@microfunguy
Quote "'my bad'"
I always felt this may have originally come from "Am I bad" (statement form, not question) with the rhetorical roots of something like "I admit to inexcusable ignorance," and when said quickly sounds like "my bad."
Quote "Zero?"
Close to it // I am certainly hoping for not zero. There are second-rate solutions that are possible, but those may impose thermo-mechanical and mounting constraints.
Quote "Sounds like astronomical spectroscopy. "
Nope // but it is possible to do if you work out the diffraction grating solution and associated optics (curved projection mirror) & mechanics. However, I would not recommend an imaging CCD in this case because it would be an overkill and not worth the effort in my opinion. You're better off with an easier and better solution: coupling a commercial spectrometer directly to your telescope.
@waterwingz
Quote "would it be possible to disassemble the lens assembly, remove the glass, insert appropriate spacers and reassemble ?"
Yes this is one of the 2nd-rate solutions. The problem is that it leaves the housing in place which limits mounting possibilities and the construction of a thermal cavity. One other solution is to remove the CCD and leave the lens in place. That depends on the mechanical architecture of each individual camera. Then to optically access the imaging surface, you have to rotate the sensor so it faces backwards. Because you cannot extend the ribbon cable (electrical signal limitations more than mechanical) there is little space left to put thermal components, in addition to having to leave the LCD screen available for most user applications. On the S90 it is possible to do, but rotating the LCD out of the way at leat to 90 deg would be required (or rotated out to 180 deg, facing forward) or better, disconnecting it. I don't know what affect that could have on error reporting. LCD rotation would affect mounting possibilities too. For example in my case I have a very tight space the hardware solution has to go into, so the LCD cannot be rotated and the CCD has to face forward. As reyalp put it nicely .. terra incognita ... all these options have to be thoroughly explored. You can see my thrust ... if you guys can figure out how to handle the error tables to override lens error reporting (a super-hack of sorts), that would be a really elegant generic solution.