive a
little before because a veto has to be set into the logic module before
the own signal it is supposed to veto).
We then took a few of the measurements foreseen in the run plan,
including elastics rates with and without EA1D veto:
P1U*P2U*A1D*A2D ~= P1U*P2U*A1D*A2D*NOT(EA1D) ~ 47 Hz.
__________________________________________________________________________
Pierrick Hanlet
Wednesday, August 29, 2001
Control lines
Yesterday afternoon, I received a call from Petra in the SCR that there was
no beam. This gave me the oppertunity to test the modifications that I made
to the pot motion control line patch panel.
Before even moving the panel back into the SCR, I tested that 5V is on
independently for each the control line, and both control lines and
emergency
lines have 5V when the emergency line is on, independent of the control line
settings.
Though I did not perform exhaustive testing on each pot, I tested and find
that one can move A when A is on, P when P is on, and the emergency switch
works to send the pots home when it is on. I also tested that you cannot
move A when A is off and P is on, and you cannot move P when P is off and
A is on.
There were some glitches, but I suspect these are software or hardware:
Once, when I had an A pot and a P pot in at 10mm, I turned on the emergency
switch and only the P pot returned. It took several attempts to get the A
pot to return. When I tried again later, it worked fine.
In light of what Lionel left us with regarding the readback and control
lines to the IIBs and RM, I would like to request an access into the tunnel
to check and perhaps rearrange the cables in our crates. Perhaps this would
greatly reduce the spurious motion of our pots.
__________________________
Wednesday, September 05, 2001
Minutes from the AFE mapping meeting
Attending:
Andrew, Jorge, Wagner, and I from the FPD group.
Fred Borcherding, Jamieson Olson, Paul Rubinov from the CTT electronics
team.
Subject:
The issues discussed delt with possible ways of mapping our detectors into
the AFEs and DFEs considering practical issues such as cabling and mapping,
how many boards of each that we need, timing issues for readout and
triggering
considering the different latencies built into our system, and mapping the
data
into the DFEs so that spectrometer information goes to the appropriate
FPGAs.
History:
Initially, we were planning on using 4 AFE boards because we have 112
channels
per detector; if we stuff multiple detectors into the MCMs on the AFE
boards,
we can fit all of our detectors onto 4 AFE boards. However, we then are
faced with timing issues, and mapping the appropriate signal into the next
board of electronics.
Issues and Recommendations:
There is a SIFT gate of 90ns < t < 120ns which according to Jon Anderson is
expected to be 1/board; though it is possible in principle to have more
than one gate. Considering the latency differences in our system, Fred
pointed out that we should simply have 1 (or 2 in the final configuration)
board(s) per timing group. This means that we will have 5 boards in the
final configuration, and that we need 3 to begin with. In this case we will
not be using all of the channels of the MCMs; this will allow us to accept
rejected AFE boards if we build our TPP (transition patch panel) in a way to
allow remapping with a piece of wire and a soldering iron.
Giving up on the idea of trying to stuff all of our channels into 4 AFE
boards, solves all of our timing problems. It simplifies the mapping of
our detectors into the AFE cards. It also simplifies the mapping of
channels into the DFEs, because we will then have 2 MCMs/detector => 1
LVDS cable/detector, and we can then take the second detector(LVDS) cable
from any of the AFE boards. This allows us the flexibility to think in
terms of 2 or 3 DFE boards with single or double wide daughter boards, and
we can stage our growth here too. We then have think in terms of whether
or not we want to process more than 1 spectrometer/FPGA, and decide
whether or not we want to use the single wide or double wide daughter
boards.
Summary:
I think that it is clear that we will request 5 AFE boards and 3 DFE boards
to the collaboration, with 3 AFEs and 2 DFE needed immediately.
Many thanks to Fred, Paul, and Jamieson for thier help in settling this
issue. I can now proceed with what I need to get done to get our readout
and triggering chains operational.
________________________
Thursday, September 06, 2001
AFE mapping
After yesterday's meeting, and based on the summary that I wrote afterwards,
I have a new drawing of the TPP. I've attached the drawing to this note.
http://www-d0.fnal.gov/fpd/FPD_News/tpp_phase.ps
A few words of description. Now that we've abandoned the idea of 4 AFE
boards, life is MUCH simpler. The left drawing is an x-ray view of the
top of the board. The horizontal 2" lines are the shrouded connectors
which will hold our flat coax ribbon cables. The lightest grey box
is the board itself. The darkest grey boxes correspond to how our
connectors will map to the MCMs; 4 cables/MCM. Note that this gives
us a spare 16 channels for each pair of MCMs (or detector); this gives
us the ability to take less than perfect AFE boards. The lettering on
the darkest grey boxes mean: AxMy => Ax->AFE board number "x" and
My->MCM number "y".
The intermediate grey colored boxes correspond to how our detectors
map onto the TPP. This particular arrangement allows us to phase in
detectors as we get the electronics. It also allows us to map similarly
timed detectors onto the same AFE board. This fixes our timing issues
for the readout.
The right drawing shows a top view of the TPP. The boxes have the same
labeling as the left drawing, and mean the same thing. The boxes are
sized according to the measured footprint of the flex cable that I have
in hand, and in these boxes will be the pads to make the connections to
the flex cables. They strattle the connectors on the bottom of the TPP,
so there shouldn't be any interference with mounting hardware. The layout
of these boxes should line the flex cables up directly below the AFE cards
so that no twisting of the flex cables will be required, other than in the
direction that they are meant to bend.
This map meets all of the requirements that I had previously stated, modulo
fitting everything onto 4 AFE boards.
__________________________________________________________________________
Michael Strang
Thursday, August 30, 2001
Quick test of pot motion
I performed a quick test of the pot motion during the access today.
I turned on the P-side 5V.
I sent a command to move P1U to 20mm at high speed.
I turned off the 5V when the LVDT read 10mm and it stopped at 10.4mm.
The software set the pot into the locked state (this is default)
I waited for approximately a minute, turned on the 5V.
The pot remained in place.
I verified that I can give a command to move the pot further in or
give the pot a command to come out. It also obeyed the individual
stop and the all stop commands.
I repeated the test for the A side with similar results.
I then inserted both A1U and P1U to 5mm and turned on the 5V so I could
study the emergency line behavior.
>From what I understand the emergency line must be turned on as follows:
Turn on both 5V command lines.
Turn on the emergency line.
When the pots are moving, turn off the control lines.
You will receive warning windows from the software that the pots are
moving and the software did not send the command. If you close this
window, it will simply reopen. It is intended to tell us when pots start
moving for an unknown external reason even though we know that we have
turned on the emergency line.
With the emergency line, both pots returned home and I turned the
emergency line off.
I verified that all pots were at home and all command lines and the
emergency lines were off.
Different shielding configurations of the pots:
A1U has the L0 tube shielded with one thickness of mu metal foil
A2U has the L0 tube shielded with two thickness of mu metal foil
A1D has the L0 tube shielded with the manufactures shield
P1U has the entire cartridge top shielded with one thickness of mu metal
foil.
___________________
Friday, August 31, 2001
Run checklist
The run checklist has now been converted to html format. Also, the HV
controls and the Camera Controls are converted. The Pot Motion Controls
still need to be converted (I'll work on that next week) but I have
included a section in the checklist details on pot motions that I think
touch on the most relevant aspects of the program.
I still need to link the LMB instructions into the list. Petra sent me
a link to what she has available. I will put it in the same format as
the other documents and link it in next week.
I have set the run checklist to be the default home page for the
Netscape browser on the computer in the SCR.
Please start with run_checklist.html, all of the other pages are linked
from there in the (to me) logical places. If you have any comments,
please let me know, especially about possible emergency conditions and
the recovery procedures. If any of the links are broken, also let me
know. Thanks.
The addresses for the various pages follow:
run checklist
http://d0server1.fnal.gov/users/strang/web/fpd/documents/run_checklist.html
checklist details
http://d0server1.fnal.gov/users/strang/web/fpd/documents/checklist_details.h
tml
emergency procedures
http://d0server1.fnal.gov/users/strang/web/fpd/documents/checklist_emergency
.html
runplan
http://d0server1.fnal.gov/users/strang/web/fpd/documents/runplan.html
camera control
http://d0server1.fnal.gov/users/strang/web/fpd/documents/camera_control.htm
hv control
http://d0server1.fnal.gov/users/strang/web/fpd/documents/HV_control.htm
__________________________________________________________________________
Carlos
Wednesday, August 29, 2001
About event size,
EVENT SIZE IS ABOUT 264 bytes.
If we have a data taking rate of 50 HZ and continuosly run for 10 hours
a day. Then we will be using
264 x 50 x 3600 x 10 = 0.5 Gbyte/day
EVENT SIZE
=========
The data buffer for stand alone daq is written in 32 bit words. The amount
of words that are used for one event are:
word
----
1 event length
2 data type
3 event number
4 scaler channel 1
.
.
51 scaler channel 48
52 tdc channels 1 and 2
.
.
59 tdc channels 15 and 16
60 packed adc data from two non zero channels
.
.
65 last packed adc data (asumming 5% of fibers fire)
66 end of event marker
======
TOTAL = 66 words * 4 bytes = 264 bytes/event
We have the chance to increase the seudorapidty coverage up to
eta of 6.9 using scintillation counters around the beam pipe. Higher
eta values will be reached with pots inserted. The scintillation
counters can be used mainly to veto on seudorapidity gaps.
The proposal is to install the scintillation counters during October
shutdown. Preparation and testing of the counters can be started once
everybody agrees on the numbers.
Please check numbers, there could be mistakes that I didnot realise.
HIGHER ETA COVERAGE FOR THE FPD
===============================
The pseudorapidity coverage of the Luminosity Monitor is:
2.7 < eta < 4.4
It will be good to have some sicntillation counters around the beam pipe
to cover a higher eta range. Two main reasons for that are:
1) to use them to veto on seudorapidity gaps.
2) to increase the acceptance of the luminosity monitor for non diffractive
events.
We have two possible places where we can mount such counters:
1) at the front face of the first low beta quad
2) right at the locations of the roman pots
1) COUNTERS AT THE FRONT FACE OF LOW BETA QUAD
================================================
The conclusion after talking to Mike McGee is the following:
There are some accelerator intrumentation devices in front of the
low beta quad, after these devices there is a 58 inch spool piece where we
can
mount the scintillation counters. The spool piece starts at about 5.2 m from
D0 interaction point.The space left by the beampipe shielding is about 18
inches
height by 18 inches wide. The beam pipe diameter at this location is 2.5
inches.
One initial proposal is to use scintillation counters surrounding the
beam pipe that can be located at 6m from D0 interaction point and with a
height of 15 cm. this would give an eta coverage of
eta_min=-ln(tan(0.5*arctan(18.2/600))) = 4.2
eta_max=-ln(tan(0.5*arctan(3.2/600))) = 5.9
POSSIBLE SCINTILLATION COUNTER DESIGN
We can use a design similar to luminosity monitor but with only two
photomultipliers and only 6 wedges. The 6 wedges will form a ring with
an inner radius of 3.2 cm and outer radius of 18.2 cm.
Each photomultiplier will read three wedges. To be able to do that we
will need one light guide per wedge, each light guide will be glued to
the middle part of the wedge, the other side of the light guide will be
smaller to fit within the face of the pmt and to allow to accomodate
other two light guides in the face of the same pmt. The pmt will run
longitudinal along the beam pipe.
The two pmts can be connected to a T connector and run only one signal
cable to the TDC board. Te same can be done for HV, just run one cable
and use a T connector to get voltage to the two pmts.
I think we can use the same type of pmts that we use in the L0 counters
(XP2282).
We will need to think in some mounting scheme to keep the counters in
place.
2) COUNTERS RIGHT AT THE LOCATION OF THE ROMAN POTS
====================================================
At each roman pot location the beta values in x and y
are different, therefore there will be some phi dependence on the
scattered angle for a particle hitting the beam pipe at that
location. Using the effective lengths in x and y for each detector location
we can determine the max eta value that a counter, just touching the beam
pipe, would reach if it is located in the vertical or horizontal plane.
MAX ETA VALUES REACHED WITH COUNTER AROUND BEAM PIPE
LEFFX(m) LEFFY(m) etamax_x etamax_y POT LOCATION
-------- ------- -------- -------- ------------
7.383 1.790 6.134 4.718 D2
8.150 2.362 6.233 4.995 D1
16.149 8.383 6.917 6.261 A2
19.326 10.785 7.097 6.513 A1
10.791 19.313 6.514 7.096 P1
8.388 16.139 6.262 6.916 P2
I think particles with eta values lower than 6 would hit the beam pipe in
the
straight section before reaching the magnets, so the lower eta values
displayed
in the table below cannot be physically possible.
If we mount cunters at A2 and P2 we will have eta coverage up to 6.9.
We can still use the same scheme with 6 wedges, the height of the wedges has
to be limited because space constraints. The height of the wedges could be
3 cm and then we can use light guides to connect three wedges to one pmt
and other three wedges to second pmt. Because the space limitation, the
light guide design has to be different than the one for counters in front
of low beta quad.
With pots inserted to different positions we can still gain higher eta
values.
For a beam emmittance of 20 pi mm-mrad, beam momentum=1000GeV and pots
located
at 8*sigma from beam we can reach a maximum value of eta=7.8.
________________________
Saturday, September 01, 2001
Veto counters
The reviewed propposal for FPD eta veto counters is:
Have two semicircles design that formed a complete 360 degree ring around
the beam pipe. Each semicircle is read out with one pmt.
I think that xp2282 counters should work fine since the region where
we will locate them should be a low magnetic field area. We will do
magnetic shielding with mu metal anyway.
To improve light collection for one semicircle we can have three light
guides connected to the upper part of the scintillation counter. The light
guides get bent and taped down to the face of the pmt (see attached
drawing).
http://www-d0.fnal.gov/fpd/FPD_News/vetocounters.ps
I think in lab 7 we can request help from experts that can bent
the light guides for us. THe pmts then are put longitudinal. The reason is
beacause we have to fit the setup in between
the ligt pipe and the D0 shielding.
The pmts from the two semicircles get connected together to have just one
signal.
We duplicate this setup at the same location. So we will have two ring
counters per location, one signal per ring. Both signals are put in
coincidence and geted with the D0 interaction clock. This will guarantee
that we look at signals correlated to interactions at D0.
I think we should be able to request help from people at lab 7 to make the
scintillation counters and light guides. We have to have someone at
Fermilab designing some structure to keep the counters in place. and to
spend time testing and mounting the system during a shoutdown.
I believe that for mounting the counters it will be three hours per
location. We need to lay down new signal an HV cables for the counters.
__________________________________________________________________________
Mario
Wednesday, August 29, 2001
Amplifier tests
The coupling to AFE is capacitive, so DC is blocked. SIFT operates
by switching the input in two moments, one when there is no signal is
expected from the VLPC or MAPMT and the other when the signal is expected
(can happen or not). The difference is send ahead to discriminator and
SVX II, so any off-set is cancelled.
See notes :
http://d0server1.fnal.gov/Projects/TriggerElectronics/WebDocs/Review990122/5
_analog/SIFT_D0Note.htm
http://d0server1.fnal.gov/users/manuel/d0_notes/sift_timing/tittle_sift.html
Pierrick
Saturday, September 08, 2001
Shift summary
Mike Strang and I (Pierrick):
We moved the pots PU and AD in to 24mm and 32mm, respectively.
We moved the pots in pairs P1U and P2U, then A1D and A2D.
Other than the higher singles rates (up to 54% higher than the last
run for P2U at the home position), there were no significant occurances.
The pots went in smoothly. For the AD pots, the singles rates did not
change after 31.5mm.
We did not affect the p-halo and p-bar-halo rates. We have not impacted
either D0 or CDF.
---====>>>>>> Note that the A1 camera is rocking back and forth
spontaeously; it appears to be in some sort of scanning mode.
With Victor and I (Pierrick)
Success in timing in early-bar signals! I have set up in the logic and
visual scalars:
P1U*P2U*NOT(EA1D)
A1D*A2D*NOT(EP1U)
P1U*P2U*A1D*A2D
P1U*P2U*NOT(EA1D)*A1D*A2D*NOT(EP1U)
Rates in kHz:
P1U P2U P1U*P2U P1U*P2U*NOT(EA1D)
---------------------------------------------
43.5 72.1 11.8 11.2
A1D A2D A1D*A2D A1D*A2D*NOT(EP1U)
---------------------------------------------
46.9 23.5 4.2 4.1
P1U*P2U*A1D*A2D P1U*P2U*NOT(EA1D)*A1D*A2D*NOT(EP1U)
-------------------------------------------------------
104 Hz 103 Hz
Victor has set up the sDAQ again to read MAPMT ADC data for P1U and P2U.
We also looked at the timing of the MAPMT signals wrt the ADC gates:
Previously: P1U arrives withing 10ns of the start of the gate
P2U arrives about 54ns of the start of the gate
Removing some of the logic to get a faster gate, P1U now arrives ~27ns
after the start of the gate. If we use the RG58/U cables, the P1U signals
will arrive just at the start of the gate. If we use LMR400 cables, we
will gain an additional 25ns from the present setup.
We are now collecting ADC data for P1U and P2U MAPMTs and scalar data for
singles, doubles, and elastics (though not with the early bars).
We collected ~30k events. The file is in:
d0ol02: /home/bodyagin/d0/fpd/camac/daq/socket/vers01/data/run000041.dat
At 03:45 we turned off all HV and sent the pots home.
__________________________________________________________________________
Michael Strang
Saturday, September 08, 2001
Pot test
This morning Eugene, Andrew and I (Mike S) inserted the pots beginning
at 10:30am. The P2 castle was set to -1600V on the L0 tubes since the
rate at P2D was high even in the home position. The other tubes were set
at -1800V. The P side pots were inserted to 24mm. The A side pots were
inserted to 32mm. We were unable to start the DAQ program.
At this position, the A-up pots seemed to be slightly impacting the
losses at D0, so we attempted to bring the pots back 0.5mm to 31.5.
Since the control line was not on, when it was enabled, the software
could not verify the position of the pots and enabled the safety
procedure of removing the pots to home. They were left there while we
studied the p-side.
Victor stopped by and showed me that the daq server on the linux machine
needs to be started from the proper node in order for the daq to be
started. At this point we were able to start the daq program and send
scaler rates to lumberjack.
We began inserting the P-side using 0.1mm steps in slow speed starting
with P1U. The initial rates were 30.5kHz for the p-halo, 1.5kHz for the
pbar-halo and 39kHz for the singles rates.
We noticed strange spikes in the losses that would only appear in a
single data point. Mike M. couldn't verify the cause of the losses and
we couldn't say if it was caused by pot motion or not.
We noticed a behavior in the d0 pbar losses with what appeared to be a
periodic increase and decrease in the losses. We didn't observe a steady
increase with increased position that we had seen in the past. Because
of this, we continued to insert the pots to see what effect we would
have on the singles rates. By 28.0mm we had a singles rate in P1U of
~71kHz. At 28.1mm it rose to ~82kHz and at 28.2mm rose to ~92kHz. We
also noticed an increase in the losses at CDF at this point. We
retracted the pots 0.5mm and the losses at CDF went away.
Based on the length of time it took to insert the P1U pot, Andrew
decided to move in 0.3mm steps to begin with on P1D. At 24mm the singles
rate for P1D was 18.7kHz. We were able to insert to 27.5mm when we began
to notice an increase in pbar halo losses at D0. We then moved to 0.2mm
insertions until 28.1mm where the losses had been increased by ~10%. At
this point, we moved to 0.1mm steps. At 29.3mm the losses at D0 reached
~20% and there was no impact at CDF. The singles rate at this time was
~29kHz. We retracted the pot by 0.5mm and noticed a decrease in the
losses but not a return to the baseline. We then removed both P1U and
P1D to 25mm and saw the losses return to normal.
At this time, Eugene left.
We then inserted both P2U and P2D at the same time. The beginning
singles rates were 39.0kHz for P2U and 60.2kHz for P2D. Based on our
observation at P1 we inserted to 26mm at high speed with no impact. We
moved to 27.0mm at high with no impact. We went to 27.5mm at high speed
with no impact. We then inserted P2U to 28.0mm at slow speed with no
change and then we moved P2D to 28.5 with no change. The singles rates
at this time were ~42kHz for P2U and ~94kHz for P2D.
We decided at this time to move on to the A-side and observe what effect
we would have there.
We reinserted A1U and A2U to 32mm. We noticed the slight increase in p
losses at D0 and removed the pots to 31.5mm where the loss went away.
During this retraction, we noticed a single spike at CDF but it was not
repeated and their losses returned to baseline after the spike.
A1D had a singles rate at 32mm of 35.8kHz. We started inserting A1D at
slow speed in 0.3mm steps to 33.3 with no noticeable changes in the
beam. We then inserted in 0.2mm steps at slow speed to 35.5mm. At 35.5mm
we observed no increase in losses at CDF or D0 and nothing in the SCR to
indicate any problems, the singles rate was ~43kHz. Once we started to
notice losses, we would change to 0.1mm steps to closely monitor the
situation. I turned off the control line so I could go and look at the
lumberjack plot. When I returned, I turned on the control line and
attempted to move the pot to 35.7mm, but the software could not verify
the location and retracted the pot to home. Upon consultation with
Andrew, we decided to reinsert the pot into the last position, fast to
34 then to 35.5 at slow. I didn't want to do the last step that fast so
I broke it down into two steps, from 34.0 to 35.0 at slow and then to
35.5 at slow.
I inserted the pot to 34.0mm at fast speed and recorded the regular data
points and noticed no problems, the rate this time was ~36.6kHz and the
previous time it had been ~39kHz. I then inserted the pot to 35.0mm at
slow speed and recorded the data points and observed nothing anomalous
in the SCR, singles rate this time was ~39kHz and the previous time at
this position had been ~40kHz. From the lumberjack plot fpda_end2.gif
you can see that at the end of the time it was sitting at 35.0mm there
were spikes in the losses and CDF and D0. Having not seen these in the
SCR and seeing as how the pot had been in this same position 10 minutes
earlier and the data I was getting didn't look any different, I inserted
the pot the final step to 35.5mm. After the software stabilized the pot
position, I went to record the rates and noticed a singles rate in
excess of 100kHz in all the counters. After this, the store was aborted.
In the past, the A1U pot has been inserted to 36mm with impact to the
beam noticed at 34.5mm back on August 16, 2001.
Attached are lumberjack plots of the pot movement:
fpdp-side is a plot tracking halo losses at CDF and D0 and the LVDT
values of the 4 p pots from 10:08 to 16:48.
http://www-d0.fnal.gov/fpd/FPD_News/fpdp-side_08sep01_1652_229.gif.
fpda-side is a similar plot for the 4 a pots.
http://www-d0.fnal.gov/fpd/FPD_News/fpda-side_08sep01_1650_229.gif.
fpda-end is a plot for the a pots from 15:38-16:08 showing how the pot
had been at 35.5mm 10 minutes earlier with no impact to the losses.
http://www-d0.fnal.gov/fpd/FPD_News/fpda-end_08sep01_1656_229.gif.
fpda-end2 is a plot for the a pots from 16:00-16:08 showing the final
movements of the pots and the jumps in the losses.
http://www-d0.fnal.gov/fpd/FPD_News/fpda-end2_08sep01_1701_229.gif.
Monday, September 10, 2001
Shorter summary of results from Saturdays tests
Based on this data, we need to try to determine the operating positions
of the pots in consultation with the BD.
For the P-side (current operating position 24mm):
We were able to insert P1U to 28.0mm when we first started to notice
large increases in singles rates with each subsequent 0.1mm step as well
as impact to losses at CDF. We did notice strange oscillatory behavior
at D0 in the pbar-halo but looking at it with Mike M. we couldn't
determine was going on.
For P1D We first notice steady losses at D0 at 27.5mm. At 28.1mm the
losses were at 10%. At 29.3mm which D0 pbar-losses had increased by 20%.
This seems to indicate that the beam is shifted upwards by as much as
0.5mm at the location of P1.
For P2U and P2D the limiting factor at this time was the singles rates
in P2D which is the hottest tube we have. We were able to get P2U to
28.0 mm with no impact to losses and no large increase in singles and
P2D to 28.5mm where the singles rates was ~95kHz. This was with a
voltage of -1600 where we don't believe the tubes to be fully efficient.
The effect of seeing higher singles rates in the P1/A1 positions vs. the
P2/A2 positions I feel is due to the fact that up to this point, both 1
and 2 positions are inserted to the same point whereas, the 2 position
need to be inserted further to reach the same sigma region of the beam.
This, I feel is leading to the lower rates we see at the 2 positions.
It seems that we can probably pick up a few mm on the P1 position and
even more on the P2 position.
A-side (current operating position 33.0mm):
We were able to insert A1D to 35.5mm at one point with no noticeable
losses. Due to human error we needed to reinsert the pot to this point
to continue the studies but the second time the store was dumped. It
appears that the position of the beam had changed within the span of
~10min. In the past, we had inserted A1U to 36mm causing a noticeable
scraping of the beam at 34.5mm.
We need to understand better what happened to try to come up with any
operating position for this side.
Tuesday, September 11, 2001
Pot insertion LVDT test
I inserted the pots to the in position (or to the limit imposed by the
software) and recorded the LVDT value for various positions on the way in
so we can better calibrate the LVDTS. I will leave the pots in these
positions overnight with the control lines on and see what happens in the
morning at which time I will retract all pots and turn off the control
line.
LVDT
Pos A1U A1D A2U A2D
5mm 5.13 4.79 4.98 3.98
10mm 10.18 9.68 9.98 8.99
15mm 15.24 14.62 14.99 13.98
20mm 20.30 19.60 20.02 18.99
25mm 25.37 24.57 25.06 23.98
30mm 30.41 29.51 30.05 29.00
35mm 35.43 34.45 35.05 33.98
40mm 37.43 36.43 37.60* 37.45*
(37) (37) (39) (39)
P1U P1D P2U P2D D1 D2
5mm 5.01 4.99 5.04 4.89 4.99 5.00
10mm 9.95 9.96 10.03 9.83 9.96 9.97
15mm 15.03 15.04 15.05 14.79 14.97 15.03
20mm 20.05 20.04 20.11 19.73 19.99 19.99
25mm 25.07 25.04 25.15 safety 24.99 25.03
30mm 30.13 30.03 30.24 29.80 30.01 29.99
35mm 31.12 31.10 33.23 32.79 35.01 35.01
40mm 39.02 36.19*
(31) (31) (33) (33) (39) (40)
NOTES
The numbers in () are limits that are set in the software. If you try
to issue a move command beyond these limits, the command is automatically
reduced to the limit value and the pot cannot be given a command to
move beyond this.
* This means that the pot accepted a command to move to the desired
postion but stopped because it hit the in limit switch.
safety- The pot could not verify the postion at 25mm after two attempts
and automatically retracted the pot to home. At this same time there was
an error communicating with FPD_RM_02-12 in the log window. The LVDT value
where it stopped was 23.02 when the desired position was 25.
Based from this information it appears that the LVDT values are
calibrated as follows:
A1U is ~0.1 to 0.4mm high with more error the more you insert the pot.
A1D is ~0.2 to 0.5mm low with more error the more you insert the pot.
A2U seems to be within the 0.05 error of the LVDT. It starts about 0.02
below and ends at about 0.05 above when inserted all the way in
A2D seems to be ~1.00mm below.
P1U starts out fine but ends up about 0.1 high when inserted
P1D starts fine but ends up 0.04 high when inserted
P2U start 0.03 high but ends up 0.2 high when inserted
P2D is about 0.2 low
D1 and D2 seem to be within the 0.05 error of the LVDT.
I will repeat the test tomorrow when I retract the pots.
Wednesday, September 12, 2001 10:54 AM
Pots overnight
The pots were left in the inserted position overnight with the control
line on. At 10:45 this morning, I checked and none of the pots had moved
from the position I left them in yesterday.
I used the "init" button on the main pot motion screen (it appears to be a
software emergency line) that sent all of the pots to home at high speed
with no problems.
So, right now, all pots are at home and the control lines are off.
__________________________________________________________________________
Captain's account (Michael Rijssenbeek)
Sunday, September 09, 2001 12:15 AM
To: d0shifters@fnal.gov
Subject: EVE Shift Summary
EVE Shift Summary. Rijssenbeek, Steinbrueck, Grunewald, Hauser, Kajfasz
Store 688 lost at 16:06 seemingly related to FPD pot insertion.
...
Captain's account (Mont)
Sunday, September 09, 2001 8:00 AM
To: d0cap@d0ol04.fnal.gov; d0shifters@fnal.gov
Subject: Owl Shift September 9, 2001
Crew:
Mont(cap), Su Yong Choi(daq), [Reiner Hauser(cal) -- 1:30], Victor
Bodyagin(muo), Markus Warsinsky(smt)
Cryo Ops Kubinsky/Frank to 06:00
Healy/McNeal from 06:00
*** FPD *** Run Coordinator, Ioannis Kourbanis, requests that pots stay
home through 12:00.
