John, that is actually a pretty big topic on its own and worthy of a separate thread. I can't be sure what the actual S/N ratio is at this point. What can say is that individual subs shot at ISO 800 are showing no noise grain that I can detect, even when I enlarge the sub. Sure I still have the odd hot pixel that was always there.
Craig Stark has done alot of research on the signal noise in Canon DSLRs and comparative performance with CCDs. Here are links to two of his papers that are great reading on this subject:
The first deals with profiling long exposure performance.
http://www.stark-labs.com/craig/reso...nLinearity.pdf
The second looks at it compared to CCDs.
http://www.stark-labs.com/craig/reso...RvsCCD_API.pdf
Unfortunately my New Moon dark site trip was rained out this month. I had intended shooting a range of target objects at different ISO levels and with and without cooling running. Now that is put back.
In my few test shots to date, I decided not bothering with darks but have limited shots to share because of this weather holding me back. I have a couple in my public gallery on Astrobin but they have to be converted to jpeg to go on there.
http://www.astrobin.com/full/171762/0/
Now as to what I used to do. I always shot at ISO 800 simply because I had past research that indicated that it minimised noise (obviously compared to 1600), and was the sweet spot for an uncooled camera. My subs lengths, generally ran from around 200" to 320", depending on the brightness of the target, whether I was on my pier or the tripod, etc.
I still have the 300" darks from my first test run with the cooled camera, and can have another look at them but am not really equipped to do any comparative analysis other than using my eyes.