I'm concerned that something is wrong with my camera. I uploaded a master sky flat taken with cooling to -35. See ftp.cyanogen.com/incoming/MasterFlatCliffK.fit. The pattern is pretty obvious at screen stretch "Medium". My concern is that this pattern is not being completely removed by calibration and is finding its way into my science images. Can you give me your opinion? best regards, Cliff
I have uploaded a master bias (and a 3 sec dark for good measure) taken on the same night at the same temperature as the flat. I do not see the dip in the bias. The major deviation I see in the bias is a fall off in the left most 500 pixels of about 8 ADU. This fall off is gradual and goes top to bottom. It has nothing like the single pixel wide matrix I see in the flat. The master dark, having been bias adjusted has a fairly even salt and pepper shape with the odd hot pixel. I'd welcome any suggestions and would be happy to collect additional data to aid diagnosis. I suppose it could be an alien mega-structure in the sky above my observatory that evening, but I suspect it has a more down to earth cause. best regards, Cliff
Doug, The shim placement document that came with the camera is dated March 4, 2016 and shows serial number X16010007. My camera is in my observatory and it will be a day or two before I go there. Let me know if I need to check this against the physical camera. best regards, Cliff
Unfortunately I don't think we have an image saved from your camera. (We do that routinely now.) It may have to go in for service.
So we've done a lot of investigation on this. We tried adjusting your camera's clock voltages and timing, changing bias levels, all sorts of things and nothing changed. We just built a batch of 6303 cameras, so we have a bunch of test subjects. Also we have a vintage 2000 camera (not SBIG) with a 6303 sensor. We also located an old (slightly damaged) sensor we had in engineering inventory. ALL of them exhibit faint lines like yours. There is one difference between the new sensors and old sensors. On old sensors the pattern is a little bit random looking, but it doesn't move around from image to image. On the new sensors the lines are always exactly every 36 rows, but they're not in the same location on different sensors. The clincher was when we swapped sensors between two of the new cameras. The position of the lines moved with the sensors. I have taken many 10s of thousands of images over the years with the old camera (it was used for Desktop Universe, among other things), and have lots flat-field frames. Proper calibration renders the lines undetectable. I have an inquiry into ON Semi about the lines, but I suspect that won't get us anywhere. These lines appear to be characteristic of KAF-6303 sensors.
Wow! I didn’t realize this would be so involved. I was digging into this because I saw the horizontal lines in some calibrated images. I am traveling until Monday and don’t have access to images so I cannot dig up an example until then. I suppose another way to approach it would be to calibrate some of the flats that go into a master flat? Anyway, I’ll wait until I hear what your inquiry turns up. Thanks for keeping me informed. Cliff
At this point we've reached the end of the investigation, barring any substantive response from ON Semi. I honestly don't hold out a lot of hope for that. The fact that a 17 year old sensor shows the same behavior suggests this isn't going to change. Also that old sensor was in a camera of completely different design, so it's definitely not something we are doing.
Yes, I took that as what you were saying. So give me until Tuesday to look through my images. Hopefully you will have a response by then, if only unsubstantial. We can figure what to do then. Best regards, Cliff
Doug, I think we should take further discussion out of the forum. I sent an email to Tim and Bill and asked them to forward it to you. Cliff