The problem is that with so high forces materials needed would need to be very 'strong' = dense = good thermal conductors :). There are two ways of increasing thermal resistance, either change material or reshape it :)

i was planning on using 6 or 8 5/32" wiborth screws

would that be enough?

When I was looking into TECs most manufacturers recommended stainless steel for clamping. I think a nylon bolt would strip it's threads trying to get even a fraction of the required pressure.

how thick should be the base of the block to support such a pressure?

because here i was told that thinner is better in this case, but how thin would be too much?

Depends on the design of the block. You can have a fairly thin base if the structure is rigid enough.

Not really if you use something like #4 bolts. I dont think nylon bolts would be able to accomidate the high clamping pressures needed for a peltier block.

as long as the structure supports the base you can go pretty thin, like if your pins on the rotor block touch the top of the block it can go thinner, I would go with about 1/16" just to be safe.

I really dont think youll be able to get adequate clamping pressure with nylon bolts. See if you can find some other high tensile strength plastic bolts. do they make polycarbonate or delrin bolts? something like that.
On my block Im using 8 6-32 screws, clamping pressure is amasing I was conserned with heat transfer, but I just decided to try it and I can always change it later if necesary.

Jon

not planning on using nylon bolts here.

dont know who bring it up, but it wasnt me :)

i guess ill use some metal screws, and isolate them too.

Water side, yes, the thinner the better but with high wattage applications large copper-coolant surface are is required so thin but with lots of surface area.

Die side is a different ball game, the thicker the better. The reason in short is spherical shape of thermal energy dispersion pattern and large capacity (thermal inertia) created by thick cold plate between TEC and die.


-J- , I think your washer idea may work, it depends upon the strenght of material washers ae made off. If they are not strong nuff they will crawl out from under your bolts, use of steel washer inbetween may help there - going from bottom to the top block-plastic washer-steel washer-bolts head. Bolt musn't touch the block (hole must be larger than bolt's diameter) and your washers need to be quite thick *you may use more than one thin one)

If theres any condensation build up, itll probably be in the bolt holes and that would totally defeat the purpose I would think.


Jon

You are right about changing thermal resistance, that's why I suggested changing materials. Reshaping is not always an option, so you should really be considering thermal resistivity, or conductivity, which are material properties and are not affected by shape.

That's not exactly a good correlation between strength and thermal conductivity, density has little to do with either. And from my calculations, by applying 300PSI of pressure to two 50mm^2 peltiers, you would need a #4 washer to withstand compressive pressure of about 10,000PSI, or ~70MPa in "those other units". According to this materials database software that I'm using, nylon ranges from 55-104MPa in compressive strength. So nylon might actually work just fine, depending on the washer's specific properties and how the pieces are assembled.

Considering both compressive strength and thermal conductivity, here are some of the viable options that the software suggests...
Polymethylmethacrylate (PMMC, Acrylic)
Polyamides (Nylon, PA)
Polyetherketones (PEEK)
Titanium - only 5-12W/m-K and definitely strong enough
And then some kind of wacky materials that would never fly due to manufacturing issues... Brick, various glasses, bamboo, stone. :D

thx for that zero

It looks like I didn't make myself clear here.
If you are stuck with particular geometry only chnaging material can do anything towards altering thermal transfer rate, quite obvious.

If change of geometry is possible it is qiote obvious that with changed contact surface (cross-sectional area of connecting device) termal transfer rate is instantly altered despite using the same material.

You are right that thermal resistance of a material is not changed by it's shape. We are not looking here at lumps of shapless raw materials but particular DEVICES made using dfferent materials.

Look at the big picture and any software is ust a tool for making some calcs not finding solutions. I use my brain to do it ;)

No, I said that thermal resistance IS changed by shape. Resistivity (taken as the inverse of conductivity) is not affected by shape. This is just like electrical resistivity, which is a material property for, let's say, copper. That does not change, but you can increase the resistance of a copper wire by increasing its length.

And I did use my brain, that's how I eliminated the obviously unrealistic materials that still meet the strength and thermal conductivity requirements. Obviously brick, bamboo or glass would not be very easy to come by in washer form, but nylon or other polymers are more realistic.

Sorry d00de if I had made any comments offending you, wasn't my intention:(
Sometimes peeps are fascinated by diff sotware doing simulations and cals and treat it as Bible almost, not mere tools which they are.
Eeach and every theory in physics was arived at through abstract mental models and equations are only universla representations of such models - just tools.
Taht's what I meant and apologies once agian if you felt offendend thinking that I implied you were not using your grey matter :) Sorry m8!

Jabo's point is valid
if you used software and it provided those candidates, then it is crap (or misused)

what you want is an AI matls program, do it correctly in a flash
or you could ask someone who knows, do it correctly in a flash too

phenolic
$0.75 ea in lots of 3000 pcs
no detectable difference in testing
punt

No no, no offense taken, I was simply trying to clear up my point. For the past 6 months I have been taking materials selection classes using that software. I know that it is not always right, that was one of the main points of my learnings- that you have to think beyond the output of the program because it cannot consider all factors like we can. For example, it often suggests stone or brick for relatively strong materials, when in reality they're no good for making small parts, however great they are for large structures. Bamboo is another strong, light material that matches many high performance metals in performance when considering those two factors. But I have yet to see airplane turbines built from bamboo :D

Bill, I don't think the software is crap. All I did was plot strength vs thermal conductivity for a set of standard materials, not a lengthy analysis involving materials indices which would have been the best way. That still wouldn't have considered the manufacturing or cost factors, which is where Jabo's "use the brain" theory comes in.
So you tested phenolic washers versus what washers? Or bolts? No difference would not surprise me, it was just a fun materials engineering question for me.

ok, understand about the software
but there are good programs, dunno who can afford them though

used nylon washers for testing (creep is why nylon itself is not a good choice)
within experimental error with or w/o
ss bolts are a poor thermal conductor to begin with

The software is Cambridge Materials Selector, it has a pretty extensive database. I don't know how much it costs, my school has some kind of site license, but I know I wouldn't be able to afford it!