Difference between revisions of "Problem reports"

From Telescope
Jump to: navigation, search
(Added text about resolution of these problems. Added info about problems with guiding.)
 
Line 19: Line 19:
 
19 Mar 2013: (after aborting and restarting a plan) autoguiding ran away in one direction, despite having a nice bright star; ACP stopped it eventually, restarted, and it went fine for a few minutes but then started systematically drifting away again. (A few days later: in retrospect, this seems like the guide camera binning problem discussed above (21 Feb 2013); and I note that we had to restart the plan at the beginning of the evening and may not have started clean.)  
 
19 Mar 2013: (after aborting and restarting a plan) autoguiding ran away in one direction, despite having a nice bright star; ACP stopped it eventually, restarted, and it went fine for a few minutes but then started systematically drifting away again. (A few days later: in retrospect, this seems like the guide camera binning problem discussed above (21 Feb 2013); and I note that we had to restart the plan at the beginning of the evening and may not have started clean.)  
  
1 Apr 2013: 2nd plan in chain did not complete; bug in ACP appears to have made the target seem unobservable and so observing was skipped and dawn plan initiated. But during the hours the telescope was waiting for dawn, it appears to have kept tracking and hit the RA limit (it was found at this limit when Eric came in in the morning).
+
1 Apr 2013: 2nd plan in chain did not complete; bug in ACP appears to have made the target seem unobservable and so observing was skipped and dawn plan initiated. But during the hours the telescope was waiting for dawn, it appears to have kept tracking and hit the RA limit (it was found at this limit when Eric came in in the morning).  ''Update 5 April 2013:  (1) Eric found the bug in the ACP code that caused it to improperly skip one target, and posted a fix to the ACP support forum.  However, this is not yet fixed in the ACP code, so take care with using WAITAIRMASS in ACP plans - will not always work as advertised.  (2)  The problem with tracking turning back on spontaneously is a bug in TheSky, according to Bob Denny.  It appears that it may be fixed in TheSky X; Eric ordered a copy of the software this week, and will upgrade TheSky and test if the fix works.''
 +
 
 +
5 April 2013: Realized on 4 April what may have been causing the guiding problems.  ACP automatically connects the camera in MaximDL, and ACP itself talks to the telescope and the dome.  But if ''MaximDL'' is not connected to the telescope (through the ACP hub; initiated by clicking the "Connect" button in MaximDL's Observatory window), then Maxim has no way to send guiding commands to the telescope.  So it will take images of the guide star, but won't be able to make any corrections.  I think this is consistent with the times where we've seen the guiding just run away, using a bright guide star.  Fortunately, ACP v. 7.0 now initiates this connection automatically at startup.  Eric installed this new version on 4 April and verified that the connection is made automatically.  Hopefully this will fix the guiding problems.

Latest revision as of 10:22, 5 April 2013

1 Feb 2013: Six good hours of observing with a chain of ACP scripts, last script ended about 1:30 AM (darks and biases), scope pointing to the west and low. Scope doesn't home automatically (should it?) so I home it via The Sky and the software reports slewing but it never reports reaching home. I go to the observatory, physically, and find the scope's RA alarm on and the scope pointing west still. The Sky seems to have lost contact with the scope.

9 Feb 2013: Ran a YETI 2x2 mosaic field. Then separately started up darks plan then an autoguided KELT field chained to a dawn flats + bias plan. In the morning, the science data were taken, but the dawn flats and biases weren't. Telescope low and west, RA alarm going off, The Sky hung (dialog box showing slewing was going).

14 Feb 2013: Had some strange guiding problems (star looked strong, pointing kept drifting away). Turned out (the next day) that the guide settings had been messed with somehow - the manual calibration speeds and angle were off, quite off in the case of x. The next evening we did a calibration and it worked and those values were reset to something reasonable.

17 Feb 2013: Smooth start up of script (R band transit with autoguiding); manually aborted the ACP plan simply to change the exposure time, but when we started the plan back up, MaximDL hung. We had to restart all the hardware (just restarting MaximDL wasn't enough - camera connection couldn't be made), and then reconnect the software (didn't restart the computer, just the scope and camera, and made sure the dome was homed).

