Minutes of the Jefferson Lab CLAS Collaboration October 14,15,16, 1999 Next meeting: Thursday - Saturday, January 20,21,22, 2000 NOTE THAT THESE DATES ARE A WEEK LATER THAN STATED AT THE COLLABORATION MEETING. PLEASE CORRECT YOUR CALENDARS. The reason is that we cannot get rooms in CEBAF Center on our first choice of days! The dates above were the second most popular times. -note that the abstract deadline for the April APS meeting is January 14th. Thus, if you want to submit an abstract to that meeting, your Working Group should act on it *prior* to the next collaboration meeting. For you to do: 1) If you are an experiment spokesperson, report any limited members on your experiment to Gail Dodge, Chair of the Membership Committee. (We are trying to assemble a complete list of members in this category.) 2) Check the two tables distributed by the Chairman (i) "CLAS Data Calibration and Cooking" and (ii) "Status of CLAS Experiments". Report corrections and updates to the Chairman. See below for discussion. The tables were handed out at the meeting. 3) If you are interested, check out the final "as approved" policy on CLAS Approved Analyses which is included below. ------------------------------------------------------------------- The meeting was held on a series of cool dry fall days, at a time when the g1(c) data-taking period was in progress. The CLAS spectrometer has been running well, in general, and the majority of the meeting was devoted to data reduction and physics analysis issues. The first day consisted of a "Physics Fest" of results. Some of these presentations were practice talks for the Asilomar DNP meeting next week. Others were analysis updates intended to bring the the Collaboration up to date on projects that are at some fairly advanced level of development, though not yet necessarily ready for conference presentation or publication. R. Schumacher distributed a table of the status of all known analysis efforts underway using CLAS data. There are over 60 projects, and over 30 PhD students working on CLAS at this time! The Chairman intends to keep this table up to date and distribute it periodically, giving the Collaboration an evolving overview of what work is in progress, who is going it, and how far along each analysis has progressed. He also distributed a related list of the names associated with the calibration and cooking of all the run periods CLAS has had. B. Mecking reported on the state of CLAS hardware and the likely run plan for the next year. The Time-of-Flight NIM paper has appeared in print; the reference is NIM 432 (1999) 265. The Drift Chamber paper has been accepted, with a demand for some revisions. These revisions will be made shortly. In Y2K several months will be devoted to the RadPhi experiment, the non-CLAS experiment in Hall B. CLAS will run parasitically with RadPhi only if there is PAC approval for a suitable measurement that could be done, and if the RadPhi group agrees. [We later discussed that there may be an argument to run the tail end of the g6 beam time (about 4 days under normal conditions) parasitically to RadPhi, if the Collaboration agrees to this at the January meeting. Carlos Salgado will make the case for this at the January meeting.] The goniometer for producing coherent linearly polarized bremsstrahlung from a crystal radiator has been installed in the beamline, 22 upstream of the tagger. Control instrumentation is not complete. Beam time to commission this piece of equipment will come from "facility development" time in 2000. Other technical upgrades are in progress: a beam profile monitor for photon beams using scintillating fibers is under development by the "Detector Meisters". An additional pair spectrometer for the Primex experiment is being designed; it may also be suitable to inject e+e- pairs of known properties into CLAS for the purpose of mapping out B-field and chamber-alignment effects. It was pointed out that the maximum beam energy of CEBAF may be less that the full 6.0 GeV called for by some proposals, at least for the near term. Guesstimates ranged around 5.8 GeV, with the exact figure depending on the tolerable rate of RF trips. The present run plan for 1999 is: January down 52 days e1 2/3 - 4/2 30 days e5 4/7 - 5/14 60 days RadPhi 5/19 - 7/31 - pol. target install 110 days eg1 9/8 - ... The DNP practice talks will not be reviewed in detail here. A few general observations follow. In more than one case it was clear that there was insufficient preparation at the working-group level to withstand the level of questioning and criticism by the whole collaboration. The most common single critique was concerning methods for extrapolating measured yields into total cross sections. It was generally felt that pure phase-space or s-wave-only event generators are not good enough, in general, for extracting total cross sections. [Eventually, the PWG's approved a number of results for presentation as long as strong caveats were attached to the plots shown at the conference.] Some plots were shown with axes that were unreadable... the usual litany of transparency sins. One talk, on kaon photoproduction, was withdrawn because of recent analysis difficulties. One talk, on eta photoproduction, was presented via live video hookup from Arizona State; this hookup was a technical success, allowing 2-way discussion with the ASU members. There were several Analysis Update talks. D. Tedeschi reported on phi photoproduction from the g1 data set, while G. Audit reported on phi and rho photoproduction at high t, mainly from the g6 data set. Each of these efforts are well advanced in their study of acceptances and systematic errors. A series of presentations then followed discussing