Working 1100D cold finger sensor cooling with TECs and water cooling

[Gina]

I now have a working cooling system for the Canon EOS 1100D DSLR camera (without condensation or interference in the image) so I am starting a new thread with the details. This will save wading through hundreds of posts in the design and development thread for anyone interested in trying the same thing.

[Gina]

The camera is mounted in an ABS plastic box to provide a way of excluding atmospheric moisture which otherwise would condense on the camera sensor and other parts as soon as the temperature was reduced below the dew point. It is not possible to seal the normal camera case as there are many controls etc. where moisture could get in. My current design makes use of a focal reducer to stop moisture entering via the light path and the focal reducer adapter ring provides the sealing at the front of the box. All connections through the box are sealed, including the water cooling part. Several silica gel (desiccant) bags were placed around the camera before sealing the box to absorb moisture from the air inside the box.

Cooling is provided by two Peltier TECs (thermally in series) inside the box with the heat being removed by a water block mainly inside the box with a foam air seal between block and box lid. Heat is transferred from the image sensor using a thin copper "cold finger" in thermal contact with the sensor and the cold side of one of the Peltier TECs. The heat from this smaller TEC is transferred to the other, larger TEC and thence from that to the water block.

The water block is cooled by a flow of water connected by soft plastic tubes to the rest of the water cooling system. I have mounted the latter on the telescope mount to reduce the water run and avoid the need for a high pressure circulating pump. The cooling system consists of a reservoir with an aquarium water circulating pump feeding water through the plastic tubes to the water block on the camera. The returning warmed water is fed to a radiator and from there back to the reservoir. The radiator has a 12v fan to blow cooling air through it.

That is the main principle of the system and I shall provide more details, diagrams and photos in succeeding posts.

[Gina]

Here is a list of the major parts used for the camera cooling modification - I'll post the other parts shortly :-

• Black ABS plastic Box :- http://www.amazon.co...ils_o00_s00_i00
  
• Peltier TEC 12v 50W 40mm square :- http://www.ebay.co.u...984.m1439.l2649
  
• Peltier TEC 5v 20W 25mm square :- http://www.ebay.co.u...984.m1439.l2649
  
• 0.7mm copper sheet for cold finger and plates between TECs and also TEC and water block :- http://www.ebay.co.u...984.m1439.l2649
  
• Water Block - I used a ThermalTake North Bridge water block as this was just slightly smaller than the 12v TEC and was the lightest I could find. Unfortunately, the source I used (eBay) seems to have dried up but there are others available.
  

The two TECs above were the ones I used and produced an EXIF T (temperature reported by the camera) of -15C with an ambient temperature of around 15C. I tried more powerful TECs but these worked less well. Later I'm planning to try lower power TECs to see if "less is more" applies at still lower power TECs.

[Gina]

When fixing the camera in a box you need to power the camera rather than using a battery as you can no longer get at the battery compartment (unless you make a trap-door). I used a proper Canon unit rather than try to connect wires directly to the camera circuitry. I provided my own regulated supply so didn't need the actual mains unit.

I got just the "Canon DC Coupler DR-E10 for Canon EOS 1100D" from Amazon :- http://www.amazon.co...ils_o01_s00_i00

This has interference suppression components inside as well as providing exactly the right contacts to power the camera. With this unit the camera reports that it is AC powered rather than Battery powered, through the USB connection.

[nytecam]

Hi Gina - got any of your progress pics to show the 'steps' ? TIA

[Gina]

Just been thinking it's time I added more to this thread I'll get on to it - sorry I've been rather slow in posting in this thread. Thanks for the prompt

For the photos I need to transfer them from another machine - this is my new (well, new to me) Win 7 box - my photos are on my Linux box. I'm having networking problems ATM so transferring via USB stick and/or drives.

[Gina]

Here are some photos which are hosted by Photobucket and scaled to 1024 pixels wide.

Inside of 1100D (back taken off) swhowing main board and image sensing assembly with metal shield.

This photo shows where the cold finger goes.

Main board removed.

Imaging assembly with metal screen removed. The square chip on the left has pins going through to the other side of the PCB - all the other components are surface mounted.. These pins need insulating from the cold finger. I used a piece of thin hard plastic sheet, more of which later. The 3 screws holding the sensor (and thence PCB) to the frame have been removed to release the sensor and PCB to aid inserting the cold finger.

This photo shows the pins protruding into the space between PCB and sensor.

In this photo the blue line shows where the cold finger will come up between main board and imaging board.

