PVDIS bi-weekly meeting minutes

• PVDIS (E08-011) main page
  
• Note: We meet every week, typically on Monday 2-4pm in Cebaf Center F228 except Nov 3, Nov 17 and Dec. 1. (Before September 2 we were meeting every other week.) Off-site people may join by phone+EVO. To receive meeting notices, please subscribe to pvdis@jlab.org.

• Here is a task list for the April-June test period, based on our discussion on 3/9/2009. Highest priority the first: (although the tests are finished, it remains at the top of this page as a reminder since many aspects will be repeated for the real run).
• Prepare for the upcoming DAQ test, make sure we have all the hardware and software ready by mid April. This includes:
1. hardware preparation, including scaler quality-check and helicity check.
2. make sure configuration "trans-pvdis" works;
3. online display of all scaler readings;
4. online analysis code to show all scaler readings and asymmetries;
5. online analysis code to study/determine discriminator threshold vs. corresponding PID cuts. THis should be based on plots of PSH/SH (adc_p) and Cherenkov spectrums with cuts on:
1. PVDIS triggers in the TS;
2. PVDIS triggers in the fastbus TDCs.
3. (spectrum should be made according to the detector grouping to facilitate diagnostics).
6. online analysis code for deadtime measurement of all 8 groups;
7. analysis code to determine detector PID performance using RHRS data. (The same code should also work for LHRS). These results, combined with #5 above, will be used to determine the final disc threshold levels.
8. communication with upcoming experiments - make sure they are okay with us;
9. (lower priority): FADC readout and analysis of pileups;
• Diancheng will help Alexandre on the new Compton electron detector readout and analysis;
• HAMC simulation, Diancheng will work on the resonance asymmetry "models", and the derivatives of the asymmetry w.r.t. all beam parameters;
• Trigger simulation code. Based on what Huaibo wrote. Still need to add the position dependence. Best thing to do is to create a table of shower/preshower hit position from HAMC, then read it into the trigger simulation code;
• FADC code. Need to compare Huaibo's code and Ole's FADC code in analyzer. Study whether the analyzer can meet our need. What is included in Huaibo's code but not in analyzer. If there is any, how to integrate into analyzer.

• 10/05/2009

• Kai reported on work done during last access.
• TBP
• Plan
• During next access:
• Add synch signal to each scaler on both arms, check all scaler copies (EN, EW from all groups and the final EN, EW, PiN, PiW should be copied, so are all signal.and.tagger signals for the left arm, and half of signal.and.tagger signals for the right arm) - Ramesh, Xiaoyan. Should get a swapping map ready before the access. And please send the swapping map to Xiaochao first.
• Add S2 to trigger superviser - Ramesh
• Fix tagger on the left arm.
• Also, cannot see scintillator rate on the left arm (though GC TDC channel shows something) - Ramesh will check this.
• Try BCM calibration to solve the bcm synch problem - Xiaoyan
• Think about the run plan and the changeover plan, assign names - all.
•  

• 09/28/2009

• Work list for the Tuesday access:
• Check what's wrong with left arm DAQ (all scalers reading zero) - Kai, Ramesh;
• Do a pulser test for right arm DAQ (use a pulser instead of the scintillator signal), then take a HAPPEX run and analyze all scaler rates. This is to diagnose why PS rates of 8 groups differ by a factor of 3-4, while TS rates are relatively close - Ramesh;
• Left arm synch check (MPS missing due to crates off?) - Kai;
• FADC work? - right arm group 5 PS (analog 740 output), TS (analog 740 output), EW (logical out from the ANDing module), and the overall eVeto (logical) - Kai, Ramesh;
• Change 100kHz to a 103.7kHz crystal clock for both left and right arms - Kai.
• Change timing signal from MPS/LNE to VME1/GMN2 from the HAPPEX timing board, see halog #289817, and test. Diancheng raised the question about int-time - Bob.
• Bob analyzed a PITA scan run from last Thursday, see PVDISLOG, also see Xiaoyan's results there. Note: It appears that PVDIS bcm asymmetries are smaller than HAPPEX. Are pedastals for both versions of BCMs corrected??
• Need report on:
• Why PAN is cutting 600 events - seems to be related to the event shift, but not fully understood yet (Xiaoyan).
• Qasym plots and comparisons (Xiaoyan)
• runplan (Xiaochao, still working).

• 09/21/2009

• Xiaochao: 
• On the left arm, see no rate. All scalers read zero except BCMs and the two high rate clocks. Even tagger is zero. Why? <- Ramesh suspects some NIM crates are off, will check this during the next access.
• Looked at PS and TS rates on right arm after the medium threshold was increased to -50mV. Still see pretty irregular pattern in PS rates among 8 groups. Don't know why. Kai's FADC data could help here.
• Start to write a run plan in html format.
  
• Reports on:
• PITA scan
• two scans were done last tuesday (9/15): the first one is by changing the Pockel cell voltage, the second one is by scanning the halfwaveplate angle such that a set point can be determined (the set point should have small Qasym but a steady and non-zero slope for feedback.) Xiaoyan will analyze those runs to extract BCM asymmetries and correlations.
• boiling test and run plan. Diancheng and Xiaoyan wrote a note of the HAPPEX boiling test (PDF). Xiaochao will read it and compile a runplan including the goal of PVDIS boiling tests.
• synch check:
• Report by Kai.
• Observations and problems spotted:
• Compared synch between HAPPEX and PVDIS, observed very good 1<->1 correlation for both left and right arms.
• PVDIS left arm shows slightly larger reading than right. This indicate timing difference between the two arms. We wonder how different the timing is, Kai will work on it. Kai will also extract the same difference for HAPPEX copies and see if is exactly zero.
• Xiaoyan commented ev_num should be >10 for all analysis.
• Our PAN cut first 600 pairs, not 16 in HAPPEX version. Xiaoyan will check why this is the case.
• See also Kai's log #2 and #3 on PVDIS LOG.
  
  
• FADCs:
• Right arm still only work for 5 channels. Crash if attempt to read more than 5;
• Kai will analyze some recent data to see if can see any PS, TS inputs on the left. If he can, it means the zero rate problem is NOT due to off/down NIM crates. <- note: in the evening he sent out FADC results which show that there is almost no signal (only a couple of events out of several runs).
• High stat Qasy check: Xiaoyan will send me results on this. The main observation is our BCM RMS is slightly smaller than HAPPEX runs. Bob commented this cannot be possible. Xiaochao suggested Xiaoyan or Diancheng to find out the V2F conversion factor and the ADC specifications used by HAPPEX, then we can see if PVDIS scaler have higher or lower sensitivities than HAPPEX ADCs.

• 09/14/2009

• Xiaoyan:
• The problem with BCM swaps were solved (problem in scalermap).
• How to run our version of pan?
• the current directory is ~apar/pvdis/pan_cvs/pan. To run pan, go to this directory and type "./usethis.sh" first, then "./pan -r [runnumber]";
• if you want to compile your version of pan, try:
1. go into /pvdis/pan_cvs/pan,  type:
           tar -cvf all.tar src
   It will generate a file called all.tar in the current directory. Copy this file to the directory you need;
2. In your own directory, type:
          tar -xvf all.tar
   A directory "src"  will be generated.
3. Before use, go to src directory, type:
      make clean; make install
   
   If the db file is properly generated and environment well set, you are
   ready to analyze runs with pvdis data.
• Posted a few HAPLOG entries: 1710, and (on double difference analysis:) 1721+1722;
• checked bcm u1 and d1 left vs. right, found multiple peaks consistently. Bob and Xiaochao note this could be from bit resolution problems of the scalers. For example, u1 counts is 8000 for 100uA. The spacing between peaks observed is about 60ppm, which is roughly 1/(2*8000). Bob said this is not a real problem since the Mean and RMS are still consistent with the input physics asymmetries.
• Bob:
• independently, solved the BCM swap problem. 
• Details: scaler9 and scaler13 should start at index 100 instead of 101.
  
  datamap   scaler     scaler9   scaler   9   0  100  28
  scaler9_0 scaler9_1 scaler9_2 scaler9_3
  
  datamap   scaler     scaler13   scaler  13  0  100  29 scaler13_0
  scaler13_1 scaler13_2 scaler13_3
  
• with Xiaoyan and Kai, made modifications to Pan to have a more desirable behaviour. To illustrate:
  
  1. On run 12158 pan will run normally;
  2. On run 12183 pan will run but print every 100 events a gentle notice that there are no PVDIS crates; (CODA was in HAPsimp mode)
  3. On run 11729 pan will exit with a loud warning that the data are out of synch (Integ=13300 problem) and that an expert should be notified.
• Bob suggested a few analysis works:
1. High-statistics charge asy:
   Although the feedback forces Qasy to zero, it still is a measurement.  Suppose, for example, that from a few days statistics we see Qasy = 0.05 +/- 0.02 ppm from both HAPPEX and PVDIS DAQs.  This would establish our ability to measure with small error bars.  If we can't we need to know soon!
   
   
2. PITA Scan:
   During PITA scans there are beam charge asymmetries like -5000 ppm, +5000 ppm and in between, with small errors (few ppm). We need to compare HAPPEX to PVDIS Qasy for these scans.    Unfortunately I think some earlier scans are useless because of the Integ=13300 problem (synch loss).  So unless someone can find a good PITA scan (please look) we may want to request another.
   
   
3. Double Difference:
   The difference between two asymmetries (asym_bcm2 - asym_bcm1) is a measure of intrinsic electronic noise.  We study these all the time for HAPPEX; should do the same for PVDIS, where bcm2(1) is either internal to PVDIS or from HAPPEX.  Should hopefully see a small width.  We sort-of already know the answer from the width of the correlation plots.
   
