ALMA Production OPT System
Last Update: JeffMangum - 11 October 2010
Some odds-and-ends related to the production OPT systems:
Flat Fielding Necessary for Daytime Measurements
Analysis of daytime measurements made during acceptance of the first production OPT system revealed a "strange background with structure". Here is some of the conversation thread associated with this issue:
The weird background at high light levels is not as stable as I had initially reported. I had only looked at two images that were taken closely in time. If one looks at images taken several apart, the bizarre background appears to warp geometrically. This could arise if the CCD temperature was not maintained accurately over this time period, or if the pattern is due to some optical effect that depends on flexure. It does, however, appear to be a multiplicative effect that can be removed if an image of nearby blank sky is taken within a few minutes of observing the target star.
asked if the centroid measurement robustness is really affected by this background:
The centroid measurements are excellent provided that the star is bright. If you need to dig deeper to fill out the grid on the sky, then the centroids could become unreliable. I didn't see any data that I could use to check on a fainter star, so I can't assign any quantitative value to the possible centroid errors. If someone can send me data over a brightness range, say doing 5-10 stars stepping 0.5-1 mag fainter each time, I'll do the measurements.
Therefore, we do need the background measurement associated with each star measurement for daytime observations.
I agree with George. If you're restricted to 4th or even 5th magnitude stars, there aren't a lot of them. It is certainly worth a try to see if this scheme works in practice. I'd recommend offsetting in azimuth about 10 arcmin just before taking an exposure on the star and use the Do Blank feature of the software. Then zero the offset and take a series of, say 10, exposures on the target star. The software will subtract the Blank before doing the centroid calculations. You'll have to use short exposures, say 0.2 sec, in order to avoid saturation in the daytime and you can tell the software to take the number of exposures you need in just one command. Taking many exposures lets you average out the seeing jitter that you get in such short exposures. (At night we were using one or two exposures of a few seconds on faint stars to average out seeing effects.)
noted that this will require some changes in the way we do all OPT observations, in practice:
This will require some changes in our standard all-sky and offset pointing observing modes, then. For example, it will effectively double the amount of time it takes to make the same number of measurements that we get now, or we will have to halve the number of measurements made. This is doable, and not a major issue. Just need to remember to implement this change when the time comes.
then asked if this was a daytime-observation-only requirement to do sky flat measurements, to which Bill Peters
Yes, it is primarily a daytime requirement. I don't think Ralph was even doing Biases and Flats at night during his all-sky pointing's. That's not to say that doing these wouldn't make a difference, only that the difference, if any, must be subtle. He was mainly trying to get his software to work, and was not too concerned with refinements. We were mainly thrilled to see how easy it was to get the RMS below 1 arcsec after the pointing fit. After you have more experience with these OPT's you'll know better how to get the most out of them.
OSF Test Measurements
- See AIV-1749 for details on initial characterization of the first production OPTS.