ATX PSU output load resistors - suggested values?

[re-atari]

Hi all,

I've been busy fixing a 300W ATX PSU, and noticed some strange things I'd like your opinion on. The PSU is branded Aopen AO300-12APNF, with active PCF. Its rated outputs are:
+3,3V - 28A
+5V - 30A
+12V - 18A
-5V - 0,3A
-12V - 0,8A
+5V SB - 2A
where +3,3V and +5V combined should not exceed 200W

The reason for opening it up was quite simple: it did absolutely nothing. My PSU tester showed it didn't output any voltages (confirmed by testing with my multimeter). No idea what had happened, it was already dead when I got it.

Judging by the PCB, it's a Fortron product (FSP Group silkscreened on underside). The reason why it probably didn't work was easy to see: a number of components were literally drenched in this horrible brown glue. You know, the stuff that's notorious for becoming conductive from heat, and this one didn't disappoint in that respect. Especially near the output coils, caps and load resistors the glue looked like burned sugar. As the caps were of a 'less reputable' brand and didn't look very fresh, I decided to replace and/or clean the questionable components. The power FETs and rectifiers checked out OK, as did the bridge rectifier.

I sketched out part of the schematic, and in doing so noticed some of the load resistors are of a quite strange value. The output voltages are connected to GND throught the following resistors:

+3,3V - 8,2R 1W
+5V - 7,5R 3W
+12V - 150R 1W
-5V - 100R 0,5W
-12V - 120R 1W
+5V SB - 470R 0,25W

The resistors were not burned and all measured OK, the values corresponding with their color markings. Especially the 3W resistor was sitting quite cosy next to a 3300uF output cap on the +12V rail, almost touching it. This PSU must have been designed for use in Siberia...

Checking this forum, I couldn't find if these values are 'normal', of if the manufacturer just put in whatever was at hand.

I'd like to know if the above are unusual values, and if so, if I can replace them with more sensible values? What values would then be valid?
I've done the math and calculated the resistors are just within spec for the wattage they have to put up with. However I did read in a recent thread, that +12V usually is loaded with 100R, +5V with 47R and +3,3R with 10R. Is there a specific rule for determining the resistance the load resistors must have?

Thanx!

re-atari

[mariushm]

Re: ATX PSU output load resistors - suggested values?

Voltage = Current X Resistance , Power Dissipation = Current x Current x Resistance

12v = Current x 150 ohm ... => Current = 12/150 = 0.08A = 80mA ,
Power = 0.08x0.08x150 = 0.96w (this is far too close to the maximum 1w rating of the resistor, you sure it's a 150 ohm resistor?)

5V = Current x 7,5 ohm => Current = 5/7.5 = 0.666 = 660mA
Power dissipated = 0.66 x 0.66 x 7.5 = 3.2 watts... more than 3w again...

The power supply needs a minimum load on the output voltages, that's what the resistors are there for. But these loads seem quite large, normally power supplies only put resistors to draw about 0.25-0.3A on each rail, not something as high as 0.66 amps.

If they're not there, you won't get regulated voltages, and you might get more than 5v and 12v when the pc starts. As the computer components start and puts some load on the rails, the power supply becomes more regulated.

You can desolder the resistors and put them on some wires and move them away from the pcb. The distance from the pcb doesn't matter.

[re-atari]

Re: ATX PSU output load resistors - suggested values?

I did exactly the same calculations and came to the same conclusions as you did. The wattage through the resistors drives them to their limit or exceeds it.

I really don't understand why the manufacturer chose these values. Judging by the looks of the solder joints the resistors were installed in the factory, no signs of resoldering.

The values I wrote are correct, both according to the color codes and my measurements. I would like to replace them with a higher resistance value, if that's possible. I mean, surely the manufacturer must have installed these particular ones for a reason

re-atari

[momaka]

Re: ATX PSU output load resistors - suggested values?

However I did read in a recent thread, that +12V usually is loaded with 100R, +5V with 47R and +3,3R with 10R.

That's possibly one of my posts ranting about the load resistors used in Deer/Allied/L&C/Solytech PSUs. These values above are typical for them, but quite small IMO.

My rule of thumb is, don't go above 0.5W on a load resistor, and use resistors rated for at least 1W or more.