   On a different note, we'll need to get good at WAC-style analysis. HAPPEX has a well-organized army of analysis experts.   We probably need to persuade them to help us.
   
   
• Bob also suggested a better "sych check" than watching the thickness of bcm asymmetry correlations. He mentioned 3 ways for HAPPEX DAQ:
1. A 12-bit DAC ramps up and down and the signal is sent to a v2f which feeds a scaler in each crate; it's easy to give the PVDIS crates a/many copy.
2. Copy of 12-bit DAC (just the voltage) is put into a voltage-mode ADC in each crate. (this cannot be adopted by PVDIS).
3. Look for physical correlations like lumi vs bcm, and detector vs bcm.  Of course none of the regression analysis would work if HAPPEX was out of synch.  For PVDIS it's bcm vs the HAPPEX bcm ("bcm1") but it's a little subtle because even if you are out of synch this may still look "sort of ok"; but the correlation is fat when it's bad and skinny when it's good, and we need to get used to what is fat/skinny. <- Xiaoyan is already doing this.
• Xiaochao: 
• Checked tagger and trigger rates, see HAPLOG entry 1716. These runs were from before last thursday (9/10) access, during which Ramesh changed medium threshold from -20mV to -30mV. Bob commented that this may explain why some groups' PS rates are doubled, tripled, or even quadrupled -- they are due to multiple after-pulsings.  Xiaochao will look at runs after 9/10.
• Kent has created an elog called "PVDIS", in parallel to HAPPEX. For everybody here, the username is the same as your jlab ID, and the password is "[firstname]iscool" with your own first name inserted. We should discuss what "Type" to include. Some issues specific for PVDIS are: DAQ timing, DAQ tagger rates, DAQ signal rates, deadtime analysis, etc.

• Remaining problems and Plan:
• We should work on our communication routines: always inform others your findings by making halog or haplog entries. If you did the hardware work, track all the way down to data+software to make sure it works!
• BCM copies in left and right PVDIS crates do not agree exactly (and have different correlations with HAPPEX bcms); <- See Xiaoyan's report above for possible explanations. Do we still worry about the different correlations????
• left arm fbTDC still not working; <- Ramesh and Bob will work on this during the Tuesday (9/15) access;
• two scaler channels are not working:<- Xiaoyan and Diancheng added these signals to other scaler channels during the Tuesday (9/15) access. BCM d10 is also working now.
• bcm d10 still not working (signal seems to be there but no scaler reading, left & right); 
• scaler9_30 not working (coin of tagger and electron narrow gr4 on the right);
• Paul is supposed to work on the paperwork, but no news so far; <- After the meeting, he said he will work on it this week.
• High stat Qasy check: Xiaoyan will analyze more runs for this;
• Have we had any experience with HAPPEX boiling tests?
• PITA scan:  is HAPPEX going to do another one?
• Tuesday (9/15) access (in addition to items listed above):
• Kai will work on FADCs;
• Kai, along with Ramesh and Luis's help, will add 4 DAC signals to PVDIS crates: gated and ungated, one copy each to each arm;
• Ramesh will raise the right arm DMedium threshold from -30mV to -50mV; 

• 09/07/2009

• Riad explained the basics of the source helicity board and the different between the new and the current boards. Note: the new helicity board will be installed on 10/26/09, the first day of the HAPPEX->PVDIS switchover. Below is a sketch of the new board timing. The helicity predictor will also be changed, see here.

• Ramesh:
• Left arm fbTDCs are still not working, need Bob's help to fix this.
• Made the first draft of run plan (tex file in the same directory), waiting for Xiaochao to read.
• Kai:
• Finished writing the timing-check macro and the online GUI for the left arm. These are based on Xiaochao's right arm codes used during the test run. But we cannot determine the various timing cuts used in the GUI, they will be checked and decided on during commissioning.
• Note: 1 week before the switch over, at least 2, preferablly 3 people should become familiar with the timing check technique;
• Xiaochao
• checked roughly the scintillator signal and found the timing reasonable. Can see some rates for all 8 groups, made from the scintillator signals, on the right arm.
• checked the tagger rate and found the 100kHz clock has a count of 50 in the right arm, which seems to be due to a 2kHz helicity flip rate or something else being wrong. Using the same flip rate can see the 100Hz tagger, but otherwise (if using 30Hz) cannot see any tagger rate. Note: HAPPEX is running at 30Hz!
• Xiaoyan did some checks on the BCM readings in the PVDIS crates. Finding:
• Left arm BCM copies have multiple swaps, for example, u1 could be u3, u10 could be u1 or d1, etc. Right now none of the six copies have the reading indicated in the scaler map, and at least 1 copy is missing. Ramesh will check the cables and possible swaps during the next access;
• Right arm BCM readings are much smaller than left arm. THis combined with Xiaochao's tagger check leads to the conclusion that the MPS signal was inverted during the access last Friday (later Mark, who made the access along with others, commented that this is possible).  At the next access Diancheng or Ramesh will go to the hall to fix this.
• Next step: Once the MPS is fixed, Xiaochao can start checking the tagger loss/deadtime. Continue checking all BCMs. Then can move on to asymmetry checks.

• 08/31/2009

• Xiaochao explained the script "group_any.cc", which she used to check the timing of all groups during the April/May testrun, and made decisions for adding/removing cables numerous times.  In the following week, 
• Ramesh will modify this script so it can be used for upcoming runs (from HAPPEX-III). It is not working now because of the different TDC range.
• Kai and Diancheng will read the script and learn how to use it.
• Kai will write a similar script for the left arm.
• The goal is that at least 2 more people know how to make decisions during commissioning.
• Diancheng was working on resonance simulations;
• Kai explained the rate scan he did using June 2nd test run data, with the pion/GC random coincidence removed. Possible explanations are:
• There are other factors not being taken into account;
• Only d10 was in the data stream and it works only up to 35uA. There was only one run that had <40uA current, and all lower current data from this run were from current ramping. THe beam may not be stable during ramping, resulting in a systematic rate change;
• Ramesh is volunteered to draft a run plan. 
• Will start the outline based on HAPPEX-III run plan;
• Nilanga will write the optics calibration part;
• Diancheng will draft the target boiling test part
• Paul R. has volunteered (or has been volunteered) to deal with the COO/ESAD/RSAD.

• 08/03/2009

• Left PVDIS crates were installed last friday after the detector stack was pulled out. The RHRS detector stack has been pulled out this morning.  
• Ramesh: 
• Working on cables for the pion rejector from the patch panel to Sum8s.
• Checked GC and T3 circuit on the left arm. The T3 should have a deadtime of ~100ns. GC circuit is almost identical to the right arm setup (i.e. deadtime dominated by the input pulse).
• Added test pulser to the GC circuit on the right arm: input the test signal as the PMT input.
• Diancheng and Xiaoyan: A short summary of the asymmetry/rate tests, in the order of when the tests were done
• with original helicity as the control signal of the asymmetry module, found:
• deadtime for the left-pvdis narrow path is reasonable. For the left-pvdis wide path and T1right has slight disagreement between the rate and the asymmetry methods. 
• Note: During this time the T1r test signal was sent to the final logic module so the T1r copy should have a deadtime of ~100ns. One problem was that the T1r test signal had double-pulsing about 36ns apart due to cable reflection.
• with reversed helicity as the control signal of the asymmetry module, found:
• Note: during this time the T1r test signal was sent to the PMT inputs of the T1 circuit so the deadtime should be the total deadtime of the T1 system;
• The narrow path was changed to 60ns wide during this test;
• Also added T3 in this test. THe test signal was sent to the PMT inputs of the T3 circuit.
• deadtime for the left-narrow and left-wide, and for T1r agree with the expectation. 
• T1r has a deadtime of ~120ns.
• T3 has a deadtime of ~180ns and the results from the rate method is different from the asymmetry method. Also, this large deadtime is not consistent with the T3 circuit. Diancheng mentioned the test signal (inputs to T3) is attenuated and he had to amplify it first. Could this distort the signal and perhaps widen the input width??  All these have to be investigated further.
• Also compared the absolute value of the asymmetry itself between "original-helicity-controlled" and "reversed-helicity-controlled" data. The difference is consistent with zero for most data points. Only the highest rate data (lowest asymmetry) is several sigma below zero. We plan to redo this point.
• Plan:
• Restore test setup so we can run tests for T1, T3, and GC simultaneously. The tests will be done by Kai with help from Xiaoyan and Diancheng.
• Scope-check the test pulser inputs to T3 (note: On tuesday we found the attenuation cannot be reproduced because both stacks are out. Instead the T3 test signal is sent directly from the lower to the upper level on the left hut).
• Scope-measure the deadtime of the T3 circuit (note: On tuesday we did this and found the deadtime should be close to 100ns. The measurement was not as clean on the right arm but the deadtime appears to be no larger than 150ns in any case).
• Kai will study the random coincidences; Xiaoyan and Diancheng will study the target boiling effect and will have a written report on what specifications we need to meet for PVDIS.
• Xiaochao will investigate further on BCM linearities from 2007 data.