[Gina]

Continued...

This photo shows the cold finger - made from a 25mm wide strip of 22swg (0.7mm) copper sheet - inserted into the space between PCB and sensor with thermal paste between cold finger and back of sensor chip. The metal screen was left off.

[Gina]

Here is a diagram of the cooling system within the camera box. Two temperature sensors are thermally connected to the cold finger (shown in blue), one (round in cross-section and coloured light blue for identification) connects to a digital display on the front of the box, the other is one of three "1-wire" digital sensors that display temperatures on the computer. The cold finger value will also be used to set the temperature of the cold finger, providing full "set point temperature" cooling control. The others are for information mainly - temperatures of intermediate plate between Peltier TECs and water block. I have yet to write the software for the control and display.

[Gina]

More photos...

The ABS box with holes for the focal reducer and tripod bush fixing.

Front of camera with FR adapter ring and draught excluder to provide moisture seal.

Camera mounted in box but no electronics in yet. The cooling system with water block can be seen

Front (top of box) showing square hole for water block and digital thermometer display for cold finger.

Front of box showing FR adapter ring and camera light path with Astronomik CLS-CCD clip filter. Also showing power and data connectors hot glued into case.

Inside of box showing camera with electronics added and sealing foam round water cooler. The USB connection can be seen on the left with the cable fed out through the side and sealed with hot melt glue.

Top of box showing MOSFET PWM control knob and data source switch. Down for manual controlled by knob, centre off and up for computer control.

Back of box in finished state showing LM317T voltage regulator chip (used fro camera power supply) attached to water block for cooling. On the other side of the water block is the digital thermometer chip in a copper clip which measures the water block temperature.

This concludes the camera part of the system - photos and diagrams of the water cooling system will follow shortly. Meanwhile, here is an image of the Veil Nubula taken with this camera with the USB camera readout reporting an EXIF T of -15C. More infor on this image was posted in the "Imaging - Deep Sky" forum.

To be continued...

[Gina]

I'll be happy to answer any questions and to clarify anything that isn't clear. Please feel free to ask

[Gina]

Now to the water cooling...

Here is a diagram of the water cooling system. We ended up on the camera with the water block - now we come to providing cooling for that. (The principle is the same as the water cooling system in a car with the water block being the engine.) The water is circulated by a pump, through the water block which heats the water then through the radiator where it is cooled by cool air blown through the fins. The cooled water then returns to the reservoir. The result is that the heat in the water block is transmitted to the air flowing through the radiator.

In this case the circulating pump is a low power aquarium water pump. It is designed to be submerged in water. Here the water is contained in a plastic jar which acts as the reservoir. The radiator is a 120mm square unit made for this sort of water cooling system (generally used to cool the CPU etc. in a computer). The fan is the same size and is the sort supplied as a computer case cooling fan.

The photo shows the reservoir, radiator and fan mounted on the mount to reduce the length of tubing needed and the power required in the pump.

[Gina]

Water cooling components :-

1. Pump :- http://www.ebay.co.u...em=390382317829
   
2. Reservoir :- Recycled plastic sweetener jar
   
3. Radiator :- Used radiator from eBay. Something like this :- http://www.ebay.co.u...=item2c6366e708
   
4. Fan :- 120mm PC case cooling fan - eBay eg. http://www.ebay.co.u...=item415e825d76
   
5. Flexible plastic tubing.
   
6. Distilled water :- eg. http://www.ebay.co.u...=item257251e256
   
7. Anti-corrosion additive :- Car Plan Blue Star Antifreeze - Independent research concluded that ordinary car antifreeze was actually a better corrosion inhibitor than the very high priced cooling fluid recommended for PC water cooling.
   

[Gina]

I've found the link to the water block I got from eBay. Unfortunately, these seem to be unavailable at present. Pity because it's ideal for the job - just the right size and very light.

http://cgi.ebay.co.u...3413006&guest=1

[Gina]

Found a water block on eBay that's just the job. 40x40x10mm. Only trouble is the long delivery from Hong Kong. I've ordered one for my second cooled camera.

http://www.ebay.co.u...=item3cc612a5a2

[Gina]

Having modified my second 1100D to fit into the box right up to the bayonet fitting, I have decided to do the same with this one. This will enable me to remove the camera and box with just a twist of 45 degrees rather than several turns of the thread. In turn I shall then be able to change clip filters much more easily, allowing me to insert an Astronomik Ha clip filter. At the same time I'm tidying up the cooling components within the box. Photos to follow...