I guess at that point, if we wish to move them in, he
would discuss BUT, I believe that he would not reinject
a store in the Tevatron before the 3 day access.
It is not definitive that FPD killed the previous store,
more likely taht the beam developed a problem then
FPD administered the 'coup de grace'.
...
__________________________________________________________________________
Mike Martens
Monday, September 10, 2001
Pot motion test 08-sep-01 breakdown
Myself and Dean Still from the Tevatron group have spent some time
looking at Saturdays quench while moving the FPD pots. Both of us feel
it was a coincidence that we quenched while moving the pots. We have not
been able to find the cause of the quench but there is evidence that the
pots were not the culprit.
The biggest piece of evidence is 3 times we had sudden DC beam before
the quench. Two of these losses occurred several minutes before the
quench while our A1-down pot was completely retracted. Secondly the
singles rates in our A1-down pot did not go up dramatically when the pot
was being moved back to its 35 mm position for the second time in 20
minutes. If we had moved the pot too close to the beam we would have
seen that reflected in our singles rate. I have looked at the data
logger and verified that none of the other FPD pots were moving during
the time of the quench.
Dean and I could not find the cause of the DC beam loss. The luminosity
was stable so we don't believe that there was any emittance blowup. We
did see the C1SV separator show up in the alarm log file near the time
of the quench and the data logger plot shows a dip in the separator
voltage. We don't know what would cause this except for a separator
spark, but we doubt it was a spark beacuse there was no spark indication
and no apparent emittance blowup. So we could not come to any
conclusions about the cause of the DC beam loss.
I have included a data logger plot
http://www-d0.fnal.gov/fpd/FPD_News/gxpa_1_lumberjack_plot.gif.
which shows our A1-down pot being moved in to 35mm, then retracted, then
moved back to the 35 mm position. On the plot you can see three steps down
in the DC beam current and two of these were when the pot was retracted.
You can also see the singles rates in the A1-down detector. The rates
decrease as the pot is being retracted and then the rates increase again
as the pot is moved back to the 35 mm position for the second time. You
can also see the dip in the separator voltage.
Speaking as acting run coordinator for the next three weeks I am
satisfied with the manner in which FPD has been moving in the pots and I
see no reason why we shouldn't continue with pot motion studies. There
was just unfortunate timing this store.
On a more humorous note, I suspect that the pots will be blamed for
every lost store unless they are proven innocent!
__________________________________________________________________________
Andrew Brandt
Wednesday, September 12, 2001
24 FPD hour shift operations to commence 9/14
In the previous mail about pot operations, I mentioned
the need for 24/7 hour operations. We really could use
help from Brazil on this, and having two more full time
people at Fermi would alleviate a lot of concerns.
We will be working toward becoming part of a routine
detector shift, but this will likely not happen until
February or so? Ideally before Lishep!!!
Accordingly, I will be setting up a shift schedule from
the period 9/14/01 to 10/7/01. The last week will be
modified based on actual shifts taken.
Shifts:
24 days * 3 shifts/day=72 hours
assuming 90% of shifts will be held (optimistic) that means 65 shifts
need to be covered.
The proposal would be 10 each for Eugene, Mike S., Christophe, Jorge,
Pierrick, and 5 each or Sergio, Mike Martens, and Victor. The demarcation
is due to the first group of people having no other shift
responsibilities.
I am working on manpower to reduce this shift load, and cancelled shifts
will be taken into account to try to reduce the load somewhat. Recall that
many people on D0 having been routinely doing shifts on this level for
many months, including owl shifts, and due to previous kinks in the system
we have been very slow in getting a shift schedule implemented.
I immediately need you to provide known conflicts so I can set up the
shift schedule for the next two weeks. After that I will make a draft
of the remaining 10 days, which will be revised based on actual shifts.
By conflicts, I mean serious conlicts such as out of town, wife out of
town so taking care of kids, etc. You can add specific requests if you
like, such as only working owl shifts, working consecutive days (or
not) etc., and I will try to accommodate where possible. Also if you can
provide specifically availability for Friday-
Sunday this week, since it is the first priority to cover.
For meetings where your attendence is necessary for short
periods, a substitute shifter can be arranged for this period (these will
all typically be between 9-5 when it is typically possible to find a
subsitute).
Pot motion will be restricted to a short time at the beginning of the
store and we will have streamlined insertion procedures that take much
less time, so I would prefer to have a 2nd person at the beginning of the
store for the 30-40 minutes while the pots are being inserted. We will also
set up the ACNET console information to be monitored by the captain during
insertion. During the end of store tests we definitely need a second
person to be there, and this should be arranged in advance. Each day there
will be an on-call person who should be able to be available between 8 am
to 12 pm on short notice (45 minutes). Stores don't typically begin after
midnight, and if studies are planned for the end of store, and the end of
store will be early in the morning, this should be arranged in advance.
Before you can be on shift, you need to meet with Mike S. today or
tomorrow to go over the full shift checklist, short checklist, setting up
ACNET documentation and emergency procedures. No more than 2 people +
Mike at once. Everyone needs to know what to do if something goes
wrong.
Please send comments, complaints, suggestions, and shift availability.
Thursday, September 13, 2001 4:07 PM
Shift Q+A and schedule through Monday
Question: Do we have enough people to cover
24 hour shifts given other work that has to be done?
Answer: Possibly. Our manpower will clearly be
stretched and I am working on this (suggestions
welcome) but it's going to have to be a very focused
period for the next few months in order to make our
experiment successful. Note: when I say 24 hour shifts
it isn't like 24 hour daq shifts where people are actually
there 24 hours. If there is no beam, at least
part of your shift is cancelled. If there is no
beam at the beginning of an owl shift, that owl
shift is cancelled, since there will not be beam
injected at 2 am. If a store aborts on your shift
for one reason or another, the rest of that shift
is cancelled.
Question:
Is it necessary to do owl shifts?
Answer:
Yes, at least some fraction of them. We need to follow
whole stores from beginning to end with pots in to
establish reproducibility and stability of pot positions
and do significant data taking. Owl shifts tend to be
one of the most reliable periods of time, and end of stores
when we can do additional pot tests will frequently occur
in the last few hours of owl shifts.
Question:
Can I combine my shift with another shift?
Answer:
Eventually this could be combined with Captain or Detector shift, but
not until operations are much more routine (perhaps December?).
DAQ shifts are not candidates for combining shifts as they require
100% of your attention.
Shift plan:
1) Ensure adequate shift training. Please arrange with Mike Strang
tomorrow for shift training (one at a time). Anyone who doesn't
get trained tomorrow should get trained Monday. This includes going
through all procedures, checklist, ACNET, etc. Many of you have
significant experience already putting in the pots, but I want
everyone to go through this training--if you're already an expert
you can spend the training trying to find shortcomings in
documetnation. I would say Mike S.+ Jorge are esentially
trained. Pierrick and Christophe need a little more work
(sorry C but you've been away). Victor, Sergio,
Mike M., Eugene and I need quite a bit more work.
2) There is still DAQ work necessary. Preliminary plan would
be for Mike, Victor and Pierrick to meet at 1:00 Friday to
insert pots, and evaluate DAQ and trigger. Mike will be gone from
2:30-4. At 4:00 Jorge can stop by, and we can keep pots in until
midnight. This will be first official shift. Note if data
being taken is not yet useful, this shift could end around 8:00 pm.
3) Sat at midnight, or when people earlier if people available--pierrick
are you available any earlier?--pots will be inserted by
Mike, Victor, and Pierrick in order to continue evaluating
DAQ and trigger. Mike can stay until 8 am taking data.
Don't forget pedestal runs and to turn off control lines.
At 8 am Sunday Jorge can take over, Eugene should come on
this shift for further training.
4) Sunday 4-12 will be Christophe if available (that would
be 24 hours of pots in). Owl shift would not be held.
Pots would be removed and reinserted next morning.
We would request end of store studies for store starting
Monday morning. I will need details of weekend by 8:00 am
so I can talk to Dmitri before 9 am meeting to request these
studies.
5) Pierrick will be on Standby for owl shifts Mon/Tues
and Tues/Weds if they may be necessary for pot insertion
studies. On Monday I will try to schedule the shifts for
the week.
Friday, September 14, 2001
Shifts
So the plan through Monday is as follows
Store will go in late morning early afternoon today.
Mike S. will meet with people today for training (except 2:30-4:30).
When Mike gets back at 4:30 he will put pots in along with
Jorge. Mike is updating run plan which includes plateau curves
and rate/trigger studies on 4-12 shift along with some DAQ studies
if Victor available. Pierrick and Mike and perhaps Victor will
continue studies on owl shift and take data.
Shift leaders for this weekend (confirmed)
Fri 4-12 Jorge +Mike for pot insertion +Victor (daq)?
Sat 0-8 Mike +Pierrick + Victor (daq)?
Sat 8-4 Christophe
Sat 4-12 Jorge
Sun no shifts
Tentative shift schedule for next week (Mike M. and Christophe
unavailable) we will discuss Monday morning
please send comments prior to then
Mon 10-12 FPD phone meeting Salle D0
Mon 11-4 Sergio + Mike M. for pot insertion
Mon 4-12 Eugene (unconfirmed)
Tues 0-8 Pierrick
Tues 8-4 Victor
Tues 4-12 Jorge
Weds 0-8 Pierrick
Weds 8-4 Sergio
Weds 4-12 Eugene
Thurs 0-8 none
Thurs 8-4 Andrew (if possible)
Thurs 4-12 Mike
Fri 0-8 none
Fri 8-4 Andrew
Fri 4-12 Jorge
Sat 0-8 Mike
Sat 8-4 Victor
Sat 4-12 Eugene
Sun 0-8 Mike
Sun 8-4 none
Sun 4-12 none
Friday, September 14, 2001
Final version of pot insertion agreement (fwd)
This agreement applies to Roman pot insertion from the period 9/14/2001
to 10/8/2001.
D0 plans to move the Roman pots in to standard positions most stores
not earlier than one hour after the beginning of the store.
The MCR and CDF will be notified before pot motion and when it is
concluded. The initial positions would be A1U=31.5 mm A1D=33.5
P1U=26 P1D=26.5 with A1=A2 and P1=P2 (previously determined as safe
positions), with the beampipe axis roughly at 45mm. Pot insertion
will be done cautiously while monitoring losses allowing time to
react to an unanticipated change in beam orbit. D0 should be notified
of any planned changes in beam position. The pots will remain at these
positions for most/all of the store retracting the pots prior to the
beam being aborted or machine studies.
Towards the end of the second store (t-4 hours) and some subsequent
stores we plan to do studies to attempt to define closer operating
positions (including initial dipole pot operating positions),
since our acceptance is critically dependence on this.
We would like to do these before any BD studies, which will require
us to have the pots out and would render our default pot positions
unreliable. Approval from BD run coordinator will be obtained
prior to these studies. An acceptable new position for now is defined
as one that does not increase the D0 losses by more than 10% and the
pots will be retracted if CDF complains about high losses related
to pot motion.
Friday, September 14, 2001 5:00 PM
Shutdown plan/manpower + task list
PRIMARY JOBS FOR OCTOBER SHUTDOWN:
=====================================
I) Installation and testing of eight new detectors
1) detectors
2) pseudodetectors
3) amplifiers+cables
Coordinator: Mike Strang
Manpower: Mike, Molina, Victor
Key issues:
1) Purchase of low voltage supplies and any other parts
2) Availability of final two detectors being constructed at UTA
3) Zip bags needed?
4) L0 signals--new cables needed?
Status:
Might be okay here. Loss of Eugene and Hans makes manpower a concern.
Needs concerted effort before shutdown to have everything ready
(more manpower here would free up some for other
efforts)
II) Pot motion upgrades
1) Installation of multiplexor
2) Software upgrade
3) IIB debugging/RM debugging
4) New 5V control circuit
5) Pot motion testing
Coordinator: Mike Martens
Manpower: Mario+Newton?, Pierrick?
Key issues:
1) Working multiplexor
2) Understanding of RM/IIB noise
3) Construction of 5V circuit
Status:
Don't have sufficient or sufficiently trained manpower
to complete this task
III) L1 Trigger and D0 integration
1) installation of AFE interface and AFE boards
and rack preparation in PW08
2) installation of DFE boards
3) testing of chain AFE through TM
Coordinator: Pierrick Hanlet
Manpower: Jia Li, Molina
Key Issues:
1) final support design
2) TPP boards
3) availability of AFE boards
Status:
Uncertain, dependent on upgrade hardware, TPP board layout,
rack prep, if these are done manpower may be adequate
IV) Beam pipe scintillator
Coordinator: Mike Martens?
Manpower: Carlos Avila?,Lab 7, Mike McGee
Key Issues:
1) Approval
2) support structure for scintillator
3) cables and access
Status:
Given manpower situation this is unlikely to get done
MANPOWER:
===========
Manpower Available
Full-time
Mike Strang
Mike Martens (after Oct. 1)
Pierrick Hanlet
Jorge Molina
Victor Bodyagin (0.7)
Half-time
Sergio Novaes
Christophe Royon (after Oct. 1)
Part-time
Jia Li
Occasional
Brian Cox
Andrew Brandt
Possible if funding
Eduardo Gregores
Needed, status uncertain
Mario Vaz
Newton Oliveira
Gilvan Alves
Helio da Motta
Electrical engineer
Injured list
Hans Willutzki
Other sources of manpower?
Saclay?
Manchester?
Naeem?
A) Before shutdown:
================
I) Detectors:
1) testing and binning 62 MAPMT's and 18 or more L0 PMT's
location: clean room roof using ICD setup
manpower: Mike S. (should be done by Sep. 26 so tubes ready
for insertion)
issues: final adaption of test setup
2) mapping 8 detectors to be done in parallel (2 will be done
in October)
location: Lab 3
manpower: Mike S. (should be done by Sep. 26 so detectors ready)
issues: using new machine, availability of detectors
3) cartridge assembly
location: Lab 6
manpower: Mike S., Molina, Sergio
issues: holes in cartridges for L0 LMB, cartridge modification
for shielding, have all parts?
4) signal cable tests for 10 detectors
location: Lab 3
manpower: Molina
issues: none
5) detectors into cartridge
location: Lab 3
manpower: Mike S.
6) PMT's into cartridge
location: Lab 3
manpower: Mike S., Pierrick
7) Assembly of cartridge top including splitters
location: Lab 3
manpower: Mike S., Molina, Pierrick
issues: purchase of parts
8) Amplifier testing
location: SCR
manpower: Victor+Shift people
issues: reliability, better testing procedure, DC offset
8) Screw machining (Martens)
9) LMB preparation and grey cables (Pierrick)
Need more restoration circuits?
10) Ribbon cable supports for tunnel? who? (Pierrick+Jia)
Note: See Mike S.'s mail for more detail
Note: Manpower seems a bit thin, given other work, data taking shifts
Pierrick available only after AFE interface design finalized
being built
II) Pot motion
1) multiplexor preparation
2) pre-prepare couplers
location: ?
manpower: Molina, Jia, or Newton
3) 5V scheme for control and emergency lines, to allow
software control (design Mario) preparation (Molina Christophe)
*can this be done in Brazil???
4) 10+15V adapters for amplifier monitoring
manpower: Pierrick,Molina (computer upgrade)
5) program for multiplexor adaptation
manpower: Eugene? Helio? Fritz?
issues: can this be done before shutdown?
6) IIB tests, sort+repair?
manpower:Pierrick,Mike M.,Newton/Mario?
issues: shielded cables from IIB to RM???
other IIB mods?
III) Trigger (Pierrick, Jia, +?)
1) AFE board tests
2) Definition of TPP
3) Construction of AFE crate, supports, cable strain relief
4) Test and select flex cables and connectors
5) prepare LVDS cables+cable from SVX to data stream
6) DFE tests
7) sDAQ (Victor+Fritz?)
8) prepare pDAQ
IV) Veto counters
1) Design, build and test veto counters
2) Design and build veto counter supports
3) Prepare veto cables
B) During Shutdown:
I) Detectors
1) Install and test detectors and pseudodetectors
2) Install and test LMB's (including last 2 green cables)
3) Install and test low voltage supplies
4) Install and test amplifier boards and cables
5) Install ribbon cable supports
6) Cameras for bottom pots
7) Replace patch panel to control room with cables
8) HV tests
9) Verify cable lengths (modify them?)
II) Pot motion
1) Install and test multiplexors
2) Install and test couplers
3) Install and test 5V system (test Emergency line)
4) Test and replace IIB's and RM's
5) Install and test circuit dividers
6) Enable in and out pot motion
7) Update software to monitor power supplies, for 5V,
multiplexor,database
III) AFE
1) Install and test AFE interface
2) Test trigger chain as much as possible
3) TDC boards?
IV) Veto
1) Install and test veto counters
2) Lay cables
*** Prepare patch pannel for LMR 400 cables? In MCH1? where?
C) After shutdown:
===================
1) Finish trigger +LM
2) Database
Priority issues:
1) finalize TPP--signal shapes?
2) which LV supplies+purchase
3) multiplexor and manpower help
4) zip bags? test with Marvin
5) cartridge holes
****************************************************
Tasks for week of 9/17-9/24
At Fermi:
1) testing and binning 62 MAPMT's and 18 or more L0 PMT's
manpower: Mike S.
issues: final adaption of test setup
2) mapping 6 detectors to be done in parallel
manpower: Mike S.
issues: using new machine, availability of detectors
3) cartridge assembly
location: Lab 6
manpower: Molina+Sergio
issues: holes in cartridges for L0 LMB, cartridge modification
for shielding, have all parts (HV wire+some splitter
connectors ordered?
4) signal cable tests for 10 detectors
location: Lab 3
manpower: Molina
5) Screw machining Martens
6) AFE stuff Pierrick
In Brazil
1) Multiplexor
2) 5V?
__________________________________________________________________________
Saturday, September 15, 2001
Sept 16 - Sept 15 runplan results
Jorge + Mike
We move the P1U, P2U, A1D and A2D pots. We insert the PU pots to 26 mm and
the AD to 33.5 mm without problem (we checked that there were no
changes in the halo rates with the scaler in D0 and with the lumberjack
plots in B0).
Jorge:
Once the pots were inserted we took the single rates of each pot and they
were: (in kHz)
P1U P2U A1D A2D P_Halo A_Halo
47.0 66.5 49.9 23.5 25.1 1.8
Then we measured the rates of the early part and they were (in kHz)
EP1U EP2U EA1D EA2D
5.4 0.74 10.8 1.3
The rates of the coincidences were (in kHz)
P1U.P2U A1D.A2D EP1U.EP2U EA1D.EA2D
11.7 4.6 0.17 0.3
Then we delay the early signal in 147ns with a width of 90ns and took the
rates of the following coincidences (in Hz):
P1U.EA1D P2U.EA1D PI1.P2U.EA1D P1U.EA2D P2U.EA2D P1U.P2U.EA2D
2800 2900 2600 85 86 49
AID.EP1U A2D.EP1U A1D.A2D.EP1U A1D.EP2U A2D.EP2U A1D.A2D.EP2U
150 90 26 22 10 3
Victor:
Took a pedestal run and found that two ADCs (slots 15 & 16) are noisy. He
will come back in the morning during Christophe's shift and try to get DAQ
running with the trigger determined by Mike and Pierrick.
Mike + Pierrick:
Jorge had problems getting the vetos on the early signals working. After
discussing the problem, we decided to verify the information we are using.
We found that the pulse widths of the L0 signals in the discriminator were
different for each tube and set them to a standard 15ns width. We also
determined that the CLK and ECLK signals we are using to determine early
and intime signal was not terminated properly in the module and terminated
them.
We verified that the intime and early clocks are delayed by 182ns with
respect to each other.
We tried to verified that the signals within the spectrometers arrive
within the 60ns width of the clock signals (except for A2D which has too
small of a rate). Based on an email from Carlos, to properly time in the
early signals for vetoes, we increased the width of the clocks to 80ns
and reverified the timing.
We verified that the timed signals are 55ns wide and that the coincidences
have a 25ns overlap.
During this test, we found that one of the outputs from A1D early timing
didn't work and put a terminator on it so it won't be accidently used. The
coincidence module that was thought to be broken, does in fact work, it
was the problem in the early module that caused the problem.
We verified that the width of the coincidence signals is 50ns.
We looked at the Luminosity Monitor signals. They look ugly (reflection
problem in the MCH??) Pierrick thinks they are still sufficient to perform
a veto.
We recalculated the rates that Jorge had measured earlier (since rates
decrease with time and it has been ~10hrs or so):
Singles (kHz)
P1U P2U A1D A2D
38.3 53.0 38.3 19.1
Intime (kHz)
P1U P2U A1D A2D
32.5 41.7 24.8 17.4
Early (kHz)
P1U P2U A1D A2D
3.6 8.1 8.6 24.9Hz
Coincidence
P1U.P2U = 9.9kHz
A1D.A2D = 3.6kHz
EP1U.EP2U = 195Hz
EA1D.EA2D = 5.6Hz
Verify veto
P1U.P2U.EA1D = 2.2kHz
P1U.P2U.NOT(EA1D) = 7.8kHz
P1U.P2U.EA2D = 0.3Hz
P1U.P2U.NOT(EA2D) = 9.8kHz
P1U.P2U.(EA1D or EA2D) = 2.1kHz
P1U.P2U.NOT(EA1D or EA2D) = 7.8kHz
A1D.A2D.EP1U = 18.3Hz
A1D.A2D.NOT(EP1U) = 3.53kHz (A1D.A2D = 3.54kHz)
A2D.A2D.EP2U = 17.7Hz
A1D.A2D.NOT(EP2U) = 3.57kHz (A1D.A2D = 3.59kHz)
A1D.A2D.(EP1U or EP2U) = 31.8Hz
A1D.A2D.NOT(EP1U or EP2U) = 3.43kHz (A1D.A2D = 3.46kHz)
Elastic rates
P1U.P2U.A1D.A2D = 81.1Hz
P1U.P2U.A1D.A2D.NOT(LUMN.LUMS) = 20.2Hz (raw elastic = 76.4Hz)
We needed to delay the LUM signals by ~50ns using cable since the delay
generator produces a min delay of ~60ns. We also narrowed the width of the
raw elastic signal and delayed it by ~16ns further because of some jitter
in the leading edge.
The full elastic trigger we came up with is:
P1U.P2U.A1D.A2D.[NOT(EP1U or EP2U) or NOT(EA1D or EA2D) or NOT(LumN.LumS)]
which gives a rate of 20.2Hz (raw elastic = 77.3Hz) so the primary
reduction in rate is the luminosity signals.
It appears that the trigger barely opens the ADC gate in time for the
MAPMT signals. There might be a slight loss in ADC counts from the
integration of the leading edge but the vast majority of the MAPMT pulse
lies within the gate.
Christophe:
Plateau tubes.
Using 4-fold / 3-fold elastic (no vetos since all logic is being used)
Using 30s integration window for good statistics
__________________________________________________________________________
Christophe Royon
Saturday, September 15, 2001
Plateau curves
All pots are now at home position. The control line is off and HV on L0
PMT has been set to 50%. Tomorrow, Victor will go on testing the DAQ and
will page me if we need to put the pots in.
Concerning the run, we had a few alarms:
- a few fake smoke detector alarms at various places (not transmitted to
firus)
- some communications errors with FPD-RM-01-02 and FPD-RM-02-12
- database entry problems (this has to be checked) ORA-01438 value larger
than specified precision allows
As a comment, I think we should also add the date and time to be desplayed
in the logbook messages to know exactly when the alarm took place.
This morning, I tried to get the plateau curves for P1 up, P2 up, A1 down
and A2 down. The idea is always to plots the ratio of the two following
triggers if we study for instance the plateau curve for P1 up:
- P1 up .and. P2 up .and. A1 down .and. A2 down
- P2 up .and. A1 down .and. A2 down (everything except P1 up)
as a function of the HV in P1 up
Here are the results:
column 1 : HV
column 2 : ratio when HV on P1 up is changed
column 3 : = P2 up =
column 4 : = A1 down =
column 5 : = A2 down =
HV | P1 up | P2 up | A1 down | A2 down
=======================================================
1000 | 0.004 | 0.000 | 0.003 | 0.000
1050 | 0.008 | 0.001 | 0.007 | 0.000
1100 | 0.010 | 0.004 | 0.012 | 0.000
1150 | 0.016 | 0.011 | 0.013 | 0.001
1200 | 0.022 | 0.023 | 0.024 | 0.004
1250 | 0.029 | 0.034 | 0.028 | 0.006
1300 | 0.040 | 0.051 | 0.036 | 0.015
1350 | 0.050 | 0.075 | 0.045 | 0.021
1400 | 0.057 | 0.094 | 0.064 | 0.030
1450 | 0.077 | 0.119 | 0.067 | 0.041
1500 | 0.089 | 0.142 | 0.092 | 0.057
1550 | 0.103 | 0.169 | 0.102 | 0.070
1600 | 0.112 | 0.185 | 0.121 | 0.088
1650 | 0.140 | 0.216 | 0.147 | 0.108
1700 | 0.164 | 0.235 | 0.185 | 0.122
1750 | 0.169 | 0.246 | 0.222 | 0.137
1800 | 0.169 | 0.242 | 0.233 | 0.156
1850 | 0.176 | 0.261 | 0.241 | 0.172
1900 | 0.177 | 0.262 | 0.233 | 0.193
1950 | 0.177 | 0.275 | 0.239 | 0.216
2000 | 0.188 | 0.292 | 0.243 | 0.236
I append as a PS file the result of the plateau curve. We can see clearly
that A2 down does not show any plateau yet, which is strange... And the
plateau for P2 up is not clear at all.
Victor Bodyagin
Tuesday, September 18, 2001
Some daq news
after Pierrick assembled fast trigger logics back I could make some
pedestal studies. Since the store was lost I was running "peds" and "daq"
with ribbon cables unplugged and plugged in while HV were off. I did it
few times in order to check stability of the pedestals' values. I'm going
to place the results into Logbook.
The conclusions are as follows:
1. When the cables are unplugged ADCs have very uniform, stable and
relatively low pedestals some ~40-50 counts, which gives us hope that ADCs
are OK.
2. When the cables are plugged in the pedestals have variation from 0 (two
channels) to 640 (one channel). Obviously, this effect is produced by
Amps. Earlier Mario has commented these variations in the same way. I
agree with him. Nevertheless, from one run to another pedestals looked
rather stable (variation ~2-3 counts).
I've activated the option of subtraction of pedestals at software
level while data taking. It will give us gain in disk space by factor
of ~2-2.5. I think with this option "ON" we can run safely without
loosing data.
Saturday, September 22, 2001 4:06 PM
Day shift report 09/22/01
I've introduced 192 new histograms into DAQ. Now one can see each
ADC
channel by channel. To see the results go to /home/d0fpd/daq_plots
and ghostview the files: adc*_all_092201.ps. The lp outputs of these
files are attached to the checklist in the folder.
For better understanding of our problems see also files:
/home/bodyagin/d0/fpd/camac/daq/socket/vers01/CUURRENT_PEDESTALS
/home/bodyagin/d0/fpd/camac/daq/socket/vers01/peds/UNPLUGGED_PEDESTALS0
or
/home/bodyagin/d0/fpd/camac/daq/socket/vers01/peds/UNPLUGGED_PEDESTALS1
I would be happy to discus these results with everybody.
Also I've stored some 20k elastic events in file:
/home/bodyagin/d0/fpd/camac/daq/socket/vers01/data/run000045.dat
Before taking data I've done a new pedestal run because since last night
pedestals of the first two ADCs shifted and data looked a little bit
"dirty".
Please, note that Pierrick made a new HV settings for L0s.
Saturday, September 29, 2001
Shift report
There is nothing much good to report about this shift. Due to the
quench in Q3 there was no beam during all shift. I was studying
pedestal problem for our FERA ADCs. The typical picture for all
ADCs can be seen in the file:
/home/d0fpd/daq_plots/ADC00_single_ped_plugged.ps
A few comments on this problem.
1. It does not look to be a problem of fast trigger, since ADC gates
had stable width.
2. It does not look to be a DAQ problem, I mean readout, packing,
unpacking and data processing.
3. The amplitude of the noise coming from Amps can explain very
well the width of the peaks (10-15 counts) but not the distances
between them which vary from 25-70 counts for different ADCs.
The growth of each particular peak is a periodic process and
simultaneous for all ADCs. It looks we have a grounding problem
unsolved.
Monday, October 01, 2001
Data taking on Oct. 1
I've taken ~10k events with zero subtracted pedestals and MAPMTs HV off
for P1 and P2 in order pedestal distributions could be studied offline.
Data are in run50.dat file.
Then I've made pedestal run after that I've downloaded new pedestals and
have taken ~40k events of elastic data. These data are run51.dat file.
Also I've taken ~10k events of elastic data with 1KHz generator off.