• 07/27/2009

• Ramesh checked the T1 and GC circuit on the RHRS upper level. Found GC deadtime should be dominated by the output width (confirmed from scope check). The T1 circuit is inconclusive: one of the modules in S1 formation has 134ns output width and another module in S1 has 122ns. The final output is 100ns. Scope check see ~120ns when the input signal is clean.
• Left PVDIS setup (on the floor) is almost ready. Right now DW, XD and KP are doing asymmetry tests:
• The input is two high rate pulser ORed. One being the output from the asymmetry module, 5KHz, 1.1%. The other (called "gate signal") is adjustable. The two provide adjustable asymmetry and rate.
• Click here for results.
• Some interpretation of the current data: 
1. Left PVDIS narrow and wide groups behave as they should; 
2. Right T1 circuit is found to have a deadtime of ~120ns. THis is measured by sending the test pulser to PMT inputs of the T1 circuit (paddle 1 L+R on S1 and paddle 5 L+R on S2). 
• More results will follow:
1. reversed-helicity;
2. made multiple T1 copies to scalers (25ns, 125 and 150ns, plus the original 100ns). Comparison of the rate or asymmetry of these should provide additional measure of the T1 deadtime.
3. add the same test to T3 on the left HRS.
• Identify tasks to be completed within the next 1-2 months:
1. Finish asymmetry tests and left PVDIS installation;
2. Install FADCs and test them (16 signals per arm can go to FADCs);
3. Come up with a plan for the deadtime rate-scan method. We are facing at least two problems:
1. BCM linearity. Should we do dedicated BCM calibration runs during our experiment? How well can we measure it?
2. accidental coincidences. How much do we expect? What are the effect on asymmetries and the rate-scan slope? Any plan to measure it in-sito, perhaps with some dedicated beam time?
4. T1 and GC deadtime: 
1. Should we analyze EDTM data from experiments this spring? 
2. Asymmetry/rate test should tell us more about T1;
3. Is it necessary to add an EDT module to GC?
5. Add GC to scalers directly, if can solve the after-pulsing problem.
6. Simulation of DIS, resonances and rate estimation for the run plan;
7. Understanding boiling tests and come up with a run plan;
8. Trigger simulation;
9. Other problems identified from the 5-pass test run:
• TDC spectrum for both electron and pion VETOs do not look correct. They both have a 2nd peak roughly 100ns after the main peak. Upon inspection we found the first two inputs of the NIM/ECL module seem to be broken and do not provide the required 50ohm termination. These are the inputs used by the two VETOs. The NIM/ECL module is replaced. But this could well not be the cause of the problematic TDC spectrum.
• RHRS gas cherenkov PMTs have severe after-pulsing. 3 PMTs checked have double -pulsing 110ns apart, each pulse being 20ns wide (this is 100% after-pulsing). THe final GC output (OR of 10 PMTs) have about 70% after-pulsing.

• 06/28/2009-07/21/2009

• PID analysis using the 5-pass test run data is done. See Kai's report (TBP) at the last parity meeting;
• Asymmetry analysis using the 5-pass test run data is done by Xiaoyan. Most asymmetries are consistent with zero (which they should be). Remaining problems:
• pion scalers for gr1-3n have problem.
• A re-visit of asymmetry analysis using polarized 3He data can be found here (Xiaoyan):
1. Transverse target:
1. T1 asymmetry from HRS scalers (after removing an unpolarized run):  PDF gif;
2. T1 asymmetry from PVDIS scalers with tagger dilution corrected: T1 (PDF gif), pull plots (PDF gif);
3. electron asymmetries from PVDIS scalers with tagger dilution corrected: asymmtry (PDF gif), pull plots (PDF gif);
2. Longitudianl target:
1. electron asymmetries from PVDIS scalers with tagger dilution corrected: asymmtry (PDF gif), pull plots (PDF gif);
2. Older results on T1 asymmetries can be found in this directory.
3. Current interpretation: 
1. Transverse target T1 asymmetries are in good agreement between HRS and PVDIS scalers;
2. All other asymmetries require close inspection.
• Came up with an explanation for the VETO deadtime (revealed by the tagger loss analysis). The cause of this deadtime is a combination of two factors:
• The so-called "updating" and "deadtime-free" PS755 is only updating when the output width is wider than the input and if the subsequent inputs happen after the initial input returns back to 0;
• The timing sequence of tagger signal is different from real physics events.
• Click here for some simple write-up on this (scroll to bottom of page). Some methods are proposed (and implements on the right PVDIS setup) to prevent both factors. 

• 06/18/2009

• We had about 16 hours of 5-pass beam and took some data. Copies of run sheets will be distributed. We had:
1. Ran LHRS at two DIS kinematics that are close to the two Q2 points of the experiment (same momentum but different angles) -- called left 14 deg runs and 18 deg runs;
2. Ran RHRS at one DIS kinematics that is close to the high Q2 point of the experiment (same momentum but different angles) -- called right 18 deg runs;
3. Ran both HRS at 18 deg and 1 GeV/c to get some high rate pions -- called low momentum or pion runs;
4. For all kinematics, took half data with HWP in and half with HWP out;
5. For all kinematics except pion runs, took half data with tagger ON and half with tagger OFF; 
6. Took HAPPEX runs throughout this period.
• A rough analysis task list for the testrun data:
1. Use onlana_R.C and onlana_L.C to analyze all HRS runs, with full statistics (all events) -- name should be assigned here. This should be done first, and send the whereabouts of the rootfile to others;
2. Use HRS data (rootfiles) to extract approximately the pion and electron rates and compare them with prediction. This will give us confidence whether the pi/e prediction for the experiment is reliable -- Xiaochao, after June 29;
3. PID: -- Kai
1. Shower and Preshower:
1. electron efficiencies: 
1. For RHRS 18deg runs, extract e-efficiencies from all runs using hardware TDC (en and ew trigger) cuts. Check if there are significant changes from run to run. If not, can chain runs together for higher statistics.
2. For LHRS 14deg and 18deg, and RHRS 18deg runs, extract from all runs (can use chain), using software cuts
1. Software cuts should be: RHRS: PS*2+SH>TScut and PS>PScut; LHRS: PRL1+PRL2>TScut and PRL1>PScut.
2. Grouping scheme for LHRS please use "both overlapping" in this updated figure (note there is a small difference here: there are 6 groups, not 5. And there are two-row overlapping between gr1 and gr2, and gr5 and gr6. Other groups have one-row overlapping).
   
3. Vary software cuts within a large range and see how results change;
4. Compare RHRS software cut results with TDC cut results from i), see if there is a significant difference; Note: TScut was about -105mV (90mV correspond to TScut 2000, scale to 105 to get the new cut), PS cut was -20mV (correspond to PScut ~350);
5. Note that the two kinematics setting on the LHRS have different momentum. Different cut ranges should be used here.
3. when presenting results, also present the Preshower vs Shower (or PRL1 vs. PRL2) spectrum for all electron samples (Cherenkov cut > large #).
2. pion rejections:
1. For RHRS 18 deg runs and low-momentum pion runs, extract pion-rej from all runs using hardware TDC cuts (PS && TS, not en or ew). Chain runs together for higher statistics.
2. For all runs (LHRS 14deg and 18deg, RHRS 18deg, and both LHRS and RHRS low-momentum runs), extract from all runs (can use chain), using software cuts. 
1. Vary software cuts within a large range and see how results change (for 14deg and 18deg runs, use same cuts as in A. electron efficiencies);
2. Compare software cut results with TDC cut results from i), see if there is a significant difference;
3. Note that for runs with different momentum settings. Different cut ranges should be used;
3. when presenting results, also present the Preshower vs Shower (or PRL1 vs. PRL2) spectrum for all pion samples (Cherenkov cut <1).
3. Combine software cut results from A-ii and B-ii to study cut dependence. Can use averaged efficiency and rejection factor within the central geometrical region.
4. Pion efficiencies:
1. Use Cherenkov cut<1 to define clean pion samples;
2. plot PS vs. SH, or PRL1 vs. PRL2 for pion samples, see if the current TSlow cut is reasonable;
3. use both TDC hardware cuts (cut on pion triggers narrow and wide), and software simulations, study pion efficiency for each group;
4. If some groups show significantly lower effiencies, find out why;
5. From PS vs. SH or PRL1 vs. PRL2 cut, check how much electrons are "contaminating" the pion triggers.
2. Cherenkov:
1. Use Preshower vs Shower (a.sum_p) to define clean electron and pion samples, and study electron efficiency and pion rejection factor from using different Cherenkov cuts.
4. Asymmetries/PAN: -- Xiaoyan
1. Check why ok_cut was not working for half of the HAPPEX runs. If this was a hardware problem with bcm1, use other bcm instead;
2. Analyze all runs, extract:
1. From electron runs (RHRS at 18 deg and 2.63GeV/c):
1. en, ew rate from all groups (if tagger was on, the true rate for gr5-8 should be the rate measured from scalers minus the rate of the tagger, the tagger rate can be obtained from its corresponding scaler).
2. en, ew rate from total triggers (OR of all 8 groups);
3. en, ew asymmetries and errors from all 8 group triggers and the total trigger (for asymmetries, do not correct for the tagger yet);
2. pi-n, pi-w asymmetries and errors from all 8 group triggers and the total trigger;
3. T1 rate and asymmetries. Note: if there are two T1s: one is T1 itself, and the other is T1.OR.tagger, extract them BOTH;
4. Tabulize all results, along with the HWP (in/out) and tagger (on/off) conditions, from all HAPPEX runs during the test period.
5. Deadtime/tagger (from PAN): 
1. Using cut on bcm3, analyzed how tagger loss depend on rate of each group, click PNG PDF for plots. It looks like there is a large (10 times expected) loss independent of the rate in each group. Suspect this loss is from the VETO input to the ANDing modules (allowing PSM & TSH to form electron triggers);
2. A HAPPEX run with VETO unplugged shows almost zero losses (within its poor statistics);
3. Another HAPPEX run without beam (31554), but with a 10kHz pulser ORing with T1 and GC (hence it looks like there is a 10kHz input to the VETO side, with random timing relative to the 100Hz tagger), shows a 0.0002 loss in tagger.
4. Will investigate this within this month using high rate pulsers in the hut. -- Xiaochao, perhaps with help from another person.
   
