NEXT
Make New Entry,
Make Followup Entry
User name Kai
Log entry time 02:20:43 on November 02, 2009
Entry number 297136
keyword=add a new 757 module to fanout the bcm u1,u3,d1,d3
As discussed in the meeting today, we found some cross talk phenomenon in
the original Phillips 726 bcm fanout module right beside the V2F in the
middle room, that is even if there is no input, there is an output 6khz
100~150mv jitter. Thus, I installed another 757 on the left side of V2F
to fanout the original V2F-bcm output as three copies. One goes to the
previous bcm input channel on the Phillips 726 module.In this way, we
will not affect other signals connecting on that Phillips 726. Alex need
to check the bcm in COMPTON data to make sure it doesn't affect. I labled
all useful cables, and you can easy to find them. The other two copies go
to the left and right arm patch panel separately. and I removed the
original left and right arm bcm fanout on the Phillips 726 that go to the
patch panel. I check the final input to the patch panel and those jitter
disappears.
I used the 350k test pulse to do some checkout runs both before and after
the setup change. I called a event whose left arm of raw scaler bcm
counting does not equal to that of right arm as a bad event, and use the
(number of bad events/total event number in one run) as the error rate.
Result is as follows:
bcm: before->after
u1: 1.94%->1.75%
d1: 5.85%->1.52%
u3: 7.81%->6.29%
d3: 13.3%->5.6%
The magnitude of the scaler counting in the test is roughly 11.7K, and
the real bcm scaler counting in the production run is between 7K~13K.
However, both errors before and after change are just like one count
difference error. I haven't found any larger difference as what I found
in the happex production run using the real bcm physics signal. So we
still cannot say it's fixed until we have the beam.
A copy of this log entry has been emailed to: camsonne, xiaochao, rom, rsubedi, dwang, deng, pkpan