DIY PSU load tester

[c_hegge]

Re: DIY PSU load tester

I'm going to be using CPU heatsinks on mine. With no fans, it takes less than a minute before they are too hot to touch. I'm hoping that the fans in the case i've knocked up can do the job.

[Th3_uN1Qu3]

Re: DIY PSU load tester

Well it turns out it isn't that bad after all. I decided to see just how much the MOSFETs will take, and managed to get 14 amps @ 12.10v for a couple minutes before they gave up. That's 170 watts! And what's even more surprising is that two of them went short but one survived - even though there is visible heat damage to the plastic case!

Since this was an absolute worst-case scenario, 170W/3 mosfets means a total of 1.7 kW, and i will rarely have to deal with power supplies >600W, i'll just go ahead and build it with PSU heatsinks. Now time for some PCB design.

[momaka]

Re: DIY PSU load tester

Not sure if this is a useful design suggestion, but perhaps you can submerge part of the heatsinks in water all the way up to the MOSFETs, keeping the MOSFETs dry of course. I think that should give you much better cooling. When the water gets too hot, change it .
Or you can also use mineral oil and completely submerge everything since it's non-conductive.

[Th3_uN1Qu3]

Re: DIY PSU load tester

There would be no problem with submerging the MOSFETs in water either due to the low voltage (well, except generation of hydrogen...), but i don't want to go that route. I just tested the 8 fans glued together - and i'm starting to get worried they might take off or something. I think they're gonna be enough.

[c_hegge]

Re: DIY PSU load tester

I would be more worried about the possibility of corrosion if they were under water. Hydrogen is awesome stuff when you have a box of matches handy

[Th3_uN1Qu3]

Re: DIY PSU load tester

Hydrogen is awesome stuff when you have a box of matches handy

Don't remind me - i was bored one night and was doing just that... One of the bubbles exploded so loudly that it was amazing it didn't wake anyone up.

[Agent24]

Re: DIY PSU load tester

How about a PC watercooling kit? Wouldn't need to immerse the heatsinks in water then. Would be expensive though.

[Th3_uN1Qu3]

Re: DIY PSU load tester

Not to mention the waterblocks wouldn't offer enough space, unless i'd be mounting the 'fets on the main radiator. This isn't a 150W quadcore, it's from 150W up to one kilowatt remember.

8x 80mm fans seems to be enough but i'll add two more anyway, AND leave provision for some on the front. I'll have to go pick up the meters and come up with the design and material for the front panel then i can commence building the rest.

Oh and if all else fails... there's no one stopping me to change the fans for 240v ones.

[PCBONEZ]

Re: DIY PSU load tester

170W x 3 doesn't mean 1.7 kW, try 510 watts.

Water cooling is just a gimmick to sell crap to gadget addicts.
There are a number of Xeon CPU's that hit 150 watts and factory air cooling works fine.

Should be able to rework some copper core coolers to get the job done.
.

[Th3_uN1Qu3]

Re: DIY PSU load tester

170W x 3 doesn't mean 1.7 kW, try 510 watts.

The nominal dissipation is intended to be 35W/mosfet, that is 105W per heatsink. There are 30 mosfets and 10 heatsinks. 105 x 10 = 1kW. I said that the dissipation limit was tested at 170W over one heatsink, before the 'fets fry. Or at least two of them did. 170 x 10 IS 1.7kW. Furthermore, i left the remaining good 'fet along with two new ones at 10.2A / 11.84v for 30 minutes. That's 121W. They held.

So it seems that this will be capable of 1.2kW load medium term - you don't need more than 30 minutes to test a PSU anyway. Though i do agree that the PSUs i will test will average at 500W, my math isn't wrong.

[PCBONEZ]

Re: DIY PSU load tester

170W/3 mosfets means a total of 1.7 kW

.