6. Tagger pileup effects (from PAN): -- Ramesh
1. Make plots of pileup TDCs from all HAPPEX runs with tagger on (individual or chained), look for pileups and extract the pileup fraction.
• Other things to discuss/announce:
• We will assemble the LHRS DAQ in the hall, instead of in the EEL building.
• HAPPEX DAQ will be under developement (Luis, Bob) and will not be available. Bob made a standalone code. Someone should take a few runs and test if we can use PAN to analyze these data (mainly scaler information and their asymmetries). This should be done within the next two weeks -- Xiaoyan or Ramesh?
• Xiaoyan analyzed asymmetries from transverse polarized 3He target running: ew trigger and T1 from both PAN and regular DAQ can be found in this directory.
• Xiaochao did the beam derivative study (x and y only) using her Fortran-based simulation, results agree with Diancheng's within a factor of two.
  
• A rough plan for next week:
• Keep RHRS PVDIS racks as intact as possible;
• Test if the standalone HAPPEX DAQ works;
• Start analyzing the testrun data following the task list above;
• Collect all items needed for the LHRS DAQ in the hall. Purchase whatever are missing within the week;
• Start putting together a cable list and labeling scheme using the "both PRL1 and PRL2 overlapping, 1->2 splitting on all blocks, 6 groups total) scheme. Again, click here for latest grouping scheme (should be confirmed by Kai's PID analysis).

• 06/08/2009

• During the 6/2 access we re-grouped Shower blocks. 
• Patch Panel between Sum8 back panel and the 2nd-level ADCs: front map (PDF);
• map of Sum8 inputs: PDF;
• cable swap that were fixed later in the analyzer database: txt;
• New grouping of Shower: PDF, PNG;
• electron efficiencies from the new grouping method: 
• Xiaochao:
• Estimated rates for the 6GeV test run with a 100mg/cm2 carbon target: PDF;
• Run plan for the 3-pass pion test run is here;
• Working on TOSP for the 5-pass test run;
• Xiaoyan:
• Asymmetries for the longitudinal running: from PAN, from T1 in end-of-run files;
• Diancheng:
• Derivatives of asymmetry w.r.t. beam properties, for pVDIS 1.1 GeV^2 point: click here. A linear relation is seen for all variables. Note: When running HAMC twice, the program picks up the same two sets of random numbers. So there is an "offset" that should be subtracted from all iterated asymmetries. With this subtracted, the linear relationship is obvious. Bob commented that we can choose to "randomize random seeds".
• Kai:
• Presentation on PID, focus on efficiencies after the shower regrouping and simulations on the LHRS: click here(PPT);
• for LHRS lead glass grouping method, found we need to overlap both PRL1 and PRL2 to avoid the "hole" in electron efficiencies.

• 06/01/2009

• General:
1. We requested a low momentum running to test the pion rejection of our DAQ. This is likely to happen before the pass change next week (around June 9th), starting just pass midnight and last the owl shift. Xiaochao will writeup a run plan, and maybe one of us need to be on that shift to ensure data taking;
2. In mid June, just after the current experiment ends, there will be a couple of days of possible beam delivery at 6 GeV and up to 100uA. This is a perfect opportunity to test the deadtime measurement of our DAQ, and any electron/pion PID performance since we can set the HRS at wherever we like. Doug has helped to setup a simple target (a carbon foil) and the polarized 3He group will install it after removing the current target. Other groups who plan to also use this beam time is the Compton crew. Xiaochao will find out if we need safety documents, arrange for shift sign-up, and think about what to put in the run plan. Meantime, it is strongly encouraged (if not required) that everyone here sign up for shifts during that period and treat it like a training of the real running.
3. There is a question whether our electronics in the RHRS will interfere with the HAPPEX detector. Jack is working with the designers (Joyce, Al, Susan) looking for a creative way of supporting the HAPPEX detector without removing our racks. But there is a possibility that we will need to remove the racks and re-install them after HAPPEX-III. The same problem exists for the left HRS: Should we install the DAQ in the left HRS and then remove it before HAPPEX-III?
• Kai:
1. Focused on electron efficiency work using both the fbTDCs and the software "simulation" using ADCs. 
1. For the right arm current setup, saw "holes" between groups.
2. Tried two different grouping methods: shifting all preshower groups by one row, and two rows, respectively.
3. Results can be found in this directory (README included)
2. For the left arm, also did software "simulation", click here for results. Do not see "holes".
3. Will do the left arm pion rejection "simulation" using transversity data. There was no pion data on the right HRS in the past 1-2 years.
4. The full presentation can be found here(PPT).
• Ramesh:
1. Made numerous accesses during these two weeks. For a list of all activities see here;
2. was RC;
• Xiaochao:
1. Latest timing and spectrum can be found from run 21755. Remaining problems are: 
1. VETOs are still cutting a bit into electron and pion narrow triggers. Electron loss can be seen from PS vs SH spectrum. Pion loss is obvious from fbTDC spectrum. But see "Plan" below.
2. Do not see clearly the TS cut in group 2, haven't found a reason.
2. Don't remember what else I did except fighting for the pion beam time.
• Diancheng:
1. Comparing HAMC simulation with Xiaochao's fortran version, may have found problem (in target<->transport transformation). 
2. Will work on the beam derivatives this week. If HAMC cannot produce the same results as the Fortran version, will need to use the Fortran version for this task.
3. Will add the electron efficiency (function of ps.trx) to the simulation.
• Xiaoyan:
1. Extracted scaler asymmetries for the transverse target running, see here; Note: we observed that there is always a 1000ppm shift between HWP OUT and IN. This can be seen from latest longitudinal running, and from the transverse running (although for transverse this is not as obvious because of the larger physics asymmetry). We did not, however, see a shift in T1 data (see for example Bob's HAPLOG entry during which group 5-8 had T1 mistakenly built in. His analysis showed almost the same value for different target spin directions). This means the problem is probablly not in the timing or charge asymmetry. We need to think about what can cause this shift. Question: during PVDIS can we use our timing board for the LUMI to check possible false asymmetries?
2. Checking the PAN data in a "microscopic" way, need to write macros to check data quality. Some examples of this kind of check-outs were provided by Bob and are posted at the HAPLOG. Xiaochao note: many of the macros here can be adopted later into an online check-out code.
• Plan and assignements:
1. Next access (tuesday) work -- Ramesh, Xiaochao, and students
1. add the 100Hz tagger to the trigger superviser (one of T3,T4,T5,T6);
2. check timing of all triggers at the TS, make sure they arrive within 50ns of each other;
3. try to implement the new preshower grouping method without affecting the current experiment (got okay from Doug). The inputs to Sum8's will be changed. Then, since a copy of all PMT signals are sent from the Sum8 (back) to a patch panel, then back to the RHRS ADC on the 2nd level, cable connections on the patch panel need to be swapped. Here (PDF,  ODS) is the cable swap list. We will take comis runs to check that all channels are restored correctly.
2. Need access but not the earliest one -- Ramesh
1. add tagger to group 1-4 (this also fix the timing problem in narrow signals);
3. Check PAN/scaler data, write general check-out macros for PAN -- Xiaoyan;
4. Get the scaler reading for the deadtime measurement -- Ramesh, Xiaoyan;
5. Simulation for beam derivatives -- Diancheng (see above);
6. Working on a summary of beam requirement (needs simulation results) and provide them to Kees by next monday -- Xiaochao;

• 05/18/2009

• Introduced Kai Pan from MIT.
• Ramesh:
1. Activities during 5/11-12 access can be found here and here. For 5/14 access see here.
2. Revising deadtime measurement setup. Here is a picture of what we used in EEL. Now we need to mix a clock/pulser signal with all three: Preshower (at the analog level), Shower (also analog), and the electron VETO (by ORing clock with both T1 and GC).
3. Next access work: list.
• Xiaochao:
1. Happex-III made their request on the beam quality (Table, and email). We had an earlier email exchange with Kees on the transverse polarization of the beam. Everything else needs to come from the simulation, which we have been given a month (that's 3 weeks from today).
2. Made a plot of Preshower hit position for events producing PVDIS electron triggers: 2D, and 1D dispversive only. Found PS block 14 ([13]) does not have ADC signal. This was corrected on 5/14 by Ramesh (bad cable from PMT to downstairs).
3. Analyzed run 21137. Regardless of multiple changes we made this week, it's still not perfect. Reasons:
1. Each group is different and now we are adjusting the delays at 8ns level;
2. Shower PMTs are not only slower by 40ns, (just guessing) each PMT is different from others and the time spread is much wider than PS;
3. But, twe found out why all Showers are 40ns behind Preshower although the cables are exactly 60ns everywhere: PMTs for the Shower are different from Preshower, they are in average slower by 30-40ns due to slower drift speed of electrons in the tube.
• Diancheng:
1. HAMC simulation results. Xiaochao's previous results.
• Plan:
1. fix timing (Ramesh's access); working on tagger system (Ramesh); think about EDTM.
2. Use PAN to extract electron asymmetries;
3. Continue with comparison of BCM asymmetries between HAPPEX and PVDIS?
4. Check HAMC simulation to find discrepancies;
5. Kai Pan will work on PID analysis on the left arm using different grouping schemes. (note: the figure linked here has a mistake: there are only 2x17 blocks on prl1 and prl2, not 2x23).
6. Next time: discuss about DAQ on the left arm, cable and module list/check.

