Executive Summary: The coma is gone. The FWHM is about 3 pixels, not as sharp as it's been in the past ( Tech Note 51 shows that Mark IV images of Serpens taken in March, 1999, showed FWHM about 2 pixels at center of frame, in both V and I). I-band is very slightly less sharp than V-band, but I don't think it's significant.
We have been having problems with coma in the Mark IV images for over a year. Dave Garnett recently postulated the cause of this coma might be that one of the lenses (number 5) might be oriented backwards. Tom Droege has modified several lenses to reverse the number 5 lens, and taken some data with them. This Technical Note is a very brief analysis of sub-sections of two images taken with the fixed lenses.
On the night of Thursday, March 30, Tom acquired images simultaneously in V and I. He then sent portions of one image from each camera to other people for analysis. Here are the relevant parts of his E-mail:
So I have cut four small images out of the clearest frame from last night. There were thin clouds. I will send these as attachments under separate cover.
The images were taken from simultaneous V and I images. There are a very small section to not take forever to send. They are:
- vcenter.fts
- vcorner.fts
- icenter.fts
- icorner.fts
While the images were taken from the same area of the image, I did not try to get the same stars.
The images had exposure times of about 100 seconds.
| Vcenter.fts | 121 x 133 | download FITS |
| Vcorner.fts | 145 x 125 | download FITS |
| Icenter.fts | 137 x 125 | download FITS |
| Icorner.fts | 149 x 125 | download FITS |
I don't know what area of the sky they cover, but the orientation of all images is North to the right, and (I'm guessing) East up.
First, here are pictures of the images, zoomed by a factor of 3:
Vcenter and Vcorner:
Icenter and Icorner:
It looks to my eyes as if the V-band corner image does show a small amount of coma (with the bright core to the lower-left and the fainter wings to the upper right), but I don't see a similar coma in the I-band corner image; if anything, the wings of the I-band corner image point straight up, or maybe a little to the left.
Using the XVista image analysis package, I plotted the radial profiles of two stars in each image. Using the data within a radius of 5 pixels of the centroid, I fit a circular gaussian to each profile.
Image FWHM (pix) avg FWHM (arcsec) --------------------------------------------------------- Vcenter 2.71, 2.68 20 Vcorner 3.19, 3.06 23 Icenter 3.19, 3.27 24 Icorner 3.53, 3.26 25
These numbers confirm the visual impression that the I-band images are slightly "softer" than the V-band images.
Plots of the radial profiles follow:
Profiles from Vcenter:
Profiles from Vcorner (in which the only stars were much fainter than those in the other images):
Profiles from Icenter:
Profiles from Icorner:
Again using the XVista image analysis package, I measured the curve of growth for two stars in each image. This is simply the amount of light (above the sky background) present in each star, in circular apertures of several sizes (3, 5, 7 and 10 pixels). I have normalized the values to the light within the largest aperture. A star with a sharp core will have values all close to 1.0, while a star with a significant amount of light in a halo will have values which grow steadily to that in the outermost aperture.
image star 1 star 2
r=3 r=5 r=7 r=10 r=3 r=5 r=7 r=10
-----------------------------------------------------------------
Vcenter 0.908 0.964 0.962 1.000 0.824 0.891 0.871 1.000
Vcorner 0.791 0.987 0.939 1.000 0.688 0.816 0.799 1.000
Icenter 0.897 1.051 1.038 1.000 0.869 1.054 1.046 1.000
Icorner 0.788 0.947 0.973 1.000 0.908 1.042 1.048 1.000
It appears that in all cases, most of the light is inside the core of the PSF. The V-band corner values are clearly less concentrated than those in the center, but the comparison isn't quite fair: the stars in the corner are fainter than those in the center. It's not so clear that there's a difference between the center and corner of the I-band image.