Thanks for all the input into the LED and laser setting debate. Here's
what I see as the concensus:
(1) We can run the LEDs at the same settings on all three uVAXes. I
want to back off a little from the maximum pulser output if we can, however, as
the present system leaves no safety margin. I think that the LED switchbox
modifications will let us do that (changing the backplane resistance should
give bigger signals from a given pulser output).
(2) Lasers... Alice and Alec tell us that the laser settings were
tuned on an individual basis in order to get things working. Like the LED
system, this leaves us with little safety margin if the laser signals
deteriorate, which we know they do. With this in mind, I'm going to try to
look at selecting start and end attenuator settings on a laser-by-laser basis,
then standardizing the algorithm that advances the attenuator in between, in
such a way that we produce more than the required number of points in each tank
so that when things deteriorate on down the line we can handle it.
Philosophical points for those who want to hear it:
We hope that the laser hardware refit will get light levels high
enough to give a considerable safety margin. The same goes for the LED
switchboxes, where the Chuck Lane modifications do the same thing. The point
is to pull the calibration hardware back from the edge, so that incremental
changes in the performance of the detector don't immediately affect the
quality of the calibrations.
We will ow attempt to do the same thing for the software... to get
things standardized in a format that provides more than enough data, so that
when pmt gains and laser light levels deteriorate in the future, we will still
be well within the quality operating range. That's not to say that
maintenance won't be necessary, but we want to reduce the call for it and to
make it as straighforward as possible.
I also want to say something more fundamental. We at MACRO have an
extremely conservative approach to changes in hardware and/or software. Much
of this stems from the fact that in the past the detector has been balanced
precariously on the edge, making any slight perturbation very dangerous.
The focus of the past year has been to move the detector away from that
state, to address the most precarious balance points and to make the detector
more robust. This is a natural process that takes place on all physics
experiments. In particular, I want to emphasize that when we decide to modify
things we are NOT saying that the work on (n-1)th revison was no good. In
fact, just the opposite is true: we only tinker with the stuff that was
made particularly well, while the stuff that was not made well gets ignored and
falls apart. None of us should expect to design the perfect system from
scratch the first time around, and even though we naturally get suspiscious
when people changes things that took a lot of effort to get running, we should
recognize that as time goes on new demands on the system and new information
about how it operates will require some changes.
Maybe I'm beating a dead horse here, but in the past I've sometimes
perceived (rightly or wrongly I can't say) some defensiveness when I've felt a
need to change someone's pet software or hardware. I just want to be very
clear that the work that's gone into the calibration system so far has been
excellent, and that the changes I'm making are really only nudges and
adjustments meant to make the earlier efforts more permanent and to keep the
system running as smoothly as possible.
- Nat