Lights and Flats - different temperature

[Luigi]

I acquired lights, bias, and dark frames at -15C. I am now wanting flat frames, but am having difficulty getting my camera (ST-10XME) cold enough as temperatures are in the 32C range during the day (90F+ in old money).

I am not convinced of using sky flats (tried some of those but some still contain stars). I am in the process of constructing a light box, but until that is complete I am going to try the t-shirt method.

If I can't get my camera cold enough, should I just take dark and bias frames at the warmer temperature and create a master flat from that?

I've read Vicent's article and if I'm understanding PI can scale for differing exposure times, but not temperature.

[georg.viehoever]

Hi,

since flats (especially using T-shirt method or EL screen) are typically taken with short exposure time (<1sec), it is my understanding that temperature is not really important for flats.

Georg

[Cleon_Wells]

Louis, adding to what Georg said  on the flats, I’ve taken flats during the day on an overcast sky. The last set of flats I took were taken at 1/800 and iso400, I also rotate the OTA, taking half the flats with the OTA rotated 180 Degrees and in the same part of the sky.  I remember that Vicent has said that the flats need to have  enough level, my Master un Calibrated Flat has an average DeBayer level of  around 1/10 of the clip level. The clip level  of the  raw Canon 14 bit  file is 0.25.
Cleon

[Juan Conejero]

PI can scale for differing exposure times, but not temperature.

Just to clarify, our ImageCalibration tool can scale (optimize) master dark frames for both differing exposure and temperature. Our optimization routine does not depend on any physical acquisition property; it is a purely numerical solution.

[Cleon_Wells]

Juan, on the flat acquisition, I rotate the OTA 180 degrees and take the second half of my sky flats to reduce any small gradients in the overcast sky.
Cleon

[Luigi]

Just to clarify, our ImageCalibration tool can scale (optimize) master dark frames for both differing exposure and temperature. Our optimization routine does not depend on any physical acquisition property; it is a purely numerical solution.

Thanks, Juan. I did not realize that.

Is physical acquisition preferable to a numerical solution?

[Nocturnal]

Hi,

flats are calibrated separately from lights. In other words the darks you shoot for your flats need to match your flats, not your lights. So you can shoot flats at a different CCD temperature than your lights. Scaling darks assumes linear responses to both temperature and duration changes. I would verify this to be true for my CCD before counting on that during calibration. For the best results you should probably have your darks within 5C of your lights or flats. Naturally this is up to you. Experiment and see what works. Of course with a light box you can shoot your flats with the same ambient and CCD temps as your lights.

[Juan Conejero]

Scaling darks assumes linear responses to both temperature and duration changes.

Not in our implementation. Our dark frame optimization routine does not make any assumption of linearity. As I noted above our routine is a purely numerical solution. It will always converge to a solution that minimizes noise induced by subtraction of the master dark frame.

Nonlinearity of sensor response has a practical consequence, though. If the computed dark scaling factor is very different from one (say 0.5 for example), hot pixels may become undercorrected after master dark subtraction. In other (rare) cases hot pixels may be overcorrected, resulting in 'dark holes'. These are very minor issues however, as these hot (or cold) pixels will normally be rejected during integration without problems.

We never take dark flat frames, as we think they are not necessary. For detailed information please refer to this tutorial on our image calibration tools.