These data are in run52.dat file they look a lot more physical, since
in this case we have really physical trigger,
i.e. P1U*P2U*A1D*A2D*nonVeto.
MAMPT HVs OFF.
Pots are at home position. Control lines are OFF.
L0 HVs %50 (Standby).
Daq is running.
Report sent to shift captain.
__________________________________________________________________________
Christophe Royon
Friday, September 28, 2001
Cable testing finished
This afternoon, Sergio and myself finished testing the short ribbon cables
at LAB 3. Most of them are found to be good and about 15% are found to be
bad. 5% had no connector at all (namely, they were put back in the box
without being done, may be because the cable was found to be bad by Bob
Johns, I do not know and have no way to test this since there is no
connector), and 10% had at least one wire not connected or grounded. These
bad cables were all labelled as bad by putting orange tape around them (it
is indeed difficult to know which channel is bad, since the numbering is
reversed for one connector compared to the other), and were put in a box
labelled BAD. The good cables are put in three different boxes at LAB 3,
labelled as GOOD.
The tester will be given back to Bob Jones on Monday if this is ok with
you.
Saturday, September 29, 2001
Shift summary
This shift was not very good. I did not manage to put the pots in, before
the shot was lost at 7:40 (quench at B1). The scalors are not working at
present (there is no longer any cycle of 10 s.), and Mike will try to
reset them.
The DAQ has always some problems (see the mail of Victor) and we do not
understand yet why the pedestals are so weird. This might be due to noise.
Monday, October 01, 2001
Python program changes
I made a short list of the changes we could make in the python
program. Please let me know what you think of them and if you have other
changes to be made.
LIST OF CHANGES FOR THE PYTHON PROGRAM:
=======================================
1) Add the time and day information in the alarm window to know
when it occured
2) Change the name of the RESET window
3) change kHz in Hz
4) put a new alarm window when wanted position is greater than
the software posiiton and say it is impossible
5) put lock or unlock in the expert program as in the HV
program
6) when the program had to be restarted or a stop command has
been issued, it should not be possible to enter a new position
except home
7) multiplexor
__________________________________________________________________________
Pierrick Hanlet
Thursday, September 20, 2001
Drawings
Hi Mario,
I've put all of my drawings on my web page:
http://www-d0.fnal.gov/~hanlet
Look under D0 Present Involvement:
I've catagorized the TPP issues into 3 catagories:
1) Pulse Shape into AFE's:
Present plan includes setting up Pulser -> Amp/Shaper -> AFE
We will look at: a) present pulse shape, if that doesn't work
b) modifying to get asymmetric shape, if that doesn't
work
c) raising gain by 10X and using diodes on OpAmp
2) TPP Circuit:
I have a suggested circuit from Anderson which has a pulse transformer
and another suggestion from Johnson using a resistor and capacitor in
parallel to ground
We need to figure this out pronto
We need to get the design approved by the ECB
3) TPP Structure:
Mario's or Pierrick's layout
Mounting TPP
Stiffening TPP
What do we do about bulkhead connectors - shall we sand down the tabs?
If we use Anderson's suggestion, do we make multiple boards?
Strain relief of our coax ribbon cables
Monday, October 01, 2001
Fixed/modified PMT test stand
How to run the tests for the L0 tubes.
Steps for use:
1) turn off HV (use HV supply on table left of computer monitor) and unplug
PIN/Diode cable in 1st ADC, second channel - do these things for taking
peds
2) remove rubber band, put new PMT in cookie, and put on rubber band
3) at prompt type "cr2249_1"
4) give name of PMT
5) a couple of CRs until start taking peds
6) plug back in PIN/Diode in channel 1 (counting from 0)
7) set HV=1500
8) type "1500" when prompted
9) CR until start taking data
10) when asked if you want to do another HV setting, answer Y or y
11) repeat steps 7-10 with different HV settings. I propose that we
take readings for HV=1500 to HV=2100 in 100V increments
12) when finished answer N or n to "another HV setting"
__________________________________________________________________________
Wagner de Paula Carvalho
Thursday, September 20, 2001
Plateauing tubes in the tunnel
Today in our local (Rio) meeting, Santoro well pointed out that
plateauing the L1 PMT's using elastics is not quite the same what we
usually do in a testing stand.
In a testing stand we set a two-fold coincidence (A*B) to trigger
the signal (usually a cosmic) and count how often the triggered particle
is also registered by a third counter C (A*B*C). The reason for the A*B
coincidence is to get rid of electronic noise on both tubes, A and B.
When we try to plateau tubes in the tunnel the arrangement is a
bit different. To study, let me say, P1U efficiency we count how many times
a P1U signal is in coincidence with a P2U signal - presumably due to the
passage of ONE particle through both - and also in time with a two-fold
coincidence in the AD arm (A1D*A2D) associated with the passage of ANOTHER
particle.
But note the conceptual difference. Apart the AD signal which
prompts us to believe the PU signal is due to the elastics counterpart,
there is no guaranty that the P2U tagging signal is really due to a
proton crossing the scintillator and not due to noise in the phototube.
That is something to think about. Maybe this plateauing procedure
works out but maybe that's just the cause of the strange "plateau" curve
we have observed for some tubes in recent attempts. Remember, the higher
the voltage, the higher is the noise and higher is the probability to have
a coincidence with some signal (real or noise) from the other tube.
An alternative would be using the approach previously suggested by
Pierrick of using the integrated signal of the MAPMT in coincidence with
one of the PMT's to tag the particle and then using this particle to infer
the efficiency of the other PMT.
__________________________________________________________________________
FPD Group account
Thursday, September 20, 2001
Pot insertion
9/20 19:03
D0 Lum = 5.3E30
P-Halo = 28.5kHz
A-Halo = 2.3kHz
Home Rates (kHz)
P1U = 23.3
P2U = 51.8
P1D = 20.7
P2D = 61.8
A1U = 0.0 (this tube has always had a low rate that doesn't appear to
change as the pot is inserted. This pot was inserted in the last run (the
rate stayed at ~30Hz both at home and inserted), but
I don't feel we should insert it until we replace the tube, Looking back
over
the logbook, the tube was fine before it was modified for mu-metal shield
but after the modification it has not given reliable rates)
A2U = 41.9
A1D = 32.8
A2D = 38.7
__________________________________________________________________________
Michael Strang
Sunday, September 23, 2001
Results of 23/Sep/01 Owl Shift
- Stats at start of Shift:
D0 Lum: 2.69E30
P-halo: 26.7kHz
A-Halo: 2.4kHz
Elastic Rate: 33 Hz
- I found that I could not open any X-windows from D0mino to the online
machine. After some research, I found that it is possible if you log on
to d0mino using ssh, but I couldn't log on with any of my passwords.
- I found many instances of netscape running on the machine. I killed
them. This could be because of the way I setup the run_checklist
webpage. I'll try to fix it on Monday.
- I modified the unpacking routines from Carlos to place the ADC channels
in variables named after the proper detector frames. I verified that the
unpacking appears to work for the most part....the first channel of all
ADCs appears to have no information and for each buffer I process I get
the error:
unpck1event: unknown event type=256
I will try to discuss this with Carlos on Monday. The Histoscopes do not
have this problem with the first channel.
- I found that paw exists on the online machine, so I was able to look at
the ADC distributions from the data. In some cases I think I can see a
separation between signal and pedestal/noise but it is not always
clear. It appears that the majority of our noise is coming in from the
amplifiers/cables and I think we need to seriously address this issue in
the shutdown to improve the quality of any data that we take.
- I was unable to move my paw ntuple filter file over from d0mino to the
online machine using any permutation of ftp that I could think of so I
was unable to attempt to reconstruct hits in the detectors, I could only
look at the adc distributions for each channel.
- Looking at the ADCs, I notice that they have 17 pins while the cables
have 16 channels. Some ADCs missed the top pin and others missed the
bottom pin. I stopped the daq program and moved all the cables to miss
the top pin. When I restarted the daq program, the histoscope would not
work. I then moved the cables so they all missed the top pin and the daq
program started working again. What is the proper way to plug in the
cables?
- I remember hearing that the MAPMTs plateau somewhere near -930V or
so. We have been running them at -900V so I took a couple of data runs
at different MAPMT voltages for comparison (~10k events each).
run000047.dat (MAPMT at -920V)
run000048.dat (MAPMT at -930V)
run000049.dat (MAPMT at -940V)
- Stats at end of shift:
D0 lum = 2.29E30
P-halo = 25.6kHz
A-Halo = 2.2kHz
Elastic = 25Hz
- I have turned off the voltage to the MAPMTs. I have removed the pots to
home and turned off the control lines. I have turned the L0 tubes to
50%. I've had the shift captain inform the MCR that the pots are
removed.
Sunday, September 30, 2001
Shift summary owl Sept 30
There were some Tev studies this shift but they just entered shot setup
at ~07:30. Probably be able to declare the store and insert pots by
12:00 assuming nothing else goes wrong (if anyone is available).
I fixed the scaler problem Christophe mentioned during his shift -- the
pulser had been disconnected from the fanout.
I made some webpages that Sergio had been bugging me about.
I retested the two MAPMTs that Jorge said were bad and found them to be
fine. Must have been a problem with the connection of the cable to the
tube.
I tested the remaining 42 tubes. Out of 63 tubes we have 63 that give
signal out of each channel.
All of the tubes were tested at -800V and at -940V. Two of the tubes
were tested in -10V steps between these values.
I performed a cross-talk test by disconnecting all but one of the 16
channels in the cookie and then looking at the output from the ADC.
There is a 1-2% cross talk between adjacent channels on the face of the
tube. This is the consistent with the documentation of the tubes.
I will take the information gathered and try to look at it later today
after I wake up to see how we might want to bin the tubes.
If anyone is available today for a few hours (Jorge, Christophe,
Victor), it might be worthwhile to insert the pots for at least a few
hours and try to collect some data. It will probably still be crap until
we can figure out the pedestals but then again, maybe it won't.
__________________________________________________________________________
Jorge Molina
Wednesday, September 26, 2001
Cartridges
This is the status of the cartridges,
In the Tunnel: 10 Top cartridges 8 w/pseudo detectors and 2 w/MAPTMs
12 Bottom cartridges 2 fully equiped and 8 w/pseudo dets.
In Lab 3: 6 Bottom parts, 2 w/ fibers already, 4 prepared to put the
detectors in.
8 Top parts
In Lab 6: 1 Top part with pseudo detector
In DAB 3: 1 Top part with pseudo detector
So in total we have:
- 20 Top cartridges, 2 w/ fibers, 10 w/ pseudo detectors, 8 to prepare
(by now)
- 18 Bottom cartridges, 4 w/fibers, 14 to prepare (6 by now)
Saturday, September 29, 2001 1
Report of evening shift
This afternoon we insert the pots with Victor in the two steps procedure
and we could do it without any problems. It accelerated a lot the
process of inserting pots so now it takes less than 40 min to insert the
PU and AD until final positions.
I could not take any data because Victor found a problem with the ped
distribution (more information coming from him).
So I spent almost all the afternoon measuring the MAPMTs, I measured 21
of them and two presented problems of giving no signals from channels
8-16. The rest of them were good. I made two measurements in each
MAPMTs: one with 800 V and the other with 940 V.
At the end of the shift we had a problem (with Victor) with the new NIM
crate that were ecpelling a burned smell. Finaly then we had a current
overload in the right rack that turned off the computer, power PC, CAMAC
crate and all the equipments pluged to that power box.
We could restart the computer and programs without any problems. The pot
motion program wanted to retract the pots to home position but the control
line forbided him.
Now Pierrick is gonna work with the pots in and tomorrow Victor will try
to fix the problem with the peds.
__________________________________________________________________________
Michael Strang
Friday, September 28, 2001
Results of fast insertion studies
We inserted P1U and P2U to 24.0mm fast. Then to 26.0mm slow. We then
inserted P2U to 28.0mm slow.
We didn't notice any significant changes in any losses. There were some
momentary (one data point) spikes in CDF pbar losses but they appear to
have had these losses before we started moving any pots.
With the P's in position, we moved A1D and A2D to 31.5mm fast. Then to
33.5mm slow and then A2D to 35.5mm slow.
Once again we didn't notice any significant changes.
In effect we turned a 60min process of insertion into a 15-20min process
with no noticeable difference between the two methods.
I suggest that for normal insertion we start with PU using the fast
insertion. If everything looks normal we continue with fast insertions
for the others. If we ever notice something strange, we revert to the
conservative insertions for the remainder of the locations.
Speaking with Victor, as of ~midnight last night, the DAQ was fixed. We
tried to collect some quick data this morning before beam studies but I
was a dohead and forgot to turn the HV on for the MAPMTs. So we
collected some pedestal data. :) Then beams division wanted to begin
their studies so we removed the pots, turned off the control lines and
turned off voltages to all tubes.
__________________________________________________________________________
Mike Martens
Friday, September 28, 2001
I have a few times when there won't be power in the tunnel. C4 won't
have power from 15:00 - 16:30 on 10/19 and D1 won't have power from
15:00-16:30 on 10/22. There is a 30 minute power outage at 0700 on
10/25, and I think a power outage at 0700 on 10/8.
There may be a few days of Tev power supply testing during the last
week. Tentatively these are penciled in on 11/14 and 11/15 but I doubt
our schedule is that really that fine tuned.
__________________________________________________________________________
Dmitri Denisov
Tuesday, October 02, 2001
D0-BD meeting
Dear All!
Below is brief summary of today's meeting between D0 and BD about beam
position at D0:
1. BD studies demonstrate that beam is in the center of D0 low beta and
there are no aperture restrictions in the vicinity of the D0 detector.
2. Available range of beam position and angle adjustment with respect to
current values are +-1mm and +-0.5mrad in both horizontal and vertical
planes.
3. BD and D0 agreed that there will be no changes to beam position at D0
done by the BD (unless mutually agreed).
4. D0 will move detector by 5.0mm up and 4.0mm horizontally (in direction
outside the ring) during first week of October shutdown.
__________________________________________________________________________
Andrew Brandt
Wednesday, October 03, 2001
FPD meeting minutes
1) everyone should read Christophe's shutdown plan e-mail
and make any comments of forgetten things to Christophe and me
by friday
we will plan 7 day 8 hour/day schedules with 1/2 shifts
scheduled from 8:30-12:30 an 12:30-4:30
so there is overlap for any tasks that might require continuity--if
someone is scheduled for 2 consecutive 1/2 shifts on same task
it goes without saying that the 2nd shift effectively becomes
1:30-5:30 so noone goes hungry!
2) everyone should send in blackout dates, when they are not available
including weekends, and also if they have other committments which
will take up some fraction of their time, with some level of detail
including any deadlines
Data taking shifts--we only have 10 1/2 shifts with beam left and plan
to cover all of them through 8 pm on Sunday
Current schedule and runplan--if Mike gets a chance he can
add this to the runplan (Mike where do you keep the runplan
data fileso someone else can edit it if you are not available?).
Store 1 4 standard pots in
End of Store 1 dipole pots in (with MCR permission)--any estimate
of positions Mike?
Store 2 assuming good data taken in store 1 and A1U behaving normal move 8
pots in instead of 4, note different operating positions for other
pots. 4 psides can get moved in at once, but more steps should be
taken for P1D after initial fast move, same for aside and A1U
Store 3 plateau any remaining pots, more data taking, 10 pots in
Store 4 more data taking
Thursday 8-16 Sergio (beam not expected until 12 or so and Sergio
DAQ shift, so will be available to keep a loose eye on data
taking only. He will page Christophe when stable beam is in
so they can schedule pot insertion (4 pots). Plan is to take
elastic no pedestal subtraction runs. Eduardo should be there to
observe insertion and should spend a little time with Christophe
on evening shift going through documentation and asking questions.
Victor should be around for data taking to make sure that
data looks ok. Prior to data run and about every 4 hours
pedestal runs should be taken (more often if any differences
are observed).
Thurs 16-24 Christophe with some Eduardo training
Take elastic no pedestal subtracted data
Fri 00-08 Pierrick
wants to do more plateauing+studies of data
assuming this is still first store, we will only have
4 pots in
Fri 08-16 Sergio
take data
Fri 16-00 Jorge (more training for Eduardo)
Sat 00-08 Mike S.
Sat 08-16 Victor
Sat 16-00 Eduardo
Sun 00-08 Pierrick if plateauing of last 4 pots and dipoles not done
he can do
Sun 08-16 Mike M.
Sun 16-20 Mike S.
Eduardo found 9 more tubes and was plateauing them today.
This is 14+9+12=35 total tubes. Last 12 will be tested on
Monday or Tuesdsay during day after detectors removed.
Christophe warns that some cartridges may be a little hot
should be checked with wallflower if counting rate is high.
Cartridge work continues. Isolated ground HV connectors not
in hand (ARGHH). Jorge will get update from stockroom tomorrow
and pass it on. Mike S. will pay for difference between
standard and overnight shipping if I have to buy these by
overnight on Friday.
If Mike is not back tomorrow, Eduardo and perhaps Jorge
can look into detector mapping. Jorge says one detector
is done by Pat Richards, who is sick.
Friday, October 05, 2001
Veto scintillation counters
Dmitri,
Let me give you an update on the veto scintillation counters
that I mentioned to you previously. I have met with Mike
Mcgee and Russ Rucinski to talk about mounting and installing the counters
on the beam pipe between end of D0 and the low beta quads.
Here is our plan:
1) On Monday and Tuesday Oct. 8+9 there will be access to the
regions in question for inspection, to verify that the drawings
accurately reflect the available space in the region. I understand
that there will be some surveying going on during this period,
so perhaps you could suggest a suitable time, I imagine 10 minutes
on each side should be fine for this. I would like to get
Mike Mcgee and Pierrick Hanlet to have a look at the area.
Also we need to inspect the signal and HV cable access routes
from the tunnel side (Mike Martens will look at this on Tuesday).
2) During the next month the counters would be built and tested.
We have the scintillator and tubes in hand. Phyllis Deering in
Lab 8 will machine the scintillator. The support structure will
be very simple and will be designed and built at UTA. The only
issue is bending the light guides to allow appropriate clearance,
and I am looking into this now. Also we will need 1-2 more HV
channels, which I will have to check into.
3) The counters would be installed Nov. 12 or 13. I talked to Russ
about this and he said it would be a simple matter to lay
scaffolding over the weekend after the detector is closed and
it should not be a problem to leave the shielding open during
this period. A couple hours on each side would be sufficient
to install the counters, and Mike Mcgee said that he can obtain
someone to help with this.
So the main issues from our point of view are
1) cable access
2) 2 extra HV channels
3) scintillator bending
Additionally, is any approval, safety reviews, or presentation of this
plan needed? From the BD point of view Mike Mcgee said there is no
problem as long as the counters weigh less than 30 pounds.
Thanks,
Andrew
__________________________________________________________________________
Sergio F. Novaes
Friday, October 05, 2001
FPD Shift 05/Oct/01 - 8h00-14h00
Readings at the start of the shift (kHz):
P1U = 63 / P2U = 69 / P1D =16 / A1D = 55 / A2D = 76 / A1U = 32
PHalo = 32.5
PbarHalo = 1.6
The 4 pots were IN at (cm):
P1U = 26.08 / P2U = 28.20
A1D = 33.03 / A2D = 34.52
High Voltage were (V):
L0PMT A1 / A2 / D1 / P1 = -1,800 / P2 = -1,600
All MAPMT's off.
Victor restarted DAQ and did pedestal run.
After that we put HV in the MAPMT P1U / P2U = -940 V and started run
Run #58 stopped with ~ 30,000 events
Around 13:00 we started (Christophe and I) to moved Dipole pots:
If PHalo = 32.5 / PbarHalo = 1.6 we considered "no loss"
D1 (kHz) D2 (kHz) Distance (mm) Velocity
Loss
7.5 0.4 15 Fast No
10. 0.3 20 Fast No
14. 0.5 25 Slow No
17. 1.0 27 Slow No
28. 2.4 29 Slow No
43. 4.5 30 Slow No
61. 5.2 30.5 Slow No
75. 7.9 30.8 Slow No
81. 8.6 31.1 Slow No
102. 11. 31.4 Slow No
129. 15. 31.7 Slow No
158. 29. 32.0 Slow No
766. 61. 32.2 Slow
Spike of loss @ 13:25
1,600 ??? 32.4 Slow Spike
of loss @ 13:30
when
D2 was moved
At 13:30 Beams Division started end of store studies and the loss increased
a lot.
The first loss (13:25) happened just in D0 p and pbar and returned to normal
after the spike.
The second loss (13:30) happened in both D0 and CDF and could/should be
related to BD studies.
We returned all pots back HOME.
HV is OFF.
Control line OFF.
Jorge, Eduardo and Victor.
October 06, 2001
Shift report evening 10/5/01
In this shift we inserted 8 pots to operating positions exept P2D wich
had a very high rate even at 1600 V.
So the position of the pots are now
P1U: 26.0 mm P1D: 26.5 mm
P2U: 28.0 mm P2D: 26.5 mm
A1D: 33.5 mm A1U: 31.5 mm
A2D: 33.5 mm A2U: 33.5 mm
When we inserted the P pots we saw an increase of ~8% for PUs and ~10% for
PDs.
For the A pots we saw an increase in the P_HALO of ~5% for ADs and for AUs
as well.
We took ~31600 events with pedestal substration and is in the 59 file run.
We also note that A2 camera is not working properly (we couldn't move it
but we could see the motor moving), so we have to check that in the next
access.
__________________________________________________________________________
FPD Group account
Saturday, October 06, 2001
Results of owl shift 10/6/2001
Beginning stats:
D0 Lum: 5.66E30
P Halo: 31.7kHz
A Halo: 2.3kHz
At beginning of shift, I spoke with Pierrick by phone and he described how
the signals going to ADCs 8 and 9 had been unplugged from the amp/shapers
and used to determine that the strange pedestal structure comes from the
amps. He proposed some studies, using the LMB to try to determine if we
can see signal from pedestal.
I performed a ped run, stored in yo1m10d06t0037.txt. It takes far too long
to collect pedestals.
I then found that the pulse generator for the LMB is also used for the
normalizing frequency of the DAQ scalers. Unfortunately, the two
frequencies vary by orders of magnitude. Therefore, I rewired the signals
so that the frequency comes from the UTA pulse generator. I set it to 1MHz
which is fanned out to Victor's scaler and to a prescaler that drops it to
1kHz.
The 1kHz is used for the visual scalers.
At this time, MCR called to say they were moving some collimators out from
the beam. There was no noticeable change in rates.
After rewiring I found that the visual scalers can act funky somtimes in
that they don't reset. If this is observed, change the reset value until
the scalers reset then return it to the value you actually want. After a
few times of being strange, it has been stable for the rest of the shift.
I set the LMB pulser to 112Hz w/ 16 microsec width and observed the PinOut
signal from the tunnel with a DC offset of 10V.
Then the fun began:
Victor had come in during Jorge's shift to start the DAQ with pedestal
subtraction. I wanted to do my studies without pedestal subtraction. I
followed the instruction I had been given by Victor to use the zero
pedestal file. I kept getting an "error reading pedestal file" until I
realized I needed to use ZERO_PEDESTAL_ADC12 to account for the extra 12
ADC channels to be used with L0.
However, this wasn't sufficient to get DAQ to work. I kept getting 1 of 2
errors regardless of the pedestal file I used:
"Error writing pedestal for slot 10, channel 0, data 2"
or
"Error setting SWR for slot 10"
The slot number or data value could change.
After a few hours of struggling with this, I gave up since Victor will be
on shift next and he can hopefully fix it. This did mean that I was unable
to perform the studies that Pierrick and I had talked about.
I took runs 50-59 and transferred them to the fpd project area on domino.
I decided to try to look at some of the data that had been collected
previously. Victor had left some instructions in my office for minor
changes to be made to the unpacking code which I did. The code compiled.
I went to unpack run59 collected during the previous shift, and during the
unpacking, all of the text on my xterm turned into something very similar
to klingon. I determined that if the run comment is longer than ~30
characters, this will happen. So don't put long run comments.
The data seemed to unpack. I still get the unknown event type error for
each buffer but this might be do to the fact that the spill[ptr+1] doesn't
correspond for the physics mask that was previously setup. Why it doesn't
match, I don't know.
I looked at the data and for Run 59, there was nothing until event 53. At
this point, I started to get data (including channel 1 which was missing
before). However, it still exhibits the same pathology of
identical repeated ADC counts in the same channel for many events in a
row. Also, every channel appears to be turned on with some number of ADC
counts. The TDC2 value is still overflowing.
I looked at Run 53 and it had some huge event numbers (it didn't start at
1) so I don't know what is up with that.
__________________________________________________________________________
Victor Bodyagin
Saturday, October 06, 2001
Shift report
I've taken two pedestals and five data runs.
Pedestals file: y01m10d06t0836.txt
Data: run60 - lmb off pampts at 940volts pedestals not subtracted 30k
events
run61 - lmb on pampts at 940volts pedestals not subtracted 20k
events
run62 - lmb off pampts at 900volts pedestals not subtracted 30k
events
Pedestals file: y01m10d06t1329.txt
Data: run63 - lmb off pampts at 940volts pedestals not subtracted 10k
run64 - lmb off pampts at 940volts pedestals not subtracted 20k
These two last files are the best. While taking 60-62 I was observing
a strange behaviour of ADC6. It had descrete distribution. Checking and
plugging of the ribbon cable did not help. It got back to normal work
by itself.
So, be warned!
Store #747 terminated, PAMPT HVs OFF, L0 HVs at 50%, Pots at home,
Control lines off.
All data files moved to scratch.
I have a good news: ADCs 7-13 (P2U), i.e. pedestals are narrow and
do not have double peak, moreover, the separation noise/signal look
promising.
I suggest to take as much data at 940volts as possible. Seems
we have a good chance to see correlated hits in P2 at least.
I recommend to take pedestal file each 4 hours.
I think we have enough of lmb data in run61 file.
__________________________________________________________________________
Pierrick
Sunday, October 07, 2001
Lost beam
Just as Eduardo and I completed inserting the pots, there was a quench at
TeV A1. We removed the pots and set the L0 HV to 1kV and MAPMT HV=0V.
I will continue studies with the LMB. Please note that it is imperative
that the next shift with beam takes data at low and high values of HV to
compare to LMB data.
Sunday, October 07, 2001
Pedestals
I was hoping that we could use LMB triggers to generate pedestals. This
would allow us to use the daq without running a separate program for
pedestals which required the daq to self-trigger. Instead, one leaves the
HV
on for the PMTs and uses the LMB with V=0V.
To verify that this works, I've generated the plots with MAPMT HV=800V and
HV=940V. I have plots for both ADC0 & ADC1 (P1U with Amp/Shapers) and
ADC8 & ADC9 (P2U w/out Amp/Shapers).
You can see the results on my web page:
www-d0.fnal.gov/~hanlet/d0_public/fpd/pedestals.html
__________________________________________________________________________
Michael Strang
Sunday, October 07, 2001
Progress on unpacking
In the unpacked data I've been looking at, I've noticed ADC channels
that do not change ADC counts for many events in a row.
Speaking with Victor, I realized that the way the unpacking works is to
only update ADC channels that actually have changed information from one
event to another and then pack them to save space.
I put some diagnostic code into the unpacking routine to dump out the
adc channels that were packed in a module as well as all the adc
channels for a particular adc module and found that the repeating adc
information corresponded to adc channels that were not packed in the
first place.
Therefore, I put some initialization (to zero) routines for my adc
arrays (as well as tdc and scaler) that I use to fill the ntuples and
this has removed the repeating adc problem.
I also verified that the unpacked information before putting it into an
ntuple corresponds to the information stored in the ntuple.
The tdc information is still messed up. tdc(1) looks fine with ~740
counts but tdc(2) either has ~4080 counts (overflow?) or in a few cases
~900 counts. I don't know what the scaler information is supposed to
tell me so I can't say if it is good or not but it does change from
event to event. And the trigger bit always gives me 128. I don't know
what that means.
Now as I look at a small sample (~100 events from run 59) of one frame I
see that in many cases we have ADC counts of under 100 in many channels
but one or two channels in some events with 200-300 ADC counts. I also
have events with only 1 channel turned on (as I looked through the other
frames for that event, it appeared that all the other frames had either
1 or no fibers turned on which could very well correspond to a single
particle passing through the detector but I have to check further and
see if it reconstructs properly using a modified geometry of the
detector).
Now I need to determine what sort of discrimination to put on the data
(it is already pedestal subtracted but due to the odd form of our
pedestal I'm sure there is still pedestal contamination that might ruin
any chance of reconstruction anything).
The dump files I'm looking at can be found at
~strang/fpd/dump
I'll try to look at the data collected over the weekend later today or
tomorrow.
__________________________________________________________________________
Michael Strang
Sunday, October 07, 2001
Run information
I've transferred runs 60-64 to d0mino.
