Peltier cooled cold plate

[TrevorW]

I'm toying with the idea of making a cold plate using a Peltier and heatsink fan from a PC etc etc that can be attached to the base of the camera to cool it during those warmer summer nights, but itstead of reinventing the wheel or be another person who have tried but found the idea useless please let me know.

Thanks

[Ian Robinson]

Bit cumbersome.

Thermoelectic cooling is much much more efficient and easier to impliment, especialy multistage thermoelectric cooling.

See http://www.rmtltd.ru/tec_modules.htm?gclid=CImSz6a705YCF QykagodCHDo3Q

[TrevorW]

From looking at the site Ian I got the impression these are Peltier devices anyhow.

The Peltier device I have measures 40 X 40 X 5mm, the bulky part is the heat sink fan etc required to remove excess heat created by the Peltier.

Correct me if I'm wrong but a lot of CCD camera's incorporate Peltier cooling devices.

Cheers

[acropolite]

Bert's (avandonk) fridge cooler (http://www.iceinspace.com.au/forum/showthread.php?t=34129) could be just the project you're looking for.

[Ian Robinson]

This will explain the principle.http://www.digit-life.com/articles/peltiercoolers/ and gives some useful info.


The attachment is a simplified sketch of how TEC of a camera CCD chip would be achieved (without pulling the CMOS CCD chip off the circuit board which is probably a very bad idea), guess it depends on how free space there is behind the circuit board the CMOS CCD chip is soldered to, the copper heat cooling sheet need not be more than mm thick (probably a lot less) to get efficient cooling.

I do not think cooling the outer body of the camera by peltier TE cooling will be very effective as the body of the camera will probably be ineffective as a medium to cool the CMOS CCD chip - a lot air spaces and gaps and air is a very effective thermal insulator.

Close contact to the CMOS CCD chip is required I think to cool effectively.

Another option may be to pump superchilled , dried and filtered air into the camera somehow directing it at the CMOS chip (which might be achieved by directing filtered and dried air over a multistage TEC and passing the very cold air over the CMOS CCD chip via some insulated plumbing that will enter the camera body at an appropriate location ). This would I think be easier to impliment and would involve only drilling two holes in the camera body.



Would I do this to my 40D? ... No not at this time or any time in foreseeable future.

[TrevorW]

I made up a temperature controller device from a kit that should allow the Peltier to acheive sub zero C temperatures but was wondering should I use thinner or thicker aluminum for the cold plate toss up between 2mm and 5mm, also I was going to use nylon bolts to hold the Peltier between the cold plate and heat sink should they be OK, any feedback welcome. I don't envisage glueing a plate to the back of the CCS circuit board until sometiem after the warranty expires.

Cheers

Cheers

[Ian Robinson]

Another concern may be how well the soldering will handle thermal shock (in summer of going from maybe mid 20s - mid 30s C to - 60 to -70C) in a few minutes or maybe a lot less (depending on how efficient and effective the heat pump is), and how the circuit board will cope with thermal shock in general.
I suspect provisions for this are made in cooled cameras at the design stage and in construction and assembly. The average DSLR is not designed to cope with extreme thermal shock.

The thermal conductivity at room temp of aluminium is 250 W/m K, cupper is much better at 401 W/m K and silver is 429 W/m K. I'd use chunk of copper as my heat drain.

[AlexN]

TEC's and Peltiers are much the same... if not identical...

I've used peltiers to cool processors of computers (back in the days when I did crazy things with computer, along with phase change evaporative cooling setups and water cooling) And Peltiers and TEC's work in much the same way..

[Bassnut]

I tried this, went nuts with it :P

Use a copper finger, aluminium is a waste of time, lousy thermal conductivity.

you will have to permanently butcher your camera, the peltier-cooled finger has to directly contact the CCD sensor.

What killed it for me, was condensation, very hard to eliminate.

Its all doable, apparently Houghys commercial mod works very well, but your in for a world of pain doing it youself, a real challenge.

[bojan]

Ian, you are right here.

In cooled cameras, only sensor is cooled. Soldering on cooled parts and circuits is avoided (crimping is preferred method)
But most importantly, all connections are made such they allow for expanding/contraction. I saw a lot of solder connections teared apart by thermal stress in the course of my work, after only couple of cycles (from +80 to -35 and back).
That is why I think Avandonk's method (fridge) is what I would do myself (however so far I do not see the need for this, since Melbourne's LP prevents me from going any longer than 1minute, so noise is not the problem for me, at the moment)... Simply and slowly lower the environment temperature camera is in.. Mimicking the natural cooling when you take camera outside from relatively warmer room.
DSLR's are NOT design to survive extreme thermal shocks.

[dugnsuz]

Another home made fridge...
http://ghonis2.ho8.com/rebelmod450d16c.html

Cheers
Doug

[TrevorW]

Thanks Dougie seen that one but didn't like the cumbsersome nature of the container.

I'm not talking about drastic temperature differences here a drop of 10 degrees or so on ambient air temp significantly reduces noise created by the CCD.

Condensation occurs when the is a signicant drop especailly sub zero in temperature relevative to ambient air temperature as it will be external to the camera and temperature controlled condensation shouldn't be an issue.

I had thought about copper but sometimes hard to get in small sizes and more expensive and the only scrap metal yard I know that may have had it closed down recently after 60 years in business.

But it's all fun trying

[Garyh]

Hi Trev,
I have been experimenting with cooling my camera with a peltier cooler and cpu heatsink for sometime. It`s definitely worth giving it a go I reckon....Either by enclosing and chilling the whole camera (like a cold winters night) or directly to the cmos..
Still contemplating how to modify the camera internally as yet to make it more efficient? but a copper plate on the back of the cmos bridged to the cold finger would be the way to go..
At he moment I just attach via the tripod socket and I get maybe 7-8 degree C drop inside the camera. But do get some condensation on the base and sometimes ice..Cold finger is a 4mm thick piece of Alluminium. Also a lightwieght foam surrounds will be in the making to insulate it more and get the temp down even more.
Attached a pic of what I have been using so far..
I would be interested in info about that kit but Trev! as I am after a way to monitor and control the peltier..
I am aiming at getting the camera a few degrees above 0 on warm nights..:)
cheers Gary

[TrevorW]

If you go to this link belwo you can get the termprature controller, I've also toyed with the idea of having a LCD incorporated to show the temp of the Peltier.

http://secure.oatleyelectronics.com//index.php?cPath=80&osCsid=2c1ed996098ba9ab78e39ca7aa45 60e2 (http://secure.oatleyelectronics.com//index.php?cPath=80&osCsid=2c1ed996098ba9ab78e39ca7aa45 60e2)

PM me your design mine looks simlar but attached to the base of the camera

:thumbsup:

[Garyh]

Here is a side by side comparison of a 400px crop from the center of the frames..
Ambient temp was 30 degrees C.
First frame: Camera was left on for 45min to allow it to reach operating temp then a 5min exposure at iso 1600 was taken.
Second frame: Peltier turned on after first exposure and left for 45min then a 5min exposure was taken..Ambient temp stayed pretty constant and within a degree or two..
I reckon a good 8-10 degree difference there..
cheers Gary

[TrevorW]

A marked difference considering it was also a 30 degree night and are talking only and 8-10 degree difference in temp overall. This difference would help produce better subs and darks.

Cheers

[TrevorW]

I've got the Peltier down to -10 will post some pocs when done

[Bassnut]

Cool Trevor, that was quick. Looking forward to pics.

BTW I had trouble finding copper, until I twiged on to flattening 3/4 copper plumbing pipe, worked well.

[TrevorW]

Ok here's the result I'm yet to field test it but want to use a 12v 5 amp regulated power supply and not my battery pack as the Peltier draws 4 amps. I would like to also attached a LCD temperature display so I can monitor the cold plates temp I've got the parts just need to work on it. I've covered the base of the cold plate with neoprene. I've incorporated a power switch with LED. A work in progress so to speak nothing original.

Cheers

[Bassnut]

OK, thats one hell of a heatsink :P.

Should get you some usefull degrees below ambient.

[TrevorW]

Ok got the temperature controller working with the help of my techo BIL now the Peltier only draws power when cold plate temp rises above a certain level just fine tuning now

[Tandum]

Looks like a very affordable and worthwhile project Trevor. Having delt with with these things before, I'd be a little concerned about condensation/ice forming inside the camera myself. They can get pretty cold especially with a heat exchanger off the Titanic as you have :)

[TrevorW]

I'm not sure about condensation as the intent would be to use the Peltier only on warmer summer nights but will have too test and see. Some stats are attached

material cost so far $60

Cheers

[TrevorW]

Taken inside ambient temperature 24 degrees C

Images all 105 seconds ISO800 camera turned off between image 1 and 2 left on between 2 and 3

After 15 minutes the cold plate temp had dropped to about 1 degree then jumped to steady 2 degrees although the Peltier draws 4amp at 12v DC only 10.2v was being fed into it. My battery supply was running low, so obviously the Peltier was not drawing the max current and running at full effeciency the heat sink hardly warmed up.

1st image is start up, 2nd after 30 minutes and 3rd after 45 minutes, no change in cold plate temp between 30 and 45 minutes.

A good reduction in noise is apparent, further testing will be done when sorted power supply issues ie: buying a 12v 5amp regulated power supply adaptor. At full current draw Peltier should acheive sub zero temperatures.

:thumbsup:

[TrevorW]

Its apparent that to get the cold plate down to sub zero temps you'll have to look at a multi-stage Peltier.

I've changed a resistor in the controller so the circuit draws 11.60V but the cold plate only gets down to around 1 degree after 15 minutes.

Another option would be to use copper for the cold plate.

[Craig_L]

Definitely try the copper. Better than 99% pure aluminium I think.

[Garyh]

Looks like you are making headways there Trev! I don`t think you shall get it much lower even with copper as I think thats about the differential limit of the module. You will have to stack them to get it lower I say..
Now if you can chill the innards to 1-2 degrees things shall be very nice!
Camera modding time!

[TrevorW]

Second set of test shots showed that although it appears there was no reduction in noise level you may notice a gradual drop in the file size after the Peltier was turned on ?? Each frame was reduced to 800X532 in PS7 saved as JPEG quality level 5

A further test run will be turning the Peltier on and leaving it on for a least a half an hour before turning the camera on and taking frames

Attached is the second run of test images ambient temp was 24C, camera was turned on over 10 minutes before 1st frame then Peltier was turned on, each frame is 15 minutes apart camera was left on during each frame,

Camera setting; Auto shutdown OFF, frames 90 seconds ISO 800 INCR OFF.

Final cold plate temp was around 2c

[MrB]

Hi Trev, you should be able to get it cooler than that.
Here's a pic of my experiment a few years ago, tho not for an astro related use(homebrew ;)).

It's a single 33W 40x40mm peltier and the cold side is at, from memory, about -10c.
Thats with a large heatsink on the cold side that can absorb a fair bit of ambient temp... tho ofcourse the ice would be insulating it. With a little finger it could go much lower.
Room temp was probably around 20ish(room rarely varies, hence used for homebrew) and the heatsink temp up around the 40 mark somewhere.

The trick is to keep the hot-side as cool as possible, and it's massively important here to reduce all thermal resistance, I don't mean the heatsink, it's only a small part of the issue and the easiest to deal with. The heatsink might feel cool but the hot-side of the TEC could be much hotter.
The max temp differential(hot to cold side) for a Peltier is approx 68c.
If your cold side is at 1c, and the heatsink is at say 30c, the differential is only 29c, where is the other 39degrees? In all the materials and joints.

Imagine every interface, and every material used, as an extra resistor in an electrical circuit... voltage loss here, voltage loss there, it all adds up. Same with TEC's, temperature loss here, temp loss there.
Even a tiny space of air(the best insulator) between two mating surfaces will cause problems, thats why it's recommended to use lithium heatsink transfer compound.

In the photo I have used an aluminium CPU heatsink, about 3"x3"x2" size that has a copper core running through it, a rod about 1" dia by 2" long, the 1" diameter was too small for direct contact to the TEC (40x40mm) so I had no choice but to use a plate between the two. This adds another material and another joint, both with their thermal resistance, each could easily add a few more degrees to the hotside temperature, which means a few more degrees to the cold side too.
Each interface had atomised-silver heat transfer compound to increase contact area(reduce air) and reduce thermal resistance.

Will be building a temp controled brew fridge soon, so will be doing this again with all copper and silver construction and fluid radiator systems instead of traditional heatsinks.
Want to get the fridge as efficient as possible and be able to get the brew to lagering temps (1 or 2 deg) and possibly go all solar/wind powered.

[TrevorW]

Mr B thanks for that

the further test as mentioned above showed an increase in the noise level the longer the camera was left on, the first image taken after the Peltier had been left on for half an hour and the camera off shows minimal noise but as the CCD heats up while the camera takes exposures 3 minutes apart the noise level rises between each image.

The best thing would be to turn camera off for a few minutes between each image which unfortunately is not practical except if long exposures are being taken.

Each of these images is the same ISO etc as previously mentioned although the time between frames is reduced.

From this it seems that the Peltier cold plate does make a difference depending on how it is applied.

Although another issue may be that the Pentax ccd has a higher noise to signal ratio than other camera's.

:thumbsup:

[TrevorW]

By removing the bottom plate I get the surface of the Peltier cold plate down to -3C after about 15 minutes with ice forming on the plate. Although I'm using a 12v power supply to run the camera will removing the batteries make any difference as batteries can cause heat???