From furget@isnhp6.in2p3.fr Thu Oct 28 10:45:27 1999 Date: Tue, 19 Oct 1999 16:52:06 +0200 From: Christophe Furget To: beise@physics.umd.edu, "Brian Quinn (bquinn@cmu.edu)" , beck@uinpla.npl.uiuc.edu, pate@lepton.nmsu.edu, kox@in2p3.fr, grave@ipno.in2p3.fr, Foglio@isn.in2p3.fr Subject: Re: French data format Dear Betsy and Brian, With some delay I give my proposal for the expected format for the 2 french electronics. christophe ************** The proposal is to have the detectors connecting in the french electronic with the following order : Octant 2 Det 1 Octant 6 Det 1 Octant 2 Det 2 Octant 6 Det 2 . . Octant 2 Det 16 Octant 6 Det 16 Octant 4 Det 1 Octant 8 Det 1 Octant 4 Det 2 Octant 8 Det 2 . . Octant 4 Det 16 Octant 8 Det 16 the detector and its buddy enter on 2 adjacent channels. *** Grenoble electronic : *** For 4 octants, there are 16 boards. One board contains 4 detectors. For the first board (0.5 kbytes) : General information on the board : 32 bits Time spectrum Octant 2 Det 1 : 24 scalers * 32 bits Individual scalers Octant 2 Det 1 : 8 scalers (25 to 32) * 32 bits Time spectrum Octant 6 Det 1 : 24 scalers * 32 bits Individual scalers Octant 6 Det 1 : 8 scalers * 32 bits Time spectrum Octant 2 Det 2 : 24 scalers * 32 bits Individual scalers Octant 2 Det 2 : 8 scalers * 32 bits Time spectrum Octant 6 Det 2 : 24 scalers * 32 bits Individual scalers Octant 6 Det 2 : 8 scalers * 32 bits ... same for the next boards ... * The scalers 25 to 32 are used for individual counting : scaler 25 : Backward Left PMT scaler 26 : Backward Right PMT scaler 27 : Forward Left PMT scaler 28 : Forward Right PMT scaler 28 : Backward meantimer scaler 28 : Forward meantimer scaler 28 : Buddy same pulse scaler 28 : Buddy previous pulse * As mentionned in the NA electronic, the scaler 1 correspond to first bit set by the clock train and correspond to the latest particule. The scaler 24 correspond to the earliest particule. The reconstructed time spectrum (with the time going in the right way !) can be obtained with : bin 1 : scaler(24) bin 2 : val_scaler(23)-val_scaler(24) bin 3 : val_scaler(22)-val_scaler(23) . . bin 24 : val_scaler(1)-val_scaler(2) All bins correspond have a width of 1 ns except the first one which correspond to about 7 ns (?) the flow rate is then 250 kbytes/sec (at 30Hz) *** Orsay electronic : *** For 4 octants, there are 8 boards One board contains 8 detectors (back+front) For the first board (4 kbytes): General information on the board (board number / slot number) : 32 bits General info on DSP 1 ( DSP number, status MPS, MPS number) : 32 bits Time spectrum Octant 2 Det 1 (128 channels * 16 bits) Buddy spectrum Octant 2 Det 1 (128 * 16 bits) Time spectrum Octant 6 Det 1 (128 * 16 bits) Buddy spectrum Octant 6 Det 1 (128 * 16 bits) General info on DSP 2 ( DSP number, status MPS, MPS number) : 32 bits Time spectrum Octant 2 Det 2 (128 * 16 bits) Buddy spectrum Octant 2 Det 2 (128 * 16 bits) Time spectrum Octant 6 Det 2 (128 * 16 bits) Buddy spectrum Octant 6 Det 2 (128 * 16 bits) General info on DSP 3 ( DSP number, status MPS, MPS number) : 32 bits Time spectrum Octant 2 Det 3 (128 * 16 bits) Buddy spectrum Octant 2 Det 3 (128 * 16 bits) Time spectrum Octant 6 Det 3 (128 * 16 bits) Buddy spectrum Octant 6 Det 3 (128 * 16 bits) General info on DSP 3 ( DSP number, status MPS, MPS number) : 32 bits Time spectrum Octant 2 Det 4 (128 * 16 bits) Buddy spectrum Octant 2 Det 4 (128 * 16 bits) Time spectrum Octant 6 Det 4 (128 * 16 bits) Buddy spectrum Octant 6 Det 4 (128 * 16 bits) ... same for the next boards ... the flow rate is then 1 Mbytes/sec (at 30Hz) To reduce by a factor 2 the flow rate, it is possible to compress the buddy spectrum from 128 channels to 1 channel (full integration). Other solution can be proposed ! The acquisition will certaintly add information on the header of the event ...