Color information in the AGASC is used to predict the brightness of a star viewed through the ACA from the known AGASC magnitude. This B-V color is recorded from the Tycho-2 Catalog. There are two special cases of color in AGASC 1.5 that impact selection of acquisition stars for the ACA. It is now possible to examine the acquisition history of each of these cases using the acquisition database and determine if an action is needed.
First is the case where no color information is available. These stars come from the HST GSC1.1 and make up a small percentage of usable stars in the AGASC. Stars with no color information are assumed to have B-V = 0.7
Second is the case of stars with B-V > 1.5. Highly red stars are difficult to transform from AGASC to ACA magnitudes (using B, V colors alone results in a large potential error). See the AGASC 1.5 re-calibration report here. Due to this difficulty and a lack of available data, these stars have been given a truncated color, B-V = 1.5.
B-V = 0.7
The ACA reviewers have, in the past, been quite wary of stars with no color information, and rightly so. Only 39 stars with B-V = 0.7000 have documented acquisition attempts. Of these, 24 (61.54%) stars failed the acquisition process. This statistic makes these the worst performing category of acquisition stars (stars fainter than 10.2 mag are next with a 21% failure rate).Below are the magnitude histograms of both successful and failed acquisition stars with B-V = 0.7000.
There is no evident correlation between the observed and expected magnitude for stars with no color information. For those successfully identified by the OBC, we were simply lucky that the stars did not exceed the faint or bright limits of the camera. Also, notice that the majority of stars not identified have no available telemetry.
The above population is admittedly small. To prove that these stars belong to a statistically different population than the majority of stars acquired by the ACA, another test was required. Below are the magnitude histograms for ALL successful and failed acquisition stars since May 2000. For this test the COLOR1 parameter changed to B-V = 0.700 and the expected ACA magnitude calculated accordingly. We would expect to see correlations between these distributions and the distributions of actual no color stars if they belonged to the same population. Clearly, there is no correlation. The 'real' OBC identified no color stars have a highly dispersed distribution in observed - expected magnitude, while the simulated distribution shows no such dispersion. The conclusion must be that stars with B-V = 0.700 in AGASC 1.5 are indeed anomalous.
Just for comparitive giggles, the magnitude distributions for all stars with the real AGASC 1.5 color correction are below.
B-V = 1.5
Our database of acquisition attempts currently contains 2461 stars with truncated color. Of these, only 1.381% (34 stars) have failed an acquisition attempt. This failure rate is on par with the population of stars that pass the starcheck criteria.Below are the magnitude histograms of both successful and failed acquisition stars with B-V = 1.5000.
Note the skew of the distribution of observed - expected magnitude. The predicted magnitude is noticeably brighter than what the ACA sees. This result is not entirely anticipated, but is most probably the result of the large errors produced by magnitude transformations on very red stars.
Conclusions
From the data above it is possible to make a few operational recommendations to improve future star catalogs.The removal of starcheck warnings regarding stars with B-V = 1.5 would have negligible impact. They are excellent performers and there is little need to be wary of them.
Currently there is no provision against stars with B-V = 0.7 in SAUSAGE. These stars should not be used for acquisition unless the star field is near completely barren. Also, the starcheck flag for this type of star should be upgraded to a red warning.
|
Questions, Comments, Recipes
Email: Brett Unks |