Make New Entry, Make Followup Entry
User name A. Camsonne
Log entry time 12:47:01 on January 13, 2009
Entry number 255838
keyword=RICH work summaryRemoved the clear of the T/H signal so only clearing gassiplex with fast
clear, this will make a bit moe of dead tim but should be ok at the rate
we ran. This makes the large event size when using the pulser go away.
We put at output on the TI firing when the number of channel fired was
more than 2000.
So we could look at the large event on the gassiplex.
One possible explanation would be as in the drawing.
It seems the fast clear has some issues at high rates, even putting the
veto on the fast clear longer did not seem to help.
And same with addind a veto on T3 ( not in place right now I have to
check with Kalyan for the timing since putting it now in would also
affect the normalization ) which was an or :
- CRAM busy
- 680 uS gate after L1A
- 300 ns gate after T3 ( odd things TH stopped to work when going more
the 300 ns ... have to think about it... )
but it did not seem to improve much the large size event, so I left this
signal out for now at this time.
I also change the fast clear instead of L1A, I put the signal which was
supposed to veto the clear in place of the Not L1A for the fast clear
signal. That should prevent any clear withing 680 uS after the L1A
leaving time for the CRAM to read the gassiplex. It seems the signal is
improved a bit but we will check and optmize the thresholds and number of
channels read a bit later today.
So looking at the event it seems we still have some issues especially at
high rates.
I guess if the explanation is right a bunch of consecutive channels
should be present in those events. We will look at the data without event
cut.
I still have to figure out why those events still happen even with the
veto. I will think about it for next access.
Figure 1
