Atik 383 gradient in bias

[jolo]

Hello!

In my Atik 383 there is a gradient in the bias frames (see attached master bias in -25C). It can be easily removed using calibration frames, but recently I have read it can be caused by not properly fixed Peltier module, so the sensor is not cooled equally. But the gradient also appears when cooler is off, so it does not seem to be the reason...
Do you also have some gradients in your 383 bias frames?

[chrisjbaileyuk]

Looks just like one of mine. Perfectly normal.

Chris

ps that is a higly stretched image. If you read the actually values they dont vary by very much!

[scirocco1]

it's not "perfectly normal". better electronics like in a FLI/SBIG/etc wont show you this kind of pattern.

[Chris]

I don't think it's possible to say much about this because we have no information about the actual pixel values.  The difference may be significant or not but a stretched jpg image doesn't tell us.

I've a radical suggestion - contact Atik support.  It's just possible that they know more about the camera they design and build than random, anonymous, internet posters.

But my guess is that this is something to do with reading the data.  It varies across the CCD because different parts of the CCD take longer or more operations to be read out.  One way to check would be to compare dark frames with different exposures, for example these bias frames with 10 second darks - at the same temperature. If you subtract them what's left should be dark current without this fixed pattern noise.  That should be more uniform although it will still have random noise of course.

I would expect that if the CCD was not being cooled equally the gradient would be less uniform.

Chris

[scirocco1]

oh but we do have those informations, we own these cameras, and we have lots of calibration (bias, dark, flat) frames.

For example, with a 15 pixel medium sample, I have the highest value in a bias field of around 360 ADU (yes, in the bottom left part of the frame), and the lowest around 330.  30 ADU might not seem a lot until you do some narrow band imaging and your nebula is just a few ADUs above the background noise.

matter of fact, I couldn't find any gradients in some friend's cameras with the same chip (QHY/FLI), though a fixed grid pattern does seem to exist.

[NickK]

Fastest way to the bottom of it - contact the support email When I stepped on a cable that ripped out the USB socket, my 383L was turned around in a day.

I can't comment on other cameras I've not experienced but google shows other makes with 8300 sensors do have gradients. The KAF8300 sensor is a full frame sensor so it's read row by row, by transferring the charge down. The last row sees every row of charge transferred over it. This makes it prone to a gradient (regardless of which vendor).

I've not found a problem unless the image is stretched heavily - I'm usually picking galaxies out of the background noise caused by the light pollution..

[Chris]

oh but we do have those informations, we own these cameras, and we have lots of calibration (bias, dark, flat) frames.

We do NOT have this information for the jpg image posted. We do not know how much it has been stretched.  This means that we do not know anything about the magnitude of the gradient it shows.

Why are people so reluctant to take this sort of question to the manufacturer's support people?   I see it all the time where people would rather ask random anonymous people on the web than contact the people who made something, who know far more about it, and whose responsibility it is to help.

Chris

[chrisjbaileyuk]

it's not "perfectly normal". better electronics like in a FLI/SBIG/etc wont show you this kind of pattern.

The SBIG8300 I had for a couple of weeks last year did. Also had a couple of column defects...