Calibration weirdness

[Forum_2018]

Hello,

Lately I've had trouble with flat fielding.
I get dark edges - like dust motes would move. But I have taken flats right after lights without moving anything.
Now I made an experiment - did normal calibration and let it save masterflat, then loaded both masterlight and masterflat and did divide.
Why do I get better/flatter result than doing normal calibration ?

Can deliver a link of masterflat and masterlight if needed.

Tapio

[Forum_2018]

Hello, maybe the dark frame for the flat field (or just the one for the images) were not correct? (different exposure time or temperature?)

[Forum_2018]

Didn't use dark frame as I have a ccd which don't need one: Sony ICX694.

[Forum_2018]

Have you tried subtracting a bias frame from the flats and lights?

Iver

[Forum_2018]

Doesn't make things right.

[Forum_2018]

Yesterday I experiment more.
Make masterflat and then loaded light and did divide with masterflat.
Only when I added flat from 20000 to 40000 it started to correct the light.
There was a lot of vignetting - originally center 24000 and edge 14000.
Does this make sense ?

[Forum_2018]

Don't confuse "Dark frame" with "Noise". Your camera may have low noise, but it has anyway a strong dark frame, at least 1000 ADU (the ADC bias signal). For this reason a proper dark frame for the flat and the image is always required, e.g.

I / F <> (I+D) / (F+DF)

By the way, there could be other reasons (non liner flat frame for excessive exposure?). To find the problem don't user masters, just make a test with a single image, single dark, single flat, single flatdark, making all steps by hand (I - D) , (F - DF) then divide I by F.

[Forum_2018]

Ok you might try setting your SX or other filter wheel to 'One Direction Only' because I have found that the filter ends up slightly different location depending on direction.
John

[Forum_2018]

Lates update to this.
Added some tape to my flat light source (to eliminate light leaks) and also put some tape in my filter wheel - filter wheel had some holes.
Don't know if this really was the cure but latest show promising.

However now I'm puzzled why I have to scale darks to get optimum calibration results (and use scale value of 2000 like).

Before I just had to use flats and maybe bias.

[Forum_2018]

Before using scaled dark frames, try this test:
take a flat field without the light source, using the skylight before full night. Do one image, dark, flat and flatdark. If it works, then the problem is just the light leaks of the artificial source.