At the moment I am dabbling in photometry - climbing the light(learning) curve. One point that its constantly stressed in the literature is the need to calibrate with flat frames.
I have been particularly unsuccesful in my efforts to date.
To give you an idea of my issue, I've attached 3 images. The first shows an uncalibrated image taken with Maxim DL. The vignetting is reasonably apparent.
The second is a copy of the flat frame I took using a t shirt at twilight on Friday.
The third is the first image calibrated in Maxim DL. As you can see, something has gone awry in the processing.
Now the flat image looks quite bright. However, when I open it in ImagesPlus and review the histogram, the curve peaks quite nicely around the 30% level. I suspect that Maxim auto-stretches the image for display.
My suspicion is that the auto-stretch is also applied when calibrating the image so that rather than removing the vignetting, the program overcompensates.
My next thought is that there is some serious reflection going on inside my scope or the extension tube of the focuser. I am using a fast schmidt-newt - f4- so I have a MPCC in front of the nosepiece to avoid coma. I am using a SBIG ST-8300 which has a big sensor.
If you look carefully at the flat, you will see that the bright patch at the centre is a narrow elipse extending almost to the top and bottom of the frame but not to the side. I suspect internal reflections are to blame.
But if that's the case, why does it only show up in the flats and not on the CCD?
Or perhaps my images are still overexposed.
I am most puzzled. Any tips would be gratefully received.
Taking flats is quite tricky until you work it out, then it is quite easy. I tried to use T shirt flats too, but had a real hard time getting them right. I ended up I bought a light box from Exfso which gave me a very stable light source. As I use a QHY9, which has the same chip as the 8300, you might have to use a potentiometer in the power supply to dim the light box off a bit to get the right length flats, etc.
In MaxIm, I take a single exposure and set screen stretch to "range" so it is an unclipped display. Using the "info" dialouge box, I experiment with the exposure time until I get between 20 and 30000 in the centre of the image on the brightness scale at the left of the info box information. If that looks ok, I save it and open it in Phototshop, adjust the black slider until it is nearly on the left hand side of the histogram and make sure it looks nice and gray and smooth in transition from centre to edges.
I also believe that you will find a "how to" here as well, if that is any help.
I can do some screen shots of how I do it, but I don't think I could get that done today. I'm also thinking that if you search the forum, this has been discussed before and should provide some good info as well.
i too had a bit of issues taking flats with maxim untill i figured one thing out...
That is making sure you know what the dynamic range of your CCD is, if its 16bit, youll have 65000 adu approximately. Now i use the 14bit 40d, which is considerably less than 65000, so i was aiming for about 20-25k adu.... wrong!
So follow what decepti said, but just bare in mind also about your dynamic range.
Oh for a really cheap and good quality light box i built one out of a Broccoli box, 2 Cold cathode 300mm fluros, a piece of white foam board from office works and a bit of duct tape. all up set me back $35 bucks and i never looked back.
Since you are using maxim, try ccd autopilot for flats (try the 1 month trial).
I also use the tshirt/sky flat method.
In ccdap you can simply give it a target range in adu and press take flats. It will adjust to make the exposure time right for the current sky brightness. I get perfect flats every time.
i had a look at ap, but it wanted some programs like pinpoint ect ect. so it was a waste of time. unless i had the coin to get the program i was out of luck
Brendan, Not so. Pinpoint is required if you want ccdap to do plate solves to check for slew errors, and for calibrating your guiding. It is an option, not a requirement. But it is very handy.
Save your money on CCDAP. Whilst a good tool for automation, if all you want to do is flats download the free MaximDL SkyFlatAssistant plug-in - http://winfij.homeip.net/maximdl/skyflats.html
Don't be fooled by the name. You don't need to use the tool exclusively for sky flats. It can also be used for tshirt or lightbox flats. Similarly to CCDAP it will measure the target ADU to ensure exposure lengths are optimal based on light conditions. You set the target ADU at the start of the run. Clearly, the exposure will vary even if the light flux remains the same as astronomic filters have different bandpasses. Narrowband filters will require a longer exposure time to reach the target ADU.
For lightbox flats and similar the light conditions don't change, thus exposure times would remain fairly constant (for a single filter that is). Sky flats however are a different story given the light alters dramatically (and rapidly) during the small dawn and dusk time windows, hence the need for measuring the target ADU and dynamically adjusting the exposure time to compensate is key to success.
What CCDAP will however provide is the ability to shift the mount ever so slightly to move any stars that maybe (likely) to be picked up during a sky flat acquisition run. As the stars are placed on different pixels across the CCD array, they are easily recognised as outlier pixels and rejected via the conventional data rejection algorithms.
So, you'll end up with accurate flats...but at different exposure times... How the hell do you calibrated the data then... easy... scale your dark frames to compensate. Take a series of 10min darks, then bias frames. Produce masters of both, then use MaximDL or CCDStack to subtract the bias from the dark to give you a thermal frame that can be scaled to match the duration of the flat frame.
