Canon 350D Peltier Mod

[MarkS]

I've done my first experiment today and the results look promising.

Here is the Canon with the CCD chip exposed:



Note the copper pipe beaten flat in the lower part of the image.  Its thickness in 2mm, which is just slightly larger than the gap between the back of the CCD and the circuit board that fits above.

Here's the conductor held in place with duck tape:



Here's the camera reassembled:



The heatsink is a standard aluminium CPU heatsink and fan.

Here is a graph of the standard deviation of noise on a series of successive 5 minute dark frames:



The noise level increases at first as the CCD warms up and then drops immediately when the cooling is switched on after frame 13.

The drop in thermal noise is consistent with a temperature drop of around 18C from just before switch on until the final frame.  The experiment was conducted at room temperature (around 20C) but no ice formed around the Peltier.

Next time I will shorten the piece of flattened copper tube and use a more efficient heatsink/fan (in copper?)

[JohnP]

Nice one Mark - Looks very good for first tests. Out of curiosity what Peltier are you using what volts/ current does it operate at?

John

[mickw]

copper pipe beaten flat

So the copper pipe is just acting as a heat sink ?

I was expecting a bi-metal "finger" with a voltage applied 

Presumably I've got it all wrong (Now there's a surprise) and need to re-read "Peltier Cooling" 

[MarkS]

So the copper pipe is just acting as a heat sink ?

The copper pipe is the cold finger - it "conducts the cold" from the peltier to the CCD.  Unfortunately the Peltier cannot be placed directly on the CCD because there is a circuit board in the way.

Out of curiosity what Peltier are you using what volts/ current does it operate at?

Just look at the photo - it's a TEC1-12703!  

I scavanged it from the 12V Maplin drinks cooler.  At 12v it consumes 2.5A and the fan adds a little bit more.

I found a datasheet for it on the net which gives the following:

Imax = 3 amps
Umax = 15.4 volts
Qc max = 26.4 watts
deltaTmax = 68 Centigrade

They can probably be sourced fairly cheaply from China.

Mark

[mickw]

Imax = 3 amps

You're going to take Imax images 

[JohnP]

Just look at the photo - it's a TEC1-12703!

'cuse me for asking......... 

I was more interested in current draw & figured you'd have specs to hand....

[MarkS]

Now I've added a bit of insulation around the copper pipe to prevent some of the earlier heat loss.  Results are on this graph - guess at which point I switched the Peltier on?



Temperature drop is now 24C but this is an underestimate because the initial temperature was still rising.  I think that with a couple more modifications a 30C drop would be quite achievable.  Although thick ice was forming on the Peltier device, the same was not true of the copper pipe - a hammer beaten copper pipe is not really flat enough to ensure good thermal contact with the Peltier (nor with the CCD).

Here is part of dark frame #17 (immediately before Peltier switch-on) and dark frame #36 (after the Peltier has stabilised):



The dark frames have been scaled in an identical manner.  I think you'll agree this is quite dramatic.

By the way - I've estimated the power consumption of the 350D - it is around 3-4 watts.  So the heat generated is not that difficult to dissipate.

Mark

[Daniel]

Excellent work Mark, the results are staggering, makes me want to pay you to mod my 40D now

[Mac]

just curious why the blip on frame 34?.

How about taking some measurements with the camera in the fridge to start with?
That way you can get some idea of what the results will be like towards the colder parts of winter,
when your camera is l really cold to start with.

The results are looking fantastic considering the initial start temp of 20~ ish.

what would the outcome be if the peltier were larger and had a greater temperature difference?

Mac.

[Ian]

Mark, you could try some thermal grease on the cold finger, both between the pelt and the CCD. I wonder if it would be possible to grind the copper flatter using a cheap whetstone or diamond knife sharpener.

Also it's important to keep an eye on the hot side of the pelt until you've got your testing sorted. They're capable of self destructing, by heating the hot side enough to melt the solder at the semiconductor junctions...

[Mike]

Marm,

Any of these any good to you....

TEC1-12709T125
TEC1-12706T125
TEC1-12703T125

If so I can get hold of them for about £3 each.

[MarkS]

makes me want to pay you to mod my 40D now

No way would I touch a 40D!  A second-hand 350D is cheap enough that a mistake won't be too expensive.

just curious why the blip on frame 34?.

No idea. It might be the dodgy fan - I had it lying around because I removed it from my current PC!

what would the outcome be if the peltier were larger and had a greater temperature difference?

It would require a larger heatsink and fan like the one in the Maplin drinks cooler - it was just too big and heavy for the camera.

you could try some thermal grease on the cold finger   ...  I wonder if it would be possible to grind the copper flatter using a cheap whetstone

Thermal grease - I've already used a bit.  Whetstone is a great idea!

By the way - here's a pic of the current prototype:



Mark

[MarkS]

Any of these any good to you....
TEC1-12709T125
TEC1-12706T125
TEC1-12703T125

Yes please!  Would you order me four TEC1-12703T125 - they're identical to the scavenged one I already have.

Mark

[Mike]

4 x TEC1-12703T125 on the way

[Mac]

wont the fan take off your left eyebrow when your framing the image? 

It's looking good.

Is there anything else you can mod on that camera?