DIRC Operation Under high background
1. The DIRC chain
Front end card
PMT tube : Gain drop vs rate ( see B. Wilson’s talk)
Analog chip : 50 ns dead time per channel -> 10% dead at 2 MHz
Digital chip : data selection. We loose hits when the time needed to process the good hit among the junk is larger than the trigger window. If we reduce the trigger window, we have less junk but less
time. Need full calculations, underway. At 200 kHz, the efficiency loss is 1% for 1 microsecond window.
DFB card: No further limitations
Conclusion: The operating limit is ~2 MHz per tube due to analog chip dead time and digital chip
Other points of the Hardware Operating Questionnaire
Current draw (see Bob Wilson)
Loss of lineary : No pileup at the entrance signal (20 ns total response time.
Increased dead time: No
Increased cross talk:No
Space charge effect : do not apply
Trips : No
Frond Electronics response
Increased crosstalk : No the thresholds are set in such a way that 8 simultaneous hits in the 8 channels
do not induce 1 extra hit. 15 simultaneous hits in the TDC can induce a 16th but negligeable probablity for 2 MHz random hits.
Loss of linearity Some pile up when measuring the charge :80 ns shaping time. Reminder : no charge read out for data taking.
Dead time : Yes :see above
Occupancy limit in FEB : No (max is 16x50=800 OK)
Read out errors : No
Overheating, overpower : confortable margin : 450 W drawn for 1000 W available. No significant occupancy/temperature sensivity
Other types : No
Spikes : No
All 168 DFbs have 1 channel discriminated output available. 24 cables available. For the moment, 16 channel scaler with 12 signals in (1 per sector) available.
What needs to be on : HV, FEE. No DAQ
High background effect: 50 ns dead time.of 10% at 2 MHz for the analog chip
Correction needed to take this in effect
Simulation : trivial
Correlation : Only local measure. Good correlation otherwise
Operation : no protection foreseen so no conflict