Minutes of the CLAS CALCOM Analysis Meeting, July 18, 1997. =========================================================== Agenda: Status of CLAS magnetic field analysis - Hovanes Egiyan Improvements in the field calculations for the CLAS magnets - Mikhail Kossov A first look at the tagger data - Luc Murphy The laser calibration system for the electromagnetic calorimeter - K. Giovanetti (no report) =============================================================== Progress has been made in the analysis of the magnetic field data (H.E.), as well as in improving the magnetic field representation in the simulation and analysis software (M.K.). The results presented on the magnetic field mapping analysis are encouraging, and we can hope to have an adequate description of the CLAS field distribution on a time scale of a few months. We also have now a more appropriate representation of the magnetic field table with trapezoids increasing in size with the radial distance from the beam line, instead of the old rectangular grid of constant size. I urge the offline analysis and simulation groups to implement the new magnetic field table and compare it with the old table. Noticeable differences may occur near the field boundaries, e.g. close to the torus or mini torus coils. The next two meetings will be organized by Bernhard Mecking. If you want to make a presentation at one of the next meetings, please let Bernhard know . My suggestions for the next meetings are: July 25: calorimeter calibration analysis - min. ionizing calibration - energy calibration - ............ August 1: background analysis - origin of beam related background (zero field tracks) - GSIM simulation of background in high energy calorimeter trigger, and possible shielding approach - ......... August 8: drift chamber analysis - background levels in drift chambers - hit analysis - status of tracking code - ........ Of course, many other topics relevant to the CALCOM effort may be discussed as well. Volker Burkert ========================================================================== Individual reports: ================== Hovanes Egiyan -------------- Hovanes very briefly reminded what equipment has been used in the CLAS torus field mapping and described the simulation program used for those measurements. The data obtained from the field mapping were compared with the simulations in order to obtain the exact positions of the coils in the cryostats. Graphs were presented showing discrepancies at the level of 2 cm between the simulation data and actual data, indicating an error in some of the geometrical parameters used in the program. However, it appears that the measured data set can be reconciled with the simulated data by shifting the center of the polar angle (theta) rotation of the field mapper by 3 cm backwards (upstream) and 2 cm downwards. In order to obtain reliable values of these parameters a more detailled analysis of the survey results of the field mapper is necessary. Mikhail Kossov -------------- To find out imperfections of the magnetic field calculations, the magnetic field which is given by the SNAKE program and by the TORUS program were tested on the "symmetry planes": on the coil plane and on the middle plane. As a consequence of the symmetry there should not be radial components of the magnetic field (Br) on these planes. The parameter Br/B was ploted on both planes in colours. It was found that the situation is not perfect especially on the coil plane. 1. As a first step the "symmetry parameter" ( n*60 deg rotations) was corrected (made to be a double precision parameter). 2. To calculate the magnetic field in one point for the CLAS Torus it is necessary to add up contributions of about 120000 elements: (6 coils) * (4*54=216 loops) * (105 elements). As the precision of REAL*4 is 10^-7 one can not get better precision than 10^-2. So the contributions were summed up in double precision. 3. The CLAS Torus had 4 rows of loops symmetric in respect to the coil plane. So the coil plane values of the magnetic field were not corrupted by singularities of the too close approach to the infinitely thin conductor. But in the Mini Torus there wos only one row of loops exactly on the coil plane. It was changed to 4 loops symmetric to the coil plane in the same way as it was done for the Big Torus. After these corrections the magnetic field calculated by the TORUS subroutine became correct (it means pass the "symmetry plane" test) However the SNAKE extrapolation was still incorrect. 4. The grid was transformed from (X,Y,Z) grid, which by chance can cross conductors and loses a half of the data file space, to the (X,Y/X,Z) grid which uses 6(12)-fold symmetry of the field and is defined on all (X,Y) plane. The steps were improved from (7cm,6cm,10cm) to (5cm,0.036,5cm). 5. The extrapolation algorithm was improved from the 11 points algorithm to the 27 points algorithm (back not only from the central point but from all points, making 3x3x3 cube). After these corrections the magnetic field calculated by the new interpolation subroutine GET27 (instead of SNAKE) became correct 6. New (4 coils) geometry was implemented and the new PTG field was calculated. 7. At the Begin Run time the PTG field can be added to the CLAS Torus field. Luc Murphy ---------- Update on the Tagger analysis: During the May run, preliminary analysis seemed to indicate that the number of events read out by the left and right PMTs on the T-counters (the thick scintillation counters which provide the trigger and timing information) differed by a full order of magnitude. The problem was resolved when it was determined that the BOS bank containing the information on the T-counters had been incorrectly deceiphered. We are now reading out the raw data banks, and the data rates on the left and right T-counters are consistent with each other. During the June running period, the tagger group was assigned a few hours of beam. The electron beam energy was 4 GeV. The main purpose of the run was to verify the timing of the tagger counters with respect to CLAS. The TDC spectra from the T-counters show clearly that a coincidence was achieved. The peaks for the T-counter TDCs are broad, on the order of 50nsec, and they reflect the expected time differences between fast/light and heavy/slow particles hitting the TOFs. The peaks are located pretty much where they were expected on the spectra, around 40 nsec after the common start provided by the Trigger Supervisor. These plots, however, indicate that the signals of the tagger to the TDCs must be delayed a bit so as to avoid cutting off the information from slow moving particles. The CLAS part of the data from this run is still being analyzed. The data will be used to calculate the delays which must be added to the T-counters before the next CLAS+tagger runs.