PartnerObserving Log for 2007 Mar 14/15

Observers: Rick Pogge, Peter Garnavich
Inst. Support: John Hill
TO: Ed Lundin

UTC Mar 14 1745: Clock board from the CCD controller sent to Tucson for emergency repair. Instrument is disabled until it arrives back on the summit.

2300 - The camera shutter is working again, but the precision exposure time readout is not working, so image headers will contain estimated exposure times only. The instrument is being remounted on the telescope and should be testable in the next half hour.

All subsequent times are UTC 2007 March 15

0013 - Back on the scope and the windows PC reinstalled in the treehoust rack. Fired up the server and interface, and it works. The exposure time measurement is in fact not working according to the logs, so only estimates will be in the TEXPTIME card. Shutter is working as advertised, images look normal.

0026 - Taking Bias and 900s Dark images with the dome dark. Biases look like bias frames, which is good (we didn't break anything else...). Weather is clear, but the satellite shows some junk heading our way from the NNW. Oh well. We'll plan on running a non-photometric program tonight.

0150 - Starting twilight flats, Dec=+27, RA=6h, beginning with Uspec: Counts range 19K-7K ADU, B: 30s range 18K-6K. V up to 60s and lots of stars. Will do the rest at dawn.

0220 - Tuning up primary mirror for the first time tonight... stopped by archiving failure, images not being transferred to the Repository disk. Logged into lbcarchive, and got a hang on typing "df -h" (some NSF mounted disk not mounted). Tried to login to lbcarchive2 -- no luck. Rebooting both machines. Back, but archive is really slow. Bfastextra images can take >1 minute to get onto the Repository. Took us a while to get the primary tuned up, it was way out and converged only slowly.

0313 - Going to M81 field to check seeing and start trying to observe. Tweaked up primary (some confusion as it took an out of focus shot of M81, then moved to the offset field and took another - something glitched). V image being started @ 0321. Seeing ~1.1 arcsec (5 pix FWHM). Star images are rather elongated - like there is a lot of astigmatism... Doing another offset field for a focus check... Going back to see if that worked. Better, but still not as round as it should be. Hmmmmm. Long bad, short good. Smells like tracking problems, but nothing obvious to characterize it yet.

0350 - Going to Dwarf/fuobj21, seeing could be <1" as requested for this program. Problems starting up because I accidentally sent the wrong OB (too many files in that directory and I clicked the wrong one - sue me), and on asking the TO to abort the move, the GUI froze. Getting it unfroze was no fun and burned up lots of time. The abort mode was non-intuitive, and the GUI lock up required me to login as root to a machine to stop/start a service. Suid would be a lot safer...

0415 - Getting back on target now, tweaking primary, started at 0422 (slow tweak up). First 300s image (SDSS g) looks awful! Aborted OB and taking a 5s test exposure at g (using G5sec.ob). This 5s image looks fine, ~0.8" seeing. Our prognosis is that something is wrong with telescope tracking/guiding on 300s images. Meanwhile John and Ed are going to look at the oil on level 4 (something about day crew working on the oil system during the day). Lost a couple of test images to the infamous "too few pixels" fault. One was a Bfastextra, the other was a G5sec image.

0450 - Looking more closely at one of the 5s g images, we're seeing some skew in image quality across the FOV - images are better on chip 1 than on chip 2, worst on chip 3. Looks like field tilt. A Ballextra.ob run confirms this. John recommended we drive to Horizon and check the seating of the cryostat (they admit they rushed it a bit this afternoon). Seems like the best thing to do. Observations of Dwarf/fuobj21 look to be a loss given the time to fix various problems.

0520 - Back to it, going back to this field since we have a baseline from the pre-adjustment configuration... G5sec image, not much change, but images on chip 1 now a little worse. Shooting a B5sec OB, images don't look so good there, either.

We have stopped science observing and are running some focus tests... Tweaking primary then running Bfinefoc.ob (started 0543)... will follow with Gfinefoc.ob and Uspecfinefoc.ob. Might as well check all the offsets while we're at it...

Focus analysis of Bfinefoc: (bessel B)
  • top of chip: focus offset 0.096, at nominal offset of +0.05, images is 4.2 pix fwhm instead of 2.8 pix.
  • bottom of chip: focus offset 0.093, fw=2.65 pix (that's 0.6" seeing kids!)

