User name Adaq
Log entry time 23:27:03 on January25,2003
Entry number 93358
This entry is a followup to: 93350
Followups:
- 93359 01/25/03 23:40 choi Wrong user name in halog entry
- 93488 01/26/03 19:57 choi Synchronization problem between ROC3 and ROC10?
It seems that the problem mentioned in the big helicity 0 peak is related to "Synchronization Problem" seen from espace replay. For runs 20548 and 20549, after "Synchronization Problem" happened, the trigger type from ROC3 and another trigger type from ROC10/ROC11 (G0 helicity) are out of sync and, I would guess, helicities are not assigned correctly giving large number of 0 helicities. For some runs like 20550, there is "synchronization problem" message during espace replay, but two trigger types agree quite well. The attached plot shows the situation. Each histogram shows trigger type from ROC3 versus trigger type from ROC10 (G0 helicity). The first one is from Run 20549 and shows inconsistency between the two trigger types. They became out of sync after ~100000 events. The second one shows "ideal" case from Run 20550. Both trigger types agree each other. It should be noted that when replaying run 20550, there is "synchronization problem" after 9000 events, but apparently it has been fixed. Since Bob is absent, after talking to Bodo, we switched to unbuffered mode for left arm DAQ and will check if the problem is still there. Until definitive answer from Bob Michaels, we need to pay attention to these problems. To help diagnosis, please use "espace_debug" instead of "espace". I have modified physics kumac to include diagnosis histogram. In ideal case, everything should be on the diagonal. Alesandre Deur just mentioned that when he was trying to switch to unbuffered mode, he has to kill coda and start from the scratch. Maybe there has been some problem with the coda. We will see how it goes.
FIGURE 1
FIGURE 2
