226w peltier running on an ANTEC psu.

Results of 226w peltier and an ANTEC psu.

I am on the quest for the ultimate psu for a 226w tec. One stipulation it needs to be watercooled. I had a lot of ideals that were, well, not good. (thanks to diggrr, he has answered many of my questions(and thanks to all of you guys and gals))

I have an Antec true 550w PSU. It is currently running my 226w tec just fine. It is not even warm to the touch. Here are a 1000 words.....
PSU stats

The setup

Wireing

Hour and a half later

snowy

voltage



Notice no minimum amperage on the any of the lines except the 12v. The 12v rail is a bit low but i believe it will be fine. Water temp is about 85f, ambient temp is about 75f, and PSU is only slightly warm on its neath side. ( I can barly feel any warmth) The wires leading from the petier to the PSU are slight warm also. I am going to gut her out and replace with extreme gauge wire. She is going to be cooled Bladerunner style. http://www.zerofanzone.co.uk/content.asp?request=liquidenermax550w&page=1



I also have a meanwell 12vdc power supply. It works great. I can adjust it to 13.7 volts. It runs very hot. Not a solution to be WCed because the way it internally contructed.

I am going to run this setup for at least another week straight before i make my mods. I will of course post the success of the burn it. I hope it doesn't decide to give me the finger and crap out.

Can i get any good amperage out of the other rails? Or is the pelt taking all of its wattage power?

What do you all think?

Seems like your doing everything right. Just one thing myself I went with a MeanWell cause my Antec didn't want to work right. After all when your runnin a pelt you want everything working the way it should.

I think you need a load on the 5 volt line to get the 12 volt up to full power. Got a couple of fans that will start reliably on 5 volts without the center hub getting warm? The rest of your motherboard, harddrives and cd's may still need some airflow, depending on the case you choose, and the cooling mods you wish to apply. You can judge by the case temps.

I used 2 ohm 35 watt resistors on my 5 volt lines in a dual psu setup (like on procooling's article), but those big resistors aren't exactly Radio Shack parts...they're built inside of a heatsink and 1/2" x 1 1/4" long.

Looking good there, bud!

but arent u running ur 226w pelt under spec? isnt it like 16v defaults? so ur running 25% lower capacity right? or?

lt. max... true, i am running under spec. (15.1v) But I hear the results in running the tec under 13.8 vs 15 are very noticable.

186w (or so) @ 13v
226w @ 15.1

Running at 15v only increases the heat the tec puts out but does decrease the cold side temp by all but a degree or two.

So i hear.

I would like to test the temps running at both voltages. It would need to be in a better stable testing environment my room i am in.
I can test it @ 11.25 ~ 13.5, i don't have a PSU capable of 15.1v @ 25A. I am ordering another antec 550, then i can combine 12v (11.5) and 3.3v (3) of the two PSU's.

Cool, I'm interested in the results of undervolting temperatures too. Just keep in mind, things will be different (vastly) once you add that +-100 watts of cpu heat.

Frosting that slab of copper is most impressive, how long did that take?

Ct. Strangelove said:

lt. max... true, i am running under spec. (15.1v) But I hear the results in running the tec under 13.8 vs 15 are very noticable.

186w (or so) @ 13v
226w @ 15.1

Running at 15v only increases the heat the tec puts out but does decrease the cold side temp by all but a degree or two.

So i hear.

I would like to test the temps running at both voltages. It would need to be in a better stable testing environment my room i am in.
I can test it @ 11.25 ~ 13.5, i don't have a PSU capable of 15.1v @ 25A. I am ordering another antec 550, then i can combine 12v (11.5) and 3.3v (3) of the two PSU's.

Click to expand...

I created some spreadsheets that are downloadable in this thread.

These spreadsheets will let you calculate the results of varying the pelt voltage. Unless you have a very good water cooling system, the best voltage to run your pelt is going to be less than the full rated voltage. The spreadsheet lets you play with C/W of the cooling system to see the impact that it has at different pelt voltages.

[Edit: I just realized that where the spreadsheet says GPU Temperature it should say Coldplate Temperature. GPU/CPU temperature will be higher than that temperature.]

howly crap! thats a slab of copper there its sitting on! i thought it was wood with white cloth over it thats lookin really good, i gotta find one of those power supplies that has 40amps on the 5v and something like that, so i can power my cpu at those high speeds!
your going with a p4 setup now instead of amd?

It took about an hour and a half to frost over.(i think) At first there was a lot of condensation, then I played some battlefield 1942, turned around and saw a white block. I had a brick!

It has been running about 19 hours now.... The PSU warmed a little more, but not even close to hot. (warm and hot, I know they are horrible methods of judging temps)


Now the copper block is iced over.....

Current ambient temperature is 82F.

dude, thats nuts, get a closeup on the block and how its frozen to it

I can't wait to setup my pelt. Pretty damn sweet man!!

Ok before I dismantle this setup to make it more of an appropriate test setup. I wanted to show you all a few last things.

I took a dehumidifer (found at goodwill) and place the evporator in a large RES. With the system running the RES temp stays about 58F (I wish -58F ) The reason why such high temps using some phase change are because the dehumidifer is only midgrade and of course the system was slaped together in a wicked haste.

huge copper block temps hit below zero! here is that close-up...



That score was made using a 100watt soldiering iron. It kept freezing over.

The money shot


Well monday is tomorrow, it is time for me to get a life now. you know school, girls, work, play, and WCing. Thx for the interest.

peace out

thats insane

My god man, that's a thing of pure beauty. I only wish I understood all specific electrical references made, oh well. . .

damn you!!

I can not get my 226w pelt to get completely cold, much less frost over!

I have tried everything but a Antec 550 or a meanwell 350. I tried a Enermax 431w (says 20A on the 12V) and the damn wires almost melted!

I was water cooling with ICED water (8-9ºC), via 1300 pump and the water never got over 17º. But the cold side of the pelt only got to about 18º and only the left side! not even the whole pelt.

then I started to smell something funny and started sniffing around and feeling things. the damn wires on the enermax were HOT and I thought for sure my molex was going to melt.

so I put that blame on the enermax.

but I bought a meanwell 150 at fry's (12v/12.5A) hoping that it would run the 226pelt at 40%. but it did not appear to even do that.

looks like I am going to have to invest in a $150 meanwell SP300 or S320 to get my results.

but do you think that there is a problem with my pelt? Why is the one half of the cold side getting hot and only the other half getting cold?

Sorry guy.

Your pelt is toast.

If the wires from the Enermax were getting hot, it was putting out plenty of current. No reason to think that there is anything wrong with the supply.

All you PSU solutions (in opinion) are not adaquete. you need more amperage. I say purchase a PSU (antec 550 $110 or so). If that does not solve the problem purchase a new TEC.

Or the problem lies withen....
The waterblock to weak
not enough flow
not enough AS2 (3)
not enough clamping pressure. TECS like about 150 pounds of clamping pressure. I tried a TEC with LITTLE clamping pressure and performance suffered. sorry no real data, only quick test.
good luck

Since87 said:

Your pelt is toast.

Click to expand...

g-d@#!$

this is weak. I even went to JAMECO last night and purchased a $130 Meanwell S-320-12 (12v/24A)

same result, except the BLACK wire on the peltier heated up, probably over 110ºF

Any idea on why my pelt did not work? do you think I burned it up with the HSF? I dunno what I was thinking. that HSF (although a good one) is made for CPUS that handle 50-70Watts and this pelt needs cooling for ~226W!!

I just tried again with icey water (3.1ºC) and the S-320-12 power supply, but no go. That black wire gets hot and the pelt is not doing much of anything.

I do not suppose I could fix this? Any chance it is a defect and I can get an exchange (do you think?)

First off explain exactly how you first tested it.

How did you decide which side of the pelt to put against the heatsink?

Are you sure you connected the supply up with the correct polarity?

How was the pelt held to the heatsink? Did you use thermal paste?

What if anything was touching the non heatsink side when you applied power? Particularly, were your fingers in contact with the heatsink, and the nonheatsink side of the pelt at the instant you applied power?

Did either side of the pelt get hot when you first applied power? How quickly after applying power did you check temperatures?

Answer these question and it will probably be pretty clear whether or not the pelt was defective when you got it.

Are you sure only one of the pelt wires is getting hot? I think that's unlikely. I wouldn not suggest testing it though. It sounds like the pelt is shorted internally. When you hook a powersupply up to it, it is likely that the supply is cycling between powered up and current limit shutdown. Ideally the supply will survive indefinitely in this state, but I wouldn't count on it.

Since87 said:

First off explain exactly how you first tested it.

How did you decide which side of the pelt to put against the heatsink?

Click to expand...

I used a method that I read somewhere which states that the side where the wires are solder to is the hot side. This method has not failed me yet.

Since87 said:

Are you sure you connected the supply up with the correct polarity?

Click to expand...

yup - postive The red wire was positive and the black was negative. This pelt did not come with any instructions, so I relied on web-site diagrams.

Since87 said:

How was the pelt held to the heatsink? Did you use thermal paste?

Click to expand...

thermal tape

Since87 said:

What if anything was touching the non heatsink side when you applied power? Particularly, were your fingers in contact with the heatsink, and the nonheatsink side of the pelt at the instant you applied power?

Click to expand...

I had a piece of aluminum on the coldside when I applied power. I did not touch it until after I made sure the power supply was on. Then I immediately touched it too see how 'cold' I imagined it to be. So no fingers on either side when power was applied. Only the HSF on the hotside and aluminum on the other. (the Al sheet was slightly larger square than the pelt, about 2-1/8" square and only about 1/8" thick)

Since87 said:

Did either side of the pelt get hot when you first applied power? How quickly after applying power did you check temperatures?

Click to expand...

I checked the temps immediately and every second (it seemed). The heatsink got very hot, but not too hot to touch, guessing 30-35ºC Again I had the Smart Case 2 fan on it... although it was not doing well because the power supply was struggling to get enough juice to it. I assumed the same was happening to the peltier. So actually, the statement I made above ' heatsink hot' were made when I used the larger enermax PSU (before that was a cheapi turbolink 410w).

Since87 said:

Answer these question and it will probably be pretty clear whether or not the pelt was defective when you got it.

Click to expand...

ok - I am interested to hear your theory :-} on 'the dead 226 watt that brodo bought'

Since87 said:

Are you sure only one of the pelt wires is getting hot? I think that's unlikely. I wouldn not suggest testing it though. It sounds like the pelt is shorted internally. When you hook a powersupply up to it, it is likely that the supply is cycling between powered up and current limit shutdown. Ideally the supply will survive indefinitely in this state, but I wouldn't count on it.

Click to expand...

the black wire was really hot 38ºC+ the red wire was hot, but not as much, gessing ~32ºC

Thanks for the replies and thought efforts, Since87