Cold Plate Design?

[reflex]

Hi all,

As you might have gathered, I am kinda new too this cooling business

I have a small problem, I need too bolt 2 heatsinks together, one on the coldside of the TEC and the other on the hotside. The problem is I need a spacer for inbetween, as the panel I am bolting through is relatively thick(20mm or so).

I was thinking just a square piece of copper of the correct thickness would work. But now I am wondering if I should sandwich the peltier between 2 copper plates and then attach to heatsinks too that.

My friend had some bad luck, his copper plates of 8 or 10mm thickness bent and cracked his 12730. The top plate bent and cracked it on both edges on opposite sides. His plate had bolts only on each corner of the TEC.

Should I have lets say, 3 bolts per side(using 40x40mm unit)? Also is there an easy way to insulate the bolt so they don't steal heat/cool from the other plate?

Thanks for any advice

[Ultrasonic2]

i find it relatively hard to believe he managed to bend such a thick bit of copper over such a small distance .. it's more likely the copper wasn't flat in the first place .. i experienced the above when using cold rolled copper of that thickness .. After closer inspection i found the copper not to be flat .. i had to get it milled perfectly flat then everything was sweet.

[reflex]

I also found it hard to believe, maybe it was bent before we started, because the bolts were also relatively close to the TEC on the corners.

[leuler]

I believe that Uncle Jimbo suggested using fiber washers to insulate the
bolts from the cold and hot plates. Plastic washers have too much give.
You could also use stainless steel bolts. Thermal conductivity for stainless
steel is a little less than steel.
http://www.engineeringtoolbox.com/th...ity-d_429.html

The copper bars that I used definitely were not flat. It took a lot of time
and sandpaper to get them flat enough.

3 bolts per side ( 8 total ) would provide more even pressure. Whether or
not that many is necessary if the copper plates are flat I don't know.
If you use only 4 bolts, it might be better to have them in the middle of
each side. It would take more force to bend the plates in those locations.

[Uncle Jimbo]

I believe that Uncle Jimbo suggested using fiber washers to insulate the
bolts from the cold and hot plates. Plastic washers have too much give.
You could also use stainless steel bolts. Thermal conductivity for stainless
steel is a little less than steel.
http://www.engineeringtoolbox.com/th...ity-d_429.html

The copper bars that I used definitely were not flat. It took a lot of time
and sandpaper to get them flat enough.

3 bolts per side ( 8 total ) would provide more even pressure. Whether or
not that many is necessary if the copper plates are flat I don't know.
If you use only 4 bolts, it might be better to have them in the middle of
each side. It would take more force to bend the plates in those locations.

Leuler is correct - there is no way an 8mm plate was bent by 150 or even 300 psi unless the mounting was a long way from the TEC. Bolting on the corners of the TEC is exactly wrong - the screws should be on the centerline.

The flatness of the plates is critical and probably why the TEC got cracked. Most manufacturers recommend flatness of 0.001 in/in with a maximum of 0.003" - that's pretty flat! It is more important that it be flat than 'shiny' although shiny is nice. If you have unfinished rolled copper, you need to clean it up. I use a regular block sander, but I put the copper bar between two pieces of material just slightly lower than the bar, to keep the sander flat and avoid a convex lap.

Use waterproof paper designed for metal (available at most hardware stores) and 'wet sand' by keeping plenty of water on the work piece. Start with 100 grit and sand until you can't see any high spots - the grinding pattern should be even all across the piece. Then use 150, sanding until the marks from the 100 grit are no longer visible. Switch to 320 and do the same again. That is actually smooth enough, but if you are patient and want a very smooth surface, go to 600 grit and then 800 grit. You will end up with a very nice finish. If it was a block designed to go on a CPU, I would go further, and do final lapping with a glass block and polishing compound. With a little 'elbow grease' you can get a mirror finish.

Next is the cooling 'stack'. I always use a spacer block between the TEC and the cold plate no matter what the installation, because it reduces loss between the hot and cold side. Don't EVER put a spacer on the hot side, it reduces efficiency. You need at least 5mm for copper and 8mm for aluminum for the spacer block. You also need insulation between the hot and cold plate. If you use neoprene to insulate, the neoprene also gives a good moisture barrier.

