Some of the details are here as well. I just thought I would put up this preliminary data as I was shocked how much faint stuff now appears and it is all due to the fridge. The faint stuff is at or lower than the noise in a Canon sensor at 20C. So no matter how many subs you collect at 20C there still will be lots of 'holes' in the data due to the noise subtraction.
Thanks all for the comments. There are three things that are working here. There could be more?
1. Camera in an environment of -10.0 C so less thermal noise .
2. The sensor temperature does not vary by more than 0.3 C so darks and lights are perfectly matched.
3. All corrections in ImagesPlus for darks, bias and flats done in 16 bit fits so all corrections are at the original pixel level not after interpolation. It is only after this is done that I convert to 16 bit tiffs by Bayer interpolation for subsequent stacking in RegiStar.
The other trick is to choose a starting frame for RegiStar somewhere in the middle of the set to be stacked. I then use IP (Local_ Smoothing and Noise Reduction_Multiresolution Smooth/Sharpen... )to get rid of any residual noise as this can affect alignment especially with short exposures. Of course this starting frame is not used for combination.
I thought I would just add that what I mean by original pixel refers to the real sensor elements that are mapped one to one to the fits image.
Beautiful, inspirational work Bert, as was your LMC mosaic.
Very nice
Do you have any condensation issues or other problems with the camera body in the fridge?
Doug
Thanks Doug. The trick is to turn the camera on before starting to cool the fridge. I use an Canon AC power supply for the camera so batteries losing voltage at low temperaturs is not a problem. Since the sensor and camera even at idle uses about two watts of power it will always be warmer than its environment. When taking long exposures I have calculated that the camera then dissipates about four watts and the sensor temperature then rises at least two degrees C from idle. What this means the camera and the sensor are the warmest object in the fridge so any condensation occurrs on the Peltier heat exchangers. So in a short version I have never had condensation problems. The other thing is there is a filter and/or a focal reducer between the sensor and the ambient warm moist atmosphere. I have carefully worked out all the temperature gradients between cold and hot so that the sensor stays within 0.3C when the camera is collecting data. And that there is no condensation on filters/focal reducer. It is a bit like aerobatics it is quite safe as long as you miss the ground (condensation).
Thanks for the explanation Bert.
In a rough 'n ready, foolhardy attempt to cool my 40D I noted that some condensation formed IN the LCD itself - frightened the crap out of me so I've never re-attempted!
Wondered if you had encountered similar?
Doug
Doug it is all about temperature differentials. Condensation only occurs if there is moisture present. The trick is to get it to condense on something robust. There must have been a higher moisture content inside the LCD screen before you started due to a previous hot moist environment the camera was exposed to. Melbourne fortunately is rather dry.
The faint stuff is at or lower than the noise in a Canon sensor at 20C. So no matter how many subs you collect at 20C there still will be lots of 'holes' in the data due to the noise subtraction.
Bert
Excellent work there, that explains why I get a lot of noise when I try going that deep. Must get a lend of Houghy's cooled camera again
Scott
Amazing depth. I probably missed something here about your cooling technique but how do you maintain such a tight control on temp. for long exposures after you take the camera out of the fridge?
nice pics 
Great stuff Bert. The field is amazing. 
