My very UN-scientific Test of Air Cooling at 172 watt TEC.

decieded to give a try at cooling a peltier with Air, simply cause I am curious and I wanted to see what would happen.

now I have a 80, and 226 sitting around here, but both of them are in my other PC, and the 172 is the easiest one to get to cuse it's on my videcard alone in my main PC.

so instead of doing a few hours of work by removing, and putting them back in later.
I decieded to take my videcard one, cause thats the easiest.

so...
I scround my Volcano 9, a spare 50x50 cold plate, and a 172 watt pelt and a clamp.

the volcano 9, then TEC, then cold plate, then a clamp holding it all together and taking temps with direct contact, with a Lian Li temp sensor.

heres what I came up with.
after running it for a half hour, these are the temps I had.
with the volcano 9 smart fan 2 running at full speed.

Room Temps = 80F

heatsink temps = 98 F

Cold plate center = 71 F.

Now, as you can see the heatsink was very hot, it was even very hot to the touch, but the cold plate was actually pretty cold, but at 71 F, it suprised me, I thought it was much colder cause the cold plate in the center was really cold to the touch.

for reference, heres the temps of the heatsink alone, with no TEC on it.

Room =80F.

Heatsink= 77F.

Heatsink copper core, the bottom of it = 76 F.

so the temps of the center of the copper cold plate, with the TEC only drop 5 degrees, from the temps of the core of the Volcano 9s copper core of the heatsink alone.

I suspect a lower wattage TEC like a 120 or 80 watt TEC, mixed with a heatsink would produce better results, cause a 172 is probably being heated by the hot side so not really allowing the cold side to really do its thing.

so later, I will try a 80 watt pelt in there, air cooled.
again, just for fun.

well it was fun to try, but I think it wouldn't be worth the effort for a few degrees only.
I also suspect a lower wattage TEC would cool better.

anyone else here try this?
if so, please post some results, it would be interesting to see.

I dont understand, the cold side was 71f???

hmmm, i think i might agree, i would also consider this a very unscientific test... room temp was 80 degrees Fahrenheit - the heatsink, sitting in room air, with a fan blowing room air onto it, was 76/77 degrees fahrenheit... my science is rusty but i venture to say thats impossible.

as for your confusion warlord, 9 degrees below ambient on the coldside makes sense to me, he probably couldnt even touch the heatsink because the hotside had it heated up so much... is there something that surprises you? i was gonna explain something but im not sure what it is that catches you, i know you know atleast an okay amount about pelts.

haha I guess Im just used to cold side pelt temps to be atleast in the negative c range lol. I could never understand pelts to well, they confuse the shiz out of me. Isnt there supposed to be a
-60c deferense between hot and cold? he is only showing about 30f.

I really don't have any experince with pelts at all, but tyhose numbers didn't make any sense to me at all. I would think that the heatsink would be 60+ C while the cold side would be doing maybe a little better or a little worse than a Heatsink alone at full load. Idle might give decent numbers, though.

Edit: Here I was thinking he had it on a CPU...

also, wouldnt this go under extreme cooling?
but no none of those numbers sound rite, the heatsink would not be below room temp, and how are you holding the temp probe to the base? if your pushing it to the base with your finger or something else that absorbs or lets off heat, then you got it warmer than it it actually is, tape it on before you start,
also, if its a round copper core heatsink then the outter part of the peltier isnt being cooled right?

Yeah, I guess I just assumed it was the 7+... This has to be a load of crap.

Edit: My appologies for the wording.

You should be getting something like the rated 68c temp difference between the sink and the coldplate. If your coldplate is 71 your heatsink will be around 120-130ish. A 120 watt running at 12v instead of 15.2 will give your roughly 95 watts of heat displacement. Asuming you have a heatsink worthy of cooling the 95 watt TEC and whatever your load wattage is you will get some decent temps. TECs are fine aircooled as long as the heatsink has massive cooling capabilities. My 675 watt TEC chiller uses 2 way oversized heatsinks and 2 120mm Comair Muffins to cool the hotside. I dont think any sinks currently used in PCs are capable of cooling a large TEC so you might start looking around.

hahha, guys, this is a Air cooled 172 watt peltier.

try it yourself.
you can't cool a 172 wattt very well.
plus, people please read the details...
this WAS NOT hooked to a computer, it was CLAMPED together with a standard house hold clamp that I got from Lowes.
the temps were taken with a Lian Li temp sensor, put with direct contact to the metal.
metal is always cooler then the air around it, atleast according to the temp sensor I have.

the difference between using the Air cooled Pelt, versus the heatsink alone, was FIVE degrees on the core of the Heatsink, again taken with direct contact sensor.

if you don't believe waht I did, go take a 172 watt pelt, a 50x50 dangerden cold plate, and a volcano 9 and clamp it all together, and do it yourself.

take a spare PSU, take the 3-4 pins of the ATX connection and a bobbypin and connect then. do this on the CLip side, to left of the clip.

plug the heatsink in, with everything clamped together, and this is what you get.

Kunaak

You are not making sense.

You said the heatsink was 98 F.