[PCBONEZ]

Re: DIY PSU load tester

@/35w each you could use some Prescott coolers you could easily put 3 or 4 MOSFETs per cooler which might save space depending on your layout.
Just drill and tap the heatsinks and you could attach them directly with machine screws.
.

[Th3_uN1Qu3]

Re: DIY PSU load tester

There i was referring to the total power. 170W over three meant 1.7k over 30. Sorry for the misunderstanding.

Anyway, i have enough of those ATX power supply heatsinks that i can do just two per heatsink. At 12A going thru 4 of them temperature settled at 95C which isn't that bad. That's with just three of 8 fans running so i guess they'll do fine. And what they can do short term is rather incredible. I got my 20A meter to read "out of range" for a minute or so before the 'fets got too hot and i had to lower the current. If four of them can do this to a 12v rail, i can just imagine what twenty will. Btw, the 420W Powerlink PSU that i am abusing for testing seems to be holding up so far.

I'd rather use the heatsinks i have now rather than buy some more, unless i can buy some Prescott coolers on the bulk i'd just use those that i already have.

[Th3_uN1Qu3]

Re: DIY PSU load tester

Sometimes boredom can be good. I was bored today and had a spare fan from a gutless wonder i can't be bothered to fix. It's a "Te Bao Metallic Plastic Co" fan. So what could be more entertaining than testing it to see how much it'll take before blowing up? I knew Intel stock coolers generally go kapow at 18 volts.

The only problem was, i couldn't blow it up. I ran it up to 27 volts which is the limit of my adjustable supply, and it just kept on working. More so, it produced an impressive amount of airflow.

So i decided to see whether the group of fans i'll be using for my load tester can take the same. Well what'd you know... each of them did. Actually, with all 8 fans running the poor little transformer gives up at 14 volts, with 4, at 20 volts. At 20v x 4 fans the amp meter reads 0.83A.

Even at "just" 20 volts, four 80mm fans make a ridiculous amount of airflow. If i was worried about them taking flight at 12v, now i actually have to hold them so they don't move. I have a huge 2x 24v transformer, it's not too practical but i guess i can use one winding to feed the fans. Now i just have to figure where would i put it. Certainly not in the load tester itself. Oh, and i also got myself something that'll do nice for a front panel.

Edit: I hooked up the big transformer. I only went up to 21 volts with all 8 to make sure i don't damage the fans. All i can say is... GOT CFM?

[PCBONEZ]

Re: DIY PSU load tester

Wouldn't it be simpler and less expensive for this project to mount them as passive cooled and use a 12-14 inch AC powered box fan to cool the whole box?
.

[Th3_uN1Qu3]

Re: DIY PSU load tester

use a 12-14 inch AC powered box fan to cool the whole box?

I'm not going to rip out the bathroom fan for this.

[PCBONEZ]

Re: DIY PSU load tester

I'm not going to rip out the bathroom fan for this.

Good idea.
Bathroom fans only move 50-100 CFM which is why I didn't suggest them.
.
The kind of fan I did suggest would cost about the same as a bathroom fan, be MUCH quieter, and get you roughly in the range of 600 CFM.
.

[Th3_uN1Qu3]

Re: DIY PSU load tester

Well an additional problem with a mains powered fan would be the 50hz field it radiates, with signals in the tens of mvs nearby, it wouldn't be such a good idea.

[PCBONEZ]

Re: DIY PSU load tester

Um, it would be rather silly to put a fan that big so close to the PCB.

And a simple grounded wire mesh grill would provide all the shielding needed.
.

[c_hegge]

Re: DIY PSU load tester

I Just ordered the last of the components for mine. For a 700W loading capability I've gone with the following resistors

3.3v: 3x 0.33ohm 50W (30A)
5v: 6x 1ohm 50W (30A)
12v1: 8x 5ohm 50W (19.2A)
12v2: "
-12: 1x 120ohm 10W (0.1A)
5sb: 1x 5ohm 10W (1A)

All 27 resistors will have a switch in series with them