A/D offset value used on Atik CCD?

[symmetal]

I've been reading about the calculations for determining the optimum sky background level for various gain and offset values on the ASI1600 to swamp the read noise by a certain factor and it was stated that the same calculations can be done for CCDs too. I have the Atik One 6 which is the same sensor as the 460ex. The gain is fixed at 0.27e-/ADU but the offset used isn't mentioned anywhere. That the offset value is correct can be checked by looking at a master bias and ensuring that the minimum ADU value is between around 100 to 500.

Here's the Fits Liberator display of the master bias which shows the min and max values are  297 and 360. I cropped off the first column on the image as that is 0 and ruins the minimum value.

This implies that the offset value is good but does anyone have an idea of what it might be. Would 300 ADU be a reasonable offset value looking at this or is there a reason it would be higher or lower than that? Hopefully the noisy knowledgeable members could help me out. 

The other reason I wanted to know was to determine the sky flux I have here without having to buy a Sky Quality Meter, as well as what it was over the past few years. You need the offset value used on your CCD to calculate this.

It's just something to do really on these long windy rainy days.  

Alan

[kens]

Use the median value - which is shown as 322. Note that this is in ADU whereas many of the ASi1600 calculators use a value in e-. If you need the value in e- it's going to be 322 ADU x 0.27 e-/ADU or 87e-

[symmetal]

Thanks Ken. In the absence of the offset of 322 the median value out of the A/D would then be 0 which is the ideal result, but as there will be noise generated in the pre-amp and A/D, I'd have thought the median value would in that case be some value above 0. Maybe not enough to have much effect, but I was just trying to get a better understanding of the noise distribution. 

Yes, I was doing the conversion between ADU and e-, and the x16 multiplication factor between the 12bit ASI and 16bit Atik OK I think. As the noise distribution on the Atik is around 64 ADU the lowest 5 bits don't really hold much signal information.

Alan

[kens]

If you've stacked many bias frames the read noise has been averaged out so the median should represent the offset. To calculate the read noise you subtract one bias frame from another one and calculate the standard deviation. I think you also need to add an offset to prevent the values clipping to zero. Heres one paper on how its done: https://www.photometrics.com/resources/whitepapers/read-noise.php

[symmetal]

Ah! That makes sense. I'll have a read through the link you've given. I can do some tests too on the ASI1600 at different gains and offset and see how the noise is affected. Don't need clear skies for that. Thanks Ken.

Alan