The CP30T-EOS is another type of DSLR cooler: a “Direct CMOS cooler” Instead of using a “box” which usually chills the air to decrease the sensor temperature, a “Direct CMOS cooler” applies a cold component directly to the CMOS (by mechanical/thermal contact).

While a “Box Cooler” typically doesn’t require any modification of the camera, a cooling system such the CP30T-EOS does require several “hard” modifications. Needless to say that if you are attempting this you will need to open up the camera body (voiding your warranty in the process) and some small tools.

Why am I doing this? Well, this is a good question because I am pretty satisfied with my realizations so far. CP80-EOS was a great start then upgraded to CP80X-EOS which was my most efficient unit up to date, and my CP50X-EOS was light weight and still efficient at removing a bunch of noise. So, I am in a position where I would like to get both great cooling and a lower weight. And that is my goal here with the CP30T-EOS.

Some history being cold finger modifications: I’ve made some researches and was able to find some cold finger modifications for DSLR about 5 to 6 years old. So this isn’t entirely new. One of the most notable cold finger modification is Dave’s CoolCam for 450D. I think this is one piece of art. Dave did a great job with his CoolCam and if I wasn’t so inclined to reducing the weight I would have probably followed his website to the letter. All of the cold finger modifications I’ve found were using some kind of hot finger or dew heater to prevent the CMOS from freezing up. This is where I set my personal limit. I didn’t want to have to take the all CMOS assembly apart. I’ve had saved up for quite a bit of time to acquire all my gear and my T1i was Baader modified by Gary Honis just a couple months ago. Luckily enough, I thought of a good way to ignore dew heaters/hot fingers all together.

My PWM temperature controller is highly responsive and precise. I thought I could use it to maintain the cold finger at a temperature above dew point. It signifies I won’t be able to bring my CMOS to freezing temperatures, but based on my Noise measurements, the noise reduction gains get more marginal at low temperatures. Plus I already know that if I can get temperature similar to what CP80X-EOS can do, I will be very happy!

A Mechanical Challenge: part the challenge with cold finger modifications is about retaining the added weight of the parts to something that is strong enough for it. We don’t want the lever from this weight to apply any side force to the CMOS or any internal part of the camera for that matter. The choice I’ve made for this job was to use this now very handy custom T-ring that I’ve designed a few months ago. It features four threaded holes that can be use to retain any type of frame and in this case it is an aluminum frame that my T-ring will retain. This frame is bent in several location and hosts on the other end the hot side heat sink.

Peltier module and Heat sink: I’ve learned from other projects that a big TEC module is not always better. Bigger TEC mean more heat generation on the hot side and for a given heat sink also means higher hot plate temperature. But remember that Peltier modules apply a temperature delta between its two faces. So the higher the temperature on the hot side, the higher the temperature on the cold side. In the end we want a TEC with enough capacity to take care of the heat coming out of the component to cool, but small enough to keep a hot side temperature as low as possible. The module I’ve chosen is 30 W and 70 C maximum delta. Considering the very low amount of mass to cool down, 30 W will be more than enough and the 70 C rating should give low temperature without much power consumption.

Cold Plate and Cold Finger:  I’ve opted for a 1/8″ cold plate attached to the TEC module. Cold finger are thin and I didn’t want to machine a too complicated part in such thin material. Using a cold plate allows me to ensure good contact between the TEC and the cold plate. And the cold finger will remain a simple rectangular piece of copper about 1.2 mm thick. I’ve machined a 1.5 mm groove in it to glue a NTC thermistor probe that is wired to my 6-pin harness and connected to the PWM controller. The top side of the cold finger was covered with a couple layers of insulation tape (yellow tape) to prevent the copper plate from shorting out any soldering point at the bottom of the controller PCB.

Here is what it looks like when completely assembled:

Actual Modifications: I’ve opened up the camera once a couple weeks before the modification to take some measurements and plan the work. I figured I would need to cut two posts, remove the metal ground shield and modify the side panel to clear enough room to route the cold finger out of the camera.

Dave suggested using a Dremel cutting wheel with a vaccum cleaner, I had this high leverage cutter that I wanted to try before going the Dremel way. I am not sure how strong it is but it had absolutely no trouble cutting these metal posts which left no chance for any small debris from messing with the electronics.

The metal ground shield was pretty much only held in place with two screws. It came out really easily.

Finally, I use a small CNC mill to cut a groove in the plastic side plate. But a Dremel tool would probably work well here too.

Final Assembly: After minor adjustments on the bends of the aluminum frame and some tweaking with the cold finger, everything was aligned, tightened and ready for the camera reassembly. I highly recommend Gary Honis’ instructions in regards to diassembly/reassembly of the camera. Everything is very detailed and you can almost follow his instructions without even thinking The most difficult part of the reassembly was for the small ribbon cables that wouldn’t snap in, they eventually went in with enough patience.