partial wave or amplitude analyses of resonance data. C. Smith reported on single pi_0 electroproduction in a model that assumes that the M1 multipole dominates the reaction, which leads to a truncation of the set of multipoles which are extracted. A W-independent fit to CLAS data then led to extraction of six products of amplitudes which can be compared to theory. Much careful work had been done to study the systematics of the results due to effects of radiative corrections and acceptance. V. Burkert discussed a different approach to fitting both pi_0 and pi_+ data, in which Breit-Wigner resonances are built into the W dependence of the partial wave fits. Born-term background and other non-resonant terms were added to improve the fits. Much effort went into tuning the model to "perfect" pseudodata before applying it to real CLAS data. The results to date were used to help study CLAS' acceptance. Also, results for E1+/M1+ from this method agreed with those of C. Smith. D. Weygand described a method of "binless" likelihood fitting, in which each event enters directly into a likelihood sum with adjustable weights for each partial wave amplitude allowed in the model. The acceptance in the spectrometer for each partial wave combination is built into the method. It is a partial wave expansion, not a resonance expansion. Initially this method is being tested on omega photoproduction data from the g1 data set. There was a series of talks reporting on the status of calibration and cooking of run groups g6(b), g2, e2, e1(c), and eg1. An impressive amount of data has been accumulated, especially now that CLAS routinely acquires events at the rate of ~2500/sec. g6 is in the early phases of calibration. g2 has 2.3 billion triggers to process. e2 has finished calibrations and will very soon begin cooking. e1c is cooking 23 million events per day and as 130 days to go using 75 dual processor Linux nodes. eg1 is also cooking, and will finish by 10/31/99; they hope to have publishable results by the middle of Y2K. Progress was demonstrated by all four of our Task Forces in the areas of (i) alignment and B-field, (ii) tracking and efficiency, (iii) acceptances, and (iv) photon normalization. It was also clear that in none of these areas were problems "solved" at the level that we can relax our efforts. Some related factoids: M. Mestayer believes all chambers are aligned to within 0.5mm, at least within sectors. S. Lassiter and W. Brooks have studied the Oxford and JLab drawings for the torus coils, and may have found a 12 mm discrepancy in the radial position of the coils, which, if truly reflecting a mistake in where the analysis software thinks the coils are, would be about right to explain the phi and theta distortions of the reconstructed momenta. The goal of the alignment task force remains: delta_p/p < 0.2% and delta_Theta < 2 mrad. J. Meuller defined the goals of the tracking task force to be: to get the efficiency measured in the data and in Monte Carlo simulation to agree "to the needed level", thus making work on tracking in some sense a subset of the acceptance work. Using Monte Carlo, various track-loss prone areas of the detector were identified. At worst, low theta, low momentum and high phi time-based tracking lost about 40% of all tracks. Overall, however, time-based tracking inefficiency was in the range of 0.6% in MC, and about 7% in real data. The sources of the 7% were discussed; one important source was due to "bad" scintillators in the TOF array. Hit-based failures are under .21%. Jim also reported in his manufacture of new and expanded "roads" for the e1 and e2 data sets. The long-target data sets (g1, g2, g6) do not yet have upgraded road files. Reports on the detector acceptance were given by G. Niculescu and W. Brooks. One important message was: more people and effort are dearly wanted in this area. Acceptance studies are in progress using MC data generated with a range of smearings in the drift chambers and the scintillators. As diagnostics, several diagnostics are being examined: charged and neutral pion mass widths, elastic peak fits, hyperon mass widths. The nominal smearing estimates in the reconstruction packages seem to be close to optimal, within 50% or so. It was shown that we may want to add random hits to the MC tracks which are generated. The CLAS Acceptance Document in taking shape. The "total acceptance" was defined as the "correlated acceptance with and accurate input model." Also, the distinction between "simple" and "full" acceptance was discussed; the simple acceptance in each kinematic bin is the ratio of event thrown dividing the number of the same events reconstructed in that bin (thus always being a ratio less than unity), whereas the full acceptance includes the effects of bin migration into and out of bins. Some sample calculations where shown which computed the bin migration matrix for a typical reaction. One conclusion was that one should not, ideally, put fiducial cuts on particles heading for the cryostat coils, since some of them scatter into the acceptance, giving tails on various distributions. Work on comparing the photon normalization as computed at Saclay and as computed at JLab was presented. Progress on understanding the photon normalization was made, in fact the methods are very similar, and differ mainly in how the tagger efficiency data are initially obtained. We still do not have a detailed evaluation of the normalization uncertainty as derived from the study of a known reaction such as pion photoproduction. V. Gyurjyan reported that the DAQ system continues to become more reliable, with a typical speed of 2,500 event/sec. A separate branch leading to analysis nodes raises