All of the data collected to date is located at:
/prj_root/810/fp_1/FPD_data
unpacked HBOOK ntuple information will be stored in
/prj_root/810/fp_1/FPD_data/unpacked
if the data file was run000059.dat the unpacked file will be fpd59.ntpl
(unpacked files will be available as soon as any other bugs are removed
from the unpacking code). We also need to determine if we want to only
have raw data in the ntpl or if we want to apply some discrimination to
the data to clean it up first.
I feel we need to start keeping track of some basic information of each
run in order to make it easier to keep track of what is what.
We need a database that keeps track of:
Run #, #events, #spills, trigger(s), where it is stored, status of daq,
comments on quality and usefulness of data collected, if pedestal
subtraction is being applied
We also need to keep track of when, in between runs, electronics
configuration changes (for instance, cables plugged into ADCs 8 and 9
are now amplifier bypassed and inverted in the card effective only
sometime this week), which pedestal file is used for the data.
__________________________________________________________________________
Christophe Royon
Monday, October 08, 2001
Agenda for 1st days
I prepared a first draft for teh agenda for the next days until next
Tuesday. Please let me know what you think of it. If this is ok for
everybody, we can try to get many of these tasks finished by next week!
Symbols:
========
S = Sergio
C = Christophe
A = Andrew
Mm = Mike M.
Ms = Mike S.
P = Pierrick
J = Jorge
G = Gilvan
E = Eduardo
Ma = Mario
N = Newton
V = Victor
============================================================================
Monday 8 October | (1) S E J C | (1) S E J C
| (10) P | (10) P
| (11) Ms | (11) Ms
============================================================================
Tuesday 9 October | (10) P | (5) P
| (11) Ms | (11) Ms
| (4) V | (12) J
============================================================================
Wednesday 10 October | (1) J C | (6) S E
| (5) P G | (3) Mm J C
| (2) Ms | (4) V
============================================================================
Thursday 11 October | (2) Ms E | (2) Ms E
| (4) V | (5) P G
| (7) P J C | (12) J
============================================================================
Friday 12 October | (2) S Ms E | (2) Ms S E
| (4) V | (3) Mm J C
| (7) P J | (4) V
============================================================================
============================================================================
Monday 15 October | (7) P J | (3) S J C
| (2) Ms S | (5) P G
| (9) J E | (4) V
| (5) P G |
============================================================================
Tuesday 16 October | (11) P J |
| (2) Ms E Mm |
| (9) J E |
| (4) V |
(1) moving all detectors out from tunnel
(2) detector preparation
(2a) cartridge top - put tubles into place with Si rubber
(2b) cables, tests... (hv splitters ready?)
(2c) fiber to L0 tests with LMB
(3) Prepare and change couplers
(4) Finish testing amplifier crates
(5) Prepare ribbon cables in collision hall for AFE
(6) L0 / MAPMT testing (tubes removed from tunnel)
(7) 5V scheme + building + installation
(8) 10 and 15 V circuits scheme + building + installation
(9) Pull last two green cables in collision hall
(10) Test LMB without restoration circuit
(11) Detector mapping
(12) Build two cables for LMBs
__________________________________________________________________________
Christophe
Monday, October 08, 2001
5V scheme
I transmit you as an attachment a scheme using AND and OR circuits which
would work for our control and emergency lines. Please let me know what
you think of it.
http://www-d0.fnal.gov/fpd/FPD_News/5vcircuit.ps.
__________________________________________________________________________
Jorge Molina
Monday, October 08, 2001
PMTs Numbers
Today we took out 8 detectors from the tunel and we disassembled them at
lab 3. We wrote the serial number belonging to each one and they are: (we
also disassembled the two top spares that we have)
Detector PMT serial number Observation
A2U 12436 double u metal
P2D 12468
D1 12?80
A1U 12465
A2D 12378
P1D 12357
A1D 12434 all taped w/u metal tube
D2 SN001 all taped no serial number
Spare 12408
Spare Broken single u metal
__________________________________________________________________________
Mike Martens
Tuesday, October 09, 2001
A1 pot vacuum leak
The A1 castle developed a vacuum leak at 10:00 this morning. The
thermocouple reads 10-3 Torr. The ion pump reads 10-5 Torr but is
drawing too much current to stay on for any length of time and keeps
tripping off.
Coincidentally(?) 10:00 is about the same time that the D0 beampipe was
let up to air pressure. This should not have had any affect on the A1
pots since there are two beam valves between the D0 beampipe and the A1
castle. (One valve between the D0 beam pipe and the low beta quads, and
one valve between the low beta quads and the A1 castle.)
Doug Allen, Mike McGee, and I will be going into the tunnel after lunch
today to investigate. At this point we will have to pump down with a
turbo pump and leak check the castle. I am not sure if a bakeout will be
necessary but we hope not.
Tuesday, October 09, 2001
A1 pot vacuum leak
It appears that we found (and fixed) the problem with the A1 castle
vacuum.
The beam valve between the castle and the low beta quads did not close.
Also on the C4 side there is no ion pump on the low beta quad beam pipe.
This means the A1 ion pump was pumping on both the castle and the low
beta quads.
We closed the valve and turned on the ion pump. We are already at a
vacuum of 5E-8 Torr so I think all is well.
__________________________________________________________________________
Michael Strang
Tuesday, October 09, 2001
Cartridge assembly schedule w/ personnel requirements
Oct 8 - 12:
Tuesday: Modify cartridges removed from tunnel (Jorge)
Test remaining L0 (Eduardo)
Wednesday: Modify cartridges removed from tunnel (Jorge/Sergio)
Test tubes removed from tunnel (assuming P1U and P2U are
removed(Eduardo)
Thursday: Begin assembling cartridge tops with binned tubes
(Mike/Jorge)
Finish modifying cartridges removed from tunnel (Sergio)
Friday: Assemble cartridge tops with binned tubes (Jorge/Sergio)
Map first detector (Mike/Eduardo)
Oct 15-19:
Monday: Assemble first detector in tube (Mike)
Finish assembling cartridge tops (Jorge/Sergio)
Tuesday: Map second detector (Mike/Eduardo)
Assemble first detector in cartridge bottom (Mike)
Assemble second detector in tube (Mike/Eduardo)
Begin wiring HV (if connectors arrive) (Jorge)
Wednesday: Map third detector (Mike/Eduardo)
Assemble second detector in cartridge bottom (Mike)
Assemble third detector in tube (Mike/Eduardo)
Wire HV (Jorge)
Thursday: Assemble third detector in cartridge bottom (Mike)
Begin adding signal cables to finished detector tops
(Mike/Sergio/Eduardo)
Finish wiring HV (Jorge)
Friday: Map fourth detector (Mike/Eduardo)
Assemble fourth detector in tube (Mike/Eduardo)
Continue adding signal cables (Jorge/Sergio)
Oct 22-26:
Monday: Assemble fourth detector in cartridge bottom (Mike)
Finish adding signal cables (Jorge/Sergio/Eduardo)
All detectors should be assembled
Tuesday: Begin testing assembled cartridge tops in PMT test stand
(Mike/Eduardo)
Install finished detectors in tunnel during rest of week
while testing continues
(Eduardo/Jorge/Mike)
__________________________________________________________________________
Andrew Brandt
Wednesday, October 10, 2001
Veto counter design
I believe we have finalized the veto counter design, and will start
on construction tomorrow. Please see the attached drawing.
Sorry for the small size, Jia has problems saving the autocad files as
eps!
http://www-d0.fnal.gov/fpd/FPD_News/lightguide.eps.
http://www-d0.fnal.gov/fpd/FPD_News/scintillator.eps.
http://www-d0.fnal.gov/fpd/FPD_News/countor.eps.
The bottom left drawing shows 1/2 of a veto counter made of 1/2 inch
scintillator. The Righthand drwaing shows the whole veto counter including
light guides (the figure under this shows how the light guides are bent.
The top left piece shows the light guide material which is bent over a 5 in
aluminum tube in order to get 2 bent light guides.
Comments welcome, but should be sent today.
I will talk with Phyllis and Pat about technical issues involving
the cutting and polishing of light guides after they are made.
The counters will be wrapped in a white reflective coating and black
material and then covered with black tape.
__________________________________________________________________________
Pierrick
Thursday, October 11, 2001
Results from yesterday
After several false starts in getting the daq up yesterday, due to changes
made during the first days of the shutdown, we finally got the daq up.
Before going into the tunnel, we (Victor, Gilvan, and I) made some tests
using the LMB to trigger the counters, without any light. When we first
turned on the Amp/Shapers, we had the high pedestals with structure in
both P1 and P2. After cycling the power, we had clean pedestals in
P2, though still on the order of 100s of ADC counts; the same was not true
for P1, it still had structure.
We then went into the tunnel. Here we measured the voltages on the
Amp/Shapers. What we got was:
Detector\V(nominal): +15V -15V +10V -10V -15V
----------------------------------------------------------------
P1 +15.1V -15.1V +12.0V -12.0V +5.4V
P2 +15.4V -15.6V +11.2V -11.4V +5.0V
Though the board is labelled -15V, the cables are labelled 5V, so the
last column is probably not of concern, though it should be verified.
The only other difference is that P1 +-10V supplies are 20% high; this
too needs to be checked if ok.
Next we took a scope to look for noise on the power lines. We found
essentially no noise on P2 (Note that here, the power supplies and RM
are well separated from the VME crate. For P1, we did see a clear 60Hz
pickup, but the amplitude was <5mV; here the RM and PS are closer to the
VME crate. I seriously doubt that this level of noise on the power lines
would propagate into the signals.
Lastly, we swapped an Amp/Shaper card from each crate.
Finally, I spoke with Tom and Gerry with regards to the TPP layout. Gerry
is nearing the end of a layout. Unfortunately it extends the width of the
board to ~7.5". For those not up on this, it is due to the 512 pulse
transformers which are ~0.3"X0.4" each. I also discussed with him the
possibility of using the package of 8 pulse transformers (8 in one unit).
He'll work on this next week when he gets back from vacation.
__________________________________________________________________________
Christophe Royon
Friday, October 12, 2001
Short status
Dear all,
Today, Molina, Eduardo and myself went into the tunnel to remove P1 up and
P2 up. The bottom parts are at LAB3 and the upper parts at D0 sothat the
MAPMTs and L0 PMTs could be tested.
We also removed the LMB cable at P1 sothat it could be used as a model to
do the other ones. We also took away the Hall sensor used to measure the
magnetic field, as well as the HV splitter box at A1 to be used also as
a model for the new ones to be made.
This morning, we changed the names of the ACNET variables with Vladimir,
and we tested that the changes did not affect the pot motion at all (we
moved all the pots we can move now) and check also that the database was
correctly updated. The other issue was to explain to Vladimir in more
detail the bit swapping we observe when we connect the inputs and outputs
to the IIB for one given detector on two different electronic cards inside
a RM.
The last topic today was the discussion with Petros Rapidis about the 1553
new cables, and he will be able to provide new cables for us. The only
path available to the tunnel will be via the penetration, which is quite
painful but not completely impossible. More to come on Monday about this
after a discussion with Ed who will do the cables for us.
We will try to revisit the agenda for next week taking all of this into
account if you agree.
__________________________________________________________________________
Michael Strang
Sunday, October 14, 2001
Cartridge assembly status
Today I collected the data for the tubes from P1U and P2U. I'll add the
new information to the spreadsheet and then we can determine how to bin
tubes on Monday morning. Tube 2 from P2U that previously showed noise
now appears to be behaving like all of the other tubes. The signal in
the scope appears the same, in time with the LMB signal. Before, there
was signal, but it was random with respect to the LMB.
Eileen says the first aluminized detector will be ready Monday, at which
time we can start mapping the detectors.
We can start the first stages of cartridge top production tomorrow.
I'll come up with a revised time schedule tonight.
Tomorrow we will modify the plates from the detector removed from the
tunnel. Using already modified plates, we will reassemble a cartridge
top with the long cables that can be used at either P1U or P2U. We will
also prepare a cartridge bottom with a prototype detector that can be
placed in the tunnel. We should have a full detector ready to go back
into the tunnel Monday afternoon or Tuesday morning. (Emphasis tomorrow
will be on getting one detector ready to go back into the tunnel, then
continuing with the normal plan)
__________________________________________________________________________
Sergio
Sunday, October 14, 2001
FPD Webpage - new version
A new version of the FPD Webpage is available at
http://www-d0.fnal.gov/fpd/Links/links.html
I have incorporated all the URL's that you sent me and I have introduced a
new section "FPD Shift" with the documentation created by Mike Strang for
our DAQ shifts.
Michael Strang
Sunday, October 14, 2001
Revised cartridge schedule
Need Jorge for the HV wiring (he has experience). Can use Eduardo/Sergio
interchangeable for other tasks. Mike S. will map and assemble CB.
Oct 15-19:
Monday: Modify cartridges at P1U and P2U. (CT and CB)
Sort PMTs.
Reassemble one CT and one CB for insertion in tunnel.
Begin assembling cartridge tops with binned tubes (BSC with
four CT)
Tuesday: Map first detector
Assemble remaining cartridge tops with binned tubes (BSC)
Finish first 4 CT with PSC
Start testing L0 bases with PMT test stand.
Assemble first detector in tube.
Wednesday: Finish last 4 CT with PSC.
Map second detector.
Finish testing L0 bases with PMT test stand.
Assemble second detector in tube.
Incorporate first detector in CB.
Thursday: Start wiring HV circuits.
Incorporate second detector in CB.
Disassemble cartridge bottoms from P1U and P2U and
map detector 1 and 2
Reassemble detector 1 in tube.
Friday: Start assembling CT with HV.
Finish wiring HV circuits.
Map third detector.
Assemble third detector in tube.
Incorporate detector 1 in CB.
Start testing CT at PMT test stand.
Add signal cables to tested CT.
Saturday: Assemble detector 2 in tube.
Incorporate third detector in CB.
Oct 22-26:
Monday: Map fourth detector.
Assemble fourth detector in tube.
Incorporate detector 2 in CB
Continue testing CT.
Begin installing completed detectors in tunnel.
Tuesday: Incorporate fourth detector in CB
Finish testing CT
Continue installing completed detectors in tunnel
Wednesday: Continue installing detectors in tunnel
Thursday: Complete installing detectors in tunnel
__________________________________________________________________________
Christophe Royon
Monday, October 15, 2001
Proposal for this week
Sergio and myself tried to do a short list of the activities
which we could do this week. I have three parts: finished,
to be done this week if possible, to be done later on.
Please let me know what you think of this.
LIST OF FPD ACTIVITIES (WEEK 10/15-21)
=======================================
(1) detector preparation: Mike S., Eduardo, Sergio, Molina
(1a) cartridge top - put tubles into place with Si rubber
(1b) cables, tests... (hv splitters ready?)
(1c) fiber to L0 tests with LMB
(2) Finish testing amplifier crates : Victor
(3) Prepare ribbon cables in collision hall for AFE: Pierrick, Gilvan
(4) 5V scheme + building + installation: Molina, Christophe
(5) Pull last two green cables in collision hall: Eduardo, Molina
(6) Detector mapping: Mike S.
(7) Build two cables for LMBs: Molina
(8) Survey information?: Mike M.
(9) New 1553 cables: Christophe
(10) Black screws: M. Mc Gee
(11) 10 and 15 V circuits scheme + building + installation: Molina,
Christophe
DONE:
=====
* moving all detectors out from tunnel
* L0 / MAPMT testing (tubes removed from tunnel)
* Test LMB without restoration circuit
* Noise study
* start building new detectors
WHAT IS MISSING?
=================
AND and OR and INVERTER circuits
5, 10, 15 V power supplies
multiplexors
HV splitters (boxes from Stock room)
Cables for scintillators
screws from M. Mc Gee
TO BE DONE LATER ON:
=====================
(*) AFE support installation (Pierrick, Gilvan???)
(*) Multiplexors ??? (Newton???, Mario???, Christophe, Molina, Mike M.???)
tests in SCR + installation in tunnel + RM scheme + tests
(*) Reinstall detectors when they are ready (one by one) (Mike S., Molina,
Christophe, Eduardo)
(*) Change in python code (Vladimir, Fritz, Christophe, Eduardo????)
(*) HV tests in tunnel (Sergio, Eduardo)
(*) flex cables tests + installation (Pierrick)
(*) Patch pannel for LMR 400 cables (Molina, Eduardo???)
(*) Install new 1553 cables if needed (Eduardo, Pierrick, Molina???)
(*) Install amplifier racks (Victor, Eduardo, Sergio???)- We need POWER
SUPPLIES
(*) Ribbon cable supports in tunnel (Pierrick, Newton???)
(*) Trigger intallation (Pierrick, Mario, Gilvan ???)
(*) Install DFE boards ??? (Pierrick, Mario, Gilvan???)
(*) Lay cables connecting AFE to data streams (Pierrick, Molina???)
(*) Tests: pot motion...., multiplexors...., python code ..., 5V scheme...
(*) DAQ program (Victor, Gilvan???)
(*) New cables to connect RM to IIB??? (shielded)
(*) Test cameras
(*) Veto counters
(*) Install new couplers (Helio)
(*) Castle checkup: coupler, encoder, switches (Helio)
__________________________________________________________________________
Gilvan A. Alves
Tuesday, October 16, 2001
Software update
Yesterday we achieved a major milestone on the software
front. Thanks to the effort of many people, we have now in the
test release, which will base the next production release, the
FPD packages:
fpddigi - defines the FPD digi chunk
l3ffpd_si - multiple interaction rejection tool
fpd_unpack_util - unpacking utilities for FPD
unpack_fpd_fe - mapping of RawDataChunk<->UnpDataChunk
l3ftrack_fpd - FPD tracking tool
I would like to thank the effort of the people involved
in the achievement of this goal and remind of the next milestone,
on Oct 29, for the reco packages. At that time we really need to
have the interface to the calibration database going. Also many
tasks like the trigger simulation and analysis package need to
be started soon.
__________________________________________________________________________
Pierrick Hanlet
Tuesday, October 16, 2001
Data
In order to get an idea of whether or not we can live without the
Amp/Shapers, I have spent some time looking at the data.
http://www-d0.fnal.gov/fpd/FPD_News/run64.ps.
What I was hoping to do was to
come up with a cut on the data, based on what existed in the ntuples. My
thought was to put a cut on the time difference between P1U and P2U.
First, the TDCs are in 250ps resolution mode. Hence the data needs to be
divided by 4 to get ns. Second, after looking at the TDC inputs in TDC1,
P1U is in channel 0, P1D is in channel 1, P2U is in channel 2, etc. As you
know, we have nothing for channels > 1, so we cannot use the TDC cut that I
was hoping for.
Using a TDC cut on P1D only, I have generated the P1U TDC distribution. The
cut is that the P1D TDC has no stop, which implies that there is no hit in
P1D.
Using a tight TDC cut on P1U and the P1D time cuts, I have generated the ADC
distributions. The "tight TDC cut" is for 176ns < t < 198 ns. I fitted for the
pedestal values on these plots; the second fit parameter is the mean, and
third is the rms of the pedestals. The distributions are a little cleaner
with these cuts, but I would still prefer to require that the proton matched
the criterium of having fired both P1U and P2U within a given time window.
__________________________________________________________________________
Michael Strang
Tuesday, October 16, 2001 7
Cartridge Status
Monday:
-Binned the tubes (both MAPMT and L0) based on gain (ADC counts) at
-800V for MAPMT and -1800V for L0.
Determined what bins would go in which castle locations based on gains,
higher gains would go to the 1 castles, lower gains to the 2 castles.
Lowest gain L0 go to veto counters, highest gain L0 go to dipoles.
For MAPMT, the bin mean follows the same prescription as above, for a
particular bin, the lowest gain goes to the mixed plane, then the X
and X' planes, then V and V' then U and U'.
-Disassembled and modified CT from P1U and P2U. Disassembled and
modified CB with Pr1 detector.
-Noticed that the old tubes that we have been using have different
labeling masks that say which channel is which on the face of the
MAPMT. I got worried that our mapping was incorrect so I used the test
stand with a sample of each type of tube (we have 3) and verified that
light sent only to channel 6 comes out of the cable for channel 6 (if
stuff was messed up, it would have come out of channel 7). So the tubes
have been fine all along.
Tuesday:
-Reassembled the CT and CB for A2U using Pr1 detector. This detector is
being installed in the tunnel at the P2U location (by Gilvan and Jorge)
for LMB/Amp tests and then it will ultimately be moved to A2U.
-After assembling the CT, we tested the tubes, in situ, using the test
stand.
-Received Dt1 and Dt2 from Eileen aluminized late in the afternoon.
-Delivered scintillator for veto detectors to Lab 8.
__________________________________________________________________________
Michael Strang
Wednesday, October 17, 2001
Tube binning and placement
A list of binned tubes and locations for the tubes as worked out by
myself and Andrew can be found at:
http://d0server1.fnal.gov/users/strang/pmtdata.xls
__________________________________________________________________________
Andrew Brandt
Wednesday, October 17, 2001
5V Components
The history of the 5V (people can and will correct me I'm sure).
1) 5V control line scheme developed and tested and installed
before Mar 1 which requires a 5V signal either to
operate the pots or to operate the emergency line. This is what we
have operated with for a long time and works fine.
Drawbacks:
a) We do not monitor the 5V which means the program doesn't know whether
the 5V is on or off
b) we have no software control which means that we can't turn
the lines on castle by castle.
c) in case of interuption of 5 V we lose ability to move the pots
even via the emergency line
Note: We did add at some point the capability of turning on the A side and
Pside separatley, and this is what we have been doing for the last few
months.
2) A scheme was proposed by Mario using diodes that would allow
for castle by castle control and to hold the emergency line at 5V
for off and 0V to remove pots, while keeping the normal control
as in the previous design. Attempts to implement this failed as
it did not allow independent castle control. It is possible that
this failure was due to the way it was implemented rather than the
design. Software monitoring also was apparently not included in the
circuit. There was then poor communication on this subject, as many
proposed phone calls never took place, emails on both sides were not
answered, etc. Plans to retest the scheme never developed, as this
is difficult during a short access due to connector changes and
jeopardizing a working system. Due to the failure to finalize
plan 2) and time pressure to get
a new working plan that meets all the criteria
christophe has proposed another plan using logic circuits.
__________________________________________________________________________
Eduardo Gregores
Wednesday, October 17, 2001
Cartridges assembling report
Today Sergio and I wrapped all the MAPMT's with the thin brown tape, sorted
then according to Mike's bining spreadsheet and started assembling the first
four detector on the list. Using the silicon rubber, we glued on the base of
the upper part of the cartridge the MAPMT's, the LO shielding tube, and the
three long narrow tubes that go around the long screws. These four detectors
are laying over the table waiting for the silicon rubber to cure. Tomorrow I
and somebody else experienced in preparing the pink rubber (Jorge?) will
continue with the assemblage.
__________________________________________________________________________
Michael Strang
Wednesday, October 17, 2001
Cartridges assembling report
Detector Mapping:
In other cartridge assembly news, I did a test run of measuring the U
frame of Dt3. It takes about 20 minutes to measure a frame, so about an
hour / detector. I've looked at the results and I'll bring them to the
lunch meeting tomorrow so we can look over them and make any changes we
think are necessary before I go ahead and map the entire detector.
The method I used today (the fibers form a channel for the area where
all four fibers overlap) resulted in the same correspondence of segments
(A particular combination of UP and U channels -> segment) as before,
but the segment widths are considerably different.
With ideal detectors, 0.80mm channels, each segment is ~0.26-0.27mm
wide. In reality, every segment is unique with overlap segments (both a
prime and unprimed channel on) ranging from 0.10mm-0.20mm, and
standalone segments (only prime or only unprimed channel on) ranging
from 0.30mm-0.45mm. This is an artifact of the way I'm defining a
channel. To moderate this, I was thinking of making two measurements for
each channel: The location of the leading fiber of the four in a channel
and location where all four first overlap and then taking the midpoint
of these two measurements as the beginning of the channel. We can
discuss tomorrow and come to a consensus.
__________________________________________________________________________
Pierrick Hanlet
Wednesday, October 17, 2001
Day's activities
Today's activities:
1- W/Marvin and Gilvan
We went into the tunnel to search for noise sources on our signal cables.
Marvin has 4 recommendations:
1) Isolate our electronic grounds from tunnel ground. Marvin has a
device that should help us to do this; he'll try to find two for us.
2) We should check that signal and HV share the same ground
3) We should use linear power supplies - *** This will impact low voltage
ps purchases
4) We should be more careful in routing our cables
2- W/Fred
I finally cornered Fred long enough to make a list of what we still need
for the AFE crate. I'll prepare this list tomorrow, and he will send it
to his group for double checking. I will also send it out to you.
I still have no commitment for time on a test stand. From what Fred
said, he doesn't expect that we'll have much oppertunity until after the
shutdown. What we need to do after preparation of the list is to
prioritize projects and send this list to management.
3- W/Gilvan and Victor
We went back into the tunnel. Here we:
1) adjusted 12V->10.3V on P1
2) Victor swapped an Amp/Shaper board
3) looked at noise on the ps alone (still noisy)
We had a mishap with a short on the P1 backplane, but other than a 3
blown
fuses, smoke, and an overheated trace, we managed to recover.
So far, it appears that we did good things to P1 as seen from the daq
I don't see any LMB signals in P2, I'll check more on that tomorrow
__________________________________________________________________________
Pierrick Hanlet
Thursday, October 18, 2001
Noise and grounding
Yesterday Gilvan and I escorted Marvin Johnson into the tunnel to help us
to track down the noise that we've been seeing on the cables. We have
been studying this for several weeks, and the problems are made worse by
the fact that there are times when the amplitude of the noise is small,
and other times when we see spikes of 1V!
Note that since the SIFT chips are so sensitive (5fC-1.5nC), the slightest
noise will mean trouble for us. We also have to be aware that if we bring
noise onto the platform by not treating our grounds correctly, we will
potentially impact the performance of other detectors in D0.
Firstly, within the SCR, we have a significant amount of noise that comes
from the crates. We have performed some tests by turning off all power
that we can, and find a reduction in noise there.
In the tunnel with Marvin we learned several things:
1) We need to isolate our grounds from that of the tunnel; i.e. we cannot
plug our power supplies into the tunnel AC. We will have to perform some
sort of isolation using either Elgar line conditioners ($5k per device);
or Sola isolation transformers.
2) Noise in tunnel:
Here we see a persistant 84kHz noise on everything. As an example, we
looked at an uplugged cable on the floor and put the probe between the
ground of the cable and a unistrut bar bolted to the tunnel floor. We
also see this 84kHz on the power supply. We see it on the AC safety
ground.
We also saw 115kHz on the P1 supplies. Marvin said that this is a typcial
switching frequency; he recommended that we use linear power supplies;
Gilvan suggests, if not linear, at least good switching supplies.
3) Marvin also expressed concern that the cables are not carefully run.
They lie on the floor, and at times are grouped in a mess of other cables.
Here we need to lift the cables off of the floor and separate them from
other cables and other potential noise sources.
4) Marvin also recommended that we be sure to tie the HV ground with the
signal ground. Clarification is required as to how this is presently done
and how exactly is ought to be done.
At today's meeting, we began to discuss a plan as to how to implement
the changes needed.
1) In order to use the Elgars, or other isolation devices, we may need
to have new AC circuits installed in the tunnel. In any case, we need
to get something installed, fast.
2) The cleaning up of cables needs to be done with a high priority.
3) We should have a map of types of cables and devices in the tunnel,
so that we can determine which types may potentially bring noise onto
the platform.
4) Whether or not we stick with the Amp/Shapers, we will go under the
assumption that we need them, and will build the system around them. As
they may also be a source of noise, we need to keep them as clean as
possible (i.e. clean power, etc).
If we can get these two jobs accomplished, then we may have already beaten
the noise problem. If not, we have simplified where we need to look for
additional sources.
__________________________________________________________________________
Andrew Brandt
Friday, October 19, 2001
Yesterdays meeting
We discussed noise issues stemming from Marvin's trip into tunnel.
Victor outlined amp testing plan.
1) test boards in SCR
2) put boards in tunnel and test with new LV supplies
including full crate (14 board) test and reading out
detectors with Fera's
3) since he needs separate system from Pierrick's tests
a new detector (one of prototypes should be installed
by next Tuesday at P1 to allow him to do these tests
in parallel with Pierricks tests on circuit (of course
since they both use daq they must coordinate who is working
when).
Victor will do test this weekend or today with amplifier
to verify the voltage dependence of pedestal problem and
get me information by Monday which may impact choice of
10-12 V 30A supply. (Marvin said supplies should be linear
not switching due to noise concerns).
Other amplifier issue is long term stability. Issue of dipping
the headers in fresh solder to improve connectivity should be
pursued.
Christophe is looking into cooling for crates and full system
test should include this.
Mike gave detector update and this is proceeding not too far
behind schedule.
