|
| 040707 |
UserPreferences |
| Magellan Telescopes | FrontPage | RecentChanges | TitleIndex | WordIndex | SiteNavigation | HelpContents | moin.sf.net |
June 30, 2004, 11:00 AM PDT
Attending: Mark Phillips, Miguel Roth, Skip Schaller (LS telephone), Frank Perez, David Osip (LCO video), Christoph Birk, Alan Dressler, Alan Uomoto, Steve Shectman (part-time) (SBS video and phone)
Contents
Frank reviewed his observations of the Z spring plunger, which was a suspect in the z-motion wandering of IMACS. Replies from Steve Gunnels suggests that the plunger, while not applying the designed preload, was working well enough that it was not obviously the source of the problem.
We left the discussion thinking that Steve Gunnels might combine a GMT trip to LCO to look at this.
There was some discussion based on an email report from Mark Phillips (copied below). The conclusion was that there is good evidence that the IMACS guiders move during long exposures when they shouldn't. A test suggested by Dressler was to mount dial indicators on the cameras and see if they move during an exposure. The S-H guider, rather than the centerfield guider, seems to be more suspect than the other. The motion could be caused either by noise on the guider encoder (a Renishaw) or by loose optics or detector. The encoder noise would cause motion because there is a servo system that locks onto an encoded position, correcting for sag in the chain drive.
Some observations suggested radial motion of a guider since the star separations changed during the observation. Mark Phillips ruled out differential refraction (0.2 to 0.3 arc seconds in these cases), which isn't accounted in the software.
Previous evidence was also seen for instrument rotator glitches, sudden large motions (a few arc seconds) that occur over periods of less than a minute. Frank thinks this could have been caused by a cam follower installed on IMACS for testing purposes; this part rubs against the rubber seals on the instrument covers and causes enough drag to affect the rotator. The cam follower has been removed.
Hi,
We decided to concentrate almost exclusively on testing rotator guiding at the IMACS port during the last two nights of engineering time (July 2-3). Skip was able to also make tests of rotator guiding at the MagIC port during the Clay engineering nights (June 30 & July 1). I think it's fair to say that he has made very good progress on getting rotator guiding to work at both ports. Indeed, during the first half of our second night on the Baade telescope, Skip compared performance with and without the rotator guiding and found much better results (i.e., the star remained with the S-H aperture much more reliably) with the rotator guiding on. I imagine that Skip still needs to make more tests (and perhaps further adjustments) before claiming victory, but this is still very good news.
On both of the Baade nights, we carried out a test of the effectivity of rotator guiding by aligning a mask and then monitoring the alignment during ~2 hours with the telescope tracking and the rotator guiding functioning. Skip has written a summary of the results of these tests which I have attached below. There are a few things to note about the tests results. First look at the 2nd night. You will notice that over the course of the test, significant offsets in all 3 quantities (dra, ddec, and dtheta) accumulated systematically. At the end of the test, when we applied the correction required to recenter the alignment stars in the mask, the S-H star was driven to the edge of its box. On the 1st night, we saw a similar behavior. However, between the time when images 13 and 14 were taken, an "event" occurred when a large error in theta appeared rather suddenly in the rotator guiding. Skip had to intervene and rotate the rotator to get the PG and S-H stars back in their boxes. Image 14 was taken after Skip had corrected the rotator, yet we see a significant offset has occurred in theta in the mask alignment solution. If this "event" had been due to Skip's rotator guiding software, then when he moved the rotator back to place the PG and S-H stars in their boxes, we should not have seen an offset in the alignment solution.
At the end of the second night, Skip carried out another test also gave interesting results. Here is his description:
"At the beginning of the night, we set the SH guider position to 938,270. We were able to move a star from the center of rotation as seen on the center field camera, to the center of the box (set at 500,500) on camera 2, by offseting the telescope 938 arcsec south. At the end of the night, we repeated this offsetting test. The star did not fall in the center of box of camera 2. It was off by 2-3 arcseconds, in both radial and tangential directions. We did the same thing, but now first rotating by 90, 180, and 270 degrees. The star was always off by slightly differing amounts and directions. These errors repeated, when we returned to the same rotation angle."
The combined results of these tests indicate that something is not quite right. One possibility is that the center of rotation of IMACS does not coincide with the center of the field rotation produced by the telescope. This could conceivably explain the systematic alignment errors that are observed to build up in the tests carried out with the mask. However, I don't see how that can be the explanation for the inconsistent (yet repeatable) offset tests as a function of rotator angle between the center of IMACS rotation and the S-H probe. Nor is it simple to explain in this way the theta "event" that occurred during the mask test on the first night. Rather, these imply that there is differential flexure between the CFG (or the slit mask in the case of the "event") and the SH probe.
Dave points out that the PG, S-H, and CFG were all properly initialized prior to the tests made on the second night, and it appears that this initialization is necessary.
I think it would be profitable to discuss these tests via telephone or videocon ASAP. Perhaps during the videocon that is already scheduled for Wed of this week?
Saludos,
Mark
We tested the rotator guiding software on both nights. Over a 2-3 hour period, the software claimed it was holding the SH star in the center of the box in the tangential direction. It measured an offset in radius. At the end of the test, we looked at the position of the star, and it appeared to the eye to be where the software said it was, although this was difficult due to bad seeing. At the same time, Mark Phillips was measuring the motion with his mask alignment software. Those data tables are attached. He aligned the mask, prior to starting the rotator guiding. At the end of the second test, he applied the offsets that were being reported by his mask alignment software. That placed the SH star at the edge of its box, some 3 arcseconds off in theta. ------------------------------------------------------------------------ 1st night: image ltime dra ddec dtheta airm 10 4:40 0.12W 0.05N 0.0079 1.154 11 5:05 0.26 0.32 0.0120 12 5:25 0.32 0.14 0.0044 13 5:40 0.80 0.22 0.0097 1.258 14 6:05 1.65 0.56 -0.0476 15 6:30 1.90 0.75 -0.0322 1.389 ] corr. applied 16 6:33 0.49W 0.06N +0.0173 17 6:39 0.86W 0.09N +0.0277 1.427 ] corr. applied 18 6:42 0.16W 0.00N +0.0116 19 6:46 0.02E 0.10S +0.0010 20 6:48 0.00W 0.03N +0.0085 21 0.01W 0.14S -0.0039 ------------------------------------------------------------------------ 2nd night: image ha dra ddec dtheta pgtheta sgradius sgtheta 218 -01:14 0.01E 0.04N 0.0039 -16.09 1.14 -15.31 219 -01:00 0.03E 0.01S 0.0039 220 -00:50 0.04W 0.07N 0.0043 221 -00:40 0.16W 0.00N 0.0114 222 -00:28 0.09W 0.07N 0.0077 223 -00:20 0.23W 0.12N 0.0149 224 -00:15 0.18W 0.17N 0.0164 225 -00:07 0.27W 0.10N 0.0163 226 -00:02 0.31W 0.16N 0.0199 227 +00:03 0.38W 0.10N 0.0182 228 +00:08 0.35W 0.08N 0.0226 229 +00:14 0.47W 0.14N 0.0308 230 +00:29 0.58W 0.14N 0.0350 231 +00:45 0.58W 0.10N 0.0365 232 +00:55 0.63W 0.20N 0.0404
Changes to the engineering schedule were discussed and the table (Schedules) updated. A few things added. We reminded ourselves to make a plan for the ADC dispersion correction testing, preparing MOE slit masks, and plan for any new IMACS work during Oemler's August observing.
