Announcement

**etnietering** · 02-03-2010, 03:11 PM

Re: Choosing resistors for loading

You are correct in your use of Ohm's law. Something else you need to consider is the power law: P = IV = (V^2)/R = (I^2)R. 15A @ 5V is 75 watts - a lot of power to be dissipated by 2 resistors that would fit in the outlines for output caps. Even with a fan and heatsinks, you should still use resistors that are rated for about twice the power they're going to dissipate. Here's what I would do: forget the circuit board altogether. Fashion some sort of mount for the inside of the PSU case out of scrap metal, and just mount power resistors directly to that mount and the metal case (with arctic silver). For the resistors I would get:

3.3v - 3x 1ohm 20-25W (they'll each be dissipating about 11W)
5v - 3x 1ohm 50W (they'll each be dissipating 25W)
12v - 4x 5ohm 50W (they'll each be dissipating 28.8W)
-12v - 1x 120ohm 2-5W (to dissipate 1.2W)
5vsb - 1x 5ohm 10W (to dissipate 5W)

so the totals would be:
3.3v - 9.9A - 32.67W
5v - 15A - 75W
12v - 9.6A - 115.2W
-12v - 100mA - 1.2W
5vsb - 1A - 5W

for a grand total of 229.07W

so 230-235W total with a fan. This will still get EXTREMELY HOT though, so be careful! If you have extra scrap metal you can fashion extra heatsinks for the inside of the case. For this much power (especially on the 12v) you might want to add a molex accessory connector or two in addition to the main ATX connector for getting power to your unit.

You should be able to find resistors on digikey. High power wirewound resistors like this for direct chassis mounting, or to-220 or to-247 package resistors like this for heatsink or direct chassis mounting would work. If you can rig a second fan onto the unit I would highly recommend it. Good luck!

edit: oops! didn't see you wanted to load the 5vsb to 2A. That would be either a 2.5ohm resistor rated for 20W, or 2 5ohm resistors at 10W each. This adds 5W to the total, making it 234.07W, so about 240 with a fan or two.

**370forlife** · 02-03-2010, 03:30 PM

Re: Choosing resistors for loading

Heat shouldn't be a problem, I have a ton of those dell/newton proprietary psu's I gutted for parts, the heatsinks in those are pretty think and so are the cases themselves. Sandwich them between the heatsinks and case should be enough to keep it pretty cool.

**etnietering** · 02-03-2010, 03:54 PM

Re: Choosing resistors for loading

Yeah that should work, just make sure there's adequate airflow - don't cramp the resistors too close to each other.

**370forlife** · 02-03-2010, 04:06 PM

Re: Choosing resistors for loading

Hmm, this may be an expensive project, maybe some other time...

I could also use my case for a lambda II elevator controller from the radisson, my brother works there and got it for me when they were throwing it out.

Anyways, the case on that is almost bulletproof, good for dissipating heat.

Spending too much, I just ordered something that may make other's raise their pitchforks at me in heresy.

**stj** · 02-03-2010, 06:32 PM

Re: Choosing resistors for loading

you can get film resistors in to218 packages for heatsink-mounting,
but you'd need a big sink and serious fan.

**momaka** · 02-03-2010, 11:00 PM

Re: Choosing resistors for loading

Just an idea - maybe instead of resistors, you can get some car light bulbs (headlight, tail light, etc.). Perhaps go around car shops and ask if they have dead headlight bulbs. The reason you want headlight bulbs is that most are 2 bulbs in one (low beams and high beams). Chances are the high beams would still work.
If not that, you can also try some type of heating element. The thin ones in hair dryers usually have pretty low resistance.
Either way, keep in mind that 230 watts is a lot of heat. 1 PSU box might not handle it too well.

**Toasty** · 02-03-2010, 11:50 PM

Re: Choosing resistors for loading

Toaster(s).