The plan started up fine. But a few minutes into it, guiding started acting funny again (much like 14 Feb) but the settings look OK.

18 Feb 2013: switching plans at 11pm, autoguiding problems upon starting the second plan (after Eric saw similar problems related to 1x1 binning, but then he got it working by bypassing ACP). Image is 1x1, and after just taking the first full-frame guide camera image, it doesn't really seem to be trying to start autoguiding.

After watching it take unguided images, continually failing to start autoguiding, for 20 minutes, I aborted the script (which MaximDL was not exposing an image, but rather while it was plate-solving the just-downloaded image. Then I simply took a guide camera image with MaximDL (it was 2x2 and looked good), started guiding, and it ran fine. After a couple of minutes, I started the Kelt target plan, and it started up fine, took the pointing image, the first target image, guiding kept running seamlessly. ACP console reported "guider check OK." Few minutes later and everything's still going fine. A few hours later and everything's going fine.

21 Feb 2013: Problems with guiding, similar to some of the stuff we've seen the last few weeks. Plan started up, autoguiding started up just fine (from within the plan). Images (both science and guide) looked fine. But after a few minutes, the guide errors got bigger and systematic -- always in one direction. ACP stopped guiding because of the big errors. I aborted the plan. Took a guide image directly with MaximDL and started guiding from MaximDL. Once I verified it was going well, I started the plan anew, but MaximDL crashed after taking the first image. Homed and shut down the hardware. Closed all software. Started hardware and software back up. Started up the plan (without getting guiding going separately in MaximDL), and everything seemed to work fine. It's 20 minutes later and everything, including the guiding, is going smoothly. While the guiding was having its problems, I verified that the guide camera image was 2x2 binned. And it was.

24 Feb 2013: YETI 2x2 observing went just fine; there was a brief power outage on campus, but while the Mac in the control room went down, everything in the dome continued to operate normally and observing was never interrupted. After the YETI observing (and dark-taking) ended around midnight, I did some recreational observing till 5pm (automated) and then dawn flats in three filters were taken, without a problem. (The last few weeks, long periods of inactivity seemed to make the system more likely to fail when ACP finally issued pointing commands to the telescope via the Sky; that didn't happen this time, perhaps because the scope was actively being used for most of that time; and also, therefore, not tracking far to the west.)

19 Mar 2013: (after aborting and restarting a plan) autoguiding ran away in one direction, despite having a nice bright star; ACP stopped it eventually, restarted, and it went fine for a few minutes but then started systematically drifting away again. (A few days later: in retrospect, this seems like the guide camera binning problem discussed above (21 Feb 2013); and I note that we had to restart the plan at the beginning of the evening and may not have started clean.)

1 Apr 2013: 2nd plan in chain did not complete; bug in ACP appears to have made the target seem unobservable and so observing was skipped and dawn plan initiated. But during the hours the telescope was waiting for dawn, it appears to have kept tracking and hit the RA limit (it was found at this limit when Eric came in in the morning). Update 5 April 2013: (1) Eric found the bug in the ACP code that caused it to improperly skip one target, and posted a fix to the ACP support forum. However, this is not yet fixed in the ACP code, so take care with using WAITAIRMASS in ACP plans - will not always work as advertised. (2) The problem with tracking turning back on spontaneously is a bug in TheSky, according to Bob Denny. It appears that it may be fixed in TheSky X; Eric ordered a copy of the software this week, and will upgrade TheSky and test if the fix works.

5 April 2013: Realized on 4 April what may have been causing the guiding problems. ACP automatically connects the camera in MaximDL, and ACP itself talks to the telescope and the dome. But if MaximDL is not connected to the telescope (through the ACP hub; initiated by clicking the "Connect" button in MaximDL's Observatory window), then Maxim has no way to send guiding commands to the telescope. So it will take images of the guide star, but won't be able to make any corrections. I think this is consistent with the times where we've seen the guiding just run away, using a bright guide star. Fortunately, ACP v. 7.0 now initiates this connection automatically at startup. Eric installed this new version on 4 April and verified that the connection is made automatically. Hopefully this will fix the guiding problems.