I have had some discussions with ADB and Fausto over the past couple
of days regarding ST/SC interactions and changes to the acquisition.
Here are some of the highlights, in stream-of-consciousness order.
The fast streamer tube memory is 10 us; the slow memory is 500 us
(actually it varies from tube to tube depending on readout order).
Currently scintillator triggers (ERP mu, CSPAM, FMT) stop the fast
chain except for CIT monopoles which stop the slow chain. According
to QTP, it appears the ERP trigger for this run was almost too late
and the ST information had almost rolled out of the fast chain before
the stop; there may even be some probability that we did lose
information (this affects Spurio-style stopping muon searches). A CIT
trigger will stop the slow chain about 500 us after the CIT trigger,
unless there is a ST monopole trigger in which case the CIT trigger is
ignored and the chain is stopped 10 us after the trigger believes the
monopole should exit the last ST plane (based on TOF seen between the
earlier planes). (No one could think of a better scenario for
handling CIT stops in the future, but I haven't yet thought very
hard.) There is (should be) a ST monopole trigger on every muon. It
triggers whenever it sees 7 planes hit with self-consistent TOFs.
Single hit time resolution is, I forget, either 1 or 2 us, but if they
get a nice 7,8,9 or 10- hit muon track they get a TOF distribution
centered near 0 with a sigma of 200 ns. The non-gaussian tails on this
distribution come and go as a function of time. So in answer to the
question of whether QTP can distinguish stopping muons (with the decay
product going into scintillator) from slow particles, the answer is
probably but it's hard to absolutely count on it.
Incidentally, they (Mario Sitta) are currently doing monte carlos to
try to determine how likely the (fast) decay products of proton decay
are to produce a ST monopole trigger and stop the slow chain when the
monopole is still in the detector. ADB points out that neutron decay
may be as important as proton decay, and it appears that the electron
channel may be more damaging than the muon channel -- he claims that
in an electron shower you can produce a low-energy gamma that can
travel a couple of meters before interacting; that would make a small
blip in scintillator but makes a big bang in ST.
When I first mentioned that we would want at least 1 ms between a
scintillator trigger and the beginning of event readout ADB was
reluctant. Apparently the current trigger pattern unit requires the
input to be active when a strobe comes; currently all triggers are
stretched in hardware (with capacitors, not gate generators) to
something over 500 us and the trigger unit is strobed 500 us after the
first trigger. Then the computer delays another 100 and something us
before asserting COMPUTER BUSY and beginning the readout. To wait 1
ms after the first trigger for additional triggers will require some
hardware changes. However, I think he has convinced himself that the
ST triggers want 1 ms also. However, since they have only 500 us of
memory they are toying with the idea of having separate stops for
attico and lower slow chains.
Also, he's not convinced we have to wait for the WFD stop before
beginning computer read-out. For example, I mentioned a worry that we
might get hits in 1T at t=0, 1C at t=400 us and 2B at t=800 us. The
event would start reading out at t=1000 us but SM2 would not stop its
waveforms until t=1800 us. However, he said the ST readout is
guaranteed to take a few ms, and we can probably arrange the CAMAC
list in such a way that the ST would always read out before the WFD.
This is fine assuming that everything works all the time, but makes
operation with a scintillator-only CAMAC list questionable.
I explained that it will take a long time to read out every box of WF,
so the ST monopole trigger input to the waveform stop system will be
low-rate. He came up with three possibilities for handling that --
o tune the parameters so that the rate of ST monopole triggers
(vetoed by Bari trigger) is low.
o prescale a higher rate ST monopole trigger so only a fraction stop
the waveforms.
o have two versions, with different parameters. The higher rate
trigger would stop the ST and the lower rate would stop the
waveforms. This way, there would be some acceptance with SC info
and some acceptance without.
There was a bit of discussion about whether single face CIT monopoles
should stop the streamer tubes. He likes the idea of having a few
muon-induced CIT stops which could allow us to convince ourselves that
we do have ability to make ST and SC info work together on real
monopoles. But probably we could get enough of those if we only
stopped ST on two-face events.
I may add more if I remember any more later.
Bob