Main INDEX
Monthly INDEX
PREV
NEXT
User name R. Michaels
Log entry time 14:13:39 on February 26, 2006
Entry number 164034
Followups:
keyword=CAEN 767 problem investigation
I investigated the problem with the CAEN 767 decoding.
There are several problems, though luckily it did not corrupt
other crates. I wrote a detailed diagnosis in the analysis
logbook : HERE
Briefly, one problem is that there are many "bad chip" errors
but this flag is ignored -- hence analysis workers are ignorant
of how many errors except that they don't see data. Second
problem is that because of the bad chip flags the decoder gets
"lost" in this crate and ends up checking the high rate scaler
clock, and since this sometimes has higher-order bits it thinks
it might be a TDC, but the data are inconsistent so it rejects
it (indeed it rejects the whole event !!). Third problem is
the decoding is very inefficient; something to keep in mind
for the future of VME is that we need clear header info for
each slot, denoting the slot# and number of hits.
I've explained techy details in the analysis logbook, but since we
are presumably going to replace 767 with Fastbus anyway (because
of the huge bad-chip rate) it is academic. I think the decoding
for Fastbus will be faster, and we know it is quite reliable.
A copy of this log entry has been emailed to: bogdanw@jlab.org,brads@jlab.org