# 05 Nov, 2018 12:12
I noticed that I am usually left with a tiny bit of "amp glow" (?) in the top right corner/side of my calibrated images and was wondering what the reason for that might be.
All images (dark, bias, light, flat) were taken with the same laptop, same chip temperature, same cables etc. Darks have the exact same exposure time as the light frames. Only difference was that for the dark/bias frames, the camera was indoors, so ambient temperature was different.
-15° Sensor temp
For creating the dark/bias frames, I followed Warren Kellers instructions in his book. Calibration of lights/flats was performed with PI batch processing.
20 images were stacked for the master dark
100 images were stacked for the master bias, then turned into a superbias using 6 layers.
No dark optimization was performed.
The fully stretched L: Source as PNG
# 05 Nov, 2018 12:45
|Hello Marc, first of all, it is a great image! Are you sure it is amp glow? To me, it looks like a minor problem with your flats.|
# 05 Nov, 2018 12:47
|Thanks! No, I am not sure it's amp glow, but it sort of fits, position-wise. But I will investigate whether this is a flat issue; I actually had a hard time finding the right exposure time, so maybe I still need to improve that…|
# 05 Nov, 2018 12:59
For me it looks like if the light source of your flats wasn't perpendicular with your optical axis. To see that, remove the first 4 layers of your image with the MultiscaleMedianTransorm, invert and do a "Boosted" automatic STF. Your flats seems to overcorrect towards the right of your lights.|
Anyway, it seems to be a minor problem, I think you can correct it with DBE, if you can't obtain new flats (if indeed this is the source of your problems!)
# 05 Nov, 2018 13:09
Interesting, yes - the flatfield was taken with a Lacerta Flatfield box, which - supposedly - has a en even illumination. But it is the 8" version, used on a 4" refractor, so maybe that creates some issues. But otherwise, the scope was pointed to the zenith and the box was put on top of the scope, pretty centered.|
I wonder if this might also be some sort of internal reflection. The Baader filters needed to be blocked at the corners (using a ring made from black cardboard) to prevent light to reflect from the uncovered sides and into the center of the image. But I suspect that this would be more noticable than this.
God, I hate this hobby .. (kidding, of course).
Sounds like I have some more detective work to do
# 05 Nov, 2018 13:17
MarcHihi I've spent hours and hours and hours to learn how correct my Samyang lens with its massive vignetting at f/2! And I'm still not sure if I'm doing it correctly!
For me, bad flats is the root of all evil!
Try rotating the flatfield box, if possible, when taking your flats. Random orientations will help averaging out any illumination unevenness!
# 06 Nov, 2018 20:15
After some fiddeling, this is what I have observed:|
Producing all calibration frames (dark, bias, flat) "by hand" actually seems to largely remove the issue. Not sure why, but my best guess is that it has something to do with how ColdMOS works for imaging. You always get some degree of (non-linear) amp glow, so "dark optimization" isn't recommended during batch processing. But as a consequence it would appear that the flat does not get properly dark substracted. I guess this would need a "dark flat", but that is, I think, not an option in BP.
# 06 Nov, 2018 20:21
|Thanks for the feedback Marc, that makes sense. I have read about using matching darks and dark flats without 'dark optimization' for such cameras. Also I haven't used BP so I didn't know what options are available. I'm waiting for your next image and I hope this thread to be useful for the rest of the community. Clear skies and flat flats !|
# 19 Nov, 2018 07:29
|Oh and another thing I noticed - despite using APT's "CCD Flat aid" function to determine exposure times (~30.000 ADU seems to work best), my flats are occasionally severely under-exposed (mean ADU in the stacked imaged around 4000). Still not sure how that happens, but that is of course another reason why one might see gradients…|
|You have no new notifications.|