__________________________________________________________________________
Eduardo Gregores
Friday, October 19, 2001
Cartride assembling
Today Sergio and I finished assembling the MAPMT's on the cartridge tops
CT1, CT2, CT3 and CT4. We noticed that the HV ground connector is too close
to the lid, and we put some tape on the inner borders in order to avoid
ground mixings. We could note also the tops will very probably need some
triming on the borders of the pink rubber disk since they are sometimes too
close to the MAPMT's connectors. We let to some specialist (Mike, Newton) to
confirm that we really would need to do this. These tops with the respective
bottom parts are now in a shelf at the clean room, in order to be kept more
protected.
We were about to start assembling new tops, when a doubt occured to us,
Molina included.
Do the bottoms have to match the tops in a univoque relation, or should they
be considered as interchangeables? When fixing the MAPMT's on the top parts,
they are aligneded by the bottom parts. How homogeneous are the bottom
parts, and how precise this is supposed to be?
Another question. The clean room is kept locked. What is the door code in
case some of us that don't know the code need to access it?
__________________________________________________________________________
Andrew Brandt
Saturday, October 20, 2001
Outline of responsibilities:
Chairman of Noise: Gilvan with help from Victor, consulting from
Pierrick+Newton
this includes obtaining Solas and any suggestions for optimizing
electronics rack organization in tunnel, interfacing with Marvin
also providing feedback on LV supplies
Amplifier Czar: Victor with consulting from Gilvan Pierrick+Newton
Get tested reliable amplifier system installed.
Detector Diety: Mike S. with help from Eduardo, Jorge, and Sergio
and consulting from Gilvan+Newton
AFE Authority: Pierrick without help from anyone! just kidding
UTA help from Jia, who will likely be out Tues-Fri consulting
with Marvin, Mario, and others
Control Line Controller: Christophe with help from Newton+Molina
and consultation from all
Cable Guy: Mike M. with help from detector guys+consulting from
Noise guy. Mike will develop plan to get cables off the floor
and any other possible isolation of cables from trays and other
cables, recabling. Also supervise laying of new cables.
LMB Implementer: Eduardo, will get other LMB's installed get 2 cables
relayed and make sure gray cables are made, operate SLP's through
3rd floor electronics. Consultation from Pierrick. One issue
is getting an LMB trigger to SCR. This could be done by using
spare channel of SLP routing cable to first floor MCH and through
NIM-ECL over rainbow cable to SCR.
Veto Facilitator: Andrew will keep this going forward, map out
what cables are needed.
I will basically be out of commission through Weds working on a DOE
proposal, will periodically check mail and attend 8:30 meeting Mon+Weds
in Farside. I need input from Gilvan and Mario on how to proceed with
LV supplies. We'd like to procure and test one set as soon as possible
if not sooner. Let me know if any parts are needed for extra MUX cable
(Newton) new control line boxes (Jorge) or other things.
__________________________________________________________________________
Michael Strang
Sunday, October 21, 2001
Proposed detector testing
As the detector cartridges are completed, they need to undergo testing
before the end of the shutdown. I propose the following testing:
L0 Base and Ground Isolation Test (Test Stand):
First, we need to test the L0 bases as they are incorporated into
the cartridge structure. At the same time, we can test that the L0
signal ground is isolated from the cartridge. The HV ground is
actually connected to the cartridge through a resistor.
To perform this test, I propose to use a previously tested tube. All
tubes were tested before with the same base. Using that base, setup
the test stand to reproduce the previous results at -1800V. Then
using that same tube, test all the bases and not any large changes in
behavior of the output of the tube. I would expect a base to either
work or not, but there could be gain changes. Each base should be
labeled in some way so we can keep track of these results.
PMT Incorporation Test (Test Stand):
After the MAPMT and L0 tubes have been incorporated in a cartridge top
along with the HV splitter, we will repeat the test to verify that all
channels of the tubes are working. This test will test the tubes and
the HV wiring. It will not test the signal cables since we will be
using a signal cable incorporated in the test stand for each tube.
Signal Cable and Light Leak test (Tunnel):
The next test will occur in the tunnel after the detector cartridge
has been installed. This will require that voltage be applied to the
tubes through HV system and then the signal will be observed from each
channel of each tube (LMB should be off). This will give us a measure
of the dark signal from each channel as well as light leaks. An analog
scope should be used for this test. If any signal cables are found to
be bad or light leaks are observed (most likely through LMB
connectors) it can be corrected at this time.
LMB and data integrity test (SCR and CH):
The final test requires voltage to be applied and LMB to be turned on
to the detector. The detector will need to be rewired in the CH to
send the signals through to the SCR. Then the DAQ program can be used
to observe the LMB signals. This allows us to probe the amplifiers,
cables and DAQ in the SCR and correct any problems that are observed.
I would expect ~10% or less of the LMB channels to not work since the
small 250micron fibers are easily broken during the detector assembly
process (polishing, aluminizing and cartridge incorporation). I will
try to determine beforehand which channels will not work, but some
might still slip through to this step.
Please comment on these plans with any changes or improvements and let
me know. We will be starting some stages of the testing plan early this
week and should be ready to start the end stages by the end of the week.
Eduardo Gregores
Tuesday, October 23, 2001
Progress
- All 8 cartridge tops with MAPMT's are ready. Some trimming perhaps still
needed. Waiting for further looks regarding this issue.
- L0 PMT's sockets tested. Results on files L0_BASES(1,2,3,4).SUM on
computer at DAB. See log book for details.
- Cartridges cables bunching and labeling under way.
__________________________________________________________________________
Michael Strang
Tuesday, October 23, 2001
Detector progress
All of the MAPMTs have been incorporated into Cartridge Tops
The L0 bases were being tested
Every fiber of Dt3 has been mapped to get a feel for the machine and the
way the fibers are located in a detector. Takes far too long to map
every fiber (~3 hours per part, 9 hours per detector if everything goes
well, which it didn't this weekend), after discussion with Gilvan, 3
overlapping fibers should be sufficient to "turn on" a channel of the
tube (~45 minutes / part, ~2.5 hrs / detector). Have had a quick look at
the data to make sure that it makes sense (i.e. the fibers are
overlapping between primed and unprimed frames the way we expect,
however, segment size varies greatly through a frame).
First new mapped detector should be ready by Thursday with another
available each day thereafter.
Pr2 (A1U) prepped and will be ready to be reinstalled in the tunnel this
afternoon.
__________________________________________________________________________
Gilvan A. Alves
Tuesday, October 23, 2001
Progress
We (Victor, Pierrick and Gilvan) spend many hours looking
for noise sources. We tested the noise levels at the ADCs
first with just cables pluged at the AMP/Shaper backplane,
but power off, then we started to insert amp boards one by
one to see if there was any effect from some board in particular.
We can see that the noise pickup from cables is very small,
even with the power on(shappers off the crate). Once we plug the
shapers then the noise levels realy grow very much, but there
are no board in particular that introduces the noise. This noise
can still be related to ground loops, since when you plug the
shapers you make a ground path between the SCR and tunnel through
the signal cables, so this could be solved with the isolation
transformers.
Another test we did was raising the +- 10V supply to 12V
and check for any effects. Actually the P2 supply can only go as
low as 11V, so the test was 11, 11.5 and 12V settings. We did not
observe any effect due to increasing of the voltage on the ADC
pedestals or noise. The only effect was a clear heating of the
shaper boards, so keeping the voltage low helps, as expected.
__________________________________________________________________________
Michael Strang
Wednesday, October 24, 2001
Cartridge status
Eduardo was trained in detector mapping. He will map the U/V planes then
I will assemble them and map the X frame and the alignment of the
pieces. He was beginning work on mapping Dt4 this evening and it will be
completed tomorrow.
Work continues on modifying the cartridge tops based on Newton's
suggestions. I'm ready to assemble Dt3 into the tube but would like to
have Newton's input on this step first to make sure I'm doing it right.
Disassembled Dt1 and Dt2 from the cartridge bottoms so they can be
mapped.
Dt5 and Dt6 are aluminized and ready to be mapped.
Jorge tried to test the CT for Pr2 (A1U) but there was a short somewhere
in the HV splitter (this is one of the mixed kind of cartridge tops). He
was fixing the problem and was going to retest the cartridge this
evening. I don't know the final results.
Cables have all been labeled in bunches of seven ready to be
incorporated into tested CTs.
__________________________________________________________________________
Eduardo Gregores
Thursday, October 25, 2001
Detectors update
Sergio and I finished finishing the cartridge tops. All 8 are ready to use.
Detectors 4U and 4V have been mapped and the data saved in the files DET4U
and DET4V. See Lab3 logbook for coments on these acomplished tasks.
__________________________________________________________________________
Michael Strang
Thursday, October 25, 2001
Cartridge update
Detectors 5 and 6 picked up from Eileen and prepped for mapping.
Detectors 1 and 2 removed from CBs and prepped for mapping.
CT for A1U fixed and tested. Jorge was going to install it in the tunnel
at the P1U position after the test.
Sergio mentioned difficulties in inserting detectors into the dipole
position. I looked at the CB tube that caused the problem and found that
it was rotated 45 degrees out of spec. I assembled a pseudo detector
bottom for a quick test to see if that is sufficient to remove the
problem or not.
A tube in one of the CTs had a broken signal pin on channel 15. It was
swapped out with an appropriate spare. It is likely that the broken tube
can be repaired.
Tomorrow Goals:
Assemble Dt4 and map the X plane (this will be 2 of 6 detectors
completely mapped).
Eduardo can start mapping Dt5 U and V.
Meet with Newton about Detector incorporation into tube and then
incorporate Dt3 and Dt4 over the weekend (and Dt5 if it gets mapped).
Jorge work on assembling HV splitter for the CTs when not working on
other things.
__________________________________________________________________________
Eduardo Gregores
Friday, October 26, 2001
Friday activities
- LMB's: cables located and started pulled half way through DZero. Looks
possible to divert them around DZero.
- Detector Mapping: Detector-6, frames U and V mapped. Comments on Lab3
logbook.
- Detector Assembling: All 8 tops with MAPMT's have been cleaned from
dust by blowing compressed air, and stored on the clean room.
Alberto Santoro
November 21, 2001
New Time
Dear All FPDers,
Unfortunately I could not go two weeks ago to Fermilab and work a bit in
site. I have followed each day work of each one of you. From my side, I am
work 24 hours per day to get funds and travels to support the necessary
manpower for our project. It is very hard to do this and be sure some time
very frustrating. We do a lot of work, contacts and end with a referee that
say no. I am optimistic for the future and near future because I believe in
our Physics and in the strong work that our group is doing. I still hope to
get the fund to buy the MAPMTs we need to complete all the Pots.
I would like to thank you each one of you, and certainly we will have good
reasons to commemorate this hard period that you did during the shutdown and
even before.
The LISHEP2002 was conceived mainly to give the opportunity to this
group to get some new and updated information about Diffractive Physics. The
first week is dedicated to those that would like to see topics never seem.
And the Workshop the Session C is dedicated to present status of our
Physics. I strongly recommend to participate of SessionC
We are entering in the period of DAQ and analysis. We did a web page
that Sergio can show to you all with bibliography which can be improved if
you use it. In principle to work together it would be much better.
I am continually working to get a new video conference room to turn
possible our discussion in the near future. This is one of my nightmare. We
will do a GRID Workshop in the end of LISHEP to see if we get some upgrade
in this way.
Please, accept my enthusiastic congratulations for your effort. Thank you
Andrew, thank you each one of you.
__________________________________________________________________________
Andrew Brandt
November 21, 2001
Shutdown Heroes!
I would like to echo Jerry Blazey's sentiment and thank Jerry, Mont, and
the entire tracker testing and installation crew for their hard work and
dedication over the past 6 weeks. I would also like to broaden the
recognition a bit. I'm not sure that the entire collaboration is aware of
the enormous effort that has been underway by a small subset of DZERO
members upgrading all of the sub-detectors during the shutdown.
I would particularly like to acknowledge the Forward Proton Detector group
who toiled in anonymity days, nights, and weekends practically non-stop from
Oct. 8 through Nov. 20. It seems like there were always several people in
the Tevatron tunnel, Collision Hall, Lab 3, and Lab 6 working on installing
and testing FPD detectors, electronics, and cables. Thanks also go to
DZERO management, technicians, and beams division for their invaluable
support.
Kudos to all the Shutdown Heroes!
__________________________________________________________________________
Victor Bodyagin
Saturday, October 27, 2001
Yesterday we together with Sergio brought a crate full of shapers into
the tunnel. Also we've repaired +10v wire terminal on P1 power supply.
The crate is powered and all signal cables plugged back. Assumingly
the correspondence between P1 crate slots and SCR ADCs is as follows:
P1 <-> RI-02-01 <-> Slot 3L <-> RC-05-01 <-> RC-32-01 <-> ADC0 SCR
P1 <-> RI-02-02 <-> Slot 3U <-> RC-05-02 <-> RC-32-02 <-> ADC1 SCR
P1 <-> RI-02-03 <-> Slot 5L <-> RC-05-03 <-> RC-32-03 <-> ADC2 SCR
P1 <-> RI-02-04 <-> Slot 5U <-> RC-05-04 <-> RC-32-04 <-> ADC3 SCR
P1 <-> RI-02-05 <-> Slot 10L <-> RC-05-05 <-> RC-32-05 <-> ADC4 SCR
P1 <-> RI-02-06 <-> Slot 10U <-> RC-05-06 <-> RC-32-06 <-> ADC5 SCR
P1 <-> RI-02-07 <-> Slot 12L <-> RC-05-07 <-> RC-32-07 <-> ADC6 SCR
Before leaving the tunnel we checked the remote control of the crate
power. Pierrick helped us. Everything worked fine. But later yesterday
night when I started to test pedestals I saw that crate is not powered!
Today Gilvan and me came to the tunnel and Gilvan immediately
found that power supply was disconnected from remotely controlled
outlet!
Now one can see the pedestals from P1 crate. The distributions have
two peaks which is not very new for P1.
__________________________________________________________________________
Victor Bodyagin
Sunday, October 28, 2001
Let me deliver you the the results of the measurements of the currents
cunsumed by our tunnel electronics:
Voltages +5v -5.2v +15v -15v +12(10)v -12(10)v Boards in
Crate
amps 0.41 0.1 0.53 0.0 0.0 TPG
amps 0.47 0.2 0.63 1.6 1.5 TPG+FDC
amps 0.53 0.3 0.73 3.2 3.0 TPG+2FDC
amps 0.48 0.2 0.63 1.6 1.4 TPG+VTX
amps 0.53 0.32 0.73 3.2 2.9 TPG+2VTX
NOTE: For Test Pulse Generator (TPG) -15v is used to produce -5.2v for
ECL logics.
It looks 600W for +/- 10v is very close to our upper limit I would
rather prefer to have some reserve.
__________________________________________________________________________
Eduardo Gregores
Monday, October 29, 2001
Activities: LMB and detector mapping
1) LMB system:
Sergio and I finished pulling the 2 green cables for LMB around DZero. The 6
green cables that come from MCH 3rd floor are now going to the back of
DZero. The two shorter cables (the unlabeled ones) ends on the CH-NW corner
and the 4 longer cables (LMB-P1, LMB-P2, LMB-A1, LMB-A2) are on the West
wall close to PW08.
Questions:
o I learned from Christophe that that are some black LMR400 cables to be
used for the LMB's. These cables are suposed to send the signal to the LMB's
or to bring back the signal from them? I think the first hypothesis is the
correct one, but I'd like to hear from people. Are those LMR400 cables
laying somewere on PW08?
o There are some green cables labeled as coming from the LMB's on the
tunnel. Where the signal from those ones is suposed to go?
2) Detector mapping:
U and V frames of detector 5 mapped.
__________________________________________________________________________
Christophe Royon
Tuesday, October 30, 2001 1
Multiplexors
Yesterday and today, we had two meetings with Vladimir, Fritz, Newton,
Eduardo
and myself about the multiplexors and the way we can use them in the python
software.
The multiplexors built by Newton will allow to use only one rack monitor per
castle instead of the two we are using now. Basically, we are going to use
two
multiplexors per castle, one to send information from the RM to the IIB and
the
other one to read back this information and to read the alarms. The RM will
send
and read information by block of eight which will be used as follows:
- rack monitor output number P8 (we send these info to the IIB): there are
four
blocks of 16 bits each transmitted to the multiplexor, namely the controls
of
the up, down, in and out pots (in that order). The words sent to the
multiplexor are the same as before, namely the number of turns where want to
go, the status command (RUN/STOP), the home command, the in command, the
speed...
- rack monitor input number P7 (we get these info from the IIB and the
multiplexor): there are eight blocks of 16 bits each transmitted, the first
four
blocks are simply the same blocks as we previous ones (we are able in that
case
to read back what we send to the multiplexors), and the four last ones
contain
all alarms as before (driver, cable, mechanical) and the positions (home,
in).
The same order is used as for P8, namely up, down, in and out.
The LVDT informations will be sent to the analogical input of the rack
monitor
as before. We also need to send in addition to the analog inputs the
amplifier
board voltages divided by two (the RM can measure up to 10 V), namely
5V/2, 10V/2, 15V/2.
We have two additional digital inputs/outputs for the rack monitors which
will
be used in the following way:
- P2 output for RM (used for control):
bit 0: reset of smoke detector
bit 1: software control of the driver power box
bit 2: software emergency (will not be used at present, but space blocked if
needed later on)
bit 3: software control of LVDT power box
bit 4: software control of amplifier power box (same as bit 3 for now)
- P3 input for RM (used for alarms...):
bit 0: smoke detector status
bit 1: software status of the driver power box
bit 2: software emergency status(will not be used at present, but space
blocked
if needed later on)
bit 3: software status of LVDT power box
bit 4: software status of amplifier power box (same as bit 3 for now)
bit 5: control line status (hardware switch status)
bit 6: emergency line status (hardware switch status)
bit 7: 5V status of the driver power box (output of the logical circuit)
bit 8: 5V status of the LVDT power box (output of the logical circuit)
bit 9: 5V status of the amplifier power box (output of the logical circuit)
(same as bit 8 now)
NB: bits 1 to 4 are not absolutely needed fr our system, they are only there
for
cross-checks
The names of the RM will also change according to this new scheme:
FPD_RM_02_01 for A1
FPD_RM_02_02 for A2
FPD_RM_02_03 for D1-D2
FPD_RM_02_11 for P1
FPD_RM_02_12 for P2
Vladimir has started creating the new epics variables and the new database
for
this new scheme. There will be big changes to be made in the python code to
allow these changes. May be it is also useful to use the EPICS variables to
do
the pot motion instead of writing directly into the RM.
__________________________________________________________________________
Christophe Royon
Tuesday, October 30, 2001
I have created a file .k5login on d0ol13 so that people will be able to
logon with ssh on d0ol13 from your desktop. You need first to logon on
your private account on the online computer (this step is absolutely
necessary), and then ssh d0ol13 -l d0fpd.
People who are authorised to do so now are Mike S., Eduardo, Victor and
myself. Please let me know if you want to be added on this list.
__________________________________________________________________________
Pierrick Hanlet
Wednesday, October 31, 2001
Tunel modifications
Yesterday's tunnel visit by Marvin Johnson and Rick Hance generated numerous
suggestions. We were told that there is a very good probability that our
setup in the tunnel will be reviewed again for operational readiness. The
suggestions are:
1) Put all devices which potentially bring ground onto the D0 platform into
a singal crate. E.g. on the P side, put a rack between 1 and 2 stations
with both Amp/Shaper crates and low voltage p.s.'s. The rack would be
floating from earth ground. A single Sola with a satiable inductor to break
safety ground would supply this rack; the satiable inductor would provide
a safe connection to ground for the rack. We would then have only 3 places
where we would bring anything onto the platform. Marvin told us that he
would supply us with the satiable inductors. Another advantage to this
setup is the need for fewer low voltage power supplies; higher power, but
fewer in number.
2) The remaining devices for motor control, etc. would then remain in
separate rack(s), and could be coupled to tunnel ground.
3) There is effort under way by Rick to find a full time senior tech who
is qualified to work in the tunnel to work with/for us for the remainder of
the shutdown.
4) In our racks, we need to close up the sides so that all smoke from
potential
fires is forcec up to the smoke alarm.
5) We need to arrange our racks more carefully to ensure that we take care
of
thermal management. The present arrangement will not work when the crates
are
full.
6) For the dipole racks, Marvin suggested that we cut existin racks and weld
back on a new base. We can then have 2 as for each other station; one
floating
with a Sola, etc. I spoke to Russ, and he said that he could get this done
for
us.
7) Rick and Marvin want to speak to BD about having them clean up the
accelerator cables to get them out of our way. They believe that BD ought
to take responsibility for that.
8) We need to lay some conduit for out cables and neaten up how they are
laid.
9) We need ~60' of 1553 to enter into the tunnel. We can then run this
ourselves. We would need to cut the shielding braid at one end. This info
can be given to Ed Arko. Once we've cut the braid between the platform and
the
tunnel, we won't need to do that with any of the other daisy chained 1553
cables.
Marvin, Rick, and I also talked to Gerry Dychakowsky to review the TPP
layout.
Both Rick and Marvin approve of the design.
__________________________________________________________________________
Pierrick Hanlet
Wednesday, October 31, 2001
Tunnel work
Gilvan, Christophe, Eduardo, Sergio, Newton, and I met this morning to
discuss
reorganization in the tunnel.
We came up with a plan which we believe incorporates the recommendations of
Marvin and Rick. The plan would be to have a separate rack for the
Amp/Shapers
and associated power. The rack is arranged to provide adequate air flow for
cooling the A/S cards. It also provides for power with a ground which is
isolated from the tunnel ground using a Sola or Elgar with a satiable
inductor on the safety ground.
The other rack contains all other devices which can use the tunnel ground
without possibly carrying the tunnel ground onto the D0 platform. The
first 1553 cable from the CH will need to have its braid cut away at one
end.
The smoke detectors can be put in parallel, so that if there is a trip, both
racks will be shut down. We do need to be careful that the output of the
RM which goes to the power box uses an isolated return.
We have a senior technician who is available to us beginning tomorrow at
13:00. One of us will have to be in the tunnel with him at all times to
observe the two man rule. I will be wearing a pager for the duration of
the shutdown so that I can be reached by the tunnel crew.
Christophe went this morning to find the additional racks that we need.
We still need to purchase some sort of conduit for our ribbon cables.
I have attached two drawings: tunnel.ps and tunnel2.ps. Both show the
rack setups, but tunnel2 shows the use of a single rack for both A/S crates.
Though at the meeting we discussed the inability to use this arrangement,
Mike S. is confident that we still can do so; if so, it would require that
the rack be placed by the "1" stations and we would have to reroute the
cables from the detectors to the rack. We still would have the advantage
of fewer power supplies, and a simpler setup. This is still to be
determined.
In this tunnel2 drawing, the satiable inductor is correctly placed; I don't
think that it works as is for the "tunnel" drawing.
What we need to do tomorrow morning is to develop a plan of action so that
we don't waste the technician's time. We need to schedule his partner,
and acquire whatever parts we need. Unfortunately, I have a dentist
appointment, so I'll be in later in the morning.
See:
http://www-d0.fnal.gov/fpd/FPD_News/tunnel.ps.
http://www-d0.fnal.gov/fpd/FPD_News/tunnel2.ps.
__________________________________________________________________________
Christophe Royon
Wednesday, October 31, 2001
Cable scheme for AFE installation
I send you the cable scheme for AFE board installation. This tells you which
cable should be used and which cable is bad.
Ribbon cable from amplifier boards to PW08
==========================================
Detector | Route to PW08
============================================================================
D1 | RC-22-2 RC-07-1
| RC-22-3 RC-07-2
| RC-22-4 RC-07-3
| RC-22-5 RC-07-4
| RC-22-6 RC-07-5
| RC-22-7 RC-07-7
| RC-22-8 RC-07-8 not used (bad contact)
============================================================================
D2 | RC-36-1 RC-10-1
| RC-36-2 RC-10-2
| RC-36-3 RC-10-2
| RC-36-4 RC-10-2
| RC-36-5 RC-10-2
| RC-36-6 RC-10-2
| RC-36-7 RC-10-2
| RC-36-8 RC-10-2 not used (spare)
============================================================================
============================================================================
A1 up | RC-03-1
| RC-03-2
| RC-03-3
| RC-03-4
| RC-03-5
| RC-03-6
| RC-03-7
| RC-03-8 not used (spare)
============================================================================
A1 down | RC-08-1
| RC-08-2
| RC-08-3
| RC-08-4
| RC-08-5
| RC-08-6
| RC-08-8
============================================================================
A1 in | RC-14-2
| RC-14-3
| RC-14-4
| RC-14-5
| RC-14-6
| RC-14-7
| RC-03-9
============================================================================
A1 out | RC-17-1
| RC-17-2
| RC-17-3
| RC-17-4
| RC-17-5
| RC-17-6
| RC-17-7
============================================================================
============================================================================
A2 up | RC-02-2 RC-04-2
| RC-02-3 RC-04-3
| RC-02-4 RC-04-4
| RC-02-5 RC-04-5
| RC-02-6 RC-04-6
| RC-02-7 RC-04-7
| RC-02-8 RC-04-8
| RC-02-9 RC-04-9 not used (spare)
| RC-04-1 not connected spare
============================================================================
A2 down | RC-21-1 RC-11-1
| RC-21-2 RC-11-2
| RC-21-3 RC-11-3
| RC-21-4 RC-11-4
| RC-21-5 RC-11-5
| RC-21-6 RC-11-6
| RC-21-7 RC-11-7
| RC-11-8 not connected spare
============================================================================
A2 in | RC-26-1 RC-15-1
| RC-26-2 RC-15-2
| RC-26-4 RC-15-4
| RC-26-5 RC-15-5
| RC-26-7 RC-15-6
| RC-26-8 RC-15-7
| RC-30-8 RC-15-3
============================================================================
A2 out | RC-30-1 RC-18-1
| RC-30-2 RC-18-2
| RC-30-3 RC-18-3
| RC-30-4 RC-18-4
| RC-30-5 RC-18-5
| RC-30-6 RC-18-6
| RC-30-7 RC-18-7
============================================================================
A2 | RC-34-1 not connected spare
| RC-34-2 not connected spare
| RC-34-4 not connected spare
| RC-34-5 not connected spare
| RC-34-6 not connected spare
============================================================================
============================================================================
P1 up | RC-05-1
| RC-05-2
| RC-05-3
| RC-05-4
| RC-05-5
| RC-05-6
| RC-05-7
============================================================================
P1 down | RC-12-2
| RC-12-3
| RC-12-4
| RC-12-5
| RC-12-6
| RC-12-7
| RC-12-8
============================================================================
P1 in | RC-19-1
| RC-19-2
| RC-19-3
| RC-19-4
| RC-19-5
| RC-19-7
| RC-19-8
============================================================================
P1 out | RC-20-1
| RC-20-2
| RC-20-3
| RC-20-4
| RC-20-5
| RC-20-6
| RC-20-7
| RC-20-8 not connected spare
============================================================================
============================================================================
P2 up | RC-01-2 RC-06-1
| RC-01-3 RC-06-2
| RC-01-4 RC-06-3
| RC-01-5 RC-06-4
| RC-01-6 RC-06-5
| RC-01-7 RC-06-6
| RC-01-8 RC-06-7
============================================================================
P2 down | RC-16-1 RC-23-2
| RC-16-2 RC-23-3
| RC-16-3 RC-23-4
| RC-16-4 RC-23-5
| RC-16-6 RC-23-6
| RC-16-7 RC-23-7
| RC-35-8 RC-23-8
============================================================================
P2 in | RC-31-1 RC-24-1
| RC-31-2 RC-24-2
| RC-31-4 RC-24-3
| RC-31-5 RC-24-4
| RC-31-6 RC-24-5
| RC-31-7 RC-24-6
| RC-31-8 RC-24-7
| RC-24-8 not connected spare
============================================================================
P2 out | RC-35-1 RC-25-1
| RC-35-2 RC-25-2
| RC-35-3 RC-25-3
| RC-35-4 RC-25-4
| RC-35-5 RC-25-5
| RC-35-6 RC-25-6
| RC-35-7 RC-25-7
| RC-25-8 not connected spare
============================================================================
P2 | RC-28-1 not connected spare
| RC-28-3 not connected spare
| RC-28-4 not connected spare
| RC-28-5 not connected spare
| RC-28-6 not connected spare
List of cables which are damaged:
==================================
D1 : D1 - RC-22-1 channels 1,5 dead
D1 - RC-22-8 bad contact channel 15, dipole side
PW08 - RC-07-6 channel 6 grounded
A1 : A1 DOWN - RC-08-7 channel 16 grounded
A1 IN - RC-14-1 channel 11 dead
A2 : A2 UP - RC-02-1 connector broken channel 16
A2 DOWN - RC-21-8 channel 16 grounded
A2 IN - RC-26-3 channel 9 dead
A2 IN - RC-26-6 channel 16 grounded
PW08 - RC-15-5 connected but connector lightly broken channel 16
RC-34-3 short circuit channel 12,13
RC-34-7 short circuit channels 14,15,16
RC-34-8 short circuit channels 14,15,16
P1 : P1 UP - RC-05-1 channel 1 dead
P1 DOWN - RC-12-1 channel 15 grounded
P1 IN - RC-19-6 channel 1 grounded
P2 : P2 UP - RC-01-1 channel 1 grounded
P2 UP - RC-01-9 channel 6 grounded
PW08 - RC-06-8 connector broken
P2 DOWN - RC-16-5 connector broken
P2 DOWN - RC-17-8 channel 8 dead
PW08 - RC-23-1 short circuit channels 10,11
P2 IN - RC-31-3 channels 6,7 dead
RC-28-2 channel 15 grounded
RC-28-7 channel 3,7 grounded
RC-28-8 channel 16 grounded
__________________________________________________________________________
Eduardo Gregores
Friday, November 02, 2001
PW08 preparation
Today Helio, Victor, Gilvan and I disantengled and rolled up the green
cables, rolled up the black cables and attached them below the trays, and
took the ribbon cables out of the rack.
