Roger N Clark said:
You cite the Sony ICX sensors in the 0.003 raneg at -10C,; the 7D2 reaches the 0.002 range at -2 C.
That is what the spec usually say on cooled CCD cameras from Atik, QSI, FLI. However, there is anarticle from Craig Stark where he tests an Atik 314L+, and experiences considerably lower dark current of 0.00076e-/s/px:
http://www.stark-labs.com/craig/articles/assets/CCD_SNR3.pdf
At this point, I haven't slept since Monday, so I am not making snense of what he said after that, or why he was measuring 0.0005e-/s/px. Regardles, these are the lowest dark current noise levels I've ever heard of.
The 314L+ uses the Sony ICX285AL, and it has a limited maximum cooling of -27C below ambient. The standard cooling temp of Sony ICX CCDs is usually -10 however unless the ambient temperature was ~62F/17C, I on't even think the temperature of the snesor during his testing would have been that cold...at 65F ambient the sensor would reach it's minimum temperature of around -8C, and at 70 F ambient the sensor would reach a minimum temperature of -6C..
anyway...whatever I was sayind.....I don't know if the newer ICX674/694/834 sensors can reach that level of dark current...but regardless, when cooled to -10C, the dark current noise is a trivial factor for most common exposure time I think for an hour long exposure, which you shouldn;t need outside of narrow band oiii filter exposures on certain targets, dark current noise barely tops 3e- at a rate of 0.003e-/s/px. So none of the apers who use the Sony sensors ever bother with any dark frames...many don;t even bother to dither I don't think.
KAF sensors (Now owned by ON semiconductor? Once Truesense, previously owned by Kodak) are definitely not as good. Its really old technology...fdamentaly I dont think it has changed in a very long time, a decade?...it's had minimal improvements, addition of microlenses, maybe a switch to AR glass and really small things like that...but it;s really old tech. So it is not surprising that it does nto have low dark current, and they don't have low read noise either (at best, around 7e- or so, most of the time it;s 10e- or higher) . The reason people still buy them is the sensors can be huge. The KAF-8300 is slightly smaller than APS-C. The KAF-11002 is FF size. The KAF-16803 is a monster 37x37mm square sensor. FLI just released a new camera with the KAF-50100, wich is a 49x37mm 50.1mp behemoth sensor. FLI and On Semiconductor apparently took the old 50100 sensor design, which used to have a mere 25% Q.E., added microlenses and apparently got the Q.E up to around 62%. That puppy is over $20k...so not exactly your hobby camera/
Anyway...yeah, KAF sensors are very old...old tech, but they are monsters. You could fit 30 of the Atik 314L+ sensors into the area of a single KAF-50100.

So, maybe not as efficient andg enerally not as goo dfor narrow band imaging...but they make the most beautiful LRGB images tou will ever see.
Ok, sorry...I think that was really sloppy writing, but I am too tired to go back and fix everything... :\
Roger N Clark said:
I use several programs for my analyses. DCRAW and RAWDIGGER for decoding the raw data without changing values. then my own custom code written mostly in Davinci from asu.edu (developed for spacecraft mission imaging data analysis).
I know of RAWDIGGEr, anf I think IRIS uses DCRAW. I was figuring I'd use iris to examin the raw data anyway.