• 05/11/2009

• Alexandre:
1. Showed progress on Compton edet: now can see Compton edge from more runs. Tried extracting asymmetries. May have some number for the beam polarization by tomorrow's Hall A meeting.
• Ramesh:
1. Activities during 5/5 access can be found here.
2. Activities during 5/8 access can be found here.
3. Today: Implementing delays to all other groups (following the delay scheme of group 4, see below).
• Xiaochao:
1. To check results of the 5/5 access, analyzed run 20792 (click for all histograms):
1. (groups 2, 4, 5, 6, 7); (note: groups 2,6,7 had 37ns removed from TS SUm8 -> DH);
2. Analyzed what's going on with the VETOs; On top of this, all VETOs arrived too late by 50ns.
2. To check results of the 5/8 access, analyzed run 20928 (click for all histograms).
1. Delays added to group 3: 0, 16, 48ns, for TS Sum8<->740, PS Sum8 <-> 740, and PS Sum8 <-> DM.
2. Delays added to group 4: 16, 32, 60ns.
3. Delay added to T1: 16ns.
4. Present conclusion: timing alignment for group4 is working fully. Will add the same delay to all other groups; Need to understand what is causing the TDC peak near 600 in the pion veto (.not.GC .and. T1).
3. A summary of changes, status, and interpretation of the TDC data is here (until run 20792).
• Diancheng:
1. Working on Compton edet with Alex;
2. Worked on group-dependent PID. Found large overcounting between groups (more to be posted);
3. Bob suggested looking at the geometrical distributions of overcounted events first to see if they indeed come from the boundary of groups.(more to be posted);
• Xiaoyan:
1. After correcting the helicity lag in software, try to compare BCM readings of HAPPEX and PVDIS again. See below for plots:
1. BCM reading, HAPPEX (bcm9) vs. PVDIS (dis): before correction, after correction. Conclusion: with correction for the helicity lag now they fully agree.
2. Asymmetry of BCMs, HAPPEX (bcm9) vs. PVDIS (dis): before correction, after correction. Note: We don't see a clear correlation between the two, either before or after the correction;
3. To check why we don't see a clear correlation, compared asymmetries from ADC18BCM and HAPPEX (bcm9): before correction, after correction, but still don't see correlations.
4. Next step: Bob suggested 1) making the asym comparison without the "OK cut" since it selects only steady beam. During beam ramping we expect to see large asymmetries, which may reveal better correlations. 2) Ask Luis to make the ADC18 vs. bcm9 asym plot and see if his results are the same;
2. PAN with the helicity lag correction seems to have problem:
1. It gives error messages once in a while;
2. It runs much slower, (not due to extra printing);
3. The "timeslot" variable after correction (PDF) is different from before correction (PDF). Bob note: This variable should always be "1" when there is no supersampling. This error may indicate data corruption, etc.. Need to print out full data set when the error occurs to diagnose.

• 05/04/2009

• Ramesh:
1. Plots for run 20264 (low threshold) and 20209 (high threshold). Both runs were taken with 1 pass beam;
2. Access work on 05/01, friday. (Note: in this entry, "GC veto" means "GC&SC", "SC veto" means ".not.GC & SC", in the PVDIS trigger diagram;
3. According to Huaibo's test earlier, all shower signals were delayed by 2x37ns, while all preshower signals were delayed by 37ns. This may explain some data below.
• Xiaochao:
1. Suspect timing mis-alignment between PS and SH signals. See here for an explaination (email on 4/29). Plots for all groups from run 20466.
2. On Friday May 1st, Ramesh added a 32-ns delay to Preshower in group 5, between the Sum8 output and the LeCroy 428F. Also added the VETO signal (Cherenkov + Scintillators) to groups 1-4. 
1. Run 20695 was analyzed and results are here: group5, while groups 6-8 remain similar to run 20466;
2. Groups 1-4 did not produce any electron or pion trigger.
3. Also looked at VETO signals using fastbus TDC. Here is a plot from run 20695. I suspect both VETO signals are too late by 124ns (if removing the extra 37ns delay in Shower signals), or 84ns (if adding extra 32ns delays to all preshower signals), but this could also be a problem in inverting both VETOs for the TDC (since the VETO themselves are 100ns wide, not reverting them could cause them showing up too late by 100ns in the TDC);
4. Current conclusions:
   
   1. Preshower signals are too early by 40ns everywhere (except groups 1 and 8 which need further evaluation). Need to add delay also between PS Sum8 and Discriminators (706 with PS-Medium threshold), such that "AND" can be formed between PS and TS triggers for electrons.
   2. VETO signals seem to be mis-aligned in timing also, need to investigate with the fastbus TDCs;
   3. (obvious) threshold should depend on p0 setting of the run.
   4. Current conclusions:
• Xiaoyan:
1. Checking helicity difference between PVDIS TIR, HAPPEX TIR, counting house (CH) TIR and PVDIS flexio. click this directory for plots. Found:
1. Flexio completely agree with CH TIR;
2. HAPPEX TIR and PVDIS TIR disagree with Flexio for about 10^-6 of events;
3. (Bob note: in-consistency could come from noise, glitches, etc.. during experiment we could cut away events with in-consistent helicities).
2. Checking helicity reading in the PVDIS scaler data stream, found reading lag by 1 pulse (in this file, the 3 columns are: event#, helicity from HAPPEX DAQ, helicity from PVDIS DAQ). Here helicity signals are sent to a v2f and then to the PVDIS scaler, hence the conversion of:

   ((scaler6_2-4000)&0x2000)/8192

3. Want to compare BCM reading between HAPPEX and PVDIS crates, but current BCM readings seem to be wrong (too small). See this file for plots. Will check the signal during Tuesday access.
• Diancheng:
1. First time see electron spectrum vs. strip number from edet commissioning. See this halog.
2. Working on PID efficiencies using run 20466. Found pretty high electron efficiencies. But pion rejection is not good because there is almost no real pions at the current running kinematics.
• Plan for Tuesday access:
  1. Check if can find the extra 37ns delay of all shower signals.
  1. If found:
  1. remove them, and remove the 32ns delay to group5 Preshower sum8 output;
  2. remove 124ns from both VETO signals (somewhere between 755 and veto inputs to 758 or 622);
  2. If not found, 
  1. will add 40ns (or 32ns) delay to the cables between PS Sum8 and LeCroy 428F, and between PS Sum8 and "DM" (706 with medium threshold), for all groups;
  2. remove 84ns from both VETO signals;
  2. Check why fbtdc[66] (GC only) does not have any signal;
  3. if there is time, change threshold to: -40mV(low), and 120mV(high), while keep the medium unchanged.

• 04/06/2009

• Ramesh: 
• slides(PDF) for the Apr 17/18 meeting. Click here for agenda.
• an open-frame rack has been ordered by Jack and has just arrived last friday;
• trigger cabling is done, but will confirm next week and restore taggers;
• will work on: online scripts to extract information for the scaler/TDC deadtime measurement and calculate the deadtime;
• will write the online analysis code to study/determine discriminator threshold vs. corresponding PID cuts;
• will show Xiaoyan and Diancheng how to write/test script for RHRS (or LHRS) PID analysis, we should have this script fully debugged so we can use it as soon as the production beam starts on Apr.23rd.
• Xiaoyan:
1. updated results on the DAC/V2F test can be found here (run #31129):
1. scaler 9 agrees well with scalers 6,7,8 after higher bits are masked;
2. scaler 8 now also agree with others after changing the input signals. Right now the setting is:
1. V2F(ECL output) -> JLab NIM/ECL level translater ->(4 identical NIM outputs)->NIM/ECL module -> scalers, where 4 channels of the last NIM/ECL module are used to make 4 identical ECL ouptuts for scaler 6, 7, 8.  One of the "4 identical NIM outputs" is sent to scaler 8.
2. Last week, the input of scaler 8 was from the JLab NIM/ECL level translater, sent by two short cables with a connector in between. Now the connector is removed and a single cable is used. This seems to solve the problem with scaler 8.
3. Xiaochao suggested using a fan-out for the input of scaler 6,7,8 such that we do not need to chain multiple channels, which could potentially cause different counts at very high rates.
2. updated results on the helicity test can be found here (run #31137):
1. The inputs are: 
• bar(NIM) -> V2F to scaler4_2 (ungated) and scaler8_2 (gated);
• [bar(NIM) .or. helicity] -> V2F to scaler4_3 (ungated) and scaler8_3 (gated);
• where the bar(NIM) is from a 740 with inputs left empty.
2. Observations and problems:
1. from the scope, found (bar(NIM)) has only a small voltage, while we expected it to be -800mV. The scaler result of scaler8_3 indicates that there is either no or a positive offset, i.e. previous 0V is now a small >0 value.
2. results from scaler8_2, on the other hand, indicate that the bar(NIM)->V2F is giving a high rate signal, i.e. the bar(NIM) is indeed -800mV.
3. To do: 
1. check on the scope to see if bar(NIM) -> V2F gives a high frequency signal;
2. then, first need to fix the bar(NIM) signal. Will terminate all unused inputs and outputs, and check on the scope for both the bar(NIM) and the [bar(NIM) .or. helicity]. Need to make sure [bar(NIM) .or. helicity] has two levels: ~ -800mV and ~ - 1.6V.
3. if can modify the setting of scaler 4, will try to use gated;
4. send [bar(NIM) .or. helicity] to either scaler 6 or 7, and use two channels: one gated and one ungated;
5. send [bar(NIM) .or. helicity] to one more channel in scaler 8 but ungated;
6. (adding the ungated allows direct comparison with scaler 4, and using scaler 6 or 7 will eliminate any difference the NIM input of scaler 8 would cause.)
7. Repeat the test.
3. After the meeting we went to Chris and asked for the specification of the JLab-made 16ch V2F module.
• Diancheng:
• Compton edet:
• not so easy to calculate Eloss in the strip. Strip thickness can be found here. Using minimal ionization loss, or the total loss from this website, the Eloss is about 0.2MeV/plane, not big enough to explain the counting rate loss from plane A to D;
• Scope check did not find any coincidence between planes C and D when using D triggering.
• HAMC:
• continue working on asymmetry sensitivity to beam position. Delta(A) seems to be linear to y_beam for large y_beam values, see here. The asymmetry from a 10K trial simulation (220 trials succeeded) are here(x-section-weighted) and here(x-section-weighted) But the linearity disappears for smaller y_beam.  Did not observe this linear relationship for changes in x_beam.
• Need to understand what is the error of the asymmetry from simulation. 
• Here is Xiaochao's prescription for how to get the simulation uncertainty. The result is consistent with what Diancheng observed. One can see that the uncertainty is determined by the acceptance width and is about 1ppm for PREX. Thus to get to the um level in beam-y or mm in beam-x one would need a lot more trials and is not be feasible. The only solution is again to keep the kinematics fixed between two iterations.
• No physics discussion today because Xiaoyan and Diancheng will work on their talks.