One important issue remains. The FIRUS alarm is not disabled and if we
diconnect it to reroute the green cables used, the FIRUS goes off. It must
be disabled. Christophe is already aware of this and was trying to do what
he could. It's also about time to make a permanent connector for the FIRUS
alarm and replace the present temporary one.
Meanwhile next crews must take most care with the present fragile conection.
__________________________________________________________________________
Pierrick Hanlet
Monday, November 05, 2001
Today's activities
Today was a VERY successful day!!!
Kudos (whatever that means) to the people who helped prepare the PW08 rack!
We now have all of the ribbon coax cables dressed neatly through the cable
trays into the rack. I have made one of the strain relief bars and will
install it tomorrow. There are two cables which need the connectors
repaired.
There are 6 sub-bundles which were originally ordered backwards; I propose
that we simply label them in the correct orientation and change our database
to reflect this. There are also a few detector bundles which do not have
a spare.
We will need to make some jumper cables for to reach the cables which go
into the SCR.
Podschweit has completed the cable labels. I spoke to Wayne about a cable
tester. He suggested coupling two cables (or sets of cables) together and
using the tester to check channel continuity in both sets simultaneously.
Once tested, we can install the labels. If we find bad channels, we will
have to test which set of cables to which the bad channel belongs. Do we
have any volunteers (one for tunnel side, one for CH/tester side)?
This morning I brought the LV power supplies to Tom Regan of Bob Jone's
group. He will make a quick and dirty mount of the ps so that we can test
them. Once tested, we will remove them, order the remaining ps, and prepare
5 boxes for the final mounting. I will check with him tomorrow morning for
progress on the temporary fixture.
With Dychakowski, we finalized the layout of the TPP. After having half of
the cables run into PW08 there was some question as to how the cables would
plug into the TPP in the correct orientation while still keeping the cables
ordered correctly. In any case, this is now clear; Gerry even came to the
CH to check out the setup, and we again verified the orientation. We now
know in which direction to plug our cables into the TPP. The only
remaining question is the package for the diode. Though I spent time with
Rubinov, we haven't settled the question. I'll look more tomorrow.
Bob Jones will make the veto counter cables. He has the connectors; we need
to supply him with cable and lengths. If we know the lengths, we can roll
and
cut them at D0 rapidly.
Tom Wesson will have someone available in two days to solder the connectors
and resistors on the MUX boards. My understanding is that this time frame
is ok.
The parts for the veto counters are almost finished. Jim Fegan believes
that
they will be ready tomorrow.
The angle brackets for the cassettes should also be ready tomorrow.
Molina picked up the blank panels for the control circuit. I asked him to
give them to Newton.
I propose that we modify the power control boxes for the Amp/Shaper crates
ourselves. It is only a change of one bulkhead connector in three boxes;
we'll get them upgraded faster than having someone else do it. Do we have
three isolated BNC bulkhead connectors?
__________________________________________________________________________
Andrew Brandt
Tuesday, November 06, 2001
Veto counter.
For the HV cables:
Using Mikes numbers we should have
4 x110ft red HV cables (one spare on each side)
that run from PW08 to veto counters (does it make sense
that these are same lenght for a side and pside???).
Then assuming that the spliiter box is mounted right next to the
veto counter on the tunnel side, we need
4 x 2ft red HV cables
4 x 5ft red HV cables
the 2ft ones go from splitter boxes to 4 closest tubes (2 on each side)
the 5ft ones go from splitter boxes to 4 furthest tubes
The green cables should be
4x16 ft (2/side)
from ODH wall to T connector
then
4 x 2 ft
4 x5 ft
2 ft go to close counters, 5 ft to far counters.
All these short cables should be findable without being made.
If 3 and 6 are better than 2 and 5, that's ok.
All cables should be labelled!
I'm not sure about the distances from ODH wall to MCH,
CHrisotphe will check these. Need 6 total (includes one
spare/side).
__________________________________________________________________________
Pierrick Hanlet
Wednesday, November 07, 2001
Modification to TPP
After discussion with Gerry Dychakowski, we agreed to change the holes
on the TPP between the connectors to through holes for 10-32 screws. This
would allow the use of pan head screws which would be safer than a small
headed screw. I spoke with Jia Li, and he was in agreement.
This needs to be made known to whoever is making the combs.
We will also widen the PMT slots in the mounting channel pieces for the
veto counters. After testing the pieces with Pat Richards, he showed me
that there is no way to guarentee a 90 degree angle between the scintillator
and the PMT. If the slot is not widened, the present slots would force the
PMTs into position, and the plastic would not line up. Jia was also in
agreement to make this modification.
Lastly, we need to widen the channel in the veto counter clamps to allow
room for all of the paper and tape.
Tom Regan has completed the LV power supply temporary mount. I brought it
into the tunnel earlier. Tomorrow, I will aid Wayne in setting up one of
the A/S racks with sola, etc which can now be used for testing. By noon
tomorrow, we should have the jumper cables for the tunnel; hopefully, this
means that we will be back online by days end.
I found a couple of candidate diodes for the TPP. With help from Dave
Huffman,
we modelled the TPP circuit. Both Marvin and Paul were pleased with the
results. I've attached the circuit diagram for the TPP, plus the diagram
used in the Spice model, and the output of the TPP circuit which will go
into the AFEs.
See:
http://www-d0.fnal.gov/fpd/FPD_News/tpp_circuit_active.ps.
http://www-d0.fnal.gov/fpd/FPD_News/Schematics.pdf .
http://www-d0.fnal.gov/fpd/FPD_News/dtest1out.pdf.
__________________________________________________________________________
Christophe Royon
Wednesday, November 07, 2001
This evening Eduardo and myself removed the amplifier power boxes from the
tunnel. Newton will replace the connector by an isolated one.
On our way, we also found three very nice new 6-fan units nobody needed
apparently... They are now at LAB6, and Eduardo will bring them tomorrow
morning to the tunnel. I would say we can probably use a three fan unit
for the dipoles (we have only two detectors there), and we still need to
find another 6-fan unit!
__________________________________________________________________________
Helio da Mota
Wednesday, November 07, 2001
Report
This is a report on some work done in the castles in the last two days.
several members of the FPD group were directly involved..
a) Steel couplers were machined to have their gauge slightly enlarged
so the worm axle could fit. The machined end is marked by a red circle
b) Original aluminum couplers have been replaced by the new steel ones.
There was no problems fitting the axles to the new couplers.
c) The D1 cartridge was sandpapered all around its cilinder to have the
external diameter reduced. The cartridge would not go into the pot
without extreme force (taking it out required an operation similar
to giving birth with the help of forceps. It does fit well now.
d) A manufacturing error was detected on the the assembled detectors that
were to be installed in D1 and D2. Basically the steel box protecting
the fiber detector was not leaving a small part of the detector
sticking out of it. This way, the detector polished bottom surface
was not laying on the pot bottom. The gap would get closer to
500 microns. We fixed this by removing the 4 screws that fixed the
box to the plastic detector, moving the steel box into the cartridge a
little bit so leaving about 1 to 2 mm of the plastic detector clear
of the box. This also has considerably improved the whole system
flexibility. We then removed all the detectors that had already been
installed and noticed that all presented the problem. They were fixed
and reinstalled. Now D1 and D2 and all the bottom pots have detectors
installed.
e) When doing the procedure for P1 and P2 up pots we found the following
conditions:
I - both pots are dirty!!! There are residues resting at the bottom
of the pot. They must be cleaned.
II - P2 cartridge has remains of tape on it (what may add more debris
to the pot.
III - The detectors do not look good (this was known, I guess). Some
channels show clearly one fiber missing and the kapton tape
around them have loosen parts.
IV - One of P1 up detector guiding rods was loosen and bad. We replace
both rods
d) We have closed P1 and P2, but no screws, nuts and led fibers are
connected.
e) We have examined means of recovering (at least partially) the pot
survey that was lost with the tampering of the switches. We propose to
the following procedure:
I - move all pots to the position where the LVDT reads 20.
II - in the tunnel measure (with the best mechanical means available,
what probably means a caliper) the corresponding pot position
relative to the castle (we can use the base of the moving arm).
III - using this known position we can known the other positions
(including POT IN and POT OUT) within some limits, of course.
This result will not be as precise as the one from a survey,
but probably can be precise to withing 100 to 200 microns. It
would be a first valued to be used with a software process that
could tell us the pots positions by using events like elastics.
This technic is doable and should not take too much time. We need,
though, to have the system operating and having LVDT readout.
__________________________________________________________________________
Mike Martens
Friday, November 09, 2001
Veto counter cables
I have the connectors for the veto counters. The cables, already cut to
length, will be ready from the stockroom after lunch today. I put them
on BD budget code. I will pick these up and deliver cables and
connectors to Bob Jones this afternoon. They should be ready by Tuesday.
We have the following cable scheme and I will ask Bob if he can attach
labels as well.
HV cables:
4 x 110 feet red HV cables Label FPD-VT-HV-Ax (x=1-4)
4 x 2 feet red HV cables Label FPD-VT-HV-Bx (x=1-4)
4 x 5 feet red HV cables Label FPD-VT-HV-Cx (x=1-4)
The 8 shorter cables run from the splitter box to the Veto Counter PMTs.
The 4 longer cables run from the Veto Counter, along the beam pipe to
the ODH wall, then along our cable tray to PW08. There are 2 cables from
the north side veto (one HV + one spare) and 2 HV cables from the south
side veto (one HV + one spare.)
Signal cables:
6 x 110 feet green RG58 cable Label FPD-VT-SIG-Ax (x=1-6)
5 x 170 feet green RG58 cable Label FPD-VT-SIG-Bx (x=1-5)
8 x 5 feet green RG58 cable Lable FPD-VT-SIG-Cx (x=1-8)
The 8 shorter cables run from the veto counter PMTs to BNC T's. This
leaves 2 signals from each side.
The 6 110 feet cables run from the Veto Counter, along the beam pipe to
the ODH wall, then along our cable tray to PW08. There are 3 cables from
the north side veto (two signal + one spare) and 3 cables from the
south side veto (two signal + one spare.)
The 5 170 feet cables (4 signal cables + 1 spare) run from PW08, towards
the north wall, along the sidewalk where our LMR cables have been run,
through the winder, and into the MCH to rack MCH-115. There exists a
penetration through the winder with a rope already pulled through. We
can use the rope to get our cables pulled through the winder.
The 110 feet section of cables and the 170 feet section of cables will
be connected at PW08. This way all signals have the same cable length to
travel and this will make the timing simpler.
__________________________________________________________________________
Pierrick Hanlet
Sunday, November 11, 2001
Almost forgot
This evening we completed the cabling of P2U and P2D. The cables are
labeled and dressed neatly into the P2 A/S rack. All that remains with
this rack are:
1) New LV PS mounting (big job)
2) LV monitoring cables to P2 CTL rack
3) preparation and installation of plenum
4) panel covering front of rack above A/S crate
5) LMB power cable
The complete cable map for P2 A/S to PW08:
Detector out A/S crate into PW08 bad channel(s)
P2U-1 --> RC01-2 --> RC06-1 -
P2U-2 --> RC01-3 --> RC06-2 -
P2U-3 --> RC01-4 --> RC06-3 -
P2U-4 --> RC01-5 --> RC06-4 19
P2U-5 --> RC01-6 --> RC06-5 -
P2U-6 --> RC01-7 --> RC06-6 -
P2U-7 --> RC01-8 --> RC06-7 -
P2U-S --> RC01-1 --> RC06-8 -
P2D-1 --> RC28-2 --> RC23-1 12
P2D-2 --> RC16-1 --> RC23-2 -
P2D-3 --> RC16-2 --> RC23-3 18
P2D-4 --> RC16-3 --> RC23-4 -
P2D-5 --> RC16-4 --> RC23-5 -
P2D-6 --> RC16-6 --> RC23-6 -
P2D-7 --> RC16-7 --> RC23-7 -
P2D-S --> RC35-8 --> RC23-8 -
P2I-1 --> RC31-1 --> RC24-1 3,13,15,17,19,23,27,29,31
P2I-2 --> RC31-2 --> RC24-2 -
P2I-3 --> RC31-4 --> RC24-3
1,3,7,13,15,17,19,21,23,25,27,29
P2I-4 --> RC31-5 --> RC24-4 all odd
P2I-5 --> RC31-6 --> RC24-5 -
P2I-6 --> RC31-7 --> RC24-6 1,3,5,7,9,11,21
P2I-7 --> RC31-8 --> RC24-7 18
P2I-S --> RC31-3 --> RC24-8 2-6,8,10-16,18-21,23,24,26-31
P2O-1 --> RC35-1 --> RC25-1 -
P2O-2 --> RC35-2 --> RC25-2 -
P2O-3 --> RC35-3 --> RC25-3 -
P2O-4 --> RC35-4 --> RC25-4 -
P2O-5 --> RC35-5 --> RC25-5 27,28
P2O-6 --> RC35-6 --> RC25-6 -
P2O-7 --> RC35-7 --> RC25-7 -
P2O-S --> RC23-8 --> RC25-8 -
__________________________________________________________________________
Michael Strang
Sunday, November 11, 2001
Detector tests
We have decided to do the full chain test with P1D and P2D only. These
are the only two detectors we will be collecting data with until we get
more power supplies and they are the only detectors that have a fully
tested chain of cables. The portable amplifier crate has not been
created and any test performed with it would need to be repeated with
the final amplifier cards anyway so it would be double effort for little
gain. The tubes at these locations will still be tested however.
The remaining detectors are being tested by looking at dark signal in
the tunnel. P2U and P2D have tested all channels good. The problem with
offset of channel one does not exist. I was unable to see the same thing
today and it must have been user error yesterday when I was tired.
P1U and P1D have also been tested. One half of a signal cable for tube 2
in the cartridge top has been plugged in backwards. I will need either
Jorge or Eduardo to remove that top for me in the morning so I can fix
the problem. It is a very easy fix. All other channels show dark signal
except for 9-11 of tube 1 from P1U. I will need to open the cartridge
top to see if the cable is partially disconnected.
HV tests:
All HV on the P side is working.
On the A side everything is working at A1. At A2 there is an overcurrent
for A1U with the MAPMTs. I will open the top tomorrow and see if I can
find anything. It could also be a problem at the HV pod since it
immediately tripped when voltage was applied leading me to believe that
the cable might be connected incorrectly at the HV pod and therefore,
when I send current it is going to ground in our HV splitter. I had
personally connected, checked and double checked the external HV
connection at A2U both before and after the test and found no problems.
On the dipoles, the MAPMT HV is correct but D2 L0 is drawing no current.
I will investigate tomorrow.
Also tomorrow, the MAPMT tests will be conducted on the dipoles and the
__________________________________________________________________________
Christophe Royon
Sunday, November 11, 2001
Pot motion
We performed the first pot motion tests after the shutdown, and we have a
few problems to solve:
- D1: we do not have LVDT infos, I checked with Newton and it seems the
voltage coming out of the LVDT is not correct, it has been taken away to
LAB 6 for repair/calibration
- D2 : ok
- A1 up: we had a mechanical failure which was due to a loose encoder, it
is solved now.
- A1 down: ok
- A2 up: ok
- A2 down: ok
- P1 up: ok
- P1 down: loose encoder, has to be changed tomorrow
- P2 up: loose encoder, has to be changed tomorrow
- P2 down: loose encoder, has to be changed tomorrow
Once this is done we could do the calibration.
Concerning the in and out pots, it is impossible to move them (bad motor
status, impossible to reach home... due I think to the RM problem). We
will try again tomorrow. If it does not work, we need to plug them out and
work with them when the multiplexers are installed.
Tomorrow, all encoders have to be checked. May be Helio could do that? All
the encoders on the P side are loose. We also need to check the LVDTs
since the one on the dipoles were also loose.
It would also be good if Molina or Eduardo could remove the DOWN upper
part at P1 to change a ribbon cable connection. May be the DOWN detectors
have to be removed partly on the P side if it is not possible to get tight
encoders otherwise.
We also have to redo the LMB power cables at A2 and P2 because they are
too short.
__________________________________________________________________________
Eduardo Gregores
Sunday, November 11, 2001
L0 and LMB cabling at SCR
Today Sergio and I connected all L0 cables in the collision hall. I'll send
later on the complete map. The destination of the green cables that reach
the small control room are:
D1 -> TD1 (light)
D2 -> TD2 (light)
A1U -> A1U (dark)
A1D -> D2 (dark)
A2U -> A2U (dark)
A2D -> D1 (dark)
P1U -> TL0P1UP (light)
P1D -> P1D (dark)
P2U -> TL0P2UP (light)
P2D -> P2D (dark)
LMB P2 -> TLMBP2 (light)
LMB for P1 is connected to the SLP. The remining LMB channels are not yet
hooked up, waiting for test setup.
The light green cables TBLMBP2, P1IN, and TL0P2IN are free to be used.
Needless to say, all light green cables, and some of the dark green cables
must be properly relabeled.
__________________________________________________________________________
Pierrick Hanlet
Monday, November 12, 2001
Bad news
Despite my pleading with the shift captian in the main control room, we
were kicked out of the tunnel for the evening. I was unable to get in
touch with Mike Martens to request his intercession on our behalf. Had
I gone to the BD meeting this morning, I would have been able to plead on
our account.
The worse news is that we are a step, or more, backwards from the last
report:
1) Fire alarms in P1, A1, A2, and D go off whenever one starts the python
script. This happens because the script automatically resets the RMs.
The
two possibilities as to the source are:
a) smoke detectors are not correctly wired
b) we have different sd in the control rack vs the A/S rack and they are
incompatible with one another
Presently, the SD has been bypassed.
2) Only the P1U pot moves. The remaining pots do not move because:
P1D - mechanical failure
A1U+A1D - unplugged
A2U - doesn't reach home
A2D - mechanical failure
D1 - bad LVDT
D2 - doesn't reach home
All of the couplers have been fixed. However, we need the pot motion
working
to test them.
A possibility source of the "doesn't reach home" problem may be that the
limit switches need to be adjusted.
Molina, Strang, and I will work on veto counter issues tonight.
__________________________________________________________________________
Mike Martens
Tuesday, November 13, 2001
Veto counter update
The HV cables are finished, and the signal cables will be finished this
afternoon.I haven't found someone who can make lables for us yet but I will
keep
trying.
Mike S reports that the counters and bases will be ready by tommorrow.
I met with Mike McGee and Doug Allen. They will get some rubber for us.
Also Jai Li's list of nuts and bolts was incorrect. We need longer bolts
which Mike McGee and Doug Allen will get for us as well.
Also we need to make modifications to the "separation plate" and the
bars that hold them together. This should involve drilling a few extra
holes in the separation plates and in the bars. (Pierrick can you page
me at 314-1734 so we can discuss these changes.)
We are not certain yet when the veto counter installation will occur.
Russ Rucinski thinks it will be on Monday, but he has to check with
Dmitri yet. He will e-mail me when he finds out the latest D0 schedule.
We are still planning as if the installation will happen Wednesday.
__________________________________________________________________________
Christophe Royon
Tuesday, November 13, 2001
Pot motion tests: success!!!
Sergio, Eduardo, Helio, Newton and myself succeeded in getting back our
full system today. D1, D2, P1 up, P2 up, P1 down, P2 down, A1 up, A2 up,
A1 down and A2 down are moving now without any problems at all.
Mainly the problems were due to the fact the pot could not reach home with
the new couplers, or a mechanical failure (loose coupler), or a bad LVDT
value at D1. The pots have been recalibrated using Helio's method by
measuring the positions at home, 20 mm, and in the in position. We have
also redone a full new correlation between position and LVDT by measuring
the LVDT value by steps of 2.5 mm (which gives 15 points of measurement
per pot). I will send out the measurements tomorrow.
All the in and out pots (except A1 in...) show the same problem, namely
they can be moved to the right position, but the status given by the rack
monitor is wrong (run when they are stopped) and thus cannot be moved
further by the software. P2 in and P2 out even did not stop at their right
position. A1 in is however moving without any problems at all, may be
because it is connected to a high quality RM?
The smoke detectors are working fine except at P1 (all the time in alarm).
We cannot mix the old and new smoke detectors and this is the reason of
our problems at p1 (we had 5 new ones and 5 old ones, and at p1, we mix
the new and the old ones). We need to ask an additional one to Rick Hance
tomorrow. Now, it is impossible to move P1 since we plugged back the
drivers on the power box as it should be, and the smoke detector will cut
the driver power. We need to change this smoke detector at the end of this
week.
Concerning remaining tests, we still need to test the alarms and the
emergency line (nothing ws changed there).
__________________________________________________________________________
Helio da Mota
Wednesday, November 14, 2001
LVDT X POT POSITION
We have used a caliper to measure the position of the arm of the switch
activating mechanism relative to a fixed part of the castle. Knowing this
and knowing the dimensions of the castle parts, we can reconstruct the pot
position relative to the beampipe center with some reasonable precision.
These values can be used as a good first aproximation in a program that
uses elastic events to find the pots position with a better precision.
We took measurements in 3 diferent positions:
a) pot in the home position
b) pot 20 mm away from the home position
c) pot in the in position
For each position we took 4 measurements with a caliper. The following
table resumes the results (the calimer measurement is the average of the
four measurements taken with the error only statistical). The tird line
for each detector shows the ratio between the caliper variation and the LVDT
readout.
Drawings showing how the measurements were taken are available.
HOME 20 mm IN
========================================================
D2 cal 55.49 +- 0.04 35.47 +- 0.07 19.29 +- 0.02
LVDT 0.0 17.57 33.62
20.02 / 17.57 16.18 / 16.05
--------------------------------------------------------
D1 cal 58.85 +- 0.15 38.76 +- 0.25 19.24 +- 0.08
LVDT 0.21 19.46 38.32
20.09 / 19.25 18.86 / 18.86
========================================================
A2U cal 59.48 +- 0.23 39.19 +- 0.06 21.66 +- 0.10
LVDT 0.0 19.93 37.58
20.29 / 19.93 17.53 / 17.65
--------------------------------------------------------
A2I cal 57.63 +- 0.05 37.46 +- 0.03
LVDT 0.04 19.11
20.17 / 19.07
--------------------------------------------------------
A2D cal 59.72 +- 0.32 39.57 +- 0.18 21.10 +- 0.06
LVDT 0.27 18.98 37.42
20.15 / 18.71 18.47 / 18.44
--------------------------------------------------------
A2O cal 59.50 +- 0.12
LVDT 0.0
=========================================================
A1U cal 59.85 +- 0.03 39.81 +- 0.07 20.13 +- 0.05
LVDT 0.0 20.20 (a) 40.0
20.04 / 20.20 19.68 / 19.8
--------------------------------------------------------
A1I cal 59.32 +- 0.19 39.35 +- 0.08 19.84 +- 0.05
0.0 18.96 37.66
19.99 / 18.96 19.51 / 18.7
--------------------------------------------------------
A1D cal 59.59 +- 0.09 39.35 +- 0.06 19.41 +- 0.08
LVDT 0.16 19.69 39.27
20.24 / 19.53 19.94 / 19.58
--------------------------------------------------------
A1O cal 59.04 +- 0.07 39.00 +- 0.03
LVDT 0.0 19.15
20.04 / 19.15
========================================================
P1U cal 58.68 +- 0.04 39.46 +- 0.03 19.88 +- 0.02
LVDT 0.0 20.04 39.77
19.41 / 20.04 19.58 / 19.73
--------------------------------------------------------
P1I cal 59.18 +- 0.03 39.58 +- 0.01
LVDT 0.0 18.68
19.60 / 18.68
--------------------------------------------------------
P1D cal 59.97 +- 0.29 40.48 +- 0.24 20.19 +- 0.09
LVDT 0.0 20.08 40.50
19.49 / 20.08 20.29 / 20.42
--------------------------------------------------------
P1O cal 58.70 +- 0.04 39.62 +- 0.35
LVDT 0.0 18.90
19.08 / 18.90
========================================================
P2U cal 58.68 +- 0.04 38.71 +- 0.03 19.67 +- 0.05
LVDT 0.0 20.09 39.41
19.97 / 20.09 19.04 / 19.32
--------------------------------------------------------
P2I cal 59.03 +- 0.02
LVDT 0.0
--------------------------------------------------------
P2D cal 59.28 +- 0.23 39.50 +- 0.12 21.27 +- 0.16
LVDT 0.0 19.75 37.39
19.78 / 19.75 18.23 / 18.04
--------------------------------------------------------
P2O cal 59.88 +- 0.02
LVDT 0.0
=========================================================
The LVDT show good linearity. The ratio between the cali[per
measurement and the LVDT readout should be the same. Most of
the times they are. There are, however, detectors where this does
not happen. In these it seems that there is some problem either with
the LVDT attachment to the castle or with the software. This issued must
be addressed. It does not, however, afctes the determination of the pot
position. We must now use the available castle survey to find out the
pots position.
Catle measuments are known, except for the gap where the two parts connect
and where we can tilt the pots for adjustment. This gap has also been
measured
with a caliper and its valu (available) shall be used in the determination)>
Most pots have their home positions remarkbly at the same relative position.
Remarkbly exceptions are D2 and A2I whose switches, probably, have been
touched.
Pots are moving. Smoke detector at P1 keeps seting an alarm.
The whole pot software seems in need of a complete overhaul. It is
unconfortable when we instruct the system to move a pot to a given
position and the system says "NO", and refuses to comply. There must be
ways of overiding the system and moving the pots anyway when and if we need.
Failure in one of the pots or subsystems should not stop the whole thing.
We were many times unable of moving one pot because something was wrong with
some other pot.
Some erradic behaviour is still present with some pots showing a very
independent and strong personality by moving at their will. Safeguards must
be implemented.
In general, the system works fine with some eventual minor problems that
shall not be a concern to us if all the control system works.
__________________________________________________________________________
Mike Martens
Wednesday, November 14, 2001
Veto counter update
All cables are ready and I have finished labeling them. They are in a
box in the SCR. I also have the BNC Tee's for the signal cable in the
box.
The HV splitter boxes are missing. I am told they needed to be re-wired
and should be ready by tomorrow.
Pierrick will meet with Mike McGee and Doug Allen this afternoon at 2 pm
for a "test fitting" of the veto counters. After they are sure they will
fit properly Pierrick will make the final modifications to the counter
hardware and we will install them tomorrow. We can begin laying the veto
counter cable tomorrow as well.
Mike S. has labeled all of the veto counters with the following scheme
based on their location in the detector:
VN1I
|||
|| ------- (I,O) Inside (or Outside) Tev radius
||
| -------- (1,2) 1 is closest to the IR, 2 is towards the ODH wall
|
--------- (N,S) North (or South) veto counter
---------------------------------------------------------------------------
HV cables:
4 x 110 feet red HV cables Label FPD-VT-HV-Ax (x=1-4)
4 x 5 feet red HV cables Label FPD-VT-HV-Bx (x=1-4)
4 x 2 feet red HV cables Label FPD-VT-HV-Cx (x=1-4)
The 8 shorter cables run from the splitter box to the Veto Counter PMTs.
The 4 longer cables run from the Veto Counter, along the beam pipe to
the ODH wall, then along our cable tray to PW08. There are 2 cables from
the north side veto (one HV + one spare) and 2 HV cables from the south
side veto (one HV + one spare.)
Signal cables:
6 x 110 feet green RG58 cable Label FPD-VT-SIG-Ax (x=1-6)
5 x 170 feet green RG58 cable Label FPD-VT-SIG-Bx (x=1-5)
8 x 5 feet green RG58 cable Lable FPD-VT-SIG-Cx (x=1-8)
The 8 shorter cables run from the veto counter PMTs to BNC T's. This
leaves 2 signals from each side.
The 6 110 feet cables run from the Veto Counter, along the beam pipe to
the ODH wall, then along our cable tray to PW08. There are 3 cables from
the north side veto (two signal + one spare) and 3 cables from the
south side veto (two signal + one spare.)
