nice to hear from you. Here is some ideas may help to answer
your questions.
------------------------------------------------------------
How is the stop actually sent to the WFD? What kind of signal is Ed
looking for (voltage, current)? There is a problem that has surfaced
recently with Gary's latching scaler. It generates TTL logic levels
for output which are viewed by TTL-NIM converters to interface with
the WFDs or anything else. It seems that when one output is driven
high (+5 V) the adjacent outputs wiggle a little bit (cross talk). It
is not uncommon for this wiggling to be enough to cause the TTL-NIM
converter to go bang and we get unwanted waveform stops and other
problems.
The stop master is a CAMAC module. Its signal is TTL type.
I assume that the cross-talk is caused by the other trigger
because this time we are going to stop every channel of the
WFD. I am only only going to send one stop signal out and
then we need a fanout to distribute the signal. However,
on the other hane, I cannot do much about this TTL-NIM adapter
problem on my module. But I will be aware of this problem
when we test the module. Clearly, to guarentee the TTL low
is less than 400mV. By the way, is the phenomenon which you
found in that TTL-NIM adapter made in Texas A&M?
So you'll stop every tank on every trigger, right? If so this makes
cross-triggering irrelevant.
For all other triggers except the Caltech Slow Monopole trigger
are just fine because they are either face coincidence trigger
such as CSPAM or single face trigger like ERP. For caltech
monopole it will be fine if we do something in the readout.
The reason is the stop signal is 1ms after a trigger. Therefore,
all info is restored. If we only read one face for single
face trigger, we will find the info for each face in different
SM's. Clearly, we are not going to miss it just read less
channels. I think this is fine since this kind of events are
smaller.
Why 14? Will they be looking at the MSBs? Are you talking about
switches you flip with your fingers?
Yes, I am talking about using external switches which will
let us change the states by finger. The 14-bit is limited
by i/o pins. You are right we can use MSBs to control the
states via CAMAC. What I worry about is the power glitch will
change the states easily.
What about hits that come during computer busy? What about hits that
come after the WFD is stopped but before computer busy, or before the
timers expire?
First, the readout is controlled by computer. Therefore, these
triggers which happen after the busy up will be forgotten by
computer even if they send a signal to the master. On the other
side, all other triggers have taken care of the computer busy
in their circuits except slow monopole. Thanks for your reminding
me.
No distinguishing between supermodules, huh? I can't think of a
reason to do so. However, sometimes you may get more than one trigger
from the same face, separated in time. If so do you just ignore the
second trigger?
That's right, the module has taken care of the multi-trigger
signals from the same face because they are not monopoles.
One reason that I make this module be same for every SM's is
that we sacrifice a long delay to keep all infomation in
the waveform. We can readout them out through readout judgement.
Also think about the readout when multiple triggers happened. Who is
going to clear the waveforms? E.g. if ERP reads out the waveforms and
saves just the buckets it is interested in into equipment 29, then
clears the waveform there will be nothing for CIT to read into
equipment 18. Gary found it necessary to have another output register
which contained the NOT of the various triggers -- so if there was NOT
a CIT trigger, ERP readout clears the waveform. We may need something
even more complicated to handle all the possible permutations.
I have a register inside the waveform as a trigger register. I
think the CAMAC can look at the bits to decide if it need to be
clear.
Thanks for your questions. It certaily helps. I shall keep you
informed.
Have fun in turnnel!
--Rongzhi