• 03/30/2009

• Ramesh: 
• worked on triggers;
• will work on: runplan for April/May running. A first draft from previous meetings is from 01/12/2009.
• will work on: online scripts to extract information for the scaler/TDC deadtime measurement and calculate the deadtime.
• Bob: 
• has the network switch but has not installed it yet;
• we need to do some helicity checks to make sure we are reading all information (scalers, helicities, etc.) at the right time. This should be happening when countings are disabled, but the helicity has not changed yet. Xiaoyan and Ramesh will work on this.
• Xiaoyan:
1. Checked scalers last week. Here are some plots from March 25, along with an email containing raw data. Did not see the same problem that we had before (that 3 adjacent events from scaler 7_26 always have the same readings.)
2. Checked scalers using the following inputs. Note: (a) scaler4 is the HAPPEX scaler; (b) scaler 6 and 7 are 3801 in the PVDIS crate with ECL input; (c) scaler 8 is a 3801 in the PVDIS crate with NIM input; (d) scaler9 is a 3820 that has the helicity and the channel information in higher bits and has not been removed yet, hence the large readings.
1. using a 100K clock, feeding to channel 0 of all scalers. Here all scalers (4, 6, 7, 8) agree well.
2. using a DAC->V2F with a ramping function, feeding to channel #1 of all scalers. Scaler8_1 shows very different readings. One reason could be that the signal is coming from a NIM output, while others are from another NIM output followed by a NIM/ECL converter. But the difference is big. Perhaps due to an un-terminated output?
3. using the helicity signal alone -> V2F, feeding to channel 2 of all scalers. Observed 3 level readings from all scalers (should be 2 if all work correctly). Bob suggested adding an overall offset to the helicity signal since the V2F may not respond to 0V.
4. Data from A-C are posted here.
5. (Note: After we solve the problems with scaler 8 and 9 and the helicities, will need to repeat this test for all scalers and all channels).
• Diancheng:
• For the Compton edet:
1. without any source, by measuring the background noise, found two problematic (too high response) strips, one in A and one in C plane.
2. used a radioative source to check the detectors. Found by not requiring a coincidence, there are single peaks in each of the 4 planes. And their positions move along with the source.  The maximum in plane A, B, C are roughly 1600, 80, and 30, respectively. 
1. Will calculate electron energy loss and see if the max vs. plane agree with the calculation. A disagreement indicate problem with single plane readouts (amplifiers, discriminators, etc.)
3. Although observed nice peaks in all 4 planes, none of the two form clean coincidences (no track!). Could have problem with the ETROC card (a black box), or single plane readouts (see above). This is a *BIG* problem and will require multiple steps of investigation (task A-1 above will be the first step). 
• For the HAMC simulation:
• Still working on the derivatives: Bob suggested starting from a large value (such as 100nm or larger in the beam position), large enough to see a change in the asymmetry (Delta_A_1) that is larger than the statistical uncertainty of the simulation [=1/sqrt(N_trials)], then double and then triple the position change, see if the change in the asymmetry is linear (i.e. Delta_A_3 = 3* Delta_A_1 and Delta A_2=2*Delta A_1, etc.).
• Beam position change might be the easiest to do, then can progress to Ebeam etc. (PREX use a look-up table for Ebeam so is not the best choice for this study).
• Physics discussions: Diancheng presented an overview/principles of the Atomic PV experiments.

• 03/23/2009

• (Note: the note-keeper lost the original paper so this is only a partial minutes.)
• Ramesh:
• Xiaoyan: 
• online scaler display?
• added two new scalers to PAN;
• checking the scalers. 
• Diancheng:
• Worked with Alexandre C. on the Compton electron detector. 
• For the HAMC, still working on the derivatives. Right now is running hamc twice with a difference in the Ebeam between the 2 iterations, but found the result in asymmetry difference to be limited by statistics. 
• Will work with Xiaochao to figure out how to "speed up" the simulation;
• Will find a way to get around the stat. problem.
• Bob: For "Trans-pvdis" to run we need a new network switch. The problem we had during d2n seems to be a dead network switch that cannot be reboot remotely, not a dead crate.
• Physics discussion: Xiaoyan presented an overview of the neutral- and charged lepton scattering.

• 03/16/2009

• Ramesh: discussed with Vince on what we have put in the R-HRS DAQ.
• The SC signals are about 80-100ns ahead of PSH and SH and is about 100ns after the gate. Although the timing gate is now 400ns (very wide), it cannot be shortened (or, at most by 80ns if really necessary). Will need to take pedestal runs to study how much noise it will cause and discuss with Vince and other spokespeople about whether this will be okay for the upcoming experiments;
• Will connect the 5th column of SH blocks so that the HRS DAQ have a copy of them.
• Diancheng:
• Worked with Alexandre on the Compton electron detector. Found that the "track" in the three planes are not always aligned. Can assume that two planes are aligned and see how much the hit position in the 3rd plane is off. Found that this "difference" are peaks centered at zero for all three planes. 
• This indicate that the reason of the non-alignment is probablly not physical shift of a particular plane.
• Perhaps also indicate that if we take the average hit position of all 3 planes, the average value will be very close to the real position where the electron hits;
• Can use the "difference" histogram to estimate the uncertainty on the hit position. 
• Will study how the hit position uncertainty relates to the final Pb result, and see where we stand on dPb.
• Xiaoyan:
• Working on online scaler display. Still don't understand how the SISproducer work with the server on the VME controller and the client code. Will ask Bryan about this;
• Need the "Trans-pvdis" configuration to run in order to test the scaler display code. This mode is now being disabled. (Bryan:  during d2n it was abandoned because there was no remote-reboot setup for the RightHRS HAPPEX crate.  It went down in the middle of the night, and I set up a standalone version.  A remote reset is possible... just need the proper modules.)

• 03/09/2009
• Physics discussion: Diancheng presented an overview of the EW Standard Model and current exp status of the weak mixing angle.

• 02/23/2009

• Xiaoyan:
• continue with integrating PVDIS into PAN; made some plots for comparing clocks and BCM readings from the HAPPEX and the PVDIS crate. Found some problems with the BCM comparison. By looking into the scaler data we found they always repeat the same numbers 3 times, need investigation;
• Diancheng:
• Integrated resonance and QE cross section into HAMC using Peter's code. The code also compute DIS x-sections and agree well with the previous DIS code (from Xiaochao);
• Will work on:
• Elastic (deuteron) x-section;
• Looking up models for Delta PV asymmetries (theory and data);
• Will start making asymmetry inputs.
• Huaibo: Trigger simulation results agree with expectation, will write a note on these results;
• Bob: Alexandre can use some help on the Compton electron detector readout, will let Xiaoyan and Diancheng talk to him.

• 02/09/2009

• Ramesh: 
• placed reqs for the electronic modules;
• Xiaoyan:
• continue with integrating PVDIS into PAN, finished about 64.2% (Bob's estimate);
• rate estimation for Ay and (e,e'd) agree with the rates from the proposal. But the right-HRS will have the same angle/p0 setting as the left HRS only for the Ay experiment, which means 1 day of running at the higher rate and many days at the lower rate. For (e,e'd), the right HRS setting has not been finalized yet (according to Todd, Vince is working on this);
• Diancheng:
• Rate from the hamc simulation look reasonable;
• Xiaochao:
• Suggest Huaibo explain all his work to Xiaoyan and Diancheng before leaving (early March);
• explained how to include resonances and elastics in HAMC. The main goal is radiative corrections.

• 02/02/2009

• Ramesh: 
• preparing req's for the left HRS. Will go down the Hall with Doug tomorrow to identify equipments we can take (borrow) from BigBite. We do not expect Hall A to take them back until the 2nd SRC experiments (after all 3 PV experiments and 1 year of DVCS);
• Will talk to Bogdan about how many cables will be available. Need to place order early if we don't have enough.
• Xiaoyan: 
• Worked with Bob on integrating PVDIS crate into PAN, finished about 1/3;
• Calculated e and pion rates for Ay and (e,e'd). First, reproduced the rates that Xiaochao calculated for the current kinematics. Results for Ay and (e,e'd) can be found here (to be posted).
• Will ask what will be the running time for the right HRS at different settings.
• Diancheng: Worked on HAMC. The acceptance cut-off seem to be better compare to last week. Xiaochao will explain more what we expect from the simulation and how to check results;
• Huaibo: Continued with trigger simulation for the deadtime.

• 01/27/2009

• Xiaoyan: 
• worked on rate estimation for the two pol3He experiments to run in March-May, so far the results look too high;
• Diancheng: working on HAMC. Seems there is acceptance matching problem because only a very small fraction of events make the Q1,2,3 cuts etc.
• Huaibo: Working on trigger simulation for the deadtime.

