The reason I did this to this camera(DSI Pro) is that I could not get a good noise performance for exposures above 60 seconds out of the unit at temperatures above 5 deg c ambient.This is my second DSIpro camera and both of them were the same ie very noisy compared to the colour version.My first camera I sent back to the dealers as it had 9 dead pixels that were permanently set at full brightness with up to 24 in total of dodgy pixels.The second camera has only two which is within spec for off the shelf imaging chips.

Cooling This camera is not a difficult job if you have access to either a lathe or a milling machine.If you remove the four allen head screws that hold the camera together you can see why.The clam shell design splits into two halves with the back being the heat sink.The heat sink has a raised metal pillar with a machined top that when assembled fits in contact with the back of the imaging chip.Assuming there is white thermal paste on the pillar and on the chip itself then the chip is kept at the same temperature as the case itself.In my case I had available a peltier device that was 20mm square in size and four millimetres thick that was much larger than the 9.5 mm square pillar.To fit it to the back of the case it was necessary to remove the pillar. I used a lathe to do this.However you need to work out the height of the pillar as the length of the cooling system needs to end up at exactly the same height.This is difficult as the bottom of the box is very uneven.To get around this I mounted the the box onto a lathe chuck so that the center of the pillar was at the lathe center in an independant type of chuck not the self centering type of chuck.I then turned the box base down around the pillar to a diameter of 30 mm or so for a depth of about 1 mm.With a depth gauge I then measured from this now flat bottom to the top of the pillar which in my case was 14.5 mm.I then carried on turning out the pillar until it was removed and I had a circular flat area on which to mount the peltier device.What was needed now was an adapter to fit between the peltier device and the bottom of the imaging chip.For this I used a short length of 19 mm dia aluminium bar 7 mm long.This I turned down for most of its length to a diameter of 9.5 mm so that it would fit into the cavity where the original metal cooling pillar went and onto the back of the imaging chip and would also be long enough to clear the surface of the printed circuit board that the chip is mounted on.The last 1.5 mm being left at 19 mm dia so as to get a good contact with the peltier cooling device.This meant that to complete the length of 14.5 mm I needed a 4 mm thick spacer between the hot side of the peltier device and the case back.This was turned in the form of a circular disc that fitted the heatsink back and was slightly larger than the peltier device.The next problem was to get power connections for the peltier device that could carry up to 2 amps of current.For this I used 2mm sockets and plugs one black and one red and NOT connected to the DSI electrically.The 2mm sockets fitted into 6mm holes drilled in the side of the heat sink case back and were an exact fit.After wiring the peltier device to these with short lengths of wire I put heat sink paste onto the adapters and the peltier device and then checked the measurements to the top of the new cooling pillar.This proved out to be less than one tenth of a millimetre too long which I left as is.The case was then reassembled.The power supply for a peltier device can be very simple such as a variable voltage power supply with a power resistor in the output of it to limit the current and to also make it approximately constant current in practice.From experience I guessed that it would only need a couple of watts or less of cooling for most cases with this assembly.The power supply I made varied from zero to 8 volts with a 4 ohm resistor in its output.The voltage drop across the peltier device varies depending on what current is passed through it in practice getting higher with current.Testing on the bench showed a current between zero and two amps for cooling.In use it was only necessary to use 0.6 amps in most cases with an ambient temperature of 18 deg C before dewing of the chip surface occurred.You can only get around dewing of the chip surface if you isolate the chip optical surface into a windowed space that has had a sealed dry atmosphere put into it.Doing this can also allow you to take the chip below zero deg C. There is a plastic gasket between the imaging chip and the DSI case front that seals the box at that point.Temperature measurement can be simply as buying an off the shelf digital room temperature gauge and removing the internal temperature sensor from it and sticking it to the back of the thermal coupler.Two wires led out to a connector can get the temperature sensor connected back into the circuit of the gauge.These sort of things can be bought for a few pounds UK. Humans can only discern large changes in things visually without references we are not noise analysers or spectroscopes. The noise improvement is visible so it was worth the effort.

The circuit is as follows for the power supply....

To finish off I recovered a microprocessor cooling fan from an old scrap PC which was about 2 inches in diameter and ran off 12 volts DC.To mount this to the back of the DSI heatsink I made a simple "U" shaped clip on aluminium panel with the fan in the center.This was the width of the DSI case with half inch sides to clip on to the DSI heatsink after bending in carefully to fit.The swiss cheese effect is to help the airflow.

This is a typical image taken of the sunflower galaxy at full size.However what you see depends on the type of computer monitor you look at it with. You need either a hi res flat panel type or a "bottle" type of monitor to see it properly.Black and white exposures always look lean in viewable information compared to colour pictures.These are two minute exposures against a heavily light polluted sky as are all of my pictures whatever the kit. The camera running at close to dew point at 3 degrees C.In my case what is important to me is what magnitude I can resolve in the arms of these objects not the core.

Tri colour imaging with bright objects gives the following.. in the standard picture everyone takes of M57..

For me Tricolor imaging of dim objects is a real challenge due to the difference in exposures needed in the three colors as green is the worst. Without the cooling I have found it extremely difficult to get a working image especially in green for the three sets of exposures needed. One tricolor image is likely to need 100 exposures in total taking weeks of work to complete with the telescope used being locked in state for that period.