[Gina]

Here are the photos of the latest modifications :-

1. Back of camera in box. Instead of a screen between Peltier TEC and cold finger, I've cut into the cold finger to form a narrow strip to which I could solder the wire that's connected to the camera frame. This can be seen at the bottom right corner of the cold finger
   
2. The front of the camera showing the pared down front bits and the front fixing holes which will be used to screw the case onto the front of the camera
   
3. The front of the box showing new holes drilled for screwing the box onto the front of the camera.
   

[Gina]

Gradually tidying up the cooling parts. These photos show how the cooling parts fit together. Next I have to work out how to hold things in place.

1. 25mm square TEC on cold finger.
   
2. Intermediate copper plate on top of 25mm TEC.
   
3. 40mm square TEC in place.
   
4. Water Block on top of 40mm TEC.
   

[yesyes]

This reminds me that I wanted to ask you whether you fasten the TECs somehow or do they just stick together using the thermal paste?

[Gina]

This reminds me that I wanted to ask you whether you fasten the TECs somehow or do they just stick together using the thermal paste?

Previously I had the lid holding everything together with plastic foam (draught excluder) applying pressure to the stack through the water block. Now I'm thinking of fastening it to the camera rather than relying on the box. The thermal paste will hold things together to some extent but tends to allow things to slide.

The 40mm square TEC shown in the photo is the 136W one that doesn't do the job - it's too powerful, generating too much heat itself and not enough cooling effect. I'm still waiting for the right replacement to arrive. I also have some lower power ones coming from Hong Kong in due course. They are 12v and 5v but take up to 2A rather than 5A. I shall be trying various combinations to see which works best. Apart from the current design with 5A TECs or the new ones with 2A, another possiblility is two 5v TECs or two 12v TECs which could mean just one MOSFET.

[astrovirus]

Don't TECs require some presure (not to much) to work correctly, at least that's what I've been reading in some other cooling mods and did so myself when I installed the TEC on my 350D.

[yesyes]

That's why I'm asking. I'm applying quite a bit of pressure on mine. I was just wondering whether it might be too much pressure... I'm still struggling getting it anywhere near cold enough. Now using a 30W and a 60W on top, both 30x30mm and both 12V. The lowest I had it was -9°C directly next to the cold side of the bottom TEC. The far side of the cold finger has never gone below zero. I've put it aside for now and continued with a different project. Maybe a fresh start in a while will give me a different perspective on it...

[astrovirus]

Is that -9°C below ambient or as absolute temp? My TEC on the 350D (45 W) gets it to -28°C below ambient within 2-3 minutes. Too much presure is not good as it gets crushed and stops functioning, but as I understand it the peltier does need some presure in order to function. What I did to ensure it's not overtight but presure is applied, is to fit somenylon washers on the bolt that connects the heat sink to the cold finger. These washers add up to the exact same thickness as the peltier and ensure it can be tightened without crushing the peltier device.

[Gina]

When I get the TECs from Hong Kong I shall have 2 of each of the following :-

1. 12v 136W 40x40mm - unsuitable for this application
   
2. 12v 50W 40x40mm - in current design
   
3. 12v 20W 30x30mm
   
4. 5v 20W 25x25mm - in current design
   
5. 5v 7W 25x25mm
   

Bearing in mind that two TECs of unequal power are considered best for stacking, I see possible combinations as :-

1. 12v 50W 40x40mm and 5v 20W 25x25mm - the current design.
   
2. 12v 50W 40x40mm and 12v 20W 30x30mm - this uses the same power as at present but doesn't use 5v, simplifying power supply - colder TEC is now larger than the cold finger.
   
3. 5v 20W 25x25mm and 5v 7W 25x25mm - the lowest power combo with one MOSFET and 5v 7A power with 12v for control and camera supply (this could be powered from the mount PSU of 13.8v). The TECs are the same physical size and therefore don't need an intervening copper plate (other than for checking the temperature).
   
4. 12v 20W 30x30mm and 5v 7W 25x25mm - same power as no.2 but dual supplies and MOSFETs - I see no virtue in this.
   
5. 12v 50W 40x40mm and 5v 7W 25x25mm - I think these are too different in power but I guess I could try it.
   

• 
  

[Gina]

Don't TECs require some presure (not to much) to work correctly, at least that's what I've been reading in some other cooling mods and did so myself when I installed the TEC on my 350D.

Yes, that's how I understand it.