Here's an assembly drawing:


The 'vapor seal' on the screw heads just means that the area where the fiber washer contacts the hot plate should be flat and smooth. Most people just cut channels in the neoprene for the wires - the wire feed-through are sealed posts which allow a perfect moisture seal on the cold side, in case it is submerged in water (as for a chiller).

The mounting screws should be along the centerline of the TEC (or TECs), about 6-8mm from the edge of the TEC. Note that you should also align the holes in parallel to the fins of the hot side sink, to get added rigidity. You can also place screws on the sides of the TECs but if you do, be sure the in line screws are tightened first. Take them all to just tight, then to minimum torque (150 psi) before tightening the side screws. Then take the sides to 150 psi, then the in line to 250 psi, then the sides to 250 psi.

Torque is done like this (from the Melcor manual):
Before bolting, best results are obtained by preloading in compression the cold plate/heat exchanger/module assembly, applying a light load in line with center of module, using clamp or weights. For two module assemblies, use 3 screws located on module center line, with middle screw located between modules. To preload, torque middle screw first. Bolt carefully, by applying torque in small increments, alternating between screws. Use a torque limiting screw driver. The recommended compression for a TEC assembly is 150 to 300 pounds per square inch of module surface area. Using the following equation you can solve for torque per screw:

*
T = (C x D x F x in2) / (# of screws)

T = torque per screw (in-lbs)
C = torque coefficient (0.20 as received, 0.15 lubricated)
D = nominal screw size (4/40 = 0.112, 6/32 = 0.138, 8/32 = 0.164)
F = Force (lbs / in2)
in2 = Module surface area (length x width)

Check torque after one hour and retighten if necessary. Use Stainless Steel Screws, fiber insulating shoulder washers, and steel spring (Belleville or split lock type) washers (see sketch in Assembly Tips).

CAUTION

1. To assure good thermal grease interfaces there should be no bowing of either surface due to torquing. To prevent bowing, apply less torque if one or both surfaces are less than 1/8 inch thick copper or 1/4 inch thick aluminum.

2. Lead wires are soldered to module tabs with bismuth/tin solder (136°C). If lead wire replacement is necessary, use bismuth/tin solder.

DO NOT use lead / tin solder (180°C).





Finally, the method of greasing and tightening is critical. From Melcor's assembly manual:
The grease thickness should be held to 0.001 ± 0.0005" (a printers ink roller works well for this). Dirt, grit and grime should be minimized; this is very important when "grease" joints are utilized due to their affinity for these types of contaminants.

The insulation/seal is often most easily provided by inserting sections of closed cell foam about the cavity. It is often desirable to provide strain relief for the input leads, not only to protect the leads themselves, but to help maintain the integrity of the seal about the modules.

The use of a "spacer block" yields maximum heat transfer, while separating the hottest and coldest parts of the system, by the maximum amount of insulation. The "spacer blocks" are used on the cold side of the system due to the lower heat flux density.

[littleowl]

8 to 10mm is to thick. you should mill that down to at least 6mm. 6mm is close to a quarter inch thick and works great!
I found that the thinner copper I had didn't work very well it was just under 1/8th of an inch. or about 4mm. If i remember right.

[reflex]

Thanks guys, very helpful replies. I think I will get the spacer machined, because by the looks of things the copper I have lying around is bent! I will most probably have too tap thread on the coldside heatsink, but otherwise this appears to be a fairly simply procedure, I must just get the correct tools/nuts and bolts.

Uncle Jimbo, for the "spacer block" do you normally use copper? It is quite a bit more effective than aluminum as far as I know.

Thanks again

[scifikg]

Thanks guys, very helpful replies. I think I will get the spacer machined, because by the looks of things the copper I have lying around is bent! I will most probably have too tap thread on the coldside heatsink, but otherwise this appears to be a fairly simply procedure, I must just get the correct tools/nuts and bolts.

Uncle Jimbo, for the "spacer block" do you normally use copper? It is quite a bit more effective than aluminum as far as I know.

Thanks again

I wouldn't tap the cold plate. Copper is really soft and you really need to crank on the bolts to get the proper heat transfer. Better to just drill through and bolt down the other side.

