Greetings and Salutations!
I presented this cooling strategy (of mine) in a Gamer's Over-clocking forum. They're final word was don't come back until you've got some data.
The idea was as follows; employ a can of circuit chiller (spray can) with a thermostatically controlled nozzle .. the exact cooling protocol is yet to be determined .. suffice to say when the action heats up .. the can delivers a drizzle of super cooling enviro-benign substance to where the action is. They're objections were legion .. primarily centered around thermal roller-coastering.
Let's examine the potential utiliity for camera cooling purposes. Allow me to suggest I'm not in the cooling industry and consequently have no hidden subversive agenda in promoting this particular approach.
In 1000 words or less here's the 'protocol'. Accessing the back of the sensor we apply a silicon grease layer I suppose it could be silicon glue as well. This substance displays reasonable thermal conductivity and more importantly is electrically inert. Don't use heat-sink compound which although having greater thermal conductivity contains zinc-oxide which has marked electrical conductivity. Electrical and thermal conductivity normally go hand in hand. So silicone grease/glue (RTV) is probably the best choice .. My suggestion is to validate all this on a less-expensive web camera first.
So, next we apply a cut out of a cpu heat exchanger .. a corner cut of the older 386 style cpu finned aluminum extrusions .. this we mount with the base flush against the silicon coupling .. probably a reasonable idea is to first apply two thin coats of silicon glue (as an electrically insulative barrier) - the heat exchanger (being made of aluminum) is very much an electrical conductor. So then we apply a layer of silicon grease and press our heat exchanger (thermal mass) against this grease. Ultimately we are going to use a moderatly fine tube to introduce the coolant in-between these fins.
My 1/2 CCD sensor is mounted away from the board (spider-like) .. it's a consideration to bore an as-large-as-permissible hole in some yet-to-be discussed fashion through the center of the sensor mount facilitating direct thermal conductance - unhindered by thermal properties of the fibre-glass circuit board - which invariably will conduct heat edge-wise into the cooling zone.
Now we spray some iso-cryonite expanding foam in a can insulation onto some scrap cardboard and use a stick to judiciously insulate the sides principally. Think in 3-D to insulate in such fashion as to discourage heat flow to the rear and the sensor and encourage heat extraction. Applying a clear perspex thickness to the front sensor-surface may not present a particular optical hindrance.
Circuit Chiller does it's job through evaporative cooling .. so it needs a little bit of space to work it's wonders. We apply it via the feed tube to the base of the finned region .. it cools the beegees out of the base of the heat exchanger. Just keep a fine 10% throttle-setting going on the can of relatively inexpensive and residue free cooler .. the can should sit upright.
If frost build-up is a problem .. keep the region positively vented with air as trickle-fed from a scuba tank. Air as found in scuba tanks is very dry .. interestingly divers suffer from dry throat as the compressed air is very much moisture-free. Perhaps a can of Dust-Off will serve this purpose as well. Note ... it's dangerous to bleed air in a single-stage fashion off a scuba tank .. requires double-staging .. be advised. Bigger experts than me re: safe low-volume bleeding of high pressure, high possible danger - compressed air (nitrogen etc) tanks.
Hope this all seems somewhat reasonable .. possible backyard home-mechanics conversion from surplus bank-style ccd cameras to high-performance, zero dark-current - astro cam. Whew : )
Terri in Toronto
anyway, instead of water, some other liquid may have been used to go below zero...
