Collaboration Meeting
Physics Analysis Kickoff

1. Status of calibration projects

• Elton: summary of tagger timing calibration
• Grisha: plans for cpv/upv studies
• Dan: tests of LGD calibrations
• Scott: procedures for extracting pedestals and time offsets
• Craig: procedure for determining a good pixel
• Mihajlo: gain and attenuation corrections for BSD

2. Czars for detector subsystems calibration

subsystem	czar
cpv/upv	Grisha
bsd	Craig
bgv	Scott
lgd	Dan
tagger	Elton

3. Road map for calibration

• pass 0:
  Pass 0 will be restricted to finding the digitization offsets in all electronics channels. For adcs this is the pedestal, to be extracted by fitting historgrams of pedestal events. For tdcs this is the time offset t₀ which has to be subtracted from each tdc value to make its spectrum line up in time with those from other similar counters. At this moment we do not have a simple way to define t=0 for a single event, but at least we can align all tdc spectra within a given subsystem. The relative timing between different detectors will have to be adjusted later. To avoid complications from who set the trigger timing, it is best to find the t₀ from subtracted tdc spectra, where all values are subtracted from a single reference counter. As discussed by Elton above, we have decided to use cpv28 as the reference counter. The t₀ value for each tdc channel will be determined by a fit to the spectrum of (t_i - tcpv₂₈).
• pass 1:
  Pass 1 will determine the next level of calibration constants, whose extraction requires the results of pass 0 to be known already. For pulse height, these include gain constants and attenuation parameters. For timing, these include pulse-height walk corrections and the effective speed of light along the length of a counter.

4. Copying data from Jlab

Scott has allocated sufficient space on the I.U. High Performance Storage System (hpss) to hold the entire raw data set from Radphi. He is now in the process of copying the data from the Jlab mass storage system (mss) over the network up to Indiana. There exists 792GB of raw data on mss from the Radphi experiment, covering runs from Mar. 1998 to Aug. 2000. Of these, 771GB are found in files larger than 200MB in size. Right now Scott is only paying attention to files 200MB in size or larger. Major bottlenecks in the copy process are the latency in fetching the files from the silo at Jlab (can take several hours for a single jget) and congestion on the internet between Jlab and Indiana. Thanks to M. Ito, he now has a procedure that can fetch several files from mss in a single request. This has greatly relieved that bottleneck, although there is still the problem of sufficiency space on the work disk. Craig will check with Jlab computer staff to obtain the promised space in the mss cache area.

Under the best conditions Scott gets 400KB/s over the internet connection, using several ftp connections running at once. Right now he is averaging 800MB/hour. He expects the job to be completed within 6 weeks. Once the data is resident on hpss, we will have much faster access than is possible on mss, a few minutes to fetch a run as opposed to a few hours. The main remaining bottleneck will be copies over the network. So far Scott has managed to achieve about 500KB/s over the internet from I.U. to UConn. This connection is over Internet-2 and should be much faster. This problem will be investigated by S.T and R.J.

5. Computer resources

6. Documentation of analysis projects

7. Physics

The eħ collider at Frascatti has suffered from a luminosity deficit since the startup of the phi factory. As a result they are not able to make much progress on CP violation yet, and are devoting a lot of running to other physics, like radiative phi decays. They have already published in 2000 a result with roughly the same statistics as the Novirsibirsk experiments. This year they expect to multiply their statistics by a factor of 10. We need to get our result out soon, or it will be eclipsed. So we continue to put the a0/f0 rare decays as our number one goal. Beyond that, there are etaprimes, but our acceptance is miserable there. We can put upper limits on c-violating decays like gamma,omega and gamma,rho. We have already seen a strong a0(980) signal in 4gamma events, which indicates that we might be able to measure radiative decays of the the a0. The decay a0->omega,gamma should be complementary to the phi channel, but that physics remains to be worked out.

8. Conference calls

We will plan a series of biweekly meetings by telephone for bringing each other up to date on our work, and discussing analysis issues. R. Jones will organize the calls and distribute the instructions by Email. It was decided that Email communications regarding these calls and analysis issues in general would be through the general collaboration mailing list rad_phi@jlab.org. The idea of creating a more resticted list as considered, but was rejected in favor of promoting the broadest possible involvement in the analysis. It will also help to spread awareness that a vigorous analysis is underway results are coming in.

The conference call will be made using regular telephone service. An international line will be booked to allow Phil to join in. The time of the call will be Thursday 9:30am EST for the fall semester, and will be renegotiated as each semester, or whenever a conflict arises. Participants will join the call by dialing a predefined number and entering an access code that will be distributed beforehand. Each caller should be seated next to a computer with a browser open so that they can view materials on the web. Each speaker with things to show should prepare them on his own web site prior to the call, and be able to point everyone to them during the presentation. A message will be sent around prior to the call giving the highlights of what will be discussed and presented.

Depending on the success of the calls, a face-to-face meeting may not be necessary before the end of the year 2000. The question will be decided after the calls get underway.