Looking at ASI1600 Pro Bias and Dark Flats

[paul]

Thinking what to do about Bias (dark Flats) and prompted by an Adam Block tutorial I decided to have a look at what they look like for the 1600. I captured 30 darks at 0.01, 0.1,0.2,0.5,1,5 and 10 seconds at gains 0,139 and 200 (camera at -15 C). 

 

Looking at the data I decided to concentrate on 0.01, 0.1 and 10 sec and use Pixelmath on the stacked images to do comparisons.

 

At Gain 0 it is possible to see the Fixed pattern noise (if stretched):

 

 

The banding variance at Exp 0.01s is small (K varied <> 3 parts in 10k) with average K being about 0.0118. At Exp 0.1s K is approx. 0.0119 and at Exp 10s K is approx. 0.0119.

 

If you subtract the 0.01s from 10 s exposure the average result is around K0.00007. The delta image shows no evidence of amp glow (even with NB my flats aren't normally longer than 10 seconds so that is good news):

But it shows more hot pixels developing.

 

I will run the test again on another night to see if these results are repeatable. If they are then I can construct a library of Bias/Dark flats. As long as the banding stays consistent even small variations from session to session in the bias values can be ignored - Flats have a typical K value of 0.5 so small variation around K0.0005 should have little impact in the profile.

 

My data for gains 139 and 200 seems suspect, there is evidence of faint vignetting and the FPN has disappeared, suggesting that even though it is a very small signal I think I have some light leaking in. Another reason to redo the test.

 

Edited December 30, 2021 by paul

[paul]

Just in case anyone is waiting on an update🤔... Well, my idea for a library of Dark Flats seems to be viable, certainly Dark Flats for a cooled camera are repeatable across multiple sessions with differences no more than you expect of frame-to-frame variation. This is true at gain 0 and 139. What I thought in my first session was LP leaking in to the camera is actually evidence that camera noise comes from different parts of the circuit and that at the higher gains, amplifier noise (I assume) dominates over the Fixed Pattern Noise seen at 0 gain.

 

I have decided not to pursue this further; I've only now spotted N.I.N.A. supports Dark Flats in its Flats Wizard and basically makes it so easy I might as well use them. And since I tend not to move my camera and leave it on the scope between sessions, I actually only need to take (Dark) Flats occasionally. So Dark Flats are a part of my process now. 

 

I need a similarly automated approach in the Asirair Pro for my wide-field setup.

 

As I suspected, my tests have verified Dark Flats subtract the bias and camera noise from Flats. Because the Dark Flats are produced for each filter it results in the correct amount of base noise being removed from each Master Flat and this results in better black point alignment between Master Lights than if I didn't use Dark Flats. Image processing is still possible if you don't use Dark Flats (it's how I did it until this month!) but you will always have an extra step to align the black points of each Light manually.