Main INDEX
Monthly INDEX
PREV
NEXT
User name R. Michaels
Log entry time 12:19:44 on February26,2001
Entry number 57004
Followups:
keyword=fixed a few scaler problems
During the access for dipole polarity, I fixed the following problems with scalers:
1. The 10Hz pulser counted in the 22nd channel of the normalization
scaler was double counting. The signal was observed to have a ringing tail.
By using an isolated output from an upstream source instead of using an
NIM/ECL to fanout the ringing went away, see 10Hz now.
2. The halog reports "deadtime from gated scalers". The deadtime is computed
from a ratio of the following two: a) signal and b) (signal.AND.not-busy)
where "not-busy" is the inverted busy signal from CODA trigger supervisor, and
signal can be anything. We use two signals for this: charge-Q10 and the
1024 Hz clock. The AND with not-busy goes into resp. chan 17 and 18 of
the normalization scalers. There were two problems, which I fixed:
2a) The Q10 used to AND with not-busy was U10. But the halog end-run
script runend.tcl assumed it was D10. Since there is an offset in frequency,
this lead to errors in end-run entries which were very noticable for cosmics
runs where it computed a negative deadtime. Fix: changed runend.tcl to use U10.
No big deal since the U10 results look good.
2b) The deadtime for clock.and.not-busy was hopelessly wrong because the
coincidence module was set to make an output with only requiring input #1.
Fix: set coinc module correctly (requires both inputs).
I think problem 2a existed since cabling changes in Jan (and is not really
a problem anyway) but 2b probably existed forever.