I've looked into the SM4 WFD problem Chris Walter described in his last
couple of emails. I took advantage of the LEDTDC calibrations in order
to check the arrival time of the WFD stop in SM4 upon CSPAM triggers.
While triggering the scope on the CSPAM input of the WFD STOP MASTER in
SM4, I looked at the STOP input to the WFDs in SM4. The STOP was always
arriving at 1000.1+-0.05 microsecs after the CSPAM. However, about
5% of the time, although CSPAM was firing (and the scope was triggering)
I was not able to see the STOP signal coming out of the WFD STOP master.
I don't think this has to do anything with the computer busy, as the
CSPAM input to the STOP MASTER is already busy-veto'd (if the signal is
taken out of the SPU) and definitely the dead time during calibrations
is much less than 5%.
I checked all the input signals to all the WFD cards in SM4, just to make
sure there is nothing obviously wrong. The clock signal is approx. 300mV
(p-p) and its period 5ns. (The clock in SM3 is 120mV p-p). The starts
have approx. 1.5usec width and 1V height while the stops have 12usec width
and a bit less than 0.8V height. I was actually surprised by the width
of the individual stop signals; the stop signal that the STOP MASTER
sends out is actually 1.5 usec wide which is somehow stretched by the
CLK/STRT/STOP fanout panel beneath the WFDs. However, this is not happening
in SM3.
I was wondering if the problem Chris is observing can be the result of having
the CSPAM actually plugged into the WFD STOP MASTERs. Since CSPAM is
neither setting the WFD_PATREG nor reading the WFD, I see no reason why
CSPAM should be plugged into these modules (and thus creating STOPs).
But if this fact is somehow related to Chris' observation, then I would
expect this to be present throughout the whole detector. Chris, have you
established the rate of this problematic events? Is it a 100% feature
of the SM4 only? I would suggest comparing ERP/CSPAM info for the events
that you described.
Starting with this run, I unplugged the CSPAM input from the SM4 WFD STOP
MASTER. This should be perfectly OK (or am I wrong?). Chris, could
you have a look at this data?
I think we should eventually unplug the CSPAM inputs from the rest of the
WFD STOP MASTERs *if WFD readout on CSPAM events is not going to be
performed*.
Let me know.
--Erik