Date: Thu, 19 Aug 2004 10:49:06 -0700 (PDT)
From: Constance Rockosi
Subject: more on spec r2 trouble on MJD 53236
There are a few more details about the debugging Steph, Pete and
I did on r2 last night.
First and perhaps most importantly, we learned some useful and
reassuring information.  The terribly compromised science frame
that caused Steph and Pete to call me shows very clearly that the
problem is digital; the histogram of the overscan on the right half
of the CCD shows multiple peaks, neatly separated by 64 DN, 512 DN,
etc.  This narrows the problem down to only about a half-zillion
places, most likely somewhere in the RS484 wiring that carries the data.
All these are much easier to work on than something inside the dewar.
The first test we did was to power-cycle the spectrographs (at the same
alt, az and rot position as they got the bad data) to see if that would
cure it and thus complicate any real debugging efforts.  The problem
persisted through the power-cycling, providing further evidence that it is
at least partly mechanical.
I didn't have any particular evidence that the DB15 piece of
the saddlebag was more likely to be the problem than any other of the
RS484 wiring, except that the saddlebag was moved around and fiddled
with more this past summer than the control chassis.  It seems to have
worked, but we'll see.
If the problem returns, my current thinking is to either replace the DB15
ribbon cable or look inside the control chassis where the RS484 lines from
the saddlebag are routed to the FOXI.  I'll ask Wendell to do the latter
anyway.
There is a PR forthcoming on this that will reference this email.
cheers,
connie