I just bought a peltier, now what?

unsmart

Distinguished
Dec 30, 2005
210
0
18,680
I will soon have a Nord TM-127-1.4-8.5 tec. It's rated for 15v 8.5amps[127.5watts] but I will be running it on 12v so the watts are 102. The temp rating is -40 to 85c, With a max temp diff of 68c on the sides.
I won this on ebay last night for $6 so It's no big deal if it's worthless, I'll just mess with it till it burns up. I don't have high hopes since the little I read online about this, was older cpu's and nothing about inline water cooling. If it was worth doing it would be more common I think.
The way I understand it.
The watts of the tec[peltier] have to be more than the heat watts of the thing to be cooled[ 102w tec -89w cpu=-13w].
You have to get rid of the heat from the tec and the heat source[102w+89w=191w]on the hot side.
With a max temp diff of 68c the hot side temp would have to be 68c to give you 0c, 55c on the hot side gives -13c and so on.
Thats what I think is true about peltiers.
The sites I've read talked about cooling right on the cpu but they where all old ones[p3s] with lower watts then my 3400+[89w stock]. So my plan is to use the tec in a water cooling system after the radiator so the waters already cooled some. I don't plan on super cooling just getting below or to room temp.
I've never used water before ether so it's all new to me, I've seen some small sealed systems that I'll base mine on. they use lower flow pumps and smaller radiators with good results.
What I need to know is.
Is what I'm think wrong or am I missing something?
What does 1c = in heat watts?
Is there a system retail or home brewed that I can work off?
Is there a site that shows systems using peltiers.
Can this work?
If you have used a peltier or know a simple way to use it in cooling please post.
 

unsmart

Distinguished
Dec 30, 2005
210
0
18,680
Thanks for the input but my tec isn't rated[102w] high enough to directly cool my cpu[89w stock but oced, so more like 100w]. If I got a higher watt one, I think it would boil the water. The question is, can I use it in a water system to chill the water after the radiator. I don't think it's strong enough to cool the water without the radiator as a buffer either.
 

RichPLS

Champion
Yeah, you could cool the water, but you need a waterblock and a heatsink to cool the peltier...
How much are you trying to squeeze out of system???
Watercooling alone will allow you to overclock more than is advisable...
The Peltier used as you describe will just burn energy and generate heat...
 

unsmart

Distinguished
Dec 30, 2005
210
0
18,680
It's more just to do it than a real need to cool that much.
I will be cooling the peltier with Thermaltake Dragon Orb 3. I know thats not how to do it but it's free.
 

_Cosmin_

Distinguished
Jan 19, 2006
424
9
18,865
I have some questions too:

1. How reliable are those Peltier Elements?
2. Why don`t you use just the CPU - PELTIER - Big Heatsink (like Zalman CU-S7000 without fan) ? Since the peltier cools your CPU and Heatsink only have to cool the peltier... i think you could get some good temps... question is ...if the peltier fails...and how often!
3. Since the peltier give sub-zero temps. why is only 6$ ? I know that thermoelectric are better than watter wich is better than air flow... so why is not more expensive?

Thanks
 

linux_0

Splendid
I would have to agree with _Cosmin_

Zalman coolers, especially the all-copper ones are very effective :D

A Peltier for $6 sounds fishy... maybe it's "hot" I dunno...

No pun intended but laugh if you wish ;-)

I hope it's not DOA. Please let us know when you get it.
 

dstaple

Distinguished
Dec 14, 2005
19
0
18,510
I've done this before and it wasn't anything special.

It's rated for 15v 8.5amps[127.5watts] but I will be running it on 12v so the watts are 102.
There is no reason to believe that this is true. For example, for ohmic devices (like a resistor etc.), power goes with P = V^2 / R, so cutting the voltage from 15 down to 12 would reduce the power to (12/15)^2 times its original value. Peltier coolers are definitely not ohmic, but they are semiconductors, so I wouldn't believe anyone who told you to calculate the power this way. The only reliable way to know would be to try it and check.

I also ran mine off 12V, and was curious what the difference was, but never bothered to check. Your (12/15) * 8.5 might be an acceptable approximation.

The watts of the tec[peltier] have to be more than the heat watts of the thing to be cooled[ 102w tec -89w cpu=-13w].
Note that just because your peltier draws 128W of power doesn't mean that it's moving 128W of heat. This is a problem in thermodynamics, and depends of the efficiency of your peltier, as well at the temperature difference between the hot and cold side.

Again, this is just a rule of thumb, because of what i stated above. In general, the higher powered the peltier, the better your results will be.

You have to get rid of the heat from the tec and the heat source[102w+89w=191w]on the hot side.
This is 100% true and the biggest reason that these rigs suck.

With a max temp diff of 68c the hot side temp would have to be 68c to give you 0c, 55c on the hot side gives -13c and so on.
100% not true. "Max temp diff" refers to the max operating range of the device. This the best you could expect if you were just cooling, say a block of metal that produced no heat.

What does 1c = in heat watts?
This is an impossible conversion. Heat and temperature are not in the same dimensions, so you really can't do this (aka, what's one hour in pounds).

On the other hand, most cooling systems have a response that the temperature above ambient on the cold side is roughly linear with the amount of heat being dumped into the system. This allows you to make estimations like: "If my 100w chip puts my heatsink 30 degrees above room temperature, then my 200w chip+peltier combo will put that same heatsink ~60 degrees above ambient."

As for info about my own system,
i seem to have lost my notes, but my peltier was 40mm * 40mm (same as the top of an Athlon 64), and was rated for 225W at 15V, 15A. Like I said earlier, I ran it at 12V, so I estimated that it's power draw was rougly 12V*12A = 144W.

First I tried sandwiching the peltier between my cpu and my water block using arctic silver. This gave a good results when my CPU was under no load, but as soon as I overclocked / ran prime 95, it only gave me, at best, a few degrees C advantage over the liquid setup. One of the biggest issues was that the <100W of extra heat dumped into my water raised it's temperature, which was pretty much worse than the extra cooling power.

I later tried cooling my water with this thing, just for fun, but it really didn't make a difference to the temperature of the water.

Anyway, the whole thing was kinda fun, and I did manage to get some neat sub-zero temperature readings when I underclocked / undervolted by chip (800MHz, 1.35V ?).

This might be more effective cooling somthing that put out less heat than a cpu.
 

Crashman

Polypheme
Former Staff
There was a guy in here using his Peltier to run a PIII 500 at 1000MHz, his pelitier died and made big heat, melted the whole socket! I'm not sure how reliable they are, but that's enough to scare me off.

He got his for $6 because he took a risk buying it on ebay.
 

unsmart

Distinguished
Dec 30, 2005
210
0
18,680
i got the peltier last week and did some testing. It is sealed and does work. I didn't have time to hook it up to one of my old psu yet,I tested with a car battery charger. I ran 12v an 4.7amp to it. It was between two xp heat sinks hot side is 40c and the cold side is about 20f below room temp, I didn't realy check just a guess.
My plan was to put the hsf out side the case using the top fan hole and place a block of aluminum on the cold side with grooves cut into it for the water. The flow would be form the pump-tec-cpu-radiator-pump with the pump at 70-80gph. that was the plan. Now I'm gitting two radiators and can't find my pump. I did find a big ass pump 332gph and a aluminum case that fits in a drive bay. So no room for the tec and small psu I was going to use. After your inputs and some research this does'nt seem wroth the trouble sence the temp drop would probley olny be a few degrees. Thanx for the help. I my stil try it after i have my water going.Right now I'm trying to git the parts picked out for a new biuld.