Max Wiring Amps
 

In summary: standard power supply wiring sucks for Peltier use, so using multiple wires, or replacing them with larger wire and better connectors is the best solution.

I keep reading about Peltier users with hot wiring. This is a waste of vital cooling power! It isn't safe either. There are some things that shouldn't be overclocked!

How many Amps can wiring handle?

I did some digging. The ATX specification is available from:
http://www.formfactors.org/
It specifies Molex connectors, which info on can be found on the Molex site:
http://www.molex.com/

Our power supplies use:
Molex 8981P - Female 4 position housing
Molex 8980P - Female crimp terminal
Molex 8981 - Male 4 position housing
Molex 8980 - Male crimp terminal

The Molex Product Design Specification states:

Current Rating:
14 AWG - 10 Amps (120 Watts at 12 Volts)
16 AWG - 10 Amps (120 Watts at 12 Volts)
18 AWG - 8 Amps (96 Watts at 12 Volts)
22 AWG - 6 Amps (72 Watts at 12 Volts)

Operating Temperature: -10 C to 85 C

A wiring-only rating of a 30 C rise yields 450 Amps/sq cm, which is about 4 Amps on 18 AWG. This is inline with the oft repeated max value of "about 5 Amps" for power supply wiring.

On a typical supply, an Enermax EG651P-VE, the wiring is 18 AWG and rated for 90 C.

Good 50 Amp continuous connectors for 13 AWG can be gotten from Radio Control hobby suppliers under the name of: Astro Flight Zero Loss 2-Pin Connectors. There are other brands too (Deans = 12 AWG, etc.). You'll also find high temperature (>150 C) silicone wire with high strand counts (more flexible and durable).

I'm shopping for a Peltier supply, and will be using thick welding cable (more strands than battery cable, so again it's more flexible) to avoid voltage drop and wasted cooling power!

No point in using wire thicker than what is attached to the pelt itself.

Yes there is!

Because of the resistance of the Peltier's wire, you should have trimmed it as short as possible, and splice in some thick wire. The Peltier is supplied with thin wire that makes installation easier (goes beteen layers of Neprene = easier to insulate).

My Peltier came with heavy-duty wiring, as I didn't get it from a PC cooling supplier.

Also, if the Pelt power source is a distance away from the Pelt (like mine is). The longer the wire, the greater the voltage drop! Using thicker wire compensates (lower resistance).

A 226W Tec only uses 24amps (Imax). Welding leads are 1/2" thick and designed to hold couple hundred amps. I just don't see your logic here. 8ga wire should be plenty good.

http://www.powerstream.com/Wire_Size.htm

Btw, this is also where a power supply with remote sensing wires comes in handy.

For those that didn't know what the S+ S- connections were for, and why they are shorted to V+ V- on most supplies ...

The higher the current draw, the greater the voltage drop for a given resistance (Ohm's law: V = IR).

The remote sense wires carry a very low current, typically 10 mA or less, so the resistance of the wire connected to them is less critical, though they have to be shielded. The remote sense connections should be hooked up as close to the Peltier as possible. Doing this causes the power supply to ignore the resistance of the high current providing wires! Essentially the remote sensing cause the supply to put out a bit more voltage to compensate, so the Peltier gets all the power it craves!

When the Peltier is close to the power supply, the remote sense connections shorted at the supply are adequate.

First, you don't have 8 AWG on the Peltier (or do you? That would be much heavier than what I have seen is common)

Second, 8 AWG is fine. I'm gonna use welding cable (2 AWG) 'cause I have it on hand, and it is nice and flexible, and I have a futher distance to go.

Third, you are assuming a single Peltier :)

Using http://www.stealth316.com/2-wire-resistance.htm at 6 feet and 24 Amps:
18 AWG - 0.9384 Volt drop (22.5 Watts lost)
8 AWG - 0.0912 Volt drop (2.2 Watts lost)
2 AWG - 0.024 Volt drop (.6 Watts lost)

Now in my case, my power supply is on my workbench, a good 15 feet away, so I'd have about 3X these losses

Sounds like your all over it. Post your progress as you go. We need some extreme cooling projects around here. :)

Well, Peltier's suck at being extreme cooling ... but they sure are a heckuva lot of fun :D

I'm currently shopping for a 50 Amp supply. I have my eye on a bench model with RS-232 control and it's 1 kW, 0-20 Volts (only 11 Amps AC). It will cost a bunch, but I always wanted a good bench supply.

FWIW, a nice supply that is on my "B" list: BK Precision #6192, 3-15 Volts, 40 Amps, $299 (at Mouser and others). http://bkprecision.com/www/np_search...g+Power+Supply

Yeah, I use 8AWG wire for my TEC, and it doesn't get hot, but the solder joints from it to the tiny leads coming out from the TEC, and those leads themselves, get quite toasty during operation. Looks like a good idea.

FWIW... those Molex connectors are rated for a max of about 9 amps, per pin (last I checked).

I've never seen the wiring amp limits, never cared to (except for my PSU array). The main issue with any PC is the Molex connector, wether it's the mobo or peripheral connectors.

Problem with the molex is that their not always quality made. All it takes is a pin to not make a good snug connection and burn it out. I had one do that and took out my hard drive with it. It didn't get a good connection and that pin heated up burning the molex including a trace across the hard drive pcb. Better to run a straight length of wire for the pelt and be safe.

Yep, exactly. I don't know why Molex is still used, I guess it just got popular. I've seen many, many burnt mobos because of bad connections.

To be avoided, or at least carefully picked. Solder is always friendlier. :D

Ben I know that you know this, but some readers may not:
Solder is not a good conductor!
Use a good physical connection (I use crimped noninsulated but splices) and then sweat the connection with good solder to fill the voids -which does boost connection efficiency.

-->just my 2c

Another reason is to avoid disaster in case of a short circuit. The larger gauge wire will heat up, but it won't explode like a fuse, or turn red hot and drip molten insulation.

FWIW, I got a MFJ-4245MV (9 to 16 Volts, 40 Amps @ 13.8 V) for experimentation. The voltage adjustment circuit is trivial (schematics are in the user manual), so I'm planning to add a constant temperature control circuit to it. The other "real" supplies are still on order.

Oh, and I picked up an Ameritron ICP-120 Inrush Protector so I don't blow fuses on power-up until I can hire an electrician to wire some high current AC circuits. The ICP-120 limits the power-up inrush to 12 Amps, which dims the light, but doesn't force another trip to the panel.

It's not that bad. Solder is 8 times less conductive than copper (4 times less than 6061 Al), but that doesn't make it unable to carry a decent current.