Thanks for all the comments.
Here's a summary of the fake-box-shuffling options under consideration:
All of these will cause some software to break.
--------------------------------------------------------------
A.
Bari box + LED in one fake box, 2 laser different fake boxes
Software changes required: need to insert code to distinguish
between Bari and LED events based on ADC value, everywhere
anything looks at that fake box.
PROS: -- probably easiest calibration software changes. (Calibration
DST and maybe other code will still break though.)
CONS: -- trickier (but perfectly doable) to set up the box and its LUT;
you need to make sure LED signal fires Ehigh,
and the Bari signal fires Elow and never Ehigh
-- leaves us open for Bari/LED calib confusion if things go wrong
(probably OK due to redundancy of calib bits in the SPU).
---------------------------------------------------------------
B.
Bari gets its own fake box, LED + laser in one box, another separate laser
box.
Software changes required: all software must learn to ignore Bari box,
calib software must use different algorithm
to distinguish laser from led runs.
PROS: -- very easy to set up
-- no Bari/calib ambiguity
-- can tell which laser fires when it fires during run
-- leaves possibility open for future combined upper/lower
laser runs
CONS: -- LED/laser ambiguity
-----------------------------------------------------------------
C.
Bari gets its own fake box, LED gets its own fake box, attico laser
+ lower laser in one laser fake box
Software changes required: pretty much same as above; here the
the two lasers must be distinguished from each other
PROS: -- also easy to set up
-- again no Bari/calib ambiguity
-- may be less confusing to have lasers together
CONS: -- laser ambiguity possible
-- trickier to set up combined laser calib running in the future
------------------------------------------------------------------
Well, none of these options leaps out at me as being the One True
Way. Alec likes A, I guess because it's slightly easier on the calib
software. Nat has expressed a slight preference for plan C
because it seems more sensible to have the lasers together, but Bob and
Alec have given a reasonable arguments for why plan B, the
laser/LED fake box, is preferable to plan C.
I am inclined towards plan B because:
1. Although plan A isn't so hard, plan B is totally trivial
for me to set up (as is plan C)
(and my remaining days here are short, and getting
shorter the longer the detector remains in its current
Hunk Of Useless Junk state),
2. It seems to me better to have any ambiguity
about what triggered a fake box always be confined to calibration runs
(for which, in principle, there should never be any ambiguity anyway),
and
3. Given that we use a dedicated Bari box, plan B is better than
plan C for the reasons given by Bob and Alec.
I think I will proceed with plan B unless someone convinces me
otherwise (which won't be so hard to do if you have a good reason).
I'm going to try to do this on around Monday or Tuesday (assuming
that the mysterious acq problems we're currently having get solved),
so that there will be a few days of running before I leave Friday.
BTW, I've been working hard to alert everyone about this change
ahead of time, so that I don't get pounced on when the change
happens and things break. I posted to Vax news, and
I also spoke with Grillo, and on his advice I
wrote personally to Spurio (responsible for ERP DST code) and
Ronga (for ERP upmu code). I've also been mentioning it to everyone
in sight. I hope this is enough. Everyone is understandably
a little distracted by the Big Dead Detector problem though. Can anyone think
of anyone else I should notify explicitly?
Kate.