Hello,
I've exploited some CCD and several CMOS sensors for years. The image is quite clean in a CCD and atrocious in a CMOS. You already know the differences. What we are learning since, say, last summer, is stacking thousands of frames with short exposures to get the best of CMOS, at a fraction of the price of a CCD.
This is limited to contrasted DSO, but, on the other hand, stacking so much short exposures results in high-resolution images of bright DSO. This is why deep-sky imaging with a CMOS sensor is a completely new technique, something standing between planetary imaging and DSO imaging. In the french review Astronomie Magazine, a superimposition of CMOS and CCD images clearly proved the two techniques to be complementary, not ennemies.
About cooling: Yes, it is by far less efficient than with CCDs. This works, but the result is not perfect at all. I added a non-regulated TEC to my ASI120MM with encouraging improvements. I don't know if the temperature was measured IN or BELOW or BESIDE the sensor but the noise decreased as the cooling lowered the temperature, undoubtly.
Then I purchased a cooled 178MM to ensure temperature stability rather than to obtain (or vainly hope) clean images. I was tired of taking new darks and series of frames every 20mn when the ambiant temperature changed during the night. The idea is, to me, to comfortably exploit a ASI CMOS like an ATIK CCD, with a library of darks. Of course the results won't be the same. But they will be complementary.
CCD = DSO (planets only with the ICX618)
CMOS = bright, contrasted DSO + Moon + planets + meteors + ultra-fast imaging
The Laboratoire d'optique of the Université de Nantes has just bought a CCD camera so study the surface of materials. The sensor is very large and extremely reliable for its repetitiveness, with no need for bias correction, but it has a very low FPS capacity. Clean or fast. This is a choice to do. The best is having a CCD plus a CMOS.
Nicolas