The 5 170 feet cables (4 signal cables + 1 spare) run from PW08, towards
the north wall, along the sidewalk where our LMR cables have been run,
through the winder, and into the MCH to rack MCH-115. There exists a
penetration through the winder with a rope already pulled through. We
can use the rope to get our cables pulled through the winder.
The 110 feet section of cables and the 170 feet section of cables will
be connected at PW08. This way all signals have the same cable length to
travel and this will make the timing simpler.
__________________________________________________________________________
Mike Martens
Friday, November 16, 2001
Veto counter update
Today we layed the cables for the veto counter from the counters to PW08
and from PW08 to MCH115. One of the HV cables (the spare) was damaged
during the pulling proccess. We still need to connect the signal cables
together at PW08.
Work on the veto counter installation went slow. We did not get any
counters installed. We only got the brackets for one side mounted.
Partly this was due to the amount of time it took to pull the cables.
I just learned that the iron will be closed during the day on Saturday
for surveying, but will be opened back up Saturday evening and stay
opened till Tuesday around noon. Therefore we will finish installing the
veto counters on Monday.
One of the veto counters (VN1I) broke and has to be re-glued. I believe
this will happen this weekend.
__________________________________________________________________________
Christophe Royon
Friday, November 16, 2001
tests
Sergio and myself performed a few tests today:
- pot motion: works fine for D1-D2, P2, A1, A2, P1 was not tested
fully because of the smoke alarm
- 5 V control line: works fine, no power to driver when off, power when on
- smoke detectors:
P1 still needs to be fixed (mixture of old/new smoke detectors)
A1 is working (we get the smoke alarm), but the alarm cannot be reset
(either from the software or in the tunnel by resetting the RMI)
D is working but we have difficulties resetting the alarm
This makes me think that at all places where we are using old smoke
detectors, we have some problems... P2 and A2 are the only places where
we have two new ones, and do not show any problems at all.
I would suggest that we move to new ones everywhere. Could Wayne do
that? We need to test more this when this is fixed and the smoke is
in the SCR for tests.
- emergency line: works fine
- the usual alarms (driver, mechanical) are working fine
Eduardo and myself removed the HV splitters from the tunnel, and put two
back at A1 and D1-D2. Moliona is repairing them.
I also brought some more cables to Bob Jones and we will get them Monday
at lunch time.
Eduardo and I also got the low voltage power supplies and they are in my
office.
The broken scintillator for the veto counters is also in my office.
__________________________________________________________________________
Mike Martens
Monday, November 19, 2001
Veto Counter update
We have finished installing the Veto counters on the North side this
afternoon. We would like to install the south side veto counters
tomorrow morning if possible.
On the North side the HV splitter box is connected to the FPD-VT-HV-A4
cable.
The detector side tubes VN1I(O) signals are tee'd together and connected
to the FPD-VT-SIG-A5 cable. The Tev tunnel side tubes VN2I(O) are tee'd
together and connected to the FPD-VT-SIG-A4 cable.
Tomorrow I will be in about 7:30 to help Mike Petkus with the detector
installation. I also need to get some BNC barrels from the stack room.
If we are going to test the North side counters we should do that early
in the morning since they may be closing up the shielding. I would need
asistance in this since I don't know what to look for nor do I know
where to hook up the HV or how to turn it on.
__________________________________________________________________________
Michael Strang
Monday, November 19, 2001
Detector tests
We have completed testing all of the MAPMT signal cables. All of the
tests are good except for cable 1 in A2U. For some strange reason, the
HV is shorting through the signal cable. Both Jorge and I have opened
the A2U cartridge top and looked at it and are at a loss as to what is
happening. Do we want to replace this cartridge top with a
pseudo-detector top so we can bring it out of the tunnel and look at it
more in depth? We can decide in the morning meeting.
Other than this, the HV is working for all the L0 tubes and all the
MAPMTs. I tried to test the North veto counter but the tubes were not
drawing current which means they are not connected.
I've dome some prelim tests on the shields. I can see dark signal from
both D1 and D2 at about -7mV. There was no one available to help with
the HV or the LMB so I tried to do some tests at the test stand on DAB3.
The tube I was using there had dark signal at -5mV without the shield.
When I put a full shield on it, the dark signal dropped to -3mV. I then
tried to ground the shield to the HV ground, but the HV connection broke
off of the base and I have to leave to get my wife from work so I will
have to continue the tests first thing in the morning.
If we decide to remove the bases, I feel that we can use the velcro as
filler for the tubes so they will be held in the cartridge top.
__________________________________________________________________________
Mike Martens
Tuesday, November 20, 2001
North side veto counter
The north side veto counters have been completely installed. Here are
some notes on the installation and cabling.
The two counters nearest the D0 detector, VN1I and VN1O, have mu-metal
shielding on them.
The two counters furthest from the D0 detector, VN2I and VN2O, do not
have mu-metal shielding.
The cables are connected as follows:
The HV for the four tubes are supplied from a splitter box at the
location of the veto counters. There are two HV cables running from the
veto counters to PW08. (One used for HV, and one spare.) At PW08 the HV
cable is connected to the Reynolds fannout box in the spigot that is
currently controlled by the HV channel MAA1I.
The two counters nearest the D0 detector have their signal cables tee'd
together and the two counters furthest from the D0 detector have their
signal cables tee'd together. There are three signal cables running from
the counters to PW08 (Two are used for the signals, and one spare.)
The signal cables at PW08 are connected to a set of five cables running
from PW08 to the movable counting house MCH115. (Two signals from north
side, two signals from south side, and one spare.)
The cables are labled
North side signal cables
Cable Cable
Counter (counter to PW08) (PW08 to MCH115)
--------- ----------------- ----------------
VN1I + VN1O (Tee'd together) FPD-VT-SIG-A5 <=> FPD-VT-SIG-B5
VN2I + VN2O (Tee'd together) FPD-VT-SIG-A4 <=> FPD-VT-SIG-B4
FPD-VT-SIG-A6 (spare)
North side HV cables
Cable
Counter (counter to PW08) HV connection
--------- ----------------- -------------
Four VN counters (splitter box) FPD-VT-HV-A4 MAA1I
FPD-VT-HV-A3 (spare
- broken terminator
at end near veto
counter)
__________________________________________________________________________
Michael Strang
Tuesday, November 20, 2001
Final tunnel status
When we left the tunnel this afternoon, the status was as follows:
All tubes in the tunnel are drawing current and show dark signal. Both
MAPMT and L0.
All L0 signal cables were pulsed and observed in the SCR.
SLP control from the SCR was sent to P2. The other locations are
connected to the SLP generator and should be controllable from software.
P2D and P1D are connected to amplifiers and passed through to the SCR.
We verified that all pots were in the home position and in/out drivers
disconnected.
Mu-metal was removed from D1. All other locations still contain
mu-metal. Based on observations with beam, we can determine if we want
to remove the shielding from the remaining tubes. It takes about 10-
20 minutes to modify a cartridge top to remove the shielding.
The full cartridge top from A2U was removed from the tunnel and replaced
with the cartridge top from A2I.
All signal and voltage cables at the cartridge tops were verified as
plugged in correctly.
All up and down pots and dipoles were moving during the day by Sergio.
__________________________________________________________________________
Sergio F. Novaes [
Tuesday, November 20, 2001
Final results of test motion
This is result of the pot motion test. All the pots were moved at the same
time. The position and error messages are listed below:
5 mm HOME 10 mm 20 mm 30 mm 35 mm 25 mm 15 mm HOME
D1 1-5 Yes 9.8 19.1 29.0 33.9 25.0 14.5 Yes
D2 0-3 Yes 7.7 17.6 27.5 32.4 27.4 25.5(*) Yes
A1U 4.8 Yes 10.1 20.2 30.3 35.3 25.0 14.9 Yes
A1D 5.1 Yes 9.9 19.7 29.5 34.3 24.7 14.9 Yes
A2U 5.0 Yes(c) 9.9 20.0 30.0 35.0 24.9 14.9 Yes
A2D 3.9 Yes 8.9 18.9 28.9 33.9 24.0 14.0 Yes
P1U (a,15.6)(b) (d,b,a) (e,b,a) (b,a) (b,a) (b,a) (b,a)
(d,a,15.6)
P1D 0.0 Yes 0.0 (a,0.0) 0.0 0.0 0.0 0.0 Yes
P2U 5.0 Yes 10.0 20.1 30.24 35.2 24.9 14.9 Yes
P2D 4.9 Yes 9.8 19.7 29.8 34.8 24.9 14.8 Yes
Error messages
(a) Initialization (yellow)
No motion detected
Impossible to reach Home
(b) Bad verification of the position :
pot probably in wrong position -
home position activated.
(c) A2 UP : Driver failure,
MOTOR STOPPED BY SOFTWARE
(d) Bad direction chosen,
Wanted Position decreased by 1/2 turns.
(e) Bad verification of the position,
new verification.
Commnets:
(*) D2 got stuck at 25.5 mm. Tried low velocity to 20mm and high to 15 mm.
At 17.6 mm got "Mechanical failure, MOTOR STOPPED BY SOFTWARE". Change to
low again. Went OK and stopped at 12.5
Conclusion: P1 are NOT moving with no driver, cable or mechanical
failure
D2 gets stuck but moves well at low velocity.
Pierrick discovered that P1 amplifier was disconnected. Repeting the test
with just P1U and P1D:
HOME 10 mm 20 mm 30 mm 35 mm 25 mm 15 mm HOME
P1U Yes 10.0 20.0 30.1 35.1 25.0 14.9
P1D Yes 10.0 20.0 30.0 35.0 25.0 15.0
Returning to the test with all pots:
HOME 35 mm 10 mm 30 mm 15 mm 20 mm HOME
D1 Yes 33.9 9.7 28.9 14.6 19.4 Yes(g)
D2 Yes 32.5 28.2(a) HOME(b) 12.8 17.6 Yes
A1U Yes 35.3 9.8 22.5(c) HOME(e) 20.2 Yes
A1D Yes 34.3 10.1 29.6 14.8 19.6 Yes
A2U Yes 35.0 9.9 30.0 14.9 20.0 16.2(h)
A2D Yes 33.9 9.0 28.9 14.1 18.9 Yes
P1U Yes 35.1 9.9 28.9 HOME(f) 20.0 Yes
P1D Yes 35.1 19.4(a) HOME(d) 15.0 20.0 Yes
P2U Yes 35.3 9.9 30.24 14.9 20.1 Yes
P2D Yes 34.8 9.9 29.8 14.9 19.8 Yes
(a) D2 and P1D stuck. Change to slow => D2 = 7.6 ; P1D = 10.0
(b) D2 IN : Mechanical failure, MOTOR STOPPED BY SOFTWARE
(c) A1 UP : Driver failure, MOTOR STOPPED BY SOFTWARE
(d) P1 DOWN : Driver failure, MOTOR STOPPED BY SOFTWARE
(e) A1 UP : Driver failure, MOTOR STOPPED BY SOFTWARE
A1 UP : POT pass the position without stopping
(f) P1 UP : POT pass the position without stopping
(g) D2 IN : Mechanical failure, MOTOR STOPPED BY SOFTWARE
(h) A2 UP : Driver failure, MOTOR STOPPED BY SOFTWARE
Got realy stuck. Restarted the software did not help.
Turn off control line
Pot moved from the tunnel till home.
__________________________________________________________________________
Jorge Molina
Tuesday, November 20, 2001
North side veto counter
I don't know if you saw this afternoon that we changed the signal cables
coming from the veto counters to cables that goes from PW08 to SCR.
The map of the cables is in the logbook at SCR (I looked for you but
you were gone by this time).
We connected 4 RG-58 green cables with different lenght that were laid
initially as L0 signal cables but then we decide not to use them due to
timing problems. Now we have to study the timing to see if we can use
this information in the stand alone DAQ.
__________________________________________________________________________
Sergio F. Novaes
Wednesday, November 21, 2001
Veto counter HV test
Here goes the result of the HV test of the veto counters:
NORTH (MAA1I) : V = 1800 V : I_read = 2024 microA
SOUTH (MAA1O) : V = 1800 V : I_read = 3 microA
It seems that SOUTH Veto counter is off and that NORTH is working fine.
Sergio F. Novaes Wednesday, November 21, 2001 Fermilab Training The ES&H Section page (http://www-esh.fnal.gov/train/itp.html) that have all the information on the Fermilab training cannot be accessed from off-site nodes. Therefore, we decided to make available a page with the relevant information about the training that is necessary to access the tunnel and the collision hall: http://d0server1.fnal.gov/users/novaes/training.htm. The due dates for several persons of our group are also shown. If your name is not in that table and you want access your information, please send me your ID number. At the end of this page we have collected some useful links and listed which training is necessary for different scenarios. This link can also be accessed via the FPD News page ("Training Due Dates"). All comments and suggestions are welcome. __________________________________________________________________________ Sergio F. Novaes Saturday, November 24, 2001 Utilities Alarm Around 11h30 I was paged about a RMI alarm in the tunnel. The alarm in the Utilities panel read: Util. BAD 02.D0-RMI-SOUTH Util. BAD 02.D0-RMI-NORTH Util. BAD 02.D0-RMI-NORTH It seemed that we were able to reset via software, but we were not sure since the alarm came back several times. In fact it oscillated between BAD and GOOD condition some times within a 30 seconds period. The last time I checked was 14h55. Molina and I decided to go to the tunnel while Eduardo stayed in the Small Control Room running the program to move the pots (see Eduardo's report about pot motion). In the tunnel, we found nothing suspicious. All the RM were OK (i.e. not tripped) and there was no sign of alarm. Molina also put another screw in A1D detector and we checked all camera position and focus. Due to the new dipole rack it is impossible to focus the camera on D1. I tried to focus on the RM instead but the connector broke. After the pot motion test we verified that all pots were in home position before leaving the tunnel. My feeling is that we SHOULD NOT LET THE PROGRAM RUNNING WHEN WE ARE NOT MOVING THE POTS. In fact the script was open in the SCR. My guess is that the communication with the RM could generate this false alarm. The captain was advised to page again if the problem remains. The should Search & Secure the tunnel tomorrow morning around 9h00 or 10h00. __________________________________________________________________________ Jorge Molina Saturday, November 24, 2001 Veto counter South side This morning Sergio and I entered to the CH to check the connection of the Veto counters and we found that the HV from South side were not connected at PW08. We connected the cable HV A2 to in the place of the P2O, and then tested drawing approximately the same amount of current that in the North Side. Also we checked the connections of the signal cables and what we have now reaching the SCR are the cables: A1, A2, A4 and A5. According to the email of Mike M A4 and A5 are connected for the North side and the only doubt that I have is wich of the remaining three is the spare one. I assumed was the A3 so that's why I connected A1 and A2 (we have to confirm that with Mike M). __________________________________________________________________________ Sergio F. Novaes Monday, November 26, 2001 Veto counter photos The photos of the new Veto Counter detector can be seen at: http://www-d0.fnal.gov/fpd/FPD_News/Pictures/November_2001/D1_veto1.jpg http://www-d0.fnal.gov/fpd/FPD_News/Pictures/November_2001/D1_veto2.jpg __________________________________________________________________________ Eduardo Gregores Monday, November 26, 2001 False alarms and FPD control software We (Sergio, Christophe, Mike, Molina and me) briefly met by chance in our office. It looks like that this alarm checks the status of the rack monitor, and by consequence the comunication with it, and it is not correlated to firus system and smoke alarms. This check has been implemented by Fritz a couple of months ago. Anyway, we could not determine the origin of these faulty alarms, and they stopped as misteriously as they started, with any plausible interference from us. We also concluded that the software should be kept running, as we could not find any possible correlation between the sotware and the alarms. __________________________________________________________________________ Pierrick Hanlet Monday, November 26, 2001 October/November shutdown summary Here is a summary of the latest activities from the shutdown: 1) Installed new 1553 cable from SCR to A1. Though we did not have the chance to test the pulse height, we did check that we could move the pots, reset the crate, and move the pots again. 2) Connected AC extension cords to A1 and P1 from A2 and P2, respectively. 3) Connected LV PSs on P1. 4) Connected P1D and P2D with jumper cables in CH. 5) Completed testing of all MAPMT signal cables (results and map follows). 6) All MAPMT cables are labeled, routed, dressed, and strain relieved into their respective crates. A mapping of the connections on the A/S crates is attached. 7) Fixed P1 station by correctly powering driver AC lines. 8) Frame for backplane support in PW08 was installed. 9) Optical cables from PW03 DFE crate to M322 were installed. There are many items which others already reported which I didn't mention here. Things that still need to be done in the tunnel: 1) LV power supply installation. For this we still need the finished products from Wayne, as well as controlled access time. 2) Pull, connect, and test 1553 cable to P1. 3) Route, connect, and test 1553 cables from A1->A2->D and P1->P2. 4) D A/S crate needs a control power box installed. 5) Complete installation of A/S cards. Things that still need to be done in PW08/platform: 1) Mount backplane support. 2) Sand tabs off of cables. 3) Install lower strain relief for cables (once determined where it will be). 4) Redress cables for final lengths into PW08 (once determined). 5) 1553 to PW08. 6) Acquire and install 3 50-conductor grey cables from PW08 to PC04. 7) Acquire and install AFE backplane 8) Install AFE cassettes. 9) Acquire and install AFE power supplies. 10) Acquire and install TPP combs and support structures. 11) Acquire, test, and install TPPs. 12) Acquire and install AFEs. 13) Acquire and install DFEs (in PW03). 14) Acquire and install Sequencers (in PC04). 15) AFE software. 16) DFE software. 17) Sequencer software. 18) Database additions (whatever that means, still to be determined). What follows is a summary of the MAPMT cable testing. The results include a map: http://www-d0.fnal.gov/fpd/FPD_News/AS_map.ps. P1 -- Detector out A/S crate into PW08 bad channel(s) P1U-1 --> RC05-1 3 P1U-2 --> RC05-2 - P1U-3 --> RC05-3 4,10 (bad cover, can fix) P1U-4 --> RC05-4 - P1U-5 --> RC05-5 - P1U-6 --> RC05-6 - P1U-7 --> RC05-7 - P1U-S --> RC05-8 2 P1D-1 --> RC12-2 - P1D-2 --> RC12-3 - P1D-3 --> RC12-4 - P1D-4 --> RC12-5 - P1D-5 --> RC12-6 - P1D-6 --> RC12-7 - P1D-7 --> RC12-8 29,30 P1D-S --> RC12-1 29,30,31,32 P1I-1 --> RC19-1 - P1I-2 --> RC19-2 - P1I-3 --> RC19-3 - P1I-4 --> RC19-4 - P1I-5 --> RC19-5 - P1I-6 --> RC19-7 - P1I-7 --> RC19-8 - P1I-S --> RC19-6 - P1O-1 --> RC20-1 - P1O-2 --> RC20-2 - P1O-3 --> RC20-3 - P1O-4 --> RC20-4 - P1O-5 --> RC20-5 - P1O-6 --> RC20-6 - P1O-7 --> RC20-7 - P1O-S --> RC20-8 - P2 -- Detector out A/S crate into PW08 bad channel(s) P2U-1 --> RC01-2 --> RC06-1 - P2U-2 --> RC01-3 --> RC06-2 - P2U-3 --> RC01-4 --> RC06-3 - P2U-4 --> RC01-5 --> RC06-4 19 P2U-5 --> RC01-6 --> RC06-5 - P2U-6 --> RC01-7 --> RC06-6 - P2U-7 --> RC01-8 --> RC06-7 - P2U-S --> RC01-1 --> RC06-8 - P2D-1 --> RC28-2 --> RC23-1 12 P2D-2 --> RC16-1 --> RC23-2 - P2D-3 --> RC16-2 --> RC23-3 18 P2D-4 --> RC16-3 --> RC23-4 - P2D-5 --> RC16-4 --> RC23-5 - P2D-6 --> RC16-6 --> RC23-6 - P2D-7 --> RC16-7 --> RC23-7 - P2D-S --> RC35-8 --> RC23-8 - P2I-1 --> RC31-1 --> RC24-1 3,13,15,17,19,23,27,29,31 P2I-2 --> RC31-2 --> RC24-2 - P2I-3 --> RC31-4 --> RC24-3 1,3,7,13,15,17,19,21,23,25,27,29 P2I-4 --> RC31-5 --> RC24-4 all odd P2I-5 --> RC31-6 --> RC24-5 - P2I-6 --> RC31-7 --> RC24-6 1,3,5,7,9,11,21 P2I-7 --> RC31-8 --> RC24-7 18 P2I-S --> RC31-3 --> RC24-8 2-6,8,10-16,18-21,23,24,26-31 P2O-1 --> RC35-1 --> RC25-1 - P2O-2 --> RC35-2 --> RC25-2 - P2O-3 --> RC35-3 --> RC25-3 - P2O-4 --> RC35-4 --> RC25-4 - P2O-5 --> RC35-5 --> RC25-5 27,28 P2O-6 --> RC35-6 --> RC25-6 - P2O-7 --> RC35-7 --> RC25-7 - P2O-S --> RC23-8 --> RC25-8 - A1 -- Detector out A/S crate into PW08 bad channel(s) A1U-1 --> RC03-1 - A1U-2 --> RC03-2 - A1U-3 --> RC03-3 - A1U-4 --> RC03-4 - A1U-5 --> RC03-5 - A1U-6 --> RC03-6 - A1U-7 --> RC03-7 - A1U-S --> RC03-8 - A1D-1 --> RC08-1 - A1D-2 --> RC08-2 - A1D-3 --> RC08-3 1 A1D-4 --> RC08-4 - A1D-5 --> RC08-5 - A1D-6 --> RC08-6 - A1D-7 --> RC08-7 31 A1D-S --> RC08-8 31 A1I-1 --> RC14-2 - A1I-2 --> RC14-3 - A1I-3 --> RC14-4 - A1I-4 --> RC14-5 - A1I-5 --> RC14-6 - A1I-6 --> RC14-7 - A1I-7 --> RC03-9 - A1I-S --> - - A1O-1 --> RC17-4 - A1O-2 --> RC17-3 - A1O-3 --> RC17-2 - A1O-4 --> RC17-1 - A1O-5 --> RC17-7 - A1O-6 --> RC17-6 - A1O-7 --> RC17-5 - A1O-S --> - - A2 -- Detector out A/S crate into PW08 bad channel(s) A2U-1 --> RC02-2 --> RC04-2 - A2U-2 --> RC02-3 --> RC04-3 - A2U-3 --> RC02-4 --> RC04-4 29,30 A2U-4 --> RC02-5 --> RC04-5 22 A2U-5 --> RC02-6 --> RC04-6 - A2U-6 --> RC02-7 --> RC04-7 - A2U-7 --> RC02-8 --> RC04-8 - A2U-S --> RC02-9 --> RC04-9 - A2D-1 --> RC21-4 --> RC11-4 - A2D-2 --> RC21-3 --> RC11-3 - A2D-3 --> RC21-2 --> RC11-2 - A2D-4 --> RC21-1 --> RC11-1 - A2D-5 --> RC21-8 --> RC11-8 1,29 A2D-6 --> RC21-7 --> RC11-7 - A2D-7 --> RC21-6 --> RC11-6 - A2D-S --> RC21-5 --> RC11-5 - A2I-1 --> RC26-1 --> RC15-1 - A2I-2 --> RC26-2 --> RC15-2 - A2I-3 --> RC30-8 --> RC15-3 - A2I-4 --> RC26-4 --> RC15-4 - A2I-5 --> RC26-5 --> RC15-5 - A2I-6 --> RC26-7 --> RC15-6 - A2I-7 --> RC26-8 --> RC15-7 - A2I-S --> - --> - - A2O-1 --> RC30-4 --> RC18-4 - A2O-2 --> RC30-3 --> RC18-3 29 (intermitant) A2O-3 --> RC30-2 --> RC18-2 - A2O-4 --> RC30-1 --> RC18-1 - A2O-5 --> RC30-7 --> RC18-7 - A2O-6 --> RC30-6 --> RC18-6 23 A2O-7 --> RC30-5 --> RC18-5 19 A2O-S --> - --> - - D1 -- Detector out A/S crate into PW08 bad channel(s) D1I-1 --> RC22-2 --> RC07-1 - D1I-2 --> RC22-3 --> RC07-2 - D1I-3 --> RC32-4 --> RC07-3 - D1I-4 --> RC22-5 --> RC07-4 - D1I-5 --> RC22-6 --> RC07-5 31,32 D1I-6 --> RC22-1 --> RC07-6 2+all odd D1I-7 --> RC22-7 --> RC07-7 1,2 D1I-S --> RC22-8 --> RC07-8 - D2 -- Detector out A/S crate into PW08 bad channel(s) D2I-1 --> RC36-1 --> RC10-1 - D2I-2 --> RC36-2 --> RC10-2 - D2I-3 --> RC30-3 --> RC10-3 - D2I-4 --> RC36-4 --> RC10-4 - D2I-5 --> RC36-5 --> RC10-5 - D2I-6 --> RC36-6 --> RC10-6 - D2I-7 --> RC36-7 --> RC10-7 - D2I-S --> RC36-8 --> RC10-8 18 __________________________________________________________________________ Andrew Brandt Wednesday, November 28, 2001 Fermi responsibilites Major tasks and responsibilities for this week and December for Fermilab resident FPDers are outlined below. Comments required! Leaders in these areas should prepare and distribute a plan of action from 11/29-1/1/02 by Thursday afternoon and present it at Friday 10:00 am meeting. We will then distribute it to rest of FPD group. 1) Data taking and analysis/run coordinator Mike S. Mike should continue work on data analysis. We worked very hard to get good data in September and now that shutdown is over, a big emphasis should be put on analyzing existing data and new data as it comes in. Molina should get involved in data analysis, and Gilvan, Jorge, and Carlos should get involved in this effort as well. Mike should keep up the daily run plan and coordinate SCR activities with Pierrick and others. The main people over the next few days will be Mike S., Pierrick, Victor, and Jorge. We will ask to start moving the pots at end of stores commencing on Monday Dec. 1 (Mike M. will be Tev run coordinator then), to reestablish default pot positions, and get back into standard running mode. In the mean time tests should be done of timing, response of veto counters, singles rates in bays, etc. Early pot tests will include plateauing the tubes, getting elastic trigger going etc. Amplifier LV should be off when not being used for now. We will start running pot motion program from MCR as soon as possible, this will help with transition towards elimination of separate FPD shifts. 2) FPD Electronics AFE/DFE/LM Pierrick Pierrick should arrange a meeting with AFE principals so we can understand the time scale for AFE integration. Perhaps next Tuesday before 4:00 at video enabled room? Safety document should be updated to reflect new rack configuration, new LV etc. Any approvals needed? He should fax Phill 815 753-8565 the comb drawings after he verifies them on Weds. Pierrick is also in charge of getting PW08 finalized and will help with commissioning in SCR. 3) Mux/pot motion Christophe Christophe needs to establish with Helio, Mike S. and Mike M. the new calibration of pots and verify meaning of pot positions. This needs to be done this week so we can have some confidence that when we start moving pots we can understand what position we are moving them to and test whether the stable operating position is the same. Once this is understood we should redefine the postion where 0 is the beam axis. This needs to be done with extreme caution to avoid any accidents--the switchover in coordinate systems probably should happen during a couple days of downtime and D0,MCR+CDF would need to be notified when we do this. Christophe should go over the list of status of proposed software mods. Christophe, Vladimir, and Victor will be working on mux tests and preparing for switchover to mux. We still need list of cable changes from Newton so that new cables can be prepared before software is ready. Also full documentaion of the control line circuit and switchover need to be provided. 4) DAQ/SDAQ/PDAQ Victor Victor should start working on SDAQ with Vladimir in addition to helping get DAQ going again. He will be aiding Christophe in Mux debugging and python software operations and may be able to help Pierrick in the AFE/DFE chain. 5) Software Sergio Sergio along with Gilvan needs to keep the software effort going smoothly. 6) L1 track equations Jorge Jorge's main responsibility will be to get the L1 equations implemented for Pdown and Dipoles. He should re-establish contacts with Mario and Mike M. on this issues. Everyone will be involved with SCR shifts and data taking, once the system is operational again. We will need more FPD shifters from Brazil and anywhere else we can find them (ideas?) and will strive to get this standardized and included in MCR and having on-call shifters. __________________________________________________________________________ Andrew Brandt Wednesday, November 28, 2001 other jobs/issues/questions/parts Aside from the major tasks of the last mail there are many smaller ones that need to be followed to completion. Please read and comment. Access Coordinator: Christophe through 12/14 Pierrick after Note: are we on list for tunnel access? We are not on D0 controlled access list--should we be Pierrick? Tunnel Issues: a) short term 1) LV installation (Pierrick/Wayne) 2) 1553 installation (Pierrick/Christophe) 3) Find missing dipole power box (Eduardo) 4) Finalize A/S installation (victor) 5) Documentation/safety (Pierrick) 6) repair D1 camera cable and move camera (Pierrick/Wayne) b) long term 1) A1U+A2U detector installation (MIke S.) 2) P1U+P2U detector installation (Mike S.) 3) Mux installation (Christophe) 4) 5V installation PW08 list: Pierrick has this Questions/Issues: Newton: any progress in understanding RM/IIB problem in tunnel? need mux+5V documentation/changeover instructions Helio: please provide final information about castle calibration if haven't already Pierrick: status of all AFE parts Pierrick+Andrew: what still needs to be purchased+cost Pierrick: status of safety review/doc/ORC Pierrick+Victor: Can you start repairing bad ribbon cable channels whenever there's an access? Mike S.: were any detector cables left not plugged through amps--can you distribute detector database information and coordinate this with cable database? Victor: provide status report on amplifiers Pierrick: Details of sensor to monitor LV including splitter--has this been coordinated with Christophe to upgrade software? Pierrick: Is all strain relief done? Is additional strain relief needed in PW08 besides the comb structure? Pierrick: What about faraday cage for TPP? Christophe: Does online system Kerberos changes affect us? Mike S: Document cable mapping procedure Pierrick+Mike S.: can we start doing TM tests even without the AFE? We need to start excercising the different elements of this chain Victor+Christophe: Backup of d0ol13, CVS of code Mike S.: Newark return for wrong connectors More Issues: 1) start running pot motion from MCR--I have enquired about this w/begel+ denisov 2) pot motion approval--likely can start from Dec. 3 (Monday) have discussed with Dmitri and sent mail to Martens 3) L0PMT shielding (mike S., Pierrick) 4) will need extra HV pod at some point for 18 pot operations (andrew) Small jobs: 1) Frame production--need to spend $0 on this (Mike S., Jorge) 2) Send detector box to UTA (Mike S.) 3) prepare detectors when they arrive (Mike S.) 4) test L0PMT shielding cut vs uncut, grounded etc. at ICD test stand (Mike S.) done already? 5) FPD examine/data monitoring histos--should be able to see hit fibers online (MIkeS.+Victor) 6) Get beam parameters to data base (Mike M. +Christophe) Parts: to obtain: 1) 1 6-fan unit (christophe) 2) short MCH1 cables (molina) 3) 6V supplies for LMB (pierrick) To order: 1) last 2 LV supplies (andrew) 2) TPP (Andrew+Pierrick) 3) pulse transformers (A+P) 4) other componenets (A+P) 5) AFE PS? (A+P) 6) LM PS? (A+P to build: 1) LMB cables (pierrick) 2) Mux cables (christophe/Jones?) 3) LMB pin dioded adapaters (Pierrick+christophe/jones) 4) short ribbon cable repairs (christophe/jones) 5) combs (Phill) 6) Short MCH1 cables (Molina) 7) more rods? How many? (Eduardo) repair: 1) A2U (MIke S.+Jorge M.) Big Issues: 1) AFE 2) DFE 3) TM 4) LM 5) Firmware for 1-4 6) Mux+Mux software 7) elastic stream 8) pot postion/data base/unpacking __________________________________________________________________________ Pierrick Hanlet Wednesday, November 28, 2001 request for safety approval I am sending you a short document describing the use, location, and description of the low voltage power supplies that we would like to install in the Tevatron tunnel. http://www-d0.fnal.gov/fpd/FPD_News/lvps.ps. __________________________________________________________________________ Andrew Brandt Sunday, December 02, 2001 FPD shifts This week the shifts are M Molina T Strang W Bodyagin Th Hanlet F Royon Let me know if any problems. I put these as day 8-4 shifts, but they should be done when they are the most useful. Two people are needed for pot motion. Let Marcia know when you are on shift so she can start learning. Mike S. should update run plan. Early studies will be rate, veto counter, and timing studies without moving in pots. From Monday or so we should be planning to move pots in at end of store (T-4) to re-verify positions. Start with P pots. Meeting with Martens+Royon+others? about pot alignment should take place before pots move and I need to talk to Mike M. who didn't respond to message last week about pot movment. We should move pots at fast to 5mm away from old postions, then slow in .5mm steps to 2mm from old positions then 4 0.3 mm steps then 0.2 mm the rest of the way. Once the new positions are established we will start moving towards an anytime fast insertion procedure, but it is good to be cautious again, recording rates and going through all details of checklists while we are getting started again. __________________________________________________________________________ Christophe Royon Sunday, December 02, 2001 Calib data Dear all, I send you the calibration data which were taken during the shutdown. We will be able to use them to have an equality between the number of turns and the LVDT info in the python program Calibration data for FPD: ========================= For each pot, we give the LVDT value and the number of turns in mm. D1: ==== turns LVDT ----------- home 0.21 5 5.08 7.5 7.47 10 9.85 12.5 12.23 15 14.68 17.5 17.08 20 19.46 22.5 21.89 25 24.29 27.5 26.58 30 29.10 32.5 31.42 35 33.94 in 38.32 D2: NB: gets stuck when goes back home ==== turns LVDT ----------- home 0.00 5 2.76 7.5 5.19 10 7.63 12.5 10.13 15 12.59 17.5 15.11 20 17.55 22.5 20.03 25 22.53 27.5 25.00 30 27.45 32.5 29.92 35 32.40 in 33.63 A1 up: ======= turns LVDT ------------ home 0.00 5 5.09 7.5 7.52 10 10.10 12.5 12.61 15 15.12 17.5 17.68 20 20.21 22.5 22.73 25 25.24 27.5 27.77 30 30.31 32.5 32.81 35 35.33 in 40.00 A1 down: ========= turns LVDT ------------ home 0.00 5 4.88 7.5 7.27 10 9.77 12.5 12.26 15 14.70 17.5 17.18 20 19.68 22.5 22.13 25 24.58 27.5 27.08 30 29.55 32.5 32.00 35 34.45 in 39.34 A1 in: ======= turns LVDT ------------- home 0.00 5 4.64 7.5 6.99 10 9.40 12.5 11.78 15 14.15 17.5 16.55 20 18.95 22.5 21.33 25 23.71 27.5 26.14 30 28.56 32.5 30.94 35 33.30 in 37.60 A1 out: NB: cannot move, wrong status (run when stopped) ========= turns LVDT ------------ home 0.00 20 19.15 A2 up: ======= turns LVDT ----------- home 0.00 5 4.88 7.5 7.33 10 9.81 12.5 12.36 15 14.87 17.5 17.35 20 19.86 22.5 22.39 25 24.88 27.5 27.41 30 29.91 32.5 32.41 35 34.95 in 37.58 A2 down: ========== turns LVDT ------------ home 0.13 5 4.03 7.5 6.48 10 9.02 12.5 11.54 15 14.03 17.5 16.54 20 19.04 22.5 21.57 25 24.03 27.5 26.54 30 29.06 32.5 31.59 35 34.06 in 37.42 A2 in: NB: cannot move, wrong status ====== turns LVDT ----------- home 0.01 20 19.11 A2 out: NB: cannot move, wrong status ======== P1 up: ======= turns LVDT ------------ home 0.00 5 5.01 7.5 7.48 10 10.01 12.5 12.52 15 15.03 17.5 17.53 20 20.04 22.5 22.55 25 25.11 27.5 27.57 30 30.11 32.5 32.63 35 35.16 in 39.75 P1 down: ========= turns LVDT ----------- home 0.00 5 5.03 7.5 7.49 10 10.05 12.5 12.53 15 15.01 17.5 17.61 20 20.07 22.5 22.56 25 25.09 27.5 27.52 30 30.14 32.5 32.62 35 35.04 in 40.00 P1 in: NB: cannot move, bad status ======= P1 out: NB: cannot move, bad status ======= P2 up: ======= turns LVDT ------------ home 0.00 5 4.85 7.5 7.34 10 9,84 12.5 12.37 15 14.84 17.5 17.37 20 19.91 22.5 22.45 25 24.95 27.5 27.49 30 30.01 32.5 32.55 35 35.11 in 39.41 P2 down: ========= turns LVDT ------------ home 0.00 5 4.92 7.5 7.37 10 9.87 12.5 12.31 15 14.80 17.5 17.33 20 19.83 22.5 22.32 25 24.83 27.5 27.32 30 29.85 32.5 32.29 35 34.83 in 37.79 P2 in: NB: cannot move, bad status ====== P2 out: NB: cannot move, bad status =======
Newton Barros de Oliveira
Wednesday, December 05, 2001
Control circuit drawings
I´m sending drawings of the control circuit box (control1.ps) , the
control circuit connections for the future (control2.ps) and the control
circuit connections we are using now (control3.ps).