My normal procedure is remove all load resistors and test PSU in a junk computer or with some light bulbs first. If voltages are unstable or whacky, I put in some load resistors, but according to my rule of thumb. In Deer/Solytech PSUs, I just swap some of the resistors around. Normally I put the 47 Ohm on the 3.3V rail, 100 Ohm on 5V rail (or completely omit it if the PSU is going in an older 5V-based PC such as Pentium 3 or older Socket A AMD), and 270 Ohms or more on 12V rail (but more often 2x 1000 Ohm resistors in parallel).

[re-atari]

Re: ATX PSU output load resistors - suggested values?

That's possibly one of my posts ranting about the load resistors used in Deer/Allied/L&C/Solytech PSUs. These values above are typical for them, but quite small IMO.

Thanx, Momaka! You're most likely right, it does ring a bell. I just can't find your post anymore. Thankfully I had made a note of the resistor values you suggested...

My rule of thumb is, don't go above 0.5W on a load resistor, and use resistors rated for at least 1W or more.

My normal procedure is remove all load resistors and test PSU in a junk computer or with some light bulbs first. If voltages are unstable or whacky, I put in some load resistors, but according to my rule of thumb. In Deer/Solytech PSUs, I just swap some of the resistors around. Normally I put the 47 Ohm on the 3.3V rail, 100 Ohm on 5V rail (or completely omit it if the PSU is going in an older 5V-based PC such as Pentium 3 or older Socket A AMD), and 270 Ohms or more on 12V rail (but more often 2x 1000 Ohm resistors in parallel).

Sounds like a solid procedure! I have yet to get to the store and buy the replacement electrolytics. I'm thinking of replacing all of the ceramic caps in the PSU as well, if only for good measure. No problem buying the power R's you suggest at that occasion in one go.

I have done the maths about the required minimum wattage, it's in the range 0,14W-0,2W. Allthough they probably are overspec'd I think I will just immediately go for 1W types all the same. No need to be greedy, and it can't hurt to be a bit overspec'd. There is enough space on the PCB to fit them anyway.

Still, makes me wonder why on earth Fortron put in load resistors with these strange values...

BTW: I don't yet have a specific intended use for this PSU at the moment, other than having a test unit at hand. If it turns out to be reliably fixed, it will come in handy for testing a 'temperamental' Asus M2N68 mainboard.

re-atari

[momaka]

Re: ATX PSU output load resistors - suggested values?

I have done the maths about the required minimum wattage, it's in the range 0,14W-0,2W. Allthough they probably are overspec'd I think I will just immediately go for 1W types all the same. No need to be greedy, and it can't hurt to be a bit overspec'd. There is enough space on the PCB to fit them anyway.

Yeah, that's what I meant in my above post. Try to keep the dissipated heat by each resistor under 0.5W, but make sure each resistor is rated for at least 1W. That way, you won't be cooking your caps.

[Pentium4]

Re: ATX PSU output load resistors - suggested values?

I had a quick experimentation with a Sun Pro PSU. I removed the minimum load resistor on the 3.3V rail (I don't remember the value) Once I did this, the 3.3V would stay rock solid @ 3.467V, load it up, wouldn't change more than 0.002V. Also, the 12V was oscillating like crazy, even with a very continuous load, it would be fluctuating constantly between 12.36-12.39V, changing about twice a second on my Fluke.

I then added just a 15Ω resistor to the 3.3V rail, and then it was steady at 3.39V, and the 12V would stay perfectly stable at 12.15V. It is an old half bridge unit but I didn't think removing the 3.3v minimum load resistor would affect the 12V so much. I taped a piece of plastic between the resistor and the cap so it wouldn't cook it as much. The fan spins VERY slow. ~400RPM

[domas]

Re: ATX PSU output load resistors - suggested values?

it seems that my 12V voltage regulation is not working well on my bench lab supply, the fsp one from this thread: https://www.badcaps.net/forum/showthread.php?t=28401

I if i do load it with >~2-3A, it drops down to <11V!

As i checked my own pictures now I can see that there is load resistor on 5V line, but it seems to be not enough.
If I add an external 100ohm resistor on 5v rail, 12V rail raises by 0.3-0.5V depending on how low it is.
If I add 20ohm resistor to 5v rail, the 12v one is raising even much more, but it is still nowhere near close to 12V. And I do not like 20ohm resistor as it is getting too hot for my taste
Probably a re-cap or something like that is needed.