[avandonk]

I built a Peltier fridge for my Canon 5DH some time ago. I can control its temperature to stay constant within 0.1C. So the camera once it is equilibrated takes darks reproducibly.

One thing I soon noticed how constant the JPG file size out of the camera was under the same conditions i.e. temperature and exposure. This file size was quite sensitive to temperature differences for the same exposure settings.

The second image below is twenty of two minute darks at constant temperature and you can see how constant the file size is especially toward the later exposures as the system equilibrates further.

What I am proposing for a given exposure dark, the file size out of the camera will give you a quite accurate match for sensor temperature.

It then is a simple matter to take a dark before and after an imaging run and use the file size to match with your dark library to get a much better match of sensor temperature.

In fact when obtaining your dark library you can monitor the temperature of the sensor until it reaches equilibrium simply by looking at the JPG file size.

It takes about forty minutes of sequential exposures before your sensor has equilibrated. My 5D's sensor is actually 17C above camera ambient when doing this. I have measured this.

Before I take light images I do at least forty minutes of dummy exposures to equilibrate the system. Obviously you should do the same for a set for a dark library.

I am sure this is why people have trouble matching darks with lights even at the same ambient temperature as your sensor is heating up slowly with constant sequential exposures until equilibrium is reached.

I have far more data so any questions are welcome.

I set the camera taking 8m minute darks continuously at 1600 ISO and let the camera's sensor temperature equilibrate at various fridge temperatures. Below is the results.

It looks like that file size variation with temperature is near enough to linear for both JPG and RAW frames.

Bert