Main INDEX
Monthly INDEX
PREV
NEXT
Make New Entry,
Make Followup Entry
User name R. Michaels
Log entry time 22:10:42 on February19,2011
Entry number 347726
Followups:
keyword=helicity checks of hardware (II)
Upon access I checked the helicity-related signals. On infinite
persistence I could see some "extra" pulses on some signals (pattern
synch), so I decided to try another fanout. I moved all the signals
and improved the timing, but things don't change qualitatively.
The fraction of events which don't miss the pattern were measured
with higher stats with another new check code, "CheckAlot(n)".
The result was 7 bad events in 2e5. I could not solve this problem
and unfortunately must go home now.
The other problem (probably related) is that almost immediately
after loading the 30-bit seed, the helicity prediction disagrees
with measurement. It is really "as if" the helicity board is not
working, but Qweak says it is. So, I'm guessing the seed isn't
loaded right (bogus data due to the noise). Note, this part of
the VME check task worked from Sept - Dec 2010, so it's not simply
a bug. I don't know what is wrong. We need more offline analysis to
find the pattern of failure. Probably the data are all recoverable,
we certainly have a lot of redundant info (clocks, multiple inputs
to the DAQ, etc), but without good offline analysis we cannot say
for sure. The recovery will probably require more intelligence
in the algorithm exploiting the rigid patterns that are expected.