• 01/12/2009

• Progress since last meeting:
• Worked on the detectors during the X'mas shutdown. By now, most of detector channels (Shower, preshower, Cherenkov and Scintillators) are working well, except a couple of channels for preshower ADCs;
• Used a 10KHz pulser as input to check all scalers. Found that 3 channels measured slightly higher rate than others (12 or 14KHz);
• PVDIS crate has been integrated into HAPPEX DAQ. Switching between combined PVDIS+HAPPEX and individual HAPPEX or PVDIS mode can be down by changing the software configuration, no hardware change need to be made;
• All PVDIS scalers are added to the regular HRS DAQ as TDC tags (these are called "Tagging TDC" in the following). These TDCs can be used to check either the PID performance, counting efficiencies, and the timing of various PVDIS signals (see run plan below);
• Diancheng has been reading Xiaochao's Fortran code for calculation DIS cross sections, rates, and asymmetries etc;
• Xiaoyan has been reading about QFS, EPC code for calculating electron and pion singles rates.
• To do list:
• Check if the higher rate scaler channels is a scaler problem (bad chip) or double-pulsing -- Xiaoyan and Diancheng;
• Modify PAN to add PVDIS scaler readings, and compare results with the PVDIS stand-alone code -- Xiaoyan;
• Will use EPC, QFS, Whitlow and Wiser's code to calculate electron and pion singles rate for upcoming runs in March/April -- Xiaoyan;
• Will integrate the basic DIS cross section and asymmetries to HAMC by linking Fortran routines to the main C++ package -- Diancheng;
• Since we are using very wide gate now for triggering, need to estimate the effect of backgound noise and inform the spokespeople of upcoming pol3He experiments -- Ramesh and Bob;
• Make a runplan for current (January) running. Below is the first draft, which is the runplan from Dec. 15 plus many good ideas that were suggested at the meeting:
1. Set HV for all detectors (S1, S2, GC, PS, SH) and gain matching (Ramesh and Huaibo):
   
1. Will start from the values used by the latest running (done);
2. Run online detector check script to make sure all channels show the pedastal and some signals (done);
3. Make sure all PVDIS discriminators are set to the same threshold within each type (DL, DH and DM); Then adjust HV for gain matching; Here DL=pion TS; DM = electron PS; DH = electron TS;
   
2. Run regular HRS and the PVDIS/HAPPEX DAQ simutaneously:
1. Use TDC to check timing (Ramesh and Huaibo??). Since the preshower (PS), the total shower (TS) are both sent to TDCs, but not the shower alone, need to do the following:
1. find out if the scintillator(SC) is in the TDC as well. IF not, add it to the TDCs;
2. Compare TDC from SC and PS to check preshower timing, adjust delay if necessary;
3. Turn off preshower HV such that now TS is equivalent to Shower alone, compare TDC from TS and SC to check shower timing, adjust delay if necessary;
2. Check rates: Can set Prescaling factors to get the rates of each trigger (online monitoring of the rates is not working right now). Currently the following 9 signals are sent to the trigger supersivor (may help if you know where to find the PNG file for the full setup) - Ramesh, please update this...
• T1: Sum8 output of group #4 preshower;
• T2: Sum8 output of group #4 shower;
• T3: Sum of 1 and 2, i.e. TS sum of group #4;
• T4: Discriminator DL 100ns output for group #4. This signal, after veto-ed by (.not.GC)&SC, is used for pion counting;
• T5: And-ing module output for group #5, i.e. PS.and.SH. This is used for electron counting;
• T6: Or-ing module output for group #4. This is signal 4 (above) veto-ed by (.not.GC)&SC, i.e. pion signals;
• T7: GC (gas cherenkov) total output;
• T8: S1.and.S2;
• T9: final electron narrow (OR of all 8 groups).
3. Use PAN to get the scaler rates and asymmetries, and compare them with the PVDIS stand-alone code (Xiaoyan). This include all 6 signals for each of the 8 groups, the BCM, and the 100KHz clock rate;
3. "Detailed" analysis -- Write the analysis code such that we can cross check the two DAQs:
1. Use software "trigger simulation" to identify electron and pions from each of the 8 detector group using HRS DAQ data, then check the tagging TDC to see if they show up in the PVDIS scalers. This will tell us if we are counting particles correctly (Huaibo, with help from Bob on the previous trigger simulation code);
2. PID analysis (Xiaoyan, Diancheng, or Huaibo)
1. Use regular HRS DAQ data to analyze PID performance of Preshower/Shower and Cherenkov, respectively. Need to quantify electron efficiency and pion rejection factor for different preshower, total-shower and cherenkov cuts. Similar analysis on previous HRS data can be found from E99117 webpage (see notes dated  10/28/2000, 03/26/2001 and 04/20/2001);
2. Plot cherekov ADC spectrum, and shower vs preshower ADC spectrum (2D), for: a) all events; and b) events with PVDIS scaler/TDC tags. Ideally (i.e. our expectation is), the difference between a) and b) should be a clear cut on each of these ADC values. The discriminator threshold would correspond to the value of the cuts. Check if this is the case;
3. If cannot observe ii, change discriminator threshold to higher values, trouble-shooting everything until we reach this expectation;
4. After completing ii, (i.e. observe the cuts), change each of the discriminator threshold by a factor of two, repeat ii and see if the cuts shift accordingly;
5. Use results of i) and requirement from the proposal, determine where we should set the cut;
6. Use results of ii), iv) and v), determine where we should set the threshold for DL, DM and DH. Do so.

• 12/15/2008

• Huaibo finished writing the libraries for the new scaler (3820). Next he will move on to programming the CAEN fADCs. But the FADC work will be at a lower priority now that we could take some real parasitic runs to check our DAQ. See below:
  
• There will be an access tomorrow during which the techs will try to turn on the cryo for the right HRS magnets. Below is a brief runplan:
1. Set the momentum to 600MeV/c. The R-HRS is now sitting at 90 deg and Ebeam=5.9 GeV. Rate estimation for HRS-R: postscript, PDF;
2. Set HV for all detectors (S1, S2, GC, PS, SH) and gain matching (Ramesh):
   
1. Will start from the values used by the latest running;
2. Run online detector check script to make sure all channels show the pedastal and some signals;
3. Adjust HV for gain matching;
   
3. Run regular HRS DAQ to check rates (Ramesh);
1. Can set Prescaling factors to get the rates of each trigger (online monitoring of the rates is not working right now). Currently the following 9 signals are sent to the trigger supersivor (may help if you know where to find the PNG file for the full setup):
• T1: Sum8 output of group #4 preshower;
• T2: Sum8 output of group #4 shower;
• T3: Sum of 1 and 2, i.e. TS sum of group #4;
• T4: Discriminator DL 100ns output for group #4. This signal, after veto-ed by (.not.GC)&SC, is used for pion counting;
• T5: And-ing module output for group #5, i.e. PS.and.SH. This is used for electron counting;
• T6: Or-ing module output for group #4. This is signal 4 (above) veto-ed by (.not.GC)&SC, i.e. pion signals;
• T7: GC (gas cherenkov) total output;
• T8: S1.and.S2;
• T9: final electron narrow (OR of all 8 groups).
  
2. Note: the above trigger setup needs more work:
• T1-T3 are analog signals, need to pass them thru discriminators and use the logical output for triggers;
• T5 and the others (T1-4, T6) have to come from the same group to make the timing check more meaningful.
  
4. Run PVDIS DAQ to check scaler rates and asymmetries (Xiaoyan);
1. Need to complete the script to plot all 6 signals for each of the 8 groups, for both the rate vs. evnumber and the asymmetries;
2. Will add BCM vs. evnumber in order to observe correlation between beam current and the rates;
   
3. Will add the 100KHz clock rate vs. evnumber for sanity check.
   
5. "Detailed" analysis -- Write the analysis code such that we can cross check the two DAQs:
1. Use software "trigger simulation" to identify electron and pions using the regualr DAQ data (Huaibo);
2. Write fastbus TDC class so we can read the PVDIS triggers from the regular DAQ (Ramesh);
3. Cross check of the two DAQs: Do electrons and pions identified by one DAQ show up in the other?

• 12/03/2008

1. Ramesh and Bob reported: still not clear when we will have the right HRS magnets. Here is a brief summary of what's wrong with it. Note that we MUST have real date to test the DAQ by the Feburary shutdown!
   
2. Huaibo: start working on the new scaler library. For the FADC, need electron data to set various thresholds in the code;
3. Xiaoyan reported on her analysis. As a reminder, here is her task list.
• made asymmetry histograms (item I-5), click here. It's good that the scintillator is showing zero asymmetry. But there is a 3-peak pattern in all asymmetry plots which is worrisome.
  
• checked deadtime of scaler readout, all seem fine (item I-3);
• Plan:
• Will check the width of all asymmetry histograms;
  
• Will learn how to do pedestal subtraction (item I-5-a);
  
• will understand (and do) item I-4 -- linearity checks;
  
4. All of us need to think about what cause the 3-peak structure in all asymmetries (the two side peaks are at +/- 0.0015).

• 11/03/2008

1. Will enter the hall to correct the DAQ timing during tomorrow's access. May have 1-2 keys for us;
2. Xiaoyan: working on the analysis project, now checking helicity signals using the G0 code; Will work on a online checking script for Lumi status that shift workers can use;
3. Huaibo: Working on the FADC analysis code for identifying pileups; Will help Bob to write the library for the new scaler (SIS 3820)
   

• 10/27/2008

1. Scintillator paddle is seeing some rates (on scope), it has been connected to a patch panel in the counting house;
2. A list of electronic modules or items needed for the left HRS can be found here. Will try to find out how much of these can be obtained from the current transversity setup (assuming they are not locking up the electronics for the 12 GeV RCS run). Paul will check how many can be loaned from Fermilab or Argonne;
3. We will need at least 1 narrow-width rack for the left HRS. The ones used by the CSR experiment have been moved to Duke (originally from Brookhaven). Xiaochao will find out if it's possible to borrow them from Duke.
   

