Subject: 2004 GRImII status
From: Mark Klaene
Submitted: Tue, 20 Jan 2004 16:24:51 -0700
Message number: 733
(previous: 732,
next: 734
up: Index)
GRIM II's single 256 x 256 NICMOS 3 chip is read out simultaneously with
four amplifiers, dividing the array into four quadrants, or "quads", which
can be considered semi-independently. For several years, the generally
reliable character of GRIM has been intermittently (and randomly)
punctuated with instances where one or more of the readout quadrants fails
to read out properly. The symptoms are either quads that return
inappropriate signal levels, or none at all, while adjacent quads function
normally. The unpredictable quality of this problem meant that the
instrument would function with all four quads operating reliably for weeks
or months, then in one night we could potentially lose an hour or more
trying to recover when a quad would drop. After some black magic, said
quad would return and GRIM would work reliably for the remainder of a
night.
In November 2003 we saw the frequency of the occurrence of this problem
spike, to the extent that quadrants were dropped in nearly every image
GRIM returned. At that point, GRIM was taken out of service while we
attempted to find a fix to the problem. Dale Sanford (UC), one of
GRIM's original architects, was brought in during December to
evaluate GRIM's condition and to assess the nature of the problem.
During his stay, the exterior (i.e., outside the dewar) electronics
were tested for obvious faults, and we characterized the performance
of what at the time were the still reliably functioning quads. Dale
concluded that the root of the problem lies within the dewar.
Although another annoying problem may be corrected with improved
ventilation. This ventilation modification will be completed within
the next week or so. Any further repair work would require opening
the dewar. An undertaking that will be very time consuming and with
risks as there are a number of other repair and maintenance items
that would have to be addressed once the dewar is opened.
Concurrently, the University of Colorado-built Near Infrared
Camera/Fabry-Perot Spectrometer (NIC-FPS) is scheduled for completion
and commissioning in mid-'04, with routine science operations to
begin around the first of the new year. The point of this message is
to address how ARC near-IR programs will be affected for the
remainder of 2004.
As of January 2004, GRIM exists in a semi-functional state. Our
characterization of the three quadrants that were "alive" in December
indicates that the presence of a bad quad doesn't appear to affect the
adjacent quads. We verified quantities such as gain, read noise,
linearity and dark current and found those values reasonably close to the
ones tabulated on the APO website
http://galileo.apo.nmsu.edu/~jcb/apo/grim/ One somewhat consistent
aspect of the problem is *which* quads tend to drop; i.e., the
leftmost two quads, as viewed in IRAF, seem consistently alive, while
the upper right quad is most flaky and the lower right quad somewhere
in between.
The main impact of this problem on science programs is, again, the
unpredictability of the quadrants. On a given night, all four may be
functional all night, one or two may drop either periodically or
permanently through a night, or some permutation of these possibilities
may exist. Non-scripted, pointed observations would not suffer
unreasonably from this effect, but considering that most GRIM observations
are scripted, the current state of GRIM has a definite and significant
impact on the majority of GRIM science. Observers considering
proposing GRIM observations later in the year should consider these
issues carefully.
Most importantly, all potential 2004 GRIM users should be prepared for the
circumstance that GRIM may be taken out of service either temporarily or
permanently, at any time and for any length of time, before NIC-FPS
becomes available. The engineers at APO hope to keep GRIM running long
enough to overlap with early NIC-FPS service and to characterize the
new instrument sufficiently to allow for correlation of its data with
GRIM data, determination of zeropoint shifts, etc. But this calendar
year will be very tentative with respect to near-IR data and the 3.5m
telescope. Users should plan accordingly. Also, any planned outages
for service and repair will be announced to the community as far in
advance as possible to keep the chaos to a minimum.
Observers faced with using GRIM in its current state should be on the
lookout for dropping quads during their observations, but there is no
evident cause to mistrust data the other functional quads are giving. If
a functional quad drops during the night, be sure to let your Observing
Specialist know. Any other unusual artifacts in the images or
otherwise odd behavior should also be noted immediately.
John Barentine/Mark Klaene/Ed Turner
--
Mark Klaene
Deputy Site Manager voice: 505 437 6822
PO Box 59 fax: 505 434 5555
Sunspot NM 88349-0059 web: www.apo.nmsu.edu
APO APO APO APO APO Apache Point Observatory 3.5m APO APO APO
APO
APO This is message 733 in the apo35-general archive. You can find
APO the archive on http://www.astro.princeton.edu/APO/apo35-general/INDEX.html
APO To join/leave the list, send mail to apo35-request@astro.princeton.edu
APO To post a message, mail it to apo35-general@astro.princeton.edu
APO
APO APO APO APO APO APO APO APO APO APO APO APO APO APO APO APO