18/09/2013

Ra trailing update

Trailing in the Ra axis was a noticeable problem with our images in the weeks leading up to the hardware failures in Tenerife. This post talks about the diagnostic work and repairs that were completed to resolve this issue.

We were not sure what was causing the apparent trailing in our images. Chris has observed that the system would point correctly, then after a period of around a minute move away from the target object. After this sometimes it would recover, others it would stay at this new location.

To isolate the possibility of the problem being one of a loose element in our imaging set-up we imaged simultaneously using both the Galaxy and Cluster cameras, watching to see if the same error was replicated on both optical paths.

video
In the video it should be noted that the Galaxy and Cluster cameras are rotated by 180 degree from each other, this results in a movement on one camera to the left appearing as right on the second.

The result of these tests were conclusive, the error was not being caused by a problem with the imaging set-up.

We contacted Software Bisque with our findings and shortly after got a reply with a number of suggestions on where to look for the error.

Our first step was to check for play in the Ra drive worm. We found a little play in the worm and tightened the castle nut to eliminate this movement. We imaged for a night with this set-up to look for changes in the trailing observed in the returned images. This work appeared to help a little, but it hadn't solved the problem.

The next step was investigating the possibility that the worm was pulling away from the Ra gear under load. We found that there was some play in the spring loading. Chris performed some tests and managed to confirm that the error in Ra tracking was very similar to the difference in pointing generated by manually moving the Ra axis from one side to the other of this play.

The Ra worm is held against the gear by a series of spring plungers. We checked the offset of these plungers and found that they were all set at in the positions suggested by Software Bisque for a mount with our instrument payload. As a last ditched attempt we pulled the spring plungers from the system and compared them to the old plungers pulled from the system when we last replaced the Ra worm. The difference between these springs was startling. The new plungers were taking considerably less force to compress that the older ones. As a result of this we have switched back to the older plungers.. and fingers crossed at the moment we are seeing a substantial reduction in the number of trailed jobs.

No comments:

Post a Comment