Found a brand new Comp USA PSU

[momaka]

Re: Found a brand new Comp USA PSU

Check the TL431 shunt 3.3V rail (or better yet, relace it with another just to rule it out from being bad).

[everell]

Re: Found a brand new Comp USA PSU

I installed the three resistor values given, and the 3.3 volt rail now comes on without shutting down the power supply. Measures only 3.1 volt. I may not have the right transistor, or there may be some other tweeking needed. But at least the power supply now works again! So what do you expect from an old Leadman???

[momaka]

Re: Found a brand new Comp USA PSU

R32: 12 Ohms (brown, red, black, gold), 1/8W likely
R33: 470 Ohms (yellow, violet, brown, gold), 1/8W likely
R40: 47 Ohms (yellow, violet, black, gold), 1/8W likely

If you need any other part numbers, let me know by tomorrow (Sunday) afternoon. It's going to be mid-May the next time I have this PSU with me again, that's why.

Back then, it looked like Leadman was better than Deer. (before 2003)

The LP-6100 was PowMax's only remotely "decent" PSU. Everything else was the same or worse than Deer. In fact, some Deer PSUs weren't so bad at all (the ones that they actually bothered to install parts in). They (Deer) just had a few quirks in their very early ATX designs where they'd always blow a small 50V 47uF cap on the secondary when the open-loop self-oscillating 5VSB decided to go sky-high. Luckily (or thoughtfully), the 5VSB was always regulated with a 7805 regulator in those designs, so it was difficult to damage the motherboard even when the 5VSB went out of control.

[larrymoencurly]

Re: Found a brand new Comp USA PSU

I think I owned almost the exact same Leadman, only it had two ball bearing fans (one outside the case) and slightly larger heatsinks with no holes on top. At least one of the 1uF or 10uF, 16V or less caps showed 50V peak spikes, but I don't know if they decreased after I replaced those caps. The fan controller consisted of a thermistor (50 ohms cold?) in series with the fan and glued in the hole of one of the donut toroid coils.

[RJARRRPCGP]

Re: Found a brand new Comp USA PSU

You know it's sad when even the average Deer is better than a Leadman (Although not the one I'm about to post in the gutless hall of shame!)

Back then, it looked like Leadman was better than Deer. (before 2003)

[everell]

Re: Found a brand new Comp USA PSU

Thanks guys, I am waiting.

[Pentium4]

Re: Found a brand new Comp USA PSU

Sorry I didn't see this sooner. I can check later

[momaka]

Re: Found a brand new Comp USA PSU

^ I can get you those values on the weekend after next (i.e. Easter weekend). Just send me a PM to remind me. I also have a LP-6100 unit, but it's not where I am right now, hence the delay.

[everell]

Re: Found a brand new Comp USA PSU

I am trying to get the 3.3 volt circuit to work properly. Can anyone tell me the resistor values of R32, R33, and R40 (on the CompUSA Leadman power supply).

[Pentium4]

Re: Found a brand new Comp USA PSU

1N5404 diodes and all JEE brand capacitors, most of them swollen. The transformer is the same size as those in 300W Antec SmartPowers.

I had an even older Leadman PSU with Leadman brand capacitors in it.

Leadman branded caps?! Never knew they even did that, do you have pictures of it?

[Behemot]

Re: Found a brand new Comp USA PSU

tom66: maybe you can read something about harmonic distorsion first, you will learn what the problems are. Obviously by talking about motor PF, which has absolutelly nothing to do with harmonic distorsion, you know very little about it. Like this guy at first http://nationalpower.weebly.com/dama...istortion.html. Sure, he is exaggerating about the cooling of everything, but yes, the problem mainly with motors is there.

Anyway, basic power meters are designed to work with power net frequency (that is 50 or 60 Hz). Even 1 kHz is already 20 times that much. What about 10 kHz… These meters would have to work (=to measure periodically) in (tens of) MHz range to be able to reliable catch all the spikes. They just don't. Now I know mine is not even close to any laboratory equipment so it just cannot measure the shit from PSUs without PFC.

But, power supplies in EU are obliged to have harmonic supression. Yet, tell that to Zalman and their ZMx00-LE PSU line.

EDIT:// ben was faster…and yes, that is very much what I think as well

[ben7]

Re: Found a brand new Comp USA PSU

Motors can't be damaged by other devices' poor power factor, that doesn't make sense. A motor has a fixed power factor which is determined by its design and construction, some times have higher PF and some have lower PF. Often, in an industrial system you'll add a power factor correction circuit, consisting of inductors and capacitors, or you'll use a VFD with built in APFC.

Why the concern over PFC? Because it costs electricity suppliers more money to have poor PF loads (which they can't charge extra for)... and they lobbied/appealed to the EU to change this so they could continue using the existing infrastructure which will need upgrading, but they can delay it or roll it out slower than if everyone kept using bridge rectifiers + big caps.

MOST of the energy (>98%) is within the first few kHz.

Your meter is crap if it cannot measure the power usage of devices accurately if they have poor PF. You should consider upgrading it. My £10 eBay device works well.

