Anybody can identify these white spots on all my images now? Started Oct 2020. Bad dew heater? Gremlin shining flashlight into my scope? My camera going bad? Water in the OTA?
It is almost like FLATS are now adding noise to my integrations.
I’ve been running SGP v3.1.0.454 and APP 1.077 since 2/22/2020. No camera drivers updated since 2/22/2020 according to the SGP directory.
No change in my imaging train during that period.
On or around October 2020, it was almost like my FLATS (from 12/2019) started to go bad, like too many photocopies of a photocopy used to do. Although that isn’t, of course, possible digitally.
I’ve updated my APP to 1.082beta version, problem persists, so not a software issue.
There are no issues with SGP sequence. Runs perfect all night. No errors.
My gain is set to 174 on all FLATS, LIGHTS, DARKS taken in SGP for the last 2 years.
I do not specify OFFSET in SGP, but in ASCOM it is defaulted to 77. I have not changed this in >2 years. I have no reason to suspect that SGP magically changed my OFFSET starting in October.
I have ruled out the moon.
I have ruled out any new extraneous lights from neighbors.
I took a new set of FLATS the other day, and the perplexing thing is that my December 2019 FLATS and December 2020 FLATS work perfectly integrating any data prior to October 2020. But any FLAT shows the same spots integrating data taken after October 2020.
Please read that statement again, I’ve posted in other forums, and sometimes it it overlooked.
If I integrate subs from 2019 and 2020 without any FLATS, to my eye, they look the same. See sample images.
I’m at a loss. I can subtract the spots in PixInsight by inverting flats and using some fancy math, but that is a pain in the but. And everything used to work just fine.
Thanks.
Other threads with more images and peoples thoughts…
My gain is set to 174 on all FLATS, LIGHTS, DARKS taken in SGP for the last 2 years.
I do not specify OFFSET in SGP, but in ASCOM it is defaulted to 77. I have not changed this in >2 years. I have no reason to suspect that SGP magically changed my OFFSET starting in October.
I recently found that my gain and offset “randomly” changed on one of my imaging rigs within the past few months, and I have no explanation for it as I never change either setting. To catch situations like this I wrote a simple app that batch checks data stored within the header of Fits and Xisf files, including gain, exposure, pedestal, filter, SSWEIGHT, and CCD temperature. You can download it here and give it a go; email me if you find any bugs:
When you run it you’ll get an initial warning regarding a missing options file; just ignore that because it will be produced after your first run. Select a folder that contains all your Fits or XISF images, select the appropriate file extension from the combobox at upper right, and press GO. The results will be shown in the “metadata” tab, and you’ll quickly be able to see whether you have mismatched temp/gain among any of your image files. Offset can’t be checked because that info is not (yet) stored in the Fits header by SGP, so it may be better to confirm the offset value you have set in the camera driver.
Speculating here but what light source did you use for your flats? Those marks look very like fingerprint smudges to me. If you are using a flat panel has it been handled a lot? The marks seem more prevalent round the edges as I think would be the case if the panel had been picked up by say a child. I suggest you capture some more flats with a different light source (or repositioned) and see if the pattern persists. If pattern is unchanged then I would look carefully at all the other optical components but if the marks were there they ought to have been calibrated out by the flats
… Similar to Ross, I found that QHY ASCOM driver was not managing its offset and gain settings robustly and would default to the first in the list when you first connected. I think it may have been corrected since but my advice is to double-check the settings before running the sequence.
Well - the problem is definitely the flat normalization of the lights. You can see the pattern matches up.
It does look like for some reason the gain has changed, you could try using PixelMath on the master flat. Scale it up or down until you get a good result. If this is CMOS, did you remember to avoid using bias files and just use dark flats for creating the master flat?