98.6 F is body temperature, yet you said that the heatsink measured 98 F and was very hot.

A piece of metal sitting in ambient air for a long time will be at ambient temperature. If your thermometer consistently measures metal objects as less than ambient, it is most likely because your thermomer is a POS. It probably uses a thermistor which is heated by the current flowing through it. When it is in the air (insulated) it reads one temp. When it is touching metal (heatsinked) it reads a lower temp.

ok, have you ever even used a TEC?

look, TECS work by a simple concept, take two dis-similar metals, charge them with electricitu, and cool the hot side and the cold side will get very cold.

now, just like when you turn on your oven burners a TEC can get well past 220, or more.

Regaurdless of ambient tempetures.

this means, you take the Hot side, and put it to the Heatsink, the Heatsink Attempts to cool it, so the cool side can cool as it's suppose to.
The heat from the Hotside, will heat the heatsink, well above ambient temps raguardless of the temps of the air around it.

if you cannot cool the hot side, the cold side can't get cold.
so what happens is, the volcano 9 is almost able to cool a 172watt pelts Hotside, giving the cool side, alittle room to actually get cooler then the air around it, so it's mildly effective, but not so much so to actually use.

does that help?
incase it doesn't, go take a butter knife, turn on your oven, stick the butter knife in the grill and watch waht happens.
it gets hot...
regaurdless of the air around it being cool.

I think what everyone is saying is that either the cold side would get too cold to touch (exageration... really cold), the hot side would get REALLY hot (regardless of how effective it is on cpus), or the TEC would self destruct.

The fact that you just have a small disk on the bottom of that heatsink that can't cover the whole TEC is a sign that something IS wrong here. I don't know what and really don't care, but in reality the temps SHOULD have been different.

...I really don't know why I'm arguing this, but hey, something just sounds fishy.

you mite want to be careful to who your asking that question... since87 is pretty dang knowledgeable in this area, probably more than you could immagine....
your readings are way off,
if that heatsink was at 98f, the cold plate should have been around 40f at least,

Also, my compunurse is reading a peice of PURE SILVER 2c above ambient just sitting here, its over 2x the conductivity of aluminum....

Kunaak said:

ok, have you ever even used a TEC?

Click to expand...

Yes, but that's kind of beside the point.

The point might be that you don't know the difference between degrees Fahrenheit (F) and degrees Celsius (C).

98 degrees Celsius is "hot" - near boiling.

98 degrees Fahrenheit is body temperature.

So perhaps you can see that most people would not consider 98 F to be as you said, "very hot, it was even very hot to the touch"

Kunaak said:

now, just like when you turn on your oven burners a TEC can get well past 220, or more.

Click to expand...

Would that be 220 degrees Kelvin?

I imagine if I plugged a pelt into a wall socket, parts of it would get well over 1000. (At least briefly.) So what?

You are providing near meaningless data. Rather than try to explain to me how TEC's work, you might consider trying to clarify your earlier confusing statements.

ok, never mind you all are right I am a liar, I had something to gain here by lying to you guys. you caught me, you all ganged up, hardly heard a word I said then give me crap, yet don't even take the time to use common sense.

put the hotside of a peltier to a heatsink and tell me it doesn't get to 100 F.

don't worry, next time I take a few minutes to try something fun, then I won't bother telling a bunch of people who have no idea how a peltier works.

you guys obviously assume you just plug it in and it gets cold automatically.

yes, your all right, you plug in a peltier and automatically it gets to 40 F, it's that easy to get good cooling, nevermind that swiftech uses the craziest heatsink on earth to get ok, cooling, it's not necessary, cause peltiers are magical.

they don't require you cooling the hotside for the cold side to work.

MOD GUY< just delete this please.
I am sitting here getting called a liar by people who don't even seem to understand how a peltier works.

sure my body heat can make a heatsink burning to the touch, yes thats possible...
if I was the human torch.

sure my temp sensors can be wrong...
like the 3 other temp sensors I have around here that all show the same temps....

obviously.

Kunaak, I am just curious.....you were talking about Fahrenheit temps and not Celsius?

Dude, calm down.... Can you understand that it makes no sense to us? Perhaps you are doing something right, but it is not coming through here.

Absolutely no one has questioned the fact that the hot side will get to 100 degrees F... We just think that it got to a much hotter 100 degrees C.

or

That the cold side got much colder... Like froast forming cold.

A lot of us are familiar with peltiers but haven't used them. I've read numerous articles and posts on the theory.

I guess what everyone is saying is that the info you have provided us does not add up to what we have read. I really don't know why. Perhaps it IS the round thing on the bottom of the heatsink. Maybe that is making the pelt less efficient. Who knows.

Anyway, all that really mattered in the first place is that you had fun. Don't let us stop you from experimenting. Just try to really watch the numbers and compare to other's findings.... At least you have the TEC's to play with!

I have none

I dont think anyone is calling you a 'liar' its more you thats confused... Since87 probably knows more about peltiers than you can ever immagine, I have had experience myself, those readings are bogus, There is something horribly rong with them.... thats no big deal untill you try to tell us they are correct, everything is corect when it is in no way correct...

It is possible that the TEC got fried due to inadequate hotside cooling. TEC's use very low melting point solder in their construction. If they are allowed to get too hot, the solder will melt and they can be damaged.

That may be an explanation for extremely poor performance.

Another thing that hasn't been mentioned is the voltage applied.

A picture is worth a thousand words...