Could you please add a new section, "FPD Control circuit" and
"Multiplexor", in our documentation (Fpd news/Newton... and/or other place)
in order to be available to the other guys?
I will send the multiplexor updated drawings in the next mails because
drrawings are large.
See:
http://www-d0.fnal.gov/fpd/Links/docs.html#control
http://www-d0.fnal.gov/fpd/Links/docs.html#multi
__________________________________________________________________________
\Mike Martens
Wednesday, December 05, 2001
Veto counter installation
This is a note documenting the veto counter installation.
At the present time I do not know where the cables at PW08 are connected
to. From Jorge e-mail the signal cables from the veto counters to PW08
were connected to cables that ran to the SCR.
----------------------------------------------------------------------------
-
There are 8 separate counters with the following tube and base serial
number
Counter Tube # Base #
------- ------ ------
VN1I 12359 1
VN1O 12365 2
VN2I 12395 4
VN2O 12378 5
VS1I 12383 3
VS1O 12376 6
VS2I 12363 7
VS2O 12382 8
-------------------------------------------
An N(S) in the counter name means the counters were installed
on the North (South) side.
A 1(2) in the name means the counters were installed on the
detector (Tevatron) side. In other words the counters with a
1 in the name are closest to the interaction region.
A I(O) means the detector was installed on the Tevatron
radial inside (outside.)
The detectors are rotated by 45 degrees. On the North side the inner
detectors
were rotated to be closer to the floor. I do not know the wether the
south side
counters were rotated by +45 or -45 degrees.
All of the counters have the mu-metal shielding except VN2I and VN2O
which do not have mu metal shielding.
Data for the Plateau curves (or lack of plateau curves) with mu-metal
shielding
are in a folder in the SCR.
----------------------------------------------------------------------------
-
All of the bases were used on the VN1I counter @1700 volts and the ADC
counts from
cosmics were measured. The average ADC counts and sigma
Base # sigma
------- ------ --------
1 342 125
2 387 151
3 441 161
4 433 133
5 366 146
6 628 227
7 553 189
8 546 223
-----------------------------------
The two counters nearest the D0 detector have their signal cables tee'd
together and the two counters furthest from the D0 detector have their
signal cables tee'd together. There are three signal cables running from
the counters to PW08 (Two are used for the signals, and one spare.)
The signal cables at PW08 are connected to a set of five cables running
from PW08 to the movable counting house MCH115. (Two signals from north
side, two signals from south side, and one spare.)
There are 5 signal cables running from PW08 to the movable counting
house MCH115
These cables are 170 feet long.
The cables are labled as follows:
PW08 to MCH115
---------------
FPD-VT-SIG-B1
FPD-VT-SIG-B2
FPD-VT-SIG-B3
FPD-VT-SIG-B4
FPD-VT-SIG-B5
There are two HV cables and three signal cables running from the north
side to PW08.
The HV and signal cables are 110 feet long.
There is the same arrangement on the south side.
North Side Signal and HV cables
--------------------------------
FPD-VT-SIG-A4 <=> connected to (VN2I + VN2O Tee'd together)
FPD-VT-SIG-A5 <=> connected to (VN1I + VN20 Tee'd together)
FPD-VT-SIG-A6 <=> spare
FPD-VT-HV-A4 <=> connected to HV splitter box.
FPD-VT-HV-A3 (cable HV-A3 is bad due to
missing connector on the side
nearest the counters.)
South Side Signal and HV cables
-------------------------------
FPD-VT-SIG-A1 <=> connected to (VS2I + VS2O Tee'd together)
FPD-VT-SIG-A2 <=> spare
FPD-VT-SIG-A3 <=> connected to (VS1I + VS20 Tee'd together)
FPD-VT-HV-A1 <=> ?? spare ??
FPD-VT-HV-A2 <=> ?? connectd to HV splitter Box ??
I am not sure which HV cable got connected to the HV splitter on the
south side.
Judging from e-mail from Jorge I believe it is HV-A2.
Message from Jorge:
>> We connected the cable HV A2 to in the place of the P2O, and then tested
>> drawing approximately the same amount of current that in the North Side.
I am not sure which spigots the HV got plugged into at PW08.
__________________________________________________________________________
Pierrick Hanlet
Thursday, December 06, 2001
AFE Review
Hi folks,
The talk that I gave this morning for the FPD Readout/Trigger Chains review
is on my web page:
www-d0.fnal.gov/~hanlet
under AFE Review Talk.
Details about the results of the meeting are forthcoming.
See also:
http://www-d0.fnal.gov/fpd/Links/docs.html#transparencies
__________________________________________________________________________
Pierrick Hanlet
Thursday, December 06, 2001
LV power supply review
Hi folks,
The low voltage power supplies for the A/S crates in the tunnel were
reviewed for safety this afternoon. Other than adding two resistors
to protect the PIN/Amp output, and some changes to my circuit diagram,
the ps passed. Wayne will make the modifications; I will update my
drawing, both by tomorrow, and then we can install the lvps's at the
next tunnel access.
When I finish the drawing, it can be found on my web page:
www-d0.fnal.gov/~hanlet
under Documentation for Tunnel Hardware.
See also:
http://www-d0.fnal.gov/fpd/Links/docs.html#5v
__________________________________________________________________________
FPD Group account (Mike, Marcia, Jorge, Christophe)
Thursday, December 06, 2001
Pots home
The pots are now at home.
The operating positions seem to be:
- p1 up : 27.2
- p1 down: 27.1
- p2 down: 30.5
- a1 up: 33.1
- a2 up: about 35
The voltage on the L0 tubes is now 0.
__________________________________________________________________________
Victor Bodyagin
Friday, December 07, 2001
Diles for CVSing
Log into any online machine and to CVS daqfiles you need to go directory
/home/bodyagin/d0/fpd/camac/daq/socket/vers01/toCVS
and there you'll see four subdirectories:
daq
logdaq
peds
vmedaq.
They contain all necessary you need for CVS archiving.
It would be also useful to put to CVS directory:
/home/bodyagin/d0/fpd/camac/lib
It contains IEEE CAMAC library for CBD8210, i.e. VME-camac branch driver,
though we are not going to develop any new version of it it would be useful
to have it within an immediate reach
__________________________________________________________________________
FPD Group account (Victor, Christophe)
Sunday, December 09, 2001
Pot motion
We moved the pots back home. The 5V are off, as well as the L0 HV.
Today, we moved successfully P2 up, A1 down, A2 down, D1 and D2.
The operating positions have been found to be:
- A2 down: 38.0 mm
- A1 down: 37.9 mm
- P2 up: 29.9 mm
- D1 and D2, at least 30 mm (we had no time to go further).
__________________________________________________________________________
Pierrick Hanlet
Monday, December 10, 2001
BD meeting
Hi folks,
At the BD meeting this morning, the schedule was announced for the next
couple
of days:
- end this store
- immediately set up for new store at 14:00 today
- end that store at 08:00 on Tuesday
- p-only store on Tuesday, studies begin
CDF has again requested a 4 hour access.
After the meeting, Mike Martens met with the machine coordinators. They
were
to decide the remainder of the schedule, and will announce it at this
afternoon's
meeting at 16:00.
Also, I went to the MCR and made a TeV tunnel work list:
1) Install P1 and P2 A/S LVPSs
2) Install P1 and P2 A/S LVPS monitoring
3) Install D CPB
4) Install D fan unit (Leon Beverly invited me to look for one with him
today)
5) Pull 1553 cable through south side D0 ODH wall
6) Route, dress, and connect 1553 cables:
A1 --> A2
A2 --> D
P1 --> P2
Those on the access list are:
Royon, Strang, Molina, Bodyagin, and Hanlet
We will need Sergio in the control room for testing. For pulling the 1553
through the ODH wall, I will need to start from the CH, with someone else
in the tunnel; once it's through, I can work from the tunnel.
There is a definite possibility that the access might be Wednesday morning;
we'll know later today.
__________________________________________________________________________
Victor Bodyagin
Monday, December 10, 2001
fans
We have problems with both new LVPSs.
PS No.1 after 10th A/S board was loaded dropped +/-15v. Meabwhile fuses
did not blow up. I can suspect this board now it is marked as suspucious
but anyway fuses should have gone before PS itself. Or probably +/-15v
was not good itself.
PS No.2 the fuse handler for -12V has a defect due to which it is
impossible to insert fuse there. The handler should be replaced.
So far all new PSs failed.
__________________________________________________________________________
Mike Martens
Wednesday, December 12, 2001
Accesses, etc.
> I was informed about D0 FPD team dropping interlocks in the tunnel today.
> As we agreed tunnel access is going under MCR rather then D0 supervision.
> If you want me to take any disciplinary actions in this case, let me know,
> please.
I am told the FPD group dropped the interlocks today (If you are falsely
accussed let me know.)
The dropping of the interlocks added about 1.5 hours to our startup time
as well as taking up valuable operator time. A single instance like this
does not personally bother me since the FPD group has a good track
record.
Acting as the run coordinator, however, I feel obliged to issue a stern
warning in hopes of preventing this type of mistake again. One mistake
will not change my mind about letting people in the tunnel. After
several mistakes I will become more reluctant to grant access priviliges
when we are trying to keep to a schedule. Also, remember to keep the
gate between the C and D sectors closed during controlled accesses.
Keeping the gate closed may have resulted in only one sector being
dropped.
Again, if my info is incorrect and it was not the fault of the FPD group
then let me know.
__________________________________________________________________________
Christophe Royon
Wednesday, December 12, 2001
Code on cvs
I put both the python updates and all the daq code to cvs.
__________________________________________________________________________
Victor Bodyagin
Thursday, December 13, 2001
My talks on FPD day
Dear Colleagues,
here are my talks on FPD day. Sorry, I was just too busy to present
them in more attractive form. Also something has happened with my pine
and I can not attach texts to this message.
Victor.
Amplifier/Shapers STATUS.
Total number of A/S of FDC and VTX type we have at our disposal: 78.
In fact we have more, I just dropped those which are not fully loaded
by headers. I've seen the supply of headers in Lab6 created by Mario but
have not tried it yet.
VTX - 40
FDC - 38
Note: VTX has slightly different from FDC shaping function which in our
case
is not sufficient.
A/S board is considered to be good if it passed two level test.
Top test: LMB signal histograms.
Lower test: preadjustment, pedestal histograms.
Let me to add here a few extra words:Preadjustment is
being
done by trimming of two potentiometers for +/-7V regulators
on each boards for all 32 channels in it simultaneously.
To do that I've sordered up two pins to channel No.25 at each
board to make the voltmeter attachment possible.
Typically, our FERA ADCs have initial pedestals ~30-50
counts at 150 nanosecond gates which corresponds to from -3
mv
to -5mv offset at 50 Ohm ADC input. Generally, 1mv offset
at 50 Ohm and 150 ns gates corresonds to 2.5 pC of charge,
i.e. to 10 ADC counts. When a particular channel fails one
can see only ADC's pedestal. Naturally, it can happen
for a number of reasons, i.e.cable, contact somewhere, and
shaper itself. So far, from time to time I could see only
one FERA channel failing, i.e. it has ZERO pedestal.
To avoid a confusion in diagnotstics I was trying to trim
system shaper+ADC to pedestal value betweeen 100 and 200.
So typical offset for system shaper+ADC is ~15mv.
This simple procedure proved to be very effective.
For instance, some boards which were considered to be bad
turned out to be normal. Simply, they have not been adjusted.
Location Status
tunnel 49 In P2 crate:14 In P1 crate:7 The other 28 are spare. Some
15 of them need to be preadjusted. Simply, they went to the
tunnel before I found the preadjusting procedure.
SCR 15+2 Good for tunnel: 15 Bad: 2 ( failed pedestal
preadjustment)
Lab6 12 have problems
Done: 1.Logbook of A/Ss, information about their location, test they
passed,
parameters of tuning, which channels have problems etc.
2.Logbook of LMB and pedestal tests results as ps-files and their
hardcopies.
3.Some testing equipment in SCR and repairing equipment in Lab6.
Plans:
1. Transform this logbook into Data Base ( with Sergio?)
2. Develop software as a submenu of Roman Pots Program for
automated
tests of A/S using Test Pulse Generator. At this point MUX
software is necessary. (with Christophe?)
Status: ready to start, i.e. I think I know enough about Python
and Tkinter to begin, the delay is caused by software
supporting RM (V.Sirotenko and F.Bartlett are working on it).
I hope to finish it in 1 month.
3. Start repairing of A/S.
New remarks: While discussion yesterday we agreed to postpone item #2 of
plans, which we believe is not that urgent and can be really
useful only in case when we'll be well behind the schedule
with detectors before July which is not expected to happen.
Also the location of boards has slightly changed:
P1 and P2 crates are fully loaded (14 boards in each).
Those preadjusted 15 from SCR went to the tunnel.
Unfortunately, due to rush cause by very limited time in the
tunnel I failed to take ~15 boards from there and to update
my
A/S logbook.
/*
* D A Q
*
* Program to take data until EOR is detected
*
* In VME crate MVME-162 processor ( Motorolla 68040) we have 3 processes
* running vxWorks real time OS:
*
* 1. TRDAQ
*
* tRdaQ - a starting process spawns two other vxWorks tasks -
* trtdaq and trundaq which do main job in the crate. All these
* processes connected through one control and two data common buffers.
*
* 2. TRTDAQ
*
* trtdaq - using CBD8210 CAMAC branch driver and A2 CAMAC crate
controller
* reads CAMAC data stores them in the buffers marks them full and checks
if
* at least one of data buffers is free, fills it etc. it also updates
* each five seconds a set of variables in EPICS Data Base.
*
* 3. TRUNDAQ
*
* trundaq1 through client socket is connected with DAQ process running
* Linux host machine. Having sent a buffer of data it marks it free for
* trtdaq.
*
* Why their names are so ugly?
*
* This comes from what vxWorks it is and from history.
* Initially, when all of your programs are loaded into VME computer
memory they are accepted by OS as "modules". All functions of your
C-program having been loaded into VME are modules. That is any
function
from the very bottom to the very top level (like main) of your program
is seen by vxWorks. In other words it can be started as a separate
task. We can not load an executable file you can load an object one
only. All actual links it needs are being adjusted while loading
these functions into VME. But to be started any module should get
attributes of a task: a separate name, actual values of control
parameters, priority etc. There is a convention in vxWorks that
the names of all tasks should begin from "t" in order system could
distinct them from modules. To follow this convention is not
necessarily but resonable. The "historical" part of the names comes
from previous DAQ-RUNDAQ version in order people who worked with it
could easily see which part of data acquisition is about.
Therefore, there are rundaq and rtdaq as names for modules and
trundaq and trtdaq for as processes.
*
*
*
* In host Linux machine we have two processes: DAQ and LOGDAQ.
*
*
* DAQ running Linux accepts data and control information from TRUNDAQ
through
* server socket. It is connected with LOGDAQ using shared memory.
*
*
* There are 2 buffers in shared memory for events, and 1
* buffer for control. DAQ writes to all 3, LOGDAQ writes
* to control only.
*
* DAQ fills buffer A and marks it FULL,
* it waits until buffer B is EMPTY and then it writes to buffer B.
* when buffer B is FULL, it waits until buffer A is empty and
* writes to that.
*
* LOGDAQ waits until buffer A is FULL, writes it to disk, fills
* histograms and then marks it EMPTY.
* LOGDAQ then waits for buffer B is FULL, writes it to disk,
* fills histograms, and then marks it EMPTY.
*
* One or more HISTOSCOPE sessions can be run to connect to the
* histograms filled by the LOGDAQ.
* Now one HISTOSCOPE session is spawned by default by LOGDAQ.
* For easier life one can use configuring files for histograms to watch
them
* grouped in one window.
*
*
* In many respects this version is based on earlier rundaq-logdaq DAQ
* written by Jon Streets for Jorway73a SCSI-CAMAC controller. In its
* sufficient part it uses IEEE CAMAC functions library for CBD8210
* written by Dave Slimmers et al. See Fermilab PN 476.
*
*
* Unfortunately, some of this library functions failed to work with us
* due to some hardware problems. More specifically, while fast
* CAMAC readout closer to the end of transfer we were loosing responds
* from FERA ADCs. So, I replaced a couple of library functions by
* custom-made. Simply, I've slowered down this too fast for us transfer.
* This is quite insignificant. Readout is fast enough.
Any questions and criticism are welcomed !!!!
__________________________________________________________________________
Andrew Brandt
Friday, December 14, 2001
Updated D0-BD pot motion agreement (fwd)
Dmitri+Mike,
Here is updated agreement, based on last one
but with values tested since shutdown.
Please let me know if there are any questions/problems
and confirm that this is ok (I have talked separately
to both of you).
thanks, Andrew for D0 FPD group
This agreement applies to Roman pot insertion from the period starting
12/14/2001:
D0 plans to move the Roman pots in to standard positions most stores
not earlier than one half hour after the beginning of the store.
The MCR and CDF will be notified before pot motion. The default positions
will be A1U=33.1 mm A1D=37.9 A2U= 35 mm A2D=38 mm
P1U =27.2 P1D=27.1 P2U=29.9 P2D=30.5
(determined since the shutdown ended as safe positions),
with the beampipe axis roughly at 45mm. Pot insertion
will be done in a similar manner as in the last weeks before the
shutdown, with the final couple mm at slow speed allowing time to react
to an unanticipated change in beam orbit. D0 should be notified
of any planned changes in beam position. The pots will remain at these
positions for most/all of the store retracting the pots prior to the
beam being aborted or machine studies.
In some subsequent end of stores we plan to do studies to define dipole
pot operating positions and test whether beam conditions have changed.
Approval from BD run coordinator will be obtained
prior to these studies. An acceptable new position for now is defined
as one that does not increase the D0 losses by more than 10% or
significantly affect CDF losses.
__________________________________________________________________________
FPD Group account (Pierrick)
Friday, December 14, 2001
HV off
Hi folks,
I was unsuccessful in getting the SLP to operate the LMBs. I want to
talk to one of the ICD folks as it is not clear to me that their's
works either. I am able to control the SLP delays as expected, but
I don't get the DC offset required to fire the LEDs. When I look at
the ICD channels, I see the same offset; hence, I'm not certain that
their's works either. It will have to wait until Monday to make
any progress.
In any case, I've turned the HV off on the L0 and MAPMTs. I still
have it on for the veto counters, as I hope to time in and plateau
at least the proton side tonight.
__________________________________________________________________________
Pierrick Hanlet
Friday, December 14, 2001
Evening's summary
Though unsuccessful in getting the SLPs to operate, I did have some success
with the Veto counters:
1) The counters are all in time. The concern that the cables from PW08 to
the SCR are all different lengths is not an issue. Both the p-side
counters came in time with one another, as did both a-side counters.
This I looked at with a p only store.
2) I cut ~100ns off of both p-side veto counter signal cables. It was quite
a job due to the rats' nest of cables that we have under the floor. This
brings the signals in ahead of the LM signal. Hence we can do a little
logic, lengthen the width, and have the signals in time with the LM
signals.
3) I re-routed the cables under the floor, and up into the relay rack to the
back of the patch panel. These are labelled VS1 and VS2, a la Martens.
Unfortunately, I did not have the correct crimping tool, so please don't
pull on the cables before I can crimp them properly.
My plan was then to plateau the counters. Because we only have 1 HV per
set of N or S counters, my original plan of using 1 to plateau 2, and vice-
versa won't work. In fact, this is true of testing the mu-metal shielding
also.
Plan B for plateauing is:
VN*VS*LN*LS
-----------
VN*LN*LS for VS counters
Unfortunately, when I killed the python script in the D0 CR in order to
restart it in the SCR, I could not do so due to IO errors with
FPD_RM_02-xxx.
I finally gave up.
We also need, or at least I need to find, better documentation on running
the python script from the D0 CR. I couldn't rlogon back to d0ol13 due
to kerberos timeout of tickets.
__________________________________________________________________________
Michael Strang
Saturday, December 15, 2001
Power outage
I just spoke with the Shift Captain at D0 and there will be a complete
power outage at D0 from 1-5 today including the computers in the main
control room.
Of course, this precludes us from controlling the pots, so the plans for
today's studies will have to be postponed until tomorrow if possible.
Victor is going to the SCR to power down our equipment in preparation of
the power outage. At 5, we can bring out equipment back online.
I am not available tomorrow for pot insertion but I have left detailed
instructions in the daily runplan and Jorge is as well trained in
insertion of pots as I am, so if it is ok with everyone (read Andrew), I
feel that Victor and Jorge could insert the pots tomorrow if they are
both available and then pursue the runplan.
http://d0server1.fnal.gov/users/strang/web/fpd/documents/runplan.html