• 10/20/2008

1. Latest news from Ed indicate that the right HRS will not be functioning this fall. Some components are broken and there is no time to fix them;
2. A scintillator paddle has been setup near the beam line (about 10m from the target) and is now giving 700Hz. Will work on feeding this signal to the PVDIS DAQ (it needs to be split/fan-out);
3. Latest Hall status: Beam will start Oct. 22nd;
   
4. Will make a list of electronics needed for the left DAQ;

1. Worked with Joe on preparing Lumi for the transversity runs, checked HV;
2. Worked with Luis and Dustin on the Lumi PMT gains, ran test runs to check pedestals of each Lumi;
3. Will checked the helicity signal of the left PVDIS DAQ this afternoon.

1. Will discuss with Paul and Bob for what to do the next, what is the best way to insure that we will test the DAQ with real electrons in the spectrometer, before HAPPEX starts?
   

• 09/22/2008

1. Xiaoyan and Huaibo did an excellent job in cabling. Now cabling for shower and preshower is almost finished;
2. Problem: One of the HV crates on the Right HRS is not working. It provides HV for the gas cherenkov, the two scintillators S1, S2, and the preshower (total ~82 channels). Alexandre said it is unlikely that a new order will be placed until FY09 (October) and thus won't be here until Nov/Dec. There are some HV crates from CSR but we may not have enough space to use them in the hut. With the HV missing, there are two ways to proceed:
• Ramesh will check with Doug and Bogdan to see if there is any spare;
• Can use some of the Shower HV channels for GC, S1, S2 and Preshower, and form enough channel for 1-2 groups of our DAQ;
• Can use a substitute high rate signal from BigBite or Lumi as suggested by Bob long time ago. But there are forseen problems with splitting the Lumi. - Xiaochao will discuss with Bob. Then Ramesh or Xiaochao will contact Dustin if we do use the Lumi to see how realistic it is (and without affecting the transversity running).
  

• 09/08/2008

1. Ramesh:
   
• Trying to read and understand PAN;
• Still don't know when Jack will install the electronics, he is too tied up by the transversity experiment;
• Will check how long the cable should be from the detectors (PSh and Shower) to our patch panel, then from our patch panel back to the regular HRS patch panel (which then go to the regular ADCs). We will need about 260 of them including spares. Will try to find them in the old storage, if not, need to make our own;
• Will check if we can fit the "minimal setup" in the test cage on the 2nd floor of Counting House. Then even if we cannot have the electronics in the hall, or if we cannot have particles in the right HRS, we can still do some work there using the true helicity signal and some high rate signals from perhaps a detector block (scintillators) placed near the target.
2. Xiaoyan:
   
•  Trying to run and understand PAN, but still haven't talked to Louis yet about analyzing some HAPPEX data. Will do so this week.
  
• (Bob note):
  
1. We want to maintain the Lumi work for transversity, for which we need to understand how to fix small problems and how to optimize the run condition;
2. For the 3 PV experiments in 2009/10, Bryan Moffit will help us to set it up, but eventually he cannot help for a long time and we need to have our own expert;
3. We also want to use PAN for the cavity monitors.
   
3. Huaibo:
• Worked on the FADC, found a problem: Currently the FADC is running with only start signals, no stop. This might be why the data look corrupted and why we cannot read out the correct pulse shape;
• Working with Bob to understand how the trigger timing works, and how to make it work for PVDIS trigger.
  
4. Diancheng:
• Have been thinking about the analysis software, and is spending some time to learn the C++ philosophy.
  
5. For the analysis software, both Diancheng and Ramesh will work on it and try to come up with some toy code (or a structure of the toy code). Then we will discuss about both and try to combine them to form a hopefully better code.
6. Xiaochao, Paul: been trying to catch up with what's going on at JLab.
   

• 08/25/2008

1. Ramesh: We have finished the cabling thanks to the hard work of Xiaoyan and Huaibo during the weekend. Unfortunately Jack will not install the electronics this week. He will probablly do it by Sept. 15.
2. The Right HRS will probablly not be cooled until this X'mas. However there will still be plenty of tests to do once the electronics are installed.
   
3. We will find out Diancheng's progress on the analyzing software, and perhaps start working from there.

• 07/01/2008

1. Xiaochao tried to understand the measurement from last week. Here is her explanation for two cases of tagger loss and how they will show up in the real measurement: basic setup, loss case #1, loss case #2, (partial) conclusion. And the 140ns overall deadtime shown from last week's measurements might come from the wide width of the input PMT signal.
2. In the past week Ramesh used a quick method to check the deadtime: Using a TDC and measure the "deadzone". The results are shown in Fig.2 on page 3 of this file. (please ignore Fig.1, it had a wrong setting). The 20ns path output shows a 20ns dead zone and the 100ns path shows a 100ns dead zone, just as what we expected. So far, so good.
3. Ramesh showed the results he did today. See Fig.3 and 4 of this file. 
• The input is: pulser (as tagger), ORed with the "digitized" PMT ("digitized" = passing it through a discriminator with 15ns width) with a linear fan-in/out (428F). 
• The discriminators in the PVDIS trigger are set to 20ns and 100ns for the narrow and wide paths. Other downstream logic modules for the 20ns path are set to 15ns output width. The delayed tagger width is 15ns (input to the final AND, the other input is the trigger output). [last week, all logic units in the 20ns path were set to 20ns, and all units in the 100ns path were set to 100ns]; 
• the timing difference between signal and the delayed tagger is about 5ns (t1=5ns).
• We do not yet understand the pileup results (Fig.3 right), the 100ns path shows a significant lower loss this time, while the 20ns path looks the same as last week;
• Fig.3 left also does not make sense in Xiaochao's model: In her model the sig/tagger loss is due to the "over threshould" effect, where the tagger comes in when the PMT has not fallen back to below threshould yet. So if input signals are wide (>100ns), then the narrow path would show a higher loss than the wide path.  For last week's measurement it might have been true, because then the input (PMT pulse ORed with pulser) was fed to a RC circuit to mimic a PMT-like pulse as the PVDIS trigger input, which was quite wide (50-100ns). But the setup for today's measurement had the PMT passing thru a discriminator to narrow it first, so there should be no "over threshould" effect.
4. For the TDC "pileup factor" measurement (two peaks), here is Bob's explanation for what each fraction should be: I1=R*t1, and I2=R*(w-t1) (note: the plot is from Xiaochao who claims that I2 should actually be R*w, and w should be the width of the trigger output).
5. Plan: We still don't understand any of the old AND the new results, except the TDC deadzone (which is nice). Next, we will try to: 1) widen the delayed tagger to ensure coincidence of the two AND inputs; 2) increase t1 to ensure clear separation of the shoulder/2nd peak and the main peak; 3) If the results still do not make sense, then choose one input rate, and vary the trigger discriminator width between 20 and 100ns to see how this change the outcome; 4) can we reproduce at least the nice pileup factor measurement from last week (which showed a 10% loss in the 100ns path and a 2% loss in the 20ns path)?

• 06/23/2008

1. Plan for the leadglass detectors in HRS-L:
• (Bob and Eugene) Design for the new frame has the lowest priority on our designers' plate. For a long down time this summer, the design may still happen, but won't be anytime soon;
• Still, Eugene plans to setup sth. in EEL122 this week to test the new blocks from HERMES. Students may work on it. Once tested, we do not need to repeat it again, and can install them as soon as the new frame is ready;
2. Plan for the electronic installation in HRS-R:
We need to understand our deadtime and pileups before intalling anything.

3. Progress on the DAQ test:
• Ramesh's earlier work was reported at the Hall A collaboration meeting on June 13. Click here for his slides;
• Deadtime and Pileups:
• In the past week, Ramesh measured pileups using a TDC and the (old definition) deadtime for various counting rates, and used a new formula to calculate the deadtime. A diagram for his deadtime measurement is shown below:

• Diagram for the pileup measurement using a TDC is:

 

• Here are plots of his results: pile-ups, deadtime (old), deadtime (new); (Note that the red and blue markers in the first two plots are reversed, red=narrow in one and red=wide in the other).  Here the horizontal axis is the PMT signal rate (adjustable by varying the HV applied to the PMT), the (old) deadtime is defined as 1-R_signal/R_tagger, and the (new) deadtime is defined as 1-[R_signal*(1-pileup)/R_tagger]. However, we do not understand his results on the deadtime (both old and new). Some TDC spectrum can be found in this directory. Questions:
• Are results from the "pileup" measurement the deadtime we are looking for?
  
• Why does the (old) deadtime has higher values in the 20ns path than the 100ns path?
• Does the new deadtime definition make sense?
  
• Asymmetries: results reported in the Hall A meeting look promising, but we need to understand where the 3% jitter comes from.
  
4. FADCs (CAEN 1720): Bob would like to start working on these FADCs at some point. One of the students could work on this;
   
5. Trigger simulations (Bob): One of the students can work on this once he/she understands more about the system;
6. Physics simulation (Bob): working on a single-arm simulation software (C++ based) for all 3 parity experiments. The missing part so far is the Physics package for PVDIS (among other things). One of the student can work on this once... (same as above);
   
7. Analysis software (Bob): same as above;
   
8. Plan:
   
• Xiaoyan and Diancheng will work with Ramesh in 122 this week, will try to understand the basics of the DAQ and the deadtime measurement.
  
• Eugene will start working on the leadglass.