User name R. Michaels
Log entry time 14:02:32 on April21,2001
Entry number 60967
keyword=Cerenkov required in T2,T4Per request of Joe Mitchell, I now require Gas Cerenkov in triggers T2, T4.
Here are the triggers:
T1......R-arm scint S1.and.S2 (as before)
T2......T1.and.Gas-Cerenkov ("or" of 10 PMTs, see below)
T3......L-arm scint S1.and.S2 (as before)
T4......T3.and.Gas-Cerenkov ("or" of 10 PMTs).
The 10 PMTs of Gas Cerenkov are an "or" (i.e. analog sum), and
this is put into the 11th channel of what the trigger manager software
calls "cherL_disc" and "cherR_disc" which are discriminators
in 2nd camac crate slot 2 of each spectrometer. The default disc level
is 30 mV, which should be fine. It can be adjusted with "expert mode"
of trigsetup, see
http://hallaweb.jlab.org/equipment/daq/trigger.html.
Notes about timing are put into page 235 of the e98108 paper log book.
One thing to add here: Because of the timing at the trigger supervisor
inputs, and because of the fact that every T2 is also a T1 (similarly T4,T3),
there can be a "event mislabeling" in events. This is NOT a big deal as
long as you understand it. Essentially T2's can appear as event 1, etc. But it can
be sorted out by looking at the trigger latch pattern, see ~adaq/evchk/README
and plot below which was created with the "evchk" code.
Explanation of plot:
X-axis is CODA event type seen in data file.
Y-axis is an event type determined by evchk from trigger latch TDC
in ROC2, slot 5, channels 0-11 are triggers 1-12 after prescaling.
What you naivly expect is trigger latch pattern = CODA event type. Point at
(1,3) is because I had pulser trigger on at start of run (then turned it off). There
weren't any T5's because ps5 was big (but see *). Points like (1,2) and (3,4) are
because some T1's are also T2's (similarly 3,4). Beware that T2 arrives at trigger
supervisor much later (340 nsec) than T1, so if ps1 = 1 and no DAQ deadtime, you
would never see T2. However T4 arrives a bit earlier than T3 (at least for cosmics),
so in this case T4's supersede T3.
(*) Ok, what is the point at (5,4) ? I'm not sure, but I guess those 30 events (out
of 10k) are from overlaps at trigger supervisor which are programmed to make event
type 5. It seems there were not any real T5's because ps5=infinite.
FIGURE 1