**Khron** · 02-04-2010, 01:43 AM

Re: Choosing resistors for loading

I've suscessfully used some car lighbulbs for load-testing an Antec i revived, a couple weeks ago

Details can be found HERE

**etnietering** · 02-27-2010, 12:25 AM

Re: Choosing resistors for loading

Well I have a sort of update for this - I made a small dummy load for testing/loading power supplies. Originally, it was just the big resistor zip tied to the heatsink. It was connected to a molex connector, and ran the 12v line at about 4A (it's a 3 ohm resistor). This worked well enough, but would crossload the PSU VERY heavily (I couldn't even draw an amp on the other rails). It's not attached to the PSU, that's just a convenient place to put it while it's in operation.

Here's a picture:

And here's another picture:

So, I ordered two 1 ohm resistors, to load the 3.3v rail and 5v rail to several amps (3.3A on the 3.3v and 5A on the 5v). They're thick film resistors in TO-220 packages, so they mounted very easily to the heatsink. I also added a 15 ohm resistor for the 5vsb, so it will draw 1/3 A from that rail. So the total power draw is as follows:
3.3v * 3.3A = 10.89W
5v * 5A = 25W
12v * 4A = 48W
5vsb * 1/3A = 1.67W
for a total of 85.56W.
Not a real heavy load, but enough to make sure things are working properly, and the load is even enough for the voltages to be happy. I also added an LED for each rail. As I was searching through all my LEDs, I found that I could match the LED color to the color of the wires that carry that voltage. So there's a yellow one on the 12v, red on the 5v, orange on the 3.3v, and purple on the 5vsb. These let me know which rails are working, and are a good indication of whether or not the heatsink will burn me if I touch it (all 4 on = hot, only purple on = cold). The whole thing is powered from the ATX connector of the PSU. I used 2 wires (18 gauge, from an old power supply) for each power rail (except the 5vsb), and 6 ground wires. The switch shorts PS_ON to ground, turning the unit on and off. I know having the LEDs unprotected like that is very unsafe, but I didn't really want to put a whole ton of effort into making some nice LED board, and I think it looks pretty cool with them all floating like that

Here is an after picture:

Here is another after picture:

Currently it's plugged into my 350W L&C that I recapped. The DC load is around 85-90W (the voltages are almost spot on), and it's drawing 124W from the wall, for an efficiency of about 70%...ouch!
What do you guys think?

Attached Files

**stj** · 02-27-2010, 04:22 AM

Re: Choosing resistors for loading

i think you should drill a couple of holes for the resistor before it melts those cable ties.

**Khron** · 02-27-2010, 05:21 AM

Re: Choosing resistors for loading

And try to keep the heatsink fins vertical; helps with the cooling

**mur** · 02-28-2010, 01:35 PM

Re: Choosing resistors for loading

Just found this:
http://www.chirio.com/electronic_load.htm

google translate helps

http://translate.google.com/translat...tm&sl=it&tl=en

I think this is a very cool load tester. But to test pc supply you would need to have at least 3 of those loaders to load all rails. Go check it out

**etnietering** · 04-12-2010, 04:25 PM

Re: Choosing resistors for loading

Well I have an update - I got a 4.7 ohm resistor, which I connected to the 12v line. 4.7 ohms @ 12v is 2.55A, and about 30.6W. So this brings my dummy load to:

3.3v - 1 ohm - 3.3A - 10.89W
5v - 1 ohm - 5A - 25W
12v1 - 3 ohm - 4A - 48W
12v2 - 4.7 ohm - 2.55A - 30.6W
5vsb - 15 ohm - 0.33A - 1.67W

for a total of 116.16W. The new resistor is connected through a molex adapter, so it can be connected independently of the rest of it. With this extra load, my logisys runs at almost 75% efficiency! Whippee....

This increases the total load, and skews it a bit more towards 12v, which will be useful for newer power supplies.

Announcement

Choosing resistors for loading

Choosing resistors for loading

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment