PDA

View Full Version here: : Peltier Cooled Fridge Mark III


avandonk
13-12-2009, 09:58 AM
I use a modified Peltier fridge to cool my Canon 5D camera for astrophotography. See image below. There are two CPU heatpipe coolers cooling two 40x40 mm Peltiers. I am powering them at a maximum of 14V and 6A each. This is about 80W+ of real cooling. I have a PID controlling the cold side of the heat exchangers to within 0.1C.

The maximum temperature drop I can get is about 24C from ambient. Ths is fine in winter as I can get the fridge to -15C but in summer on a hot night lucky to manage 0C.

What I propose to do is put CPU water blocks on the fridge and run cooled water or water/antifreeze through them. I was going to use a chest freezer to cool the coolant. This way I can get -20C in the fridge even in summer.

A better more elegant idea is to sandwich another Peltier between a water block and a CPU heatpipe cooler and use this to precool the coolant. Two of these will most probably be needed to get a suitable temperature drop. See the rough image two.

I will have to look at the cost and convenience of either method.

The chest freezer will have far more grunt. A 50% antifreeze water mixture is good for -37C. So in theory the Peltiers are only needed for fine contol of the temperature. The usual small chest freezer gets to about -18C. This means a camera fridge temperature of -30C is a doddle with this configuration.

It then occured to me all you extreme CPU cooling people may be interested.

The problem with Peltiers is if they are overloaded or fail you may lose your CPU.

If you insert a water block Peltier and CPU cooler sandwich in your coolant line just before your CPU the coolant will be below ambient.

If you want more cooling just add another.

With this configuration even if the Peltiers failed you still have your coolant at ambient due to the radiator.

Yes it is cloudy!


Bert

DavidU
13-12-2009, 11:17 AM
That's a great idea Bert.:thumbsup:

leon
13-12-2009, 03:07 PM
Nice Going Bert I bet that thing gets pretty cold inside.

Now I'm not trying to hijack your thread Bert, but the one I'm working on is coming along nicely as well, it is at present having a bench test, and seems to be dropping well in temp.

But I suppose the best test will be in real life on the scope, anyway have a look and see what you think, thanks for your help as well. :thumbsup:

Leon :thumbsup:

lacad01
14-12-2009, 10:51 AM
Both you guys' engineering prowess is an inspiration, top stuff :thumbsup:

leon
14-12-2009, 11:26 AM
Thanks Adam, :D but i suppose i had not say to much, :) :) as this is actually Bert's thread ;)

Leon :thumbsup:

avandonk
14-12-2009, 03:35 PM
You can go for broke Leon. I do not own this thread any more than I possess all knowledge. After all we are here to exchange information not to stake out territories.

I am going to go with the chest freezer and 50% antifreeze. The 50% antifreeze is liquid down to -37C.

What is driving me that data at -10C is noticeably better than data at -5C.

I hope to routinely have the Canon 5DH in a -30C environment. You can get away without using darks at this temperature with a Canon CMOS.

Bert

mldee
14-12-2009, 05:15 PM
Sorry for a stupid question from a learner but: Wouldn't it be simpler just to use darks and forgo the additional equipment complexity?

avandonk
14-12-2009, 07:08 PM
When the noise is about the size of your signal after you subtract the noise you get lovely holes in your faint data.

All of science is about signal to noise ratio. If you suppress or eliminate noise then obviously your signal to noise ratio is far better

Bert

mldee
14-12-2009, 10:20 PM
OK, makes sense.

As a total novice, most comments I read in this forum tend to suggest that darks + PS are the all -singing, all-dancing answer to an astro-maiden's prayer. Obviously not!

Physics does reign supreme. Sorta glad of that. :)

Paul Haese
17-12-2009, 10:11 AM
Bert,

what do you intend to do about dew forming around the sensor and the subsequent issues relating to that? Many dedicated camera manufacturers use silicate or gas or both to prevent dewing. Your system will attract dew unless you take steps to reduce its impact.

avandonk
17-12-2009, 11:13 AM
Paul I have measured the temperature of the sensor while it is doing continuous long exposures and it is 17C above ambient or fridge temperature. In other words it is the hottest object in the fridge so as long as it is exposing continuously no condensation. I have run the fridge at -20C in winter and had no condensation.

When cooling down I set the camera taking dummy exposures and you can tell when the sensor is at equilibrium when the jpg size straight out of the camera levels out.

I thermostatically control the temperature of the lens barrel or FR in the case of the 100ED to 20C. If I do not do this the lens or FR gets condensation.

If I run the fridge at -10C the sensor is at 7C. At -30C fridge temp the sensor is at -13C.

On a hot summer night this will be very useful.

Bert

AlexN
29-12-2009, 02:22 AM
Bert - Considering you are using computer cooling gear to set up most of this, have you considered a vapo style setup for chilling the water? Cost effective? probably not.. but functionality wise I dare say you could achieve great cooling results..

I had a phase change cooling setup on one of my AMD Athlon XP systems back in the day, under the stock air cooler, the CPU would have idle temps around 60°c, and up to 85°c under load... With a water cooling setup I got it down to 30° idle and 37° under load (whilst also maintaining a 1.5ghz overclock with + 0.9v to the core)

I eventually wanted more, went to a phase change refridgerated cooling system and had the CPU sitting at -60° under full load with the same 1.5ghz overclock...

Granted, the heat loads etc are different, as is the application of the cooling to the fridge.. However I'm sure you could figure something out..

avandonk
29-12-2009, 02:19 PM
Alex I have managed to get a second hand 100l freezer. It can get down to -25C. I will use this to cool a 50% anti freeze solution to about -20C and pump this through the cpu water block to cool the hot side of the Peltier. This should get at least -35C or more as an environment to cool the camera.

The freezer has a 190W compressor which means it can pump at least twice this in the form of heat. This is equivalent to about 800W of Peltiers.

Bert

AlexN
29-12-2009, 02:59 PM
That oughta do it mate.. :)