User name R. Michaels
Log entry time 20:11:17 on October 4,1998
Entry number 2340
keyword=cause of DAQ slow-down
I've been called a few times because the event rate dropped,
and/or the deadtime was high. On these occassions the event
size was also high. Normally it is about 0.8-0.9 kbytes per event,
and if it becomes substantially bigger, the DAQ will obviously
slow down. Look at the event size = differential kbytes/sec divided
by rate. This is almost certainly NOT a CODA problem, but
a detector noise problem. My guess is the VDCs are oscillating
again. The experiment should isolate and analyze these big events
to determine the cause. I wonder how well they reconstruct.
Also re: the CODA crashes. You might try asking MCC to shut
off beam before pressing End Run. If G. Heyes is right, this will
make the probability to crash at start-of-next-run negligible.