1. Those big oil filled transformers can be damaged by the high frequencies.
2. You beat me to it. Yes, poor power factor leads to problems with current measurement, and the current waveform can become very distorted. Just because of that, measuring the mains voltage with a poorly functioning meter can give you crazy results. As far as I know, its because the meter has a sampling rate at which it samples the voltage, and calculates the RMS value based on those samples. If it has a slow sampling rate, then it is less precise. It may sample just when there is a noise spike on the waveform, and it will change the ultimate reading of the meter. Don't take that word for word, it's only my theory :P

Edit: a pdf from EATON talking about harmonics in the power grid.

Edit 2: Another PDF, from Danfoss, has more diagrams and math

[tom66]

Re: Found a brand new Comp USA PSU

Motors can't be damaged by other devices' poor power factor, that doesn't make sense. A motor has a fixed power factor which is determined by its design and construction, some times have higher PF and some have lower PF. Often, in an industrial system you'll add a power factor correction circuit, consisting of inductors and capacitors, or you'll use a VFD with built in APFC.

Why the concern over PFC? Because it costs electricity suppliers more money to have poor PF loads (which they can't charge extra for)... and they lobbied/appealed to the EU to change this so they could continue using the existing infrastructure which will need upgrading, but they can delay it or roll it out slower than if everyone kept using bridge rectifiers + big caps.

MOST of the energy (>98%) is within the first few kHz.

Your meter is crap if it cannot measure the power usage of devices accurately if they have poor PF. You should consider upgrading it. My £10 eBay device works well.

[Behemot]

Re: Found a brand new Comp USA PSU

I'm not sure what your point is, yeah you may have some energy past a few kHz but it's pretty much insignificant.

And 5VSB circuits using killer caps do suck big time. But they don't have to, it is possible to make them more reliable, but my point is you may as well use a controller IC because it'll end up cheaper by the time you've made all the modifications.

I guess you have no idea what you are talking about. I have first-hand experience with this matter you see. Power supply without PFC made my otherwise relatively precise wattmeter measure with 50 % error. Yeat PSUs with at least passive PFC are just fine.

So come again, what some energy on higher harmonics? And once again, if you were true and this was just some marginal problem, why all that hystery about PFC in EU? I tell you why: it is not marginal, it results in actual problems, for example motors can be badly damaged by heavily distorted power.

[larrymoencurly]

Re: Found a brand new Comp USA PSU

Those are actually nice heatsinks ... big rectifier diodes, probably 1N540x type? The transformer isn't one of those really tiny ones, and there is proper EMI filtering ... only bad thing is the caps .. rulycon on the output? :P

1N5404 diodes and all JEE brand capacitors, most of them swollen. The transformer is the same size as those in 300W Antec SmartPowers.

I had an even older Leadman PSU with Leadman brand capacitors in it.

And at least yours has a little heat sink on the 5VSB transistor

I added that because every other ATX power supply I saw had that transistor mounted on a heatsink.

Back in the olden days, CompUSA got its brand of PSUs from many suppliers, including Win-Tact (used by PC Power & Cooling), but there was no way to tell from the cardboard box, and CompUSA wouldn't let us open it unless we paid the 15% restocking fee.

[Pentium4]

Re: Found a brand new Comp USA PSU

Oh JEE, you're right!

I had an ATX that may have originally been an AT design because its 5Vsb was on a separate circuit board. PC Power & Cooling featured it in a magazine ad to show how bad a PSU could be.

Here's a Leadman identical to Pentium4's CompUSA unit, right down to the April 13, 2001 production date, only the +3.3V has a choke in its output, and the heatsinks are different.

And at least yours has a little heat sink on the 5VSB transistor

[ben7]

Re: Found a brand new Comp USA PSU

Can't be Rulycon,,,,,,,,,,,they aren't bloated. Every Rulycon on my CompUSA psu was bloated!

Oh JEE, you're right!

[everell]

Re: Found a brand new Comp USA PSU

T... only bad thing is the caps .. rulycon on the output? :P

Can't be Rulycon,,,,,,,,,,,they aren't bloated. Every Rulycon on my CompUSA psu was bloated!

[tom66]

Re: Found a brand new Comp USA PSU

If they are present… It clearly has *some* impact if the THD is such problem, right?

I'm not sure what your point is, yeah you may have some energy past a few kHz but it's pretty much insignificant.

And 5VSB circuits using killer caps do suck big time. But they don't have to, it is possible to make them more reliable, but my point is you may as well use a controller IC because it'll end up cheaper by the time you've made all the modifications.

[ben7]

Re: Found a brand new Comp USA PSU

I had an ATX that may have originally been an AT design because its 5Vsb was on a separate circuit board. PC Power & Cooling featured it in a magazine ad to show how bad a PSU could be.

Here's a Leadman identical to Pentium4's CompUSA unit, right down to the April 13, 2001 production date, only the +3.3V has a choke in its output, and the heatsinks are different.

Those are actually nice heatsinks ... big rectifier diodes, probably 1N540x type? The transformer isn't one of those really tiny ones, and there is proper EMI filtering ... only bad thing is the caps .. rulycon on the output? :P