the prospect of eventually being able to reconstruct the events as fast as they come in. The limiting element in DAQ speed are the 1872/1875 TDC's, which have a 10 microsec + 2.5 microsec/hit conversion time. There are no plans to upgrade these TDC's. The Level 2 trigger and readout scheme were discussed by V. Gyurjyan and D. Doughty. The DAQ part is OK, but the actual segment linking hardware has a few remaining bugs which render it still too unreliable to use routinely. The hardware "never" loses good tracks (unless more than 2 superlayers fail), but lets too many bad tracks leak through as various superlayer elements latch up. The goal of the group is to run Level 2 for one week without a sector failure. E. Smith reported on the Quality Constant group. Most detector elements now have well-defined calibration procedures with CLAS-note documentation. Also, there is a new CLAS-note by J. Manak et al, 1999-016, on "e1, g1, and g6 Data Processing Procedures". [This looks like a great introduction to the whole process!] It was announced that Mark Ito will be the new coordinator of the QuaCons group. One of the remaining tasks is to ensure that the first-pass calibration procedures are consistent across all run groups. M. Ito reported on the off-line computing system. We presently have a main farm of machines incorporating 75 dual processor Linux nodes, with 720 gigabytes of work space and 650 gigabytes of cache. The cache space is for raw data files that otherwise reside in the silo. Access to data on the silo seems to be the main bottleneck presently. The tape drives are being upgraded to handle 350 mount/hour, up from 195. About 50% of the CPU cycles on this system are reserved for CLAS cooking. Note that IBM/AIX support is being withdrawn by the computer center, or rather, these machines will no longer be actively maintained and upgraded. Mark is planning to propose some off-line software coding standards to help streamline some of the overall software effort. K. Dhuga and associates outlined progress on installing the coherent bremsstrahlung system. The goniometer is in place, 22m upstream of the tagger, but not fully instrumented yet. More instrumentation work and installation of the 384 E-counter scalers are scheduled for the January down period. For polarimetry, G. Feldman discussed a technique using Moller electron formation in the field of an atomic electron, wherein the recoiling atomic electron carries the information about the incoming photon polarization. Tests of this technique are in progress at LEGS. Commissioning time for this whole setup will proceed in Y2K during suitable machine development periods. Among the reports from the Physics Working Groups there were some highlights such as: B. Niczyporuk showed (as reported by Latifa) some good-looking first results for forward (low-t) pseudoscalar electroproduction above the resonance region. For e1(c), the EC timing including the TOF timing is now 250 ps (!!). This is good, especially for investigations looking for neutrons. The existence of BRAG (Baryon Resonance Analysis Group) was advertised; this group includes CLAS members but also theorists from outside the Collaboration who are willing to help the Collaboration in coming to grips with handling the resonance data. There is also a workshop planned, NSTAR2000, for next year, which will be another venue for discussing baryon resonance physics. The next PAC meeting is likely to be during the last week of January. Any proposals CLAS submits would have to go through the Physics Working Groups only (without plenary review), by mid December. No group within CLAS expressed a completely firm plan to send something to the PAC. Three groups are thinking about it. On the topic of how CLAS processes drafts of publication, the Chairman pointed out that the CLAS Charter is (intentionally) vague on details. Since we are within a few months of starting the publication process, it seemed that we need a more detailed policy. A policy review committee was defined which should report a draft of such a publication policy at the next meeting. The members are: R. Minehart, B. Berman, B. Raue, and R. Schumacher (Chair). The questions which can be addressed include: mechanisms for internal review, feedback to authors, global criteria for results quality, and author list questions. Please give inputs to these persons about this process if you are interested. There are expectations that CEBAF will be operating at 12 GeV by about 2006. What physics can CLAS do at 12 GeV? It behooves the CLAS User community to consider this question, since work to establish the necessary upgrades must be started well ahead of that time. It would be timely build a focus group to consider the physics issues. At the next meeting of the Collaboration we will have a small-group session devoted to CLAS physics at 12 GeV. Paul Stoler has graciously agreed to stimulate some discussion and thought in the community prior to that meeting. If you have any ideas brewing, he would be most happy to hear from you. A discussion of the new policy on CLAS Approved Analyses then followed. A draft which had been developed by the Coordinating Committee was presented. Several changes of wording were proposed, and some were accepted and some were rejected. After about 40 minutes of debate the following draft was adopted as a new by-law to the Collaboration by a vote of 33 to 5. -------------------------------------------------------------------- "CLAS Approved Analysis" Procedure for CLAS Members to Undertake New Analyses of Existing CLAS Data. Draft: 10-18-99 Given that there are reaction channels or observables present in the CLAS data sets which are not covered by any