Does anyone have some references or such for imaging with an AC4040 and processing images from an AC4040 (with filters)? Transferred from an STL6303E CCD which I have used for years, and got a lot of fine images. But the new camera is a different sort of beast. I know processing for the STL...the new CMOS chip is quite a bit different. Any direction would be appreciated. Working in a Bortle 1.5 sky with a 17" scope. Thanks
Yeah it's a different beast. Don't use bias frames, and don't use exposure scaling. Take darks exactly matching your lights, and flat-darks exactly matching your flats. Use a large dither amplitude to remove any residual fixed pattern. Meanwhile I'm tinkering with some new algorithms... I think there's some opportunity to improve the residual fixed pattern.
Flat darks exactly matching flats are impossible with ACP's flats algorithms. There has to be a better way. Exposure scaling, at least within a limited range, has to be in the future. Taking darks exactly matching lights is 100% impossible when doing photometry in varied filters, of varied objects, per night. There has to be a better way.
You might be able to get away with scaling within certain exposure ranges. You will have to test it for yourself. I'm not recommending it, and here are my reasons: Most CMOS sensors change operating modes at different exposure durations. Even worse, internal sensor heat generation changes radically because the sensors generate a LOT more heat when they are reading out. So the rate at which you are taking exposures significantly impacts the observed dark current, and that variation typically has structure to it (much hotter top/bottom of the array). The amount of variation and its patterns of behavior differs significantly between sensor models. If you are scaling for exposures between, say 10 and 60 seconds, you could potentially take a set of 10 second darks and manually assign them to a bias group (MaxIm DL lets you do that - it just won't do it automatically). As long as they're taken at the same temperature as the dark frames you may be able to scale them. For exposures under 10 seconds you could try doing the same thing with 1 second "bias frames". For sub-second exposures you might be able to use an actual bias frame. A final word of warning - the behavior of CMOS sensors vary wildly from one sensor model to another. In fact, one of the biggest differences we've encountered is between the FSI and BSI versions of the GSENSE4040. These differences are often complex and can be difficult to explain in layman terms. We do not document it, and please don't ask, because that would make a lot of work for me. And I don't think it's a good idea. So... exposure scaling might work within limited ranges, if done with care and attention, but you will need to fully validate it. I am sticking to my recommendation to never, ever do it with CMOS sensors. I think it would be safer to generate a list of your flat-field exposure times and write a script to collect darks for each exposure time. It might be a pain if your acquisition software is automatically selecting different flat times, but that is IMHO the only fully reliable way to do it.
I agree that there may well be a lot of work involved. If I'm going to do proper photometry with my new scope and AC4040 FSI, it seems I'm already obligated to do that work. The current situation for photometry is untenable without some temporary workarounds until we have proper protocols. For flat darks (with a flat panel) of 2-10 seconds exposures taken with ACP, workarounds might be: (1) adjust the lamp intensity per filter so that exposures times all group near just a few times, and hope the more-or-less matching darks work. (2) just don't use darks (or use a bias, as you suggest) for the flats. ---This will have to be decided with data. (Woe to those trying to use a CMOS and panel for flats in narrow band filters.) For the photometry darks, I suppose there might be two workarounds to test and from which to choose one: (1) settle on (many) standard exposure times, say every 60 seconds from 60 to 900 seconds, and strictly match flats and dark exposure times. This might be adequate for my present asteroids. But woe to variable-star observers who will need 20-50 standard exposure times in a more-or-less exponential scale. Such could easily require 500 flat images and 500 matching dark images, the whole set retaken every few weeks or months. File handling alone would give an aspirin a headache. (2) take flat frames at (many) standard exposure times, and narrowly scale calibration of working frames between them. ---Agreed that proving that (2) works with confidence is unlikely to merit the effort. So I guess CMOS strait-jackets us on exposure times.
Well this is interesting. There was an issue Doug, some time ago where I tried to bin 2x2 and there was an issue with the driver or camera hardware where I was getting really interesting patterns (literally patterns) in stars and such. The advice at the time was to NOT bin. Has there been any progress regarding this issue? And, a couple of days ago I posted where ACP could not call on Maxim for an exposure. Bob Denny also commented on it here and offered that it seemed to be a Maxim issue, anything on this? 04:09:56 **Script Error (Tracking has been stopped)** 04:09:58 Source: MaxIm.CCDCamera 04:09:58 Message: **Failed to start the exposure 04:09:58 Location: line 1533 column 33.
There was an issue with StackPro and binning. I'll have to check with @Adam Robichaud whether that's been fixed. He's working remote today, and it's late in the day; hopefully he will be able to respond before the weekend.