Focus analysis of Gfinefoc: (SDSS g):
  • top of chip: focus offset 0.096 w/FWHM=2.8 pix (0.64")
  • bottom of chip: focus offset 0.094 w/FWHM=2.38 pix (0.54")

Focus analysis of Uspecfinefoc: (new Uspec filter):
  • top of chip: focus offset 0.486 w/FWHM=3.15 pix (0.71")
  • bottom of chip: focus offset 0.475 w/FWHM=2.67 pix (0.61")

Focus analysis of Vfinefoc: (Bessel V):
  • top of chip: focus offset 0.0089 w/FWHM=2.89 pix (0.66")
  • bottom of chip: focus offset -0.013 w/FWHM=2.75 pix (0.62")

Focus analysis of Rfinefoc: (SDSS r):
  • top of chip: focus offset 0.069 w/FWHM=3.18 pix (0.72")
  • bottom of chip: focus offset 0.056 w/FWHM=2.51 pix (0.60")

Focus analysis of Ufinefoc (Bessel U):
  • top of chip: focus offset 0.318 w/FWHM=3.31 pix (0.75")
  • bottom of chip: focus offset 0.308 w/FWHM=2.93 pix (0.67")

Entered new offsets, and then ran 5s G and B images to check. ... shutter stuck open. Logs show "timeout error..." for both.

I really really want to type a long stream of obscenities right now...

0625 - Power cycling blue channel, the fix from Sunday night... this better work... didn't...

0640 - LBT down to horizon pointing and bacck out to prime focus... John and Ed saw it was sticking, but got better results power cycling the shutter controller manually from the hub. We can do this remotely with testpower.

0715 - Trying a test shot. Shutter worked this time, doing a mirror tweak shot

0723 - Started Lockman Block 2. First 360s image looks OK, stars aren't very round, though. Seeing ~1" or so. 2nd image in the _VLA_V_1 OB is about the same. Third image in sequence lost the guide star early and image quality was degraded as a result. 4th image in sequence lost guide star 32s into the 360s exposure. 5th image never got guider lock at all. This is persistent enough that John is IssueTrak-ing it since once it loses the star, it never recovers, which seems wrong given the seeing is ~1" or less. The autoguiding algorithm could be better.

0759 - Primary mirror tweak, then Block 2, _VLA_V_2.ob. Seeing 1" or so. Collimation OK. Started at 0806 - got a nice guide star this time. Looks better, but also could be a better image. Nominal seeing still ~1". #2 started off with guide star OK, and kept it. Seeing good (0.8-0.9"). Looks good so far for this sequence. Had to abort the last exposure about 1 minute in at ~0835 because elevation tracking stopped (torque maxed out). Ed & John working to recover...

1037 - After some work, moving to Dwarf/fuobj11, a completion target. First pointing attempt led to an elevation drive failure. Reset and coords resent. Seems to have reached target. After the primary tweak, we started at 1057 with fuobj11_1_r, first image looks beautiful - 0.7" seeing all over the 4 chips. Second image also good. Third image taking forever to get into the repository. Data from the technical chips are flowing just fine - no problems in the log. Finally emerged - showing crazy lines - apparently a shutter failure, but moving to the next dither during readout (problem is, it took so long to come down we didn't know until the 4th was done). Fourth and final image shows shutter failure. Taking 5s snapshot... shutter failure. Logs show "timeout error on communication" in both cases.

Only the first 2 images of the 4 in this OB are good, the last 2 are compromised by shutter failure

1129 - cycled power on the shutter controller using the testpower program on lbccontrol, trying another 5s snapshot... timeout error in log. Repeated attempt, same failure to work.

1136 - That's it, we're calling it a night. John tells us we have broken the record for the most number of IssueTraks during a single night of SDT observing (10 or so). At least we get something for the night.

Night Summary

PartnerObserving, problems, and lots of partially-completed OBs due to telescope and instrument failures. A shame since the seeing was clearly subarcsecond. Long images were compromised by what could be some tracking problem (maybe), and occasional shutter failures. The one good thing we got from the night during a period of good seeing was the offsets for all the main filters in the SDT programs, especially the Uspec and remounted SDSS g filters. Some oddities, like the change in the B filter offset.

Overall very frustrating, with much time lost.

Morning Update: The tracking problem we experienced appears to have been caused by the fact that one of the elevation drive motors literally burned out (the right-front motor), and so the remaining 3 drive motors were struggling against an effectively unbalanced telescope. When John and Ed added weight last night to rebalance the scope, they were effectively preloading it to help out the remaining three working motors. That's why the last image, before the shutter failure, were so nice, whereas before long exposures were not producing good results compared to very short (5s) exposures. The burned up motor brushes we were shown this morning were spectacular. The day crew is installing a replacment brush assembly (the commutators were apparently not damaged) and will be testing it soon.

Of course, this means we haven't yet had a chance to address the shutter problem...

-- RichardPogge - 14 Mar 2007
Topic revision: r9 - 21 Jul 2009, NormCushing
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback