PSU to power peltier?

Jakalwarrior

2[H]4U
Joined
Aug 6, 2005
Messages
3,189
I am wanting to power a peltier with an old psu. The peltier is ~270 Qmax but I will probably run it at half or 3/4 power for the little project I am putting it in. I just have a few questions about using a dedicated PSU to power a pelt.

Can I cut 3 molex lines off and combine them to make 15v? or are they all on the same line and would still just be 5v? If not can I combine the 12 volt rail + the 5 volt rail and use a potentiometer to scale back the voltage? The thing supposedly maxes out at 16 volts and 32 amps but I plan on running it lower than that since I would have a hard time keeping the hot side cooled if I went that high.
 
Can I cut 3 molex lines off and combine them to make 15v? or are they all on the same line and would still just be 5v?
Doing that will cause the different cables to run in parallel and you'll end up with 5V.
If not can I combine the 12 volt rail + the 5 volt rail and use a potentiometer to scale back the voltage?
No, you'll end up just forcing current from the +12V wire down the +5V wire.

The only way to get more than 12V is to ground the +12V rail to the -12V rail, which would give you a total voltage of 24V. Also, be aware that most PSUs won't take very kindly to being loaded only on the +12V rail, and they might not work altogether or they could become damaged by doing so. The best way to go would be to get a dedicated single-voltage DC supply or to connect the TEC to a PSU with other computer components connected to it already.
 
While 24v does sound good I haven't seen any PSUs with more than just a tiny bit of amperage on the -12 :( I'm going to be running ~15-20 amps depending on what PSU I can dredge up.

Would running a couple fans on the 5v lines be enough to keep the PSU happy? I was going to cool the hot side with water but I'm sure I can find something for the fans to do.
 
Get a meanwell psu. they sell them at dangerden.
 
They're readily available on eBay as well. Just do a search for '24v switching power supply' or similar and you'll come up with lots of suitable units like these:

http://cgi.ebay.com/15V-DC-23-2A-35...in_0?hash=item5633dfb867&_trksid=p4999.c0.m14
or
http://cgi.ebay.com/24V-DC-14-6A-35...in_0?hash=item5634b90710&_trksid=p4999.c0.m14

The same 12V unit sold at dangerden is also available at a much better price:

http://cgi.ebay.com/12V-DC-29A-350W...in_0?hash=item3a51f941a4&_trksid=p4999.c0.m14

I also wouldn't suggest running such a load on a PC PSU, though if you were okay with just running at 12V, adding a load resistor to the 5V rail to waste a couple W is usually enough to get stable operation. It is apparently fairly easy to modify most PC PSUs for floating operation, and you could then connect two of them with the 12V rails in series to get 24V, but since it means modifying PSUs and buying two of them, the cheap dedicated supply makes a lot more sense.
 
As an eBay Associate, HardForum may earn from qualifying purchases.
I will probably be ok with just a 12v since the thing maxs out at 16 and I have no intention of running it full tilt 545watts. I am still learning though. Is there a way to use the 12v line to pump out full 12v but limit the amps? A pot would reduce both the voltage and amps right?
I want to feed the thing 12v just not max amps since the PSU would likely blow up. I'm probably going to be using a cheap POS, maybe even a pull from an old OEM. I have lots laying around. Don't care if I blow them up eventually but instantaneous flames or shutdown would kinda suck.

I'm just building a novelty drink cooler :D you know, for when your beer gets hot. Instead of drinking pisswater or throwing it out. You set it in a short aluminum cylinder attached to a pelt and wait a minute for frosty beer once again. If it works great I might one day buy it a dedicated DV power supply but I'm trying to make this from spare parts and a pelt.
 
I don't know exactly what it is, but there will be some formula for peltier devices that defines how I (current, amps) and V (voltage) are related. For example a purely resistive device will exhibit Ohm's law, or V=IR (where R is the constant resistance), and I believe Peltier devices follow a similar relationship. More complex relationships are possible, but basically, no you can't just limit current without changing the average voltage as well, they are directly related. And a pot is absolutely not suitable here, the beefiest ones available are designed to dissipate a few W at most. You'd need a giant rheostat. You can just run the peltier at a lower voltage, which should also lower its power consumption; if you have a datasheet on the peltier it may give you an IV curve and you can get some more information out of that.

If the voltages available in a PC PSU are not suitable, you could build a simple PWM controller with a few dollars worth of parts, or get some power resistors and burn off lots of energy as heat (probably counter-productive, eh :p).
 
Will a dimmer switch for AC lighting work on DC?

My only concern though is that if I plug it straight in to the 12v line on the PSU it will draw waaay to many amps and either shut it down or blow it up
 
Will a dimmer switch for AC lighting work on DC?
No, these will only work on AC.

My only concern though is that if I plug it straight in to the 12v line on the PSU it will draw waaay to many amps and either shut it down or blow it up
If it's rated for 16V, you might overload your PSU, but that should just cause it to shut down immediately or overheat and shut down. You can run it on 7V using the 5V rail as the low side.
 
you know you might just consider buying a 120-to-12V transformer and a small bridge or full-wave rectifier. Pelter plates will function perfectly well on an ugly waveform.
 
I got a server PSU that is made to run on 12v line and only has a 5v as "backup". 575 watts all on the 12v line so I guess that means ~48 amps. Cost 17 or 18 bucks shipped.

Now I just have to figure out what to do with the nintendo cartridge looking connector and hope the AC side plug is the same as an ATX one.
 
Do you have a link to pics or a datasheet?

Probably you're going to want to just cut off the existing output connector and I'd wire some quickconnect terminals to the 12V rail (if you don't know which wires are which you'll need to get a $5 multimeter) and matching terminals on the peltier. Make sure to use beefy wires & QCs or more likely bundle several of the existing ones together; the ~17A you should be drawing with this setup is nothing to sneeze at and you'll want to keep the wire resistance as low as possible.
 
Ah, looks pretty simple, just a 3 terminal edge connector. The fat terminal in the centre is almost certainly ground; you'll need a multimeter (or it might be silkscreened) to figure out what the other pads are.

Once you've got that mapped out, assuming the top and bottom match, I would drill a hole through (don't use good bits, the fiberglass isn't good for them) and bolt ring terminals to them to attach your wires.

If it's 270W nominally at 16V, P=IV so I=16.875. Since V=IR (approximately, apparently this holds for peltiers) where R is some constant, R=V/I or R=16/16.875. R=0.95. Plug in V=12 and you get only 13A (looks like my math was wrong too). And P=IV, so P=150W. Initially I didn't consider that power follows a square rule with voltage and just used a linear interpolation. Probably worth verifying with a meter though.
 
If all I have to do is drill holes in it and attach leads that will be great Much easier than taking the thing apart. Less chance of death by electrocution too :D

Its rated for 545 at 16v while using 32 amps. Which I know makes no sense... Ebay seller math gone wrong? The real max wattage should be rated at 512 watts power consumption. The 270 number I mentioned is the Qmax, the amount of heat it can move in the best case scenario. Ive seen several Qmax numbers mentioned for this model though. Ranging from 270 - 340. Depending on whos data sheet I want to believe. I wont really be able to test that though since I have no way of telling how many watts of heat its moving. The 340 seems doubtful though since most peltiers only get around 50% efficency.

We are both using the same formula though and your math is right, just plug in 512 in place of 270. I = V / Vmax * Imax
 
Back
Top