already-approved CLAS experiments, this policy regulates how existing full members of the CLAS Collaboration may gain approval of the whole Collaboration to undertake the relevant analysis. Since all full members of the CLAS Collaboration have free and equal access to all CLAS data (Charter Section IX), this policy regulates the issue of priority of working on an analysis with a specific stated physics goal. This policy covers work that will NOT be separately presented to a Jefferson Lab PAC. 1) Each CAA must be approved by the relevant Physics Working Group (PWG) and by the full CLAS collaboration. The PWG will maintain a list of active CAAs. 2) Proposers must submit a written proposal to the relevant PWG. It must contain - list of spokespersons with 1 contact person. - specific data set and reaction channel to be analyzed. - discussion of the potential physics results to be gained from the analysis. - discussion of the analysis methods planned with sample results from existing data where possible. - discussion of the relationship to any existing CAA or PAC- approved experiment. The PWG coordinator should solicit comments from the spokespersons of any overlapping CAA or approved experiment. 3) The PWG coordinator will follow the standard methods for evaluation of PAC proposals of that group. There will be an evaluation of the CAA proposal by the PWG followed by a vote. The vote must occur at a publicly announced meeting with this CAA listed on the agenda. The PWG can vote to deny or approve the proposal or request resubmission with specified changes. 4) After the CAA is approved by the PWG, it will be presented to the full CLAS collaboration at a scheduled collaboration meeting. The collaboration can make it an official CAA with a 2/3 majority of the voting members present. The vote shall be by secret ballot. 5) The PWG will list all approved CAA's in a publicly accessible place. 6) The Chairperson of the Coordinating Committee will communicate the roster of approved CLAS analyses to the Chairperson of the PAC to make the PAC cognizant of physics analyses in progress within CLAS which have not been separately presented to the PAC. RIGHTS and RESPONSIBILITIES: 1) Thesis students can be assigned to the analysis, with notice to the Coordinating Committee, just as for any other CLAS PAC-approved experiment. 2) The proposers (spokespersons) are covered by the same "talks" policy for presentations at workshops and conferences as apply to any PAC-approved CLAS experiment. 3) The CAA will normally be approved after the first time a run group takes beam. The CAA spokespersons will have no right to specify run conditions (e.g. beam energy, target...) in any future run period. 4) Progress in a CAA must be reported to the relevant PWG's on a regular basis. --------------------------------------------------------------------------- K. Hicks reported for the Service Work Committee. He defined the goals of this Committee as: (1) To encourage all institutions to take on a fair share of the service work, and (2) To identify service work tasks that need more manpower. He reported that the Committee presently feels that all except 2 institutions are carrying a reasonable service load; the exceptions are being handled through the Membership Committee. There ARE tasks, however, which need additional manpower, and the examples given were: - doing the shift scheduling - help with beam line and slow controls - help in the on-line group - GSIM running, especially organizing the "large" runs - help with calibration and cooking, especially for small run groups G. Dodge reported for the Membership Committee. There were no new members to vote on at this meeting; there will be next time. There was a discussion of whether non-members, typically young graduate students, should be allowed to take worker shifts. [We have a bylaw, "A1. Data-Taking Shift Policy" from October 1996, which actually does not spell this out.] The recent de-facto policy has been that only Full and Term members may take shifts. Not everyone has adhered to this rule, however. It was decided that a non-committee of present and former COPs (Experimental Coordinators), COP_COM, look into this and recommend a policy. Included are, at least, Arne, Elton, Will, and Mac. The sentiment in the room was [I believe], to support R. Miskimen's suggestion to ALLOW non-members who are affiliated with institutions, e.g. new graduate students, to take worker shifts as long as they have passed the minimal training: ODH, Rad Worker I, and general safety. It was felt that these extra shift workers should not count towards the member total used to determine to total number of shifts assigned to each institution. This would provide an extra headache for Arne's [already truly excellent] shift scheduling software. R. Miskimen reported for the Speakers Committee that about 4 talks have been approved since the last Collaboration meeting. He requested that postscript versions of talks be made available to the Committee for linking to their web pages. P. Stoler reported the task force on Working Groups. The recommendation of this group was NOT to change the Physics Working Group structure at the present time, when we are trying to get first publications out. He made a personal suggestion for a future structure that would include both physics and technical working groups. This suggestion was discussed briefly, but clearly the overall issue is complicated, and will require more careful thought. Any changes, additions or other clarifications of these minutes by the Collaboration members will be cheerfully considered. Amended minutes will be sent out only in case of egregious errors or omissions. [...]=chairman's editorials