ANDICAM Rescue

January 17-22, 2005

Ramon and Bruce drove to CT on Tuesday 1/17/05, arriving about 11AM.
Andicam was dry but coldish (ccd temp at -40C).  Ramon hooked up his
leak detector and confirmed that the Dewar still leaked in the same
area.  We filled Andicam and found the problem with the bad quad, which
is:

While we were waiting for the Andicam-Aladdin arrays to be obtained the
Head Electronics Box (HE) and cabling for Andicam were built.  Since the
Aladdin array has 32 outputs we used the 62 position connector for the
HE to Dewar cable.  When we gave up waiting for the Aladdin arrays and
began installing the Hawaii arrays we fabricated the Hawaii Clock Bias
Board (CBB) on a full triple wide CBB (since that is what was designed
into the system ).  This meant that the outputs of the Hawaii array were
connected to the CBB on the mid connector of the CBB and on P3, the
connector reserved for 32 input channels.  Since the HE was wired for 32
channels the outputs from the Hawaii array were also connected, via
traces on the CBB printed circuit, to the cables for the 32 output
configuration.  One of these unused but connected cables developed a
short.  I cut the traces on the CBB that were connecting the outputs to
the unused cables and repaired the faulty coaxial line.  All four
quadrants now work.

We warmed the system.  Both the 2-171 Oring for the IR plate and the
2-033 Oring for the reverse mounted connector plate were replaced.
Unfortunately the leak remained. Careful leak checking revealed that the
leak was actually in the weld between the vacuum jacket outer shell and
the mounting ring, directly above the IR port.  The leak position was
misidentified because while the IR port was being tested the helium
wafted upwards, as it will do, and the weld leak was mistaken for a
leaky Oring.

The warmup controller problem was caused by erroneous parameters in the
CN8500 (among other things it was set for a J type thermocouple instead
of an RTD).  Reprogramming it according to the documentation solved the
problem

On Wednesday we removed the CCD and IR detectors and the vacuum jacket.
An oily or greasy spot was found on the inside of the shell.  A track of
the oil extended to the bottom of the vacuum jacket giving the
appearance that oil or grease had been sucked into the vacuum vessel for
some time, running down the inside of the jacket.  There was also a
short trace in the tangential direction, as if the Dewar had been on its
side for some time.  While the CCD was out I set the ccdfilt to U and
opened the shutter.  The filter glass did not appear to be broken and
since no replacement U filter was on had I did not disassemble the
instrument to replace it.

We took the vacuum jacket to LS where Fabian Collao ground out the
offending area and re-welded it.  We returned to CT, re-assembled the
system only to find that the leak was worse.  Back to LS, this time with
the leak detector (one and one half trips to CT in one very long day).

On Thursday the shop fabricated covers for the vacuum jacket so we could
leak test the vacuum shell without taking it to CT.  Fabian also tried
again to weld the offending area.  Late Thursday we tested this second
repair attempt and found that it was still worse. We also discovered a
smaller leak at about 90 degrees from the original one, very near the
end of the longitudinal seam in the outer shell.

On Friday Fabian re-welded both bad areas.  Ramon had now departed on
his much delayed vacation.  Atwood tested the thrice repaired Dewar
Friday afternoon and found it much better.  Exactly how much better was
hard to tell since we had been forced to use different vacuum pump
configuration for each test.  I decided to backfill the gap between the
ring and the shell with 5 minute epoxy.  The epoxy seemed to seal the
leak and I left the shell under vacuum on the leak detector Friday
night.

Saturday

Upon rechecking there was a still smaller leak.  I added more epoxy,
which seemed to reduce the leak rate even more.  I took the shell and
leak detector back to CT and re-assembled the system (again).  Leak
testing showed that there was still a very small leak in the same area
(the second leak had been fixed with repair number three.).  The total
pressure was 20 microns after about 5 hours of pumping.  The pressure
went to zero immediately with a whiff test.  The instrument was cooled.
Except for some erratic reading on the ccd temperature while cooling the
system now appears to be normal.

Juan Espinosa informed me that the reason I had not been getting reports
on the pressure was because the ionization gauge tube was bad.
Unfortunately I did not know it was bad and did not bring a spare.  If
the ion gauge were available we could predict exactly when the Dewar
will need to be pumped as well as confirm that the leak does not get
worse. I am sure the instrument had no leaks when installed and have no
explanation for how the oil/grease came to be in the gap between the
ring and the shell.  It has been suggested that it was residual oil from
the machining process but I could not detect any oil or grease anywhere
except in the area of the leak.

As always, the CTIO staff was very helpful and a pleasure to work with.

--
Bruce Atwood
Research Scientist