Dear CBers, I am releasing a new set of Crystal Ball programs. As usual, they can be found as .tar.gz files at: http://www.public.asu.edu/~comfort/CB.html Don't forget the twiddle, and I think the address is case sensitive. The files will unpack into directories that have version numbers. For convenience, the directories can then be renamed without version numbers if you wish. Please review the README files for each program. The primary new features are 1) particle masses updated to RPP 2000; 2) incorporating the E927 experiment into the Monte Carlo code and the analyser; and 3) incorporating improved code (and parameters) and kinematic-fit routines from S. Prakhov into the analyser. The particle masses were changed only in the beamline, dec_kin, and evanalyser codes. It is much harder to do that in the cball code, due to the way GEANT builds in the masses. A proper treatment would require regenerating at least the geant321 library. Because the masses are used mainly in energy-loss calculations rather than for kinematics in our applications, the effects of using the "wrong" masses is minor. I had been maintaining parallel versions of cball and evanalyser for the E927 experiment, and it had become too much of a burden. Merging the treatments needed for the E927 experiment into the main codes was not too difficult. The routines are thus available for anyone, but further modifications and development can be expected. There were some structural changes in the cball code. The coding for the veto barrel was moved to a separate routine. A new ffcard parameter VBAR was added, with 'true' as the program default (for compatability with most previous usage). Also, the Spinka et al. additions for the VB tubes was included in the new routine, and all of their LH2 coding was included in the trgt_lh2.F routine. These latter changes can affect the results slightly, but represent better the physical situation. The kinematic-fit routines have been of general interest, but were not easily available. I have not done anything more at this time than to link them into the evanalyser and to make some adjustments so that they conform to the rest of the code. Documentation is very limited, but people may review them and try to adapt them to their own applications. At the same time, I added Serguei's modified treatments of beam tracking and cluster-finding algorithms, as well as the treatment of smearing for Monte Carlo data. These changes require a new gains file and a file of crystal resolutions, which are included. There are also new values of several cluster and MC parameters. Please see the comments in the ctp_cb.config file for details. It is important to understand that previous values are not good. In particular, all previous gains files are obsolete! Tony and I have run a number of tests, and I am happy to say that the results have many signs of considerable improvement over the past. Finally, I have restructured the CTP files a bit, most notably breaking up the ctp_cb.config file into several distinct files. I have done a pretty good job of documentating changes within each modified routine. Here is a brief discussion of changes that are relevant to users. Program Beamfile ---------------- The version number is 1.1 1) The table of particle masses was updated to RPP 2000 and documented. 2) A minor adjustment to the default iseed calculation for ranlux was made. Program deckin -------------- The version number is now 2.3 1) The table of particle masses was updated to RPP 2000 and documented. 2) A minor adjustment to the default iseed calculation for ranlux was made. Also, a new code 'modify_matrix' was added that can be used to modify the dec_kin_matrix.dat file. Program cball ------------- The version number is now 3.2 Changes to version 3.1 -- 1) The code that overwrote the branching ratios for Lambda and K0_S particles, introduced in v3.0, was removed. It led to confusions. People can insert their own code as needed. 2) A new routine trgt_ke3.F was added to handle the E927 (Ke3) setup. Suitable variables were added to materials.inc, with values defined in block_data.F and materials.F. Several processing routines were modified to incorporate the additional option. The code is subject to future modification for analysing data from the test run(s) and the future experiment. 3) The code has been set to maintain compatability in the ntuple files for reading into the standard analyser code. Thus, variables for the veto barrel, neutron counters, and some other things have been retained for the E927 option, but are not used. An additional block is written for the E927 option. 4) A minor adjustment to the default iseed calculation for ranlux was made. Changes to version 3.2 -- 1) Moved the setup of the Veto Barrel geometry to a new routine vbarrel.F, which is controlled by a new ffcard VBAR. 2) Moved the code for the Veto Barrel tubes (i.e., shields and flanges), previously in ugeom2.F, to the vbarrel.F routine. 3) Moved all code related to the LH2 target from the ugeom2.F routine to the trgt_lh2.F routine. The Analyser ------------ The version number is now 2.0. (A major upgrade.) 1) The particle masses in the file ctp_physics_param.config were updated to RPP 2000. 2) Features were added to handle the processing of data (real and MC) for the E927 (Ke3) experiment. A routine e927_process.F was added. At this time, it is set up for handling data from the 1998 test run and is subject to further development. 3) Serguei Prakhov's code for kinematic-fit analysis was added. The main code is in the KinFit directory, but some changes (mainly to define and set up some new variables) were also made to routines in the Main and Physics directories. 4) The handling of beam tracking in the drift chambers was modified, from code developed by S. Prakhov. 5) The algorithm for finding clusters was slightly modified, from code developed by S. Prakhov. 6) Code was incorporated to handle the "afterglow" of pedestals for Monte Carlo events, from code developed by S. Starostin. The feature is turned on if mc_en_smear < 0. If used, a file of pedestal data must be read in (called from mcarlo_analysis.F). A sample file is available. 7) The code for smearing the crystal energies for Monte Carlo data was slightly modified. Previously, we used dE/E = mc_en_smear / E^0.25 (with Gaussian randomization). Now, the power 0.25 is replaced by the ctp variable pw_en_smear, set in the ctp_cb.config file. A recommended value is 0.315. 8) The energy values ehv and ewv for the halo-veto and wave-veto counters, generated in Monte Carlo, were unpacked into the proper sections of the detectors. Ditto for the energies and times for the neutron counters. No changes in linkage from the MC code were required. No processing is done, but CTP variables (arrays) are available. 9) The structure of analysis directories and ctp files was modified. Now, it is easy to set up multiple analysis directories for different application, each with a separate set of ctp files in a subdirectory CTPconfig. Some examples are included in the distribution. 10) A variable for the CB threshold voltage was added, and the kinematic fit routines were adjusted for proper treatment. 11) Various cosmetic changes for improved documentation were made. It may be necessary to tweek some makefiles for the local installation. Not all system configurations can be anticipated. Let me know if you find any problems. Regards to all, Joe