Alec writes:
> Rich pointed out that a number of these hits were near the ends of tanks, where
> a small error in the position can make for a big error in energy.
>I agree with his suggestion for a tdcz vs adcz cut.
A tdcz vs adcz cut (requiring |ADCZ-TDCZ|< 150 cm) has been done by
the monitor since the upgrade in early September. (There are further
things that I want to understand about this issue however; I have done
some calculations of what the expected rate of ADCZ vs TDCZ mismatches
due to accidental coincidences of radioactivity is, based on rate
measurements and response curves. However, my answers come out lower
than the actual mismatch rates in many tanks, and in November I'm
going to make some measurements to try to resolve the discrepancies.)
>He also likes the ideas
> of:
> a) lop of the ends of the tanks (I feel this is a bit drastic)
> b) use adcz to get energy (I feel this is a bit circular)
> c) if an event is near the end of a tank, use the other end's energy
> values (this seems ok)
I also feel uncomfortable with (b), and I agree that (c) is the best
of these options. You lose out on pe statistics for low energies
though. I think there might be a better way -- I suspect that our
standard response curves tend to be artificially steep near the ends
of the tank, and that we could probably do a better job of mapping it
out in the near-end region (or at least, avoid some of the artificial
steepness). Needs some thought. But probably (c) is better than what
we're doing now -- I agree. I'll put something into the next monitor
upgrade.
Kate.