CERN 99 Test Beam Results

During the September 1999 Test Beam Run at the GIF facility at CERN the data from CSC P2" were taken at different conditions including various rates of the GIF (Gamma Irradiation Facility) source. It is important to measure the rate of the photons detected by P2" at different absorption factors K of the GIF source filter. The measurements were done with use of P2" anode front end electronics. For this the data were taken with multihits LeCroy 3377 TDC which were set to 32 microseconds time range. Since the average energy

deposition from 662 keV photon from Cs137 source is several time larger than one from 1 MIP the cross-talks between anode channels can contribute to the rate measurements yielding the overestimated rate. The results of the offline analysis taking into account the anode cross-talks are presented here. It is shown that the LHC gamma rate of 20 kHz/wire group corresponds to an absorption factor of the GIF gamma source filter of 10. Also given are the beam rate profile and chamber+anode front end noise rate at GIF source off.

The observation of the anode cross_talk in the CERN 99 Test Beam data was reported on EMU meeting in November, 1999 at Fermilab (Talk). Here is the dump of one event from run 488 (GIF K=5, HV=3.6 kV) for chip #1 from board #1 and board #4 . Each time slice corresponds to 250 TDC counts (125 nsec). The element of the table is the hit arrival time given as (TDC count)/10 and it decreases with slice number. One can see an evident excess of the hit combinations when the hits are grouped into 2,3,4 and more adjacent channels in the time interval of 125 nsec. Chip #1 in board #4 has even 15 channels ON within one slice. For comparison here is the event dump for hits simulated for the GIF rate of 40 kHz per anode wire group. A simple simulation was used where the number of hits per wire group was Poisson distributed (with mean of 1.28) and each hit has a flat distribution in the time interval of 32 microseconds.

The overall picture for all 24 chips (run 488, 1000 events) is presented in Fig.1 (maximum number of hits per channel) and in Fig.2 (maximum number of hits per slice). Some chips in board 2 and especially in board 4 have distributions which are quite different from chips in boards 1 and 3. It also can be seen in Fig.3 where the maximum number of hits per channel vs chip number is presented.

To correct the number of observed hits for cross-talks the simple algorithm was used for each from 2 sets of 8 channels in the chip.

The first set includes channels 1 thru 8 and the second - 9 thru 16. The reason for such choice is in the layout - within the group of 8 channels the first 4 read signals out of one layer of CSC and the next 4 channels - from the next layer. Therefore the cross-talks between channels in one layer are seen as adjacent hits within group of 4 and cross-talks between layers - as hits in the both groups of 4 channels each. In the algorithm:

- skip the beam hits in 1 microsecond wide window, look at the hits in remaining 31 microseconds window;

- if there are 2 and more hits within one 125 nsec slice then count them as one hit;

- if there are hits in adjacent time slices then count one hit only with the earliest time if the difference in time is less than 125 nsec;

- skip the whole event (for given chip) if there are more than 14 hits per channel in 32 microseconds window or more than 6 hits in any 125 nsec slice.

It is obvious that such algorithm underestimates the number of hits. The corresponding correction can be taken from results of simulation and it is rather small for the rate of 40 kHz per wire group - about 6% and for the 200 kHz detected rate- about 15%. The results below were NOT corrected for such bias.

- The rates per anode wire group from raw data (online, no correction for the cross-talks) and after the data were corrected offline are given in Fig.4 for each layer of CSC (run 488, GIF K=5, HV=3.6 kV). The rates in wire group #23 in each CSC layer vs GIF attenuation factor K are in Fig.5 . The wire group #23 was chosen as having the max. occupancy of the beam hits in the muon beam spot. One can see that for wire group #23 at K=5 the detected GIF rate is 30-45 kHz. The LHC gamma rate of 20 kHz/wire group corresponds to an absorption factor of the GIF gamma source filter of 10. Online results (no correction for the cross-talks) are 1.5-2 times higher.

- Beam rate profile at GIF OFF, run 449, HV=3.6 kV (Fig.6)

which is about 0.2-0.5 kHz in the beam spot region.

- Chamber and anode electronics noise rate (run 450, GIF OFF, HV=3.6 kV, out of beam spill raw data, no correction for cross-talk, Fig.7) is less than 100 Hz per wire group (excluding some noisy channels).

- The rates above were measured with AD96 boards at the threshold of 15-25 fC. Here are the GIF rates measured for the AD16 (8+8) boards, wire group #23, threshold is about 20 fC and HV=3.6 kV (Fig.8). Note that at GIF K=1000 about 50% of the observed rate is due to the accidental beam rate (~0.5 kHz).