February 11, 2015 Calibration
GlueX Calibration Meeting
Wednesday, February 11, 2015
11:00 am, EDT
JLab: CEBAF Center, F326
Connection Using Bluejeans
- To join via Polycom room system go to the IP Address: 22.214.171.124 (bjn.vc) and enter the meeting ID: 630804895.
- To join via a Web Browser, go to the page  https://bluejeans.com/630804895.
- To join via phone, use one of the following numbers and the Conference ID: 630804895
- US or Canada: +1 408 740 7256 or
- US or Canada: +1 888 240 2560
- Upon connection all microphones are automatically muted. To unmute your mike on a Polycom or equivalent unit, enter *4
- More information on connecting to bluejeans is available.
- Software Review, February 10-11, 2015
- Collaboration Meeting February 19-21 at Jefferson Lab
- Timing Offset Calibrations (Mike S.) - slides
- Calibration status/updates
- Data Monitoring
Attending: Sean (NU); Simon, Paul, Mike S., Manuel, Nathan, Will M., Kei, Noemi, Eric, Mark D., Alex B., Eugene (JLab); Matt S., John Z. (IU); Justin (MIT); Naomi (CMU); Mahmoud (FIU); Regina (Tegan)
Timing Offset Calibrations
Mike gave a brief update on his automated timing offset procedure. We looked at his current results for time resolutions between the start counter and various detectors. The observed resolution is a combination of the start counter resolution and that of a particular detector. Right now, Mike is testing his procedure with different runs, and fixing the bugs that have been exposed. He's found some differences in how the fADC reports results and error conditions between mode 7 and mode 8, and has accounted for this. He's also worked with Simon to make time-based tracks available in runs with zero magnetic field, where they were not available before. The quality and detector matching of these tracks are not as good as when the magnetic field is on. He will also compare these results with those obtained from simulations.
Matt gave an overview of the current FCAL efforts:
- The current primary goal is gain calibration. The time offsets depend on the high voltage setting, so that needs to be fixed first. Recent studies have shown that the current timing resolution (~3-5 ns) is good enough for this point in the experiment.
- The primary gain calibration procedure is based on minimizing the width of the pi0 peak. This was used for the RadPhi experiment, and tested using GlueX MC.
- Matt then presented a series of plots that showed that the first iteration of this algorithm improved the peak, but after many iterations and converging on a final result, it appeared that background events were being pulled into the peak.
- This effect is still being studied, but it could be ameliorated either with more data or with a more pure sample.
- The amount of data needed is being quantified.
- Another possibility is to use the LED monitoring system under the assumption that the response should be smooth across the FCAL.
- Two different smoothing functions were tested, and their results were found to be correlated.
- Using the average gain factor of the two, a smoother FCAL response is seen.
- The effectiveness of this method (i.e. does it reduce the pi0 mass resolution) is being tested. It may at least give a better starting point for other calibration methods.
- A final option is to use the fact that photons have an E/p ~ 1. Matt is working on this, but is having problems finding electrons, either due to analysis bugs or skim requirements.
Work is progressing in BCAL gain calibrations using pi0's. A similar algorithm as that used by the FCAL is being used, with the added complication that each BCAL cell is read out on 2 sides (upstream/downstream), instead of the 1-sided FCAL readout.
Simon used a few long cosmic runs taken in the fall to determine the first wire alignment constants and add them to the CCDB. The shifts in the location of the wire ends that he obtains are ~100 um, though they can be as large as ~500um.
Simon also uploaded a new time-to-drift distance table for 1200 A magnetic field, with a 56/44 Ar/CO2 mix. A similar table already existed for magnet-off data. The new table results in a position resolution of 150 um for long drift distances. (In both cases, near the wire the position resolution is ~250 um, consistent with results from the prototype).
Mike is working on the efficiency calculation.
Lubomir continues to work on the package-by-package alignment.
Simon needs to make some new tables to hold the FCAL gains, since the current tables don't take into account the half-sized strips. The hit factories will need to be updated as well.
Sean noted that it was communicated to him that progress is being made in timing calibration, and paddle alignment is currently being looked at.
After the meeting: It was noted that the tdc_shift calibration (which removes the 6-fold uncertainty) for later runs was not always set correctly. Kei and Sean are working on updating this constant and new values should be in later today
Constants from the test stand for attenuation length and propogation speed have been put in the CCDB. Current efforts are in determining these and other time calibrations. The reconstruction code will also need to be updated to use these.
Nathan gave an update on his tagger studies.
- He uses PS-trigged data to cross-calibrate the PS and Tagger, to understand the correlations between their energy measurements, and to investigate tagging efficiencies.
- He looks for a e+/e- pair coincident in the PS and a photon in the tagger, subject to various quality cuts.
- He only uses ADC times, as he is still working on understanding the ADC/TDC matching.
- He sees a clear correlation between tagger energy and PS energy, though not a linear one (note that for an electron beam energy of 10.1 GeV the tagger microscope covers the energy range 6.7-7.7 GeV). The shape of this correlation and the apparent tails in the PS energy need to be understood. Note that the PS energy reconstruction depends on the combination of the e+ and e-, so that might need improvement.
- Using events with a loose tag in the PS, he gets a tagging efficiency as a function of photon energy of ~85% in the hodoscope, and ~<70% in the microscope. The drop in efficiency in the microscope is understood to be related to fibers with lower light yield. The drop in efficiency in the hodoscope at high photon energies is not yet understood. It could be due to backgrounds in either detector, or other issues with the reconstruction.
- He is currently working on using this data for timing calibrations.
No new news - resolving hardware issues has been the priority.
Paul has implemented a new constant for the RF frequency, since we'll be running at a lower frequency this spring (~250 MHz versus ~500 MHz).
Sean has implemented a scheme to automatically load the correct magnetic field for different runs using the CCDB. This scheme works with the fall data; procedures for setting the correct constants for new runs are still being discussed.
Sean reviewed the Calibration Dashboard and noted that he is collecting information on detector resolutions and efficiencies.
Sean and Simon are still running checks on MC they have generated. Calibrations should be checked and an issue with missing CDC hits should be resolved before more simulations are run.
It was agreed that the situation with offline monitoring is stable enough that status reports can be left to the offline meeting. Instead, once data taking resumes, it would be useful to review some of the output of the monitoring. To do so, as calibration procedures are developed, they can also provide useful output (histograms, etc.) to the monitoring as well.
Generally, detector groups should review what they are contributing to the offline monitoring to see if they want any changes made.