I have just started using an f/10 system for photometry. Unlike my previous f/2.8 system using a camera lens, the profile across a flat is quite different from the profile of the sky background across a science image. See attached screen shots.
Flats were taken at dawn with a white acrylic plate to act as a diffuser across the front of the telescope which was aimed at the zenith.
My guess at the explanation is that light ‘rays’ entering the telescope from the sky (at infinity) are parallel, whereas those entering the 'scope from the under surface of the acrylic diffuser would do so at various angles. Other explanations?
Not sure what the two images showing. The top one looks to me like a flat-field image for a system that has a lot of vignetting, while the bottom one looks basically flat. In principle the flat should take out the gradient, though you may find there’s still a small zero-point gradient in the photometry from center to edge that requires an additional radial correction.
Since most of the signal in an image is at the low values (near black), it may help to display things with light sky and black stars so faint stuff is more readily visible.
The top image is a flat field taken as described - at dawn, acrylic plate diffuser over telescope which is aimed at the zenith. There is severe vignetting. The graph at the top left is the profile along the yellow vertical line.
The bottom screenshot is of a non-callibrated science image and the corresponding line profile.
My conclusion is that the described method of taking flats with this system does not seem to be valid. If the system produces vignetted images, the effect should be seen in the line profile of the sky background.
Here are the corresonding screen shots from my f/2.8 system, comprising an astro camera imaging through a camera lens. Vignetting is visually obvious in both the flat and the science image. I’ve drawn line profiles across only half of the images to reduce the chance of the line crossing images of stars (which would alter the Y scale of the plot)
The flat was taken using a different method - a white digital screen directly in front of the lens.
To me, the contrast with the f/10 system is very obvious.
The sky background image was non-callibrated. Exposure was 30 sec. I should have explained the camera was a ZWO ASI294MM imaging through a V filter and an 8" f/10 Schmidt-Cassegrain telescope.
Now that I have done some more experimentation, gaps in my knowledge are being filled.
The line profile across the diagonal of a master flat is determined by the exposure. Attached are two master flats. The one with the severe vignetting is the master from the set of flats of which I showed one in the initial post. My guess is the exposures were too long. The master flat with not much vignetting was made from images with a much shorter exposure duration, way below 50% of saturation. The exposure was selected so that the individual flats had a line profile almost identical to the sky background in the science image in my initial post.
Flats with intermediate exposures will have intermediate degrees of vignetting.
I think the guideline for making flats is that the mean ADU across the image should be about 50% of the ADU at saturation. Correct me if I’m wrong. I guess that makes sense, if one adheres to the other guideline that exposures for science images should have peak ADUs of the brightest comp stars about 50% of saturation.
Therefore, my suggestion in a previous post that there may be a problem with making flats using a diffuser just in front of an f/10 telescope was not correct. Live and learn.
Chuckle. I spent much of my professional career in the X-ray and radio astronomy domains. It was only after I retired that I started serious optical work. I remember one of my first reactions was: what the $#&@ is a flat!
The flat you first showed of your f/10 system looks a lot like mine: about 42k in the middle and 39k on the wings. Also, the software by default displays the results in “extended” mode so things that the brightest are white and the darkest black even though the difference in absolute scale is only about 8%