Personally, I don't use CCDAP anymore. Good tool, but what ACP delivers is far more flexible in automating sky flats, plus it has a web interface. Anyway, I'll keep on topic as comparisons may lead to greater discussions (different thread).
Brett, I think I've contributed my opinionated views on the two products here before. A search may reveal an insight.
For those wanting additional reading material on sky flats (don't be afraid of them as they're not as complex as you think), you can download a white paper on the topic.
Having just started using Maxim DL, I still have the trainer wheels on.
When I have a chance to do some processing, I'll give it all a whirl. The SkFlatAssistant plug-in sounds promising as my main issue is getting the exposure length right.
I'm off to buy a lightbox (figuratively speaking) from Exsfo now.
James, the screen stretch function is just that - a screen stretch. It is by no means the best method in judging the accuracy of a flat. FITS files are numerical. When processing data, it helps to understand the figures.
In MaximDL, on the status bar to the lower right you'll find an item 'I: xx,xxx'. Where ever the mouse pointer hovers on the image, the 'I' value (Intensity) displays the pixel's numerical value. Towards the center of the flat you should be hitting the target ADU. +/- 1,000 isn't an issue btw, doesn't need to be dead on.
If you want to get truly analytical on the flats, use MaximDL's line tool and draw a diagonal line across the frame. This will present a graph showing values along the line. There are a few other ways to make an assessment.
Save your money on CCDAP. Whilst a good tool for automation, if all you want to do is flats download the free MaximDL SkyFlatAssistant plug-in - http://winfij.homeip.net/maximdl/skyflats.html
Don't be fooled by the name. You don't need to use the tool exclusively for sky flats. It can also be used for tshirt or lightbox flats. Similarly to CCDAP it will measure the target ADU to ensure exposure lengths are optimal based on light conditions. You set the target ADU at the start of the run. Clearly, the exposure will vary even if the light flux remains the same as astronomic filters have different bandpasses. Narrowband filters will require a longer exposure time to reach the target ADU.
For lightbox flats and similar the light conditions don't change, thus exposure times would remain fairly constant (for a single filter that is). Sky flats however are a different story given the light alters dramatically (and rapidly) during the small dawn and dusk time windows, hence the need for measuring the target ADU and dynamically adjusting the exposure time to compensate is key to success.
What CCDAP will however provide is the ability to shift the mount ever so slightly to move any stars that maybe (likely) to be picked up during a sky flat acquisition run. As the stars are placed on different pixels across the CCD array, they are easily recognised as outlier pixels and rejected via the conventional data rejection algorithms.
So, you'll end up with accurate flats...but at different exposure times... How the hell do you calibrated the data then... easy... scale your dark frames to compensate. Take a series of 10min darks, then bias frames. Produce masters of both, then use MaximDL or CCDStack to subtract the bias from the dark to give you a thermal frame that can be scaled to match the duration of the flat frame.
Personally, I don't use CCDAP anymore. Good tool, but what ACP delivers is far more flexible in automating sky flats, plus it has a web interface. Anyway, I'll keep on topic as comparisons may lead to greater discussions (different thread).
Jase
thanks for this. I've downloaded and installed the plug-in.
Now for the dumb question - how do I drive it? I can't find anything on the web that tells me how to initiate it.
I am sure there is an easy answer - just dunno what it is.
Since you are using maxim, try ccd autopilot for flats (try the 1 month trial).
I also use the tshirt/sky flat method.
In ccdap you can simply give it a target range in adu and press take flats. It will adjust to make the exposure time right for the current sky brightness. I get perfect flats every time.
Brett
Brett
some tips would be handy here.
I've downloaded CCD autopilot and gave it a run with my lightbox. t took 2 flats - 7.6 sec and 3.4 sec and then decided it was too bright and refused to play ball any more.
Am I missing something? I assumed that it would run until it hit the target ADU setting.
I have tried to run the Sky Flat Assistant plug in in Maxim DL. All I have is a greyed out menu item in the Plug Ins listing. It won't run or even make itself available.
Am I missing something?
I am taking somequite nice shots but have a devil of a time getting flats to adjust for colour imbalance across the field. Flats with my lightbox that match the target ADU still give me an unusually light strip acros my images. You can see this with the flats I've included in my first post.
You need to take one image manually and have it open. Then the menu becomes active. It's a very cool little tool :-)
James
Thanks James - thats good progress.
Now I have an error message that says MSSTDFML.dll is missing. A quick google search yields lots of sites urging me to do a system scan with their software - which gives me the willies.
In yr original post, you had stars in the flat, thats not good at all, thats going to mess with the stars in the light (probably eliminate them, as in yr final image). Do you still have stars in yr flats?.