[Uncle Jimbo]

I wouldn't tap the cold plate. Copper is really soft and you really need to crank on the bolts to get the proper heat transfer. Better to just drill through and bolt down the other side.

Especially with a thin plate. You need at least 10 threads in copper, which with a 3mm 50 pitch is about 5mm. If you use a 6mm plate you might just manage it threading the copper.

[Uncle Jimbo]

Thanks guys, very helpful replies. I think I will get the spacer machined, because by the looks of things the copper I have lying around is bent! I will most probably have too tap thread on the coldside heatsink, but otherwise this appears to be a fairly simply procedure, I must just get the correct tools/nuts and bolts.

Uncle Jimbo, for the "spacer block" do you normally use copper? It is quite a bit more effective than aluminum as far as I know.

Thanks again

You don't get much loss on the aluminum because of the size, but still, Copper is better - it has better conductivity, and aluminum tends to react with copper.

[reflex]

Ok, finally got the heatsink I ordered.
It has 2 threaded holes along what should be the centre of the TEC. I havent measured yet, but they appear to be about 2.5 to 3mm in diameter. The heatsink base is 8mm thick. Will this be able to take the torque?
Also the copper spacer I am using will probably be about 25 - 30mm in thickness. Will this cause a large drop in efficiency?

Thanks for any help

[Uncle Jimbo]

Ok, finally got the heatsink I ordered.
It has 2 threaded holes along what should be the centre of the TEC. I havent measured yet, but they appear to be about 2.5 to 3mm in diameter. The heatsink base is 8mm thick. Will this be able to take the torque?
Also the copper spacer I am using will probably be about 25 - 30mm in thickness. Will this cause a large drop in efficiency?

Thanks for any help

8mm should be enough to take the torque. Can you send a picture of what you are doing? That would be useful.

[reflex]

I hope these pictures help. The reason for the rather thick copper spacer is that both heatsinks are flat on the bottom and there will most probably be about 20mm of polystyrene isolation inbetween them. I would rather not countersink the whole heatsink into the polystyrene
The spacer is going to be 20-30mm thick.
There is no isolation or coldside heatsink in the drawing yet.(The drawing is of the cooler box in one of my other posts)
Thanks

[Uncle Jimbo]

That's a pretty thick spacer. I use 10mm or so. If it is well insulated it should be fine, but the thicker it is, the more thermal resistance to the other side. You should only use what you need to get through the panel.

Also, what is on the other side of this?

[reflex]

On the other end of the spacer is a heatsink for the coldside. The pressure to the TEC is going too be provided by the 2 heatsinks.
The material I am using for the inside of the cooler is about 5mm thick and then I am going to insulate that with 20mm polystyrene. Maybe I should go for 10mm polystyrene and reduce the thickness of the spacer.
I was planning to insulate the copper spacer with neoprene and the rest with polystyrene.
Here is another sketch, hope this helps.

[reflex]

Hi Uncle Jimbo, in formula for the Torque of the screws: There is the "D" "nominal screw size", is this the diameter of the screw? I am using a metric 3mm screw (0.1181102inches). I use 150lbs/in2 for "F" and using a 40x40mm TEC calculated "in2" was 2.48.

T = (0.2)(0.1181102)(150)(2.48)/(2)

I got a final answer of 4.394
Thanks for any help once again

[Uncle Jimbo]

Hi Uncle Jimbo, in formula for the Torque of the screws: There is the "D" "nominal screw size", is this the diameter of the screw? I am using a metric 3mm screw (0.1181102inches). I use 150lbs/in2 for "F" and using a 40x40mm TEC calculated "in2" was 2.48.

T = (0.2)(0.1181102)(150)(2.48)/(2)

I got a final answer of 4.394
Thanks for any help once again

I use 3mm stainless and usually torque to 7, aiming for 250 lb/in2 so your number looks right.

Now that I see what you are doing, I think the 30mm spacer will be fine.

[reflex]

Ok great! I just wanted too confirm that I was calculating the torque correctly. I think I will use around 200lb/in2.
Thanks for the help, I will probably give updates on how it is going either here or in my cooler box post. Just waiting for the spacer to come back from the engineers