Thanks for those links. The paper by the designers is so much better than the incredibly dumbed down fluff in the article.

As I read the article I kept thinking things like "what Canon lens?", and "what are the details of the coating?". The paper has a wealth of technical information on all of this.

The paper mentions "science grade CCD sensors" - are these really better than some of the new ultra low light CMOS cameras from Sony and Canon?

The sensors used are Kodak KAF8300, which are a stand-alone sensor that you hook up to a laptop. They are designed specifically for astronomy, and perform better in some ways than typical consumer cameras. DSLRs that are marketed as tuned for astrophotography often have better sensitivity to certain key wavelengths than their standard counterparts (especially hydrogen-alpha at 656nm), but are missing other benefits of 'science-grade sensors'.

"Astronomical CCD cameras are a whole different world from DSLR cameras. These are cameras designed exclusively for AP, and most offer active cooling. The nature of a CCD vs a CMOS sensor in a DSLR results in a camera that is higher sensitivity, but also higher noise. Many Astro CCD cameras have mono sensors (no bayer color layer) which means they are used with filter wheels with different colored filters to capture a natural color image. They can also be used with various narrow-band filters to capture light outside the visible spectrum."

from http://soggyastronomer.com/the-best-dslr-cameras-for-astroph...

Nice. Did a bit of reading on this, looks like high as possible quantum efficiency (across a broad spectrum) is what is desired.

You are right, most DSLR cameras lose about half of the QE to the rgb filters.

More than likely. Most "science" grade CCD sensors, at least in the Astronomy world, are in fact shaved down to make them even more sensitive.

My understanding is that "science grade CCD sensors" have thermal controls to ensure the thermal noise is constant.

Footnote 20 is wonderful.