I've had my share of both ends. But in all cases, no components were damaged except for the faulty PSU itself. So yeah, blow it with air, look at its caps and put a load on it (e.g. an hdd you don't care much about) and measure. If all is good, I don't see why dump them.

wbahnassi wrote on 2024-10-22, 05:04:

I've had my share of both ends. But in all cases, no components were damaged except for the faulty PSU itself. So yeah, blow it with air, look at its caps and put a load on it (e.g. an hdd you don't care much about) and measure. If all is good, I don't see why dump them.

Yeah, I've got oodles of old hard drives I can cable up to potentially take one for the team. Never tested an AT PSU, should be a good learning experience under my belt. That makes me feel better about just the PSU probably going. It took a year and a half of eBay alerts before I found a tested/working version of this particular mobo brand/model. I'd hate to slag it by just going "hey, let's just try this real quick!" 😀

Antec PSUs were famed for best-in -class design, but were badly hit by capacitor plague. These probably predate that period, but at this age anything electrolytic (or tantalum or RiFa) is suspect.

If exceedingly cautious, you could completely re-cap the beast with good but above all brand-new parts. I still intend to do that one of these days with one of my better AT PSU (and an excellent 550W 2005-era Antec that almost oozes bad cap fluid).

However you can probably get away with using a good pre-plague PSU so long as you test it first. Note that a single HDD provides enough load to at least get a constant figure, but doesn't load the PSU anywhere near fully. So I'd recommend a two-step approach - first hook up that old drive to at least see that it doesn't spectacularly short out or deliver way out-of-spec voltages. If that goes well (i.e. voltages well within spec), hook up a system that will challenge it and measure again. Only once that is also proven in-spec would I trust it.

Sometimes even apparently awful devices pass that test. On my test bench I use an Octek 009B for AT systems, which is a nominal 200W PSU, but in terms of weight and component count I'd judge it to be very low-end. However it reliably delivers whatever I ask of it, so it's good enough for me - and if it happens to get shorted and die because I hooked up something bad, at least it's not a rare fancy PSU I just smoked.

DeathRabbit679 wrote on 2024-10-22, 04:51:

Was going over an old IT closet at work that was fixing to be e-wasted, found 4 NOS antec pp300v PSUs. I've got a baby AT build coming up, thinking of using one of these. Is that a horrible idea, I've heard antec could sometimes be dodgy in the early 90s, so was gonna open it up, do a visual inspection for obvious nonsense (leaky caps and what all), then test the voltages coming out of it to make sure they seem reasonable. But if these are notoriously awful or something, probably go a different way. Thnx for the advice.

If/when you do open it up, perhaps post some pictures of its internals over here. I'd be happy to provide my PSU analysis based on that.

I don't have much experience/encounters with Antec PSUs from the 90's, but their early 2000's PSUs were mostly Channel Well Technology (CWT) builds - pretty good builds overall, but with terrible Fuhjyyu caps most of the time that were almost guaranteed to go bad in those particular PSU platform/designs.

So open it up and post pictures here first.

As dionb noted, a single HDD might not really be a large enough load to show if the PSU has any issues. Also, being an AT PSU, it would probably prefer to have most of its load on the 5V rail rather than the 12V rail... so hooking up a bunch of HDDs might even generate / trigger a false-positive failure (likely due to 12V rail load being too high and 5V rail not being high enough, causing a voltage dis-balance.) On top of that, "stable" voltage readings with a multimeter don't really mean much, because a multimeter cannot measure the (high frequency / AC) ripple and noise present on each voltage rail. For that, you'll need a oscilloscope... and a proper load for each rail... so it can get challenging to proper properly test the PSU. Indeed a sacrificial (and relatively high-load / high TDP) system would be the best option.
So best approach really is to just open the PSU and look for anything out-of-whack before powering it up.
I would recommend recapping only if the original caps are no-name brands and/or brands more likely to fail. Any Japanese -branded caps should still be good, despite the age (with a few exception noted, like Nichicon PR series, which are known to leak their electrolyte from the bottom.)
Otherwise, for AT and 5V-heavy ATX PSUs, I typically first try them out with several 20W 12V halogen light bulbs connected on the 5V rail and one more of the same on the 12V rail. This gives the PSU about 2-3 Amps of load on the 5V rail (~10-15 Watts) and 20 Watts on the 12V rail. Most old PSU's I've tested with this load will happily power on and give stable voltages on a multimeter. If/when I have the time, I use a large resistive load bank made from the internals of a 3kW dryer heating element (thick runs of NiChrome wire). With this, I can test my 5V-heavy PSUs with up to or even over 40 Amps on the 5V rail, if desired.

Thanks for offering to take a look. Should have time to crack one open tomorrow to poke around.

So I finally got time to take one of these apart. This one was the one I was absolutely sure was NIB, or if it's a fake, it's the best I've ever seen, including the 30 years of yellowing on the outer box and antec manuals inside. The others had the seal broke, even though they looked brand new, it's possible the say some brief usage at some point. I cant even begin to comment on the design, I skipped Power Systems to take VHDL Design in college, haha.

The whole thing

Close up 1

Close up 2

The caps look perfect, other than the usual glue mess, but what caught my eye was that "extra wire" you can see in the 2nd close up. That seems...ungood. I thought maybe it had come unsoldered due something I might have done when I cracked it open but it looks like it was never used, possible manufacturers defect? "Extra wire" and "power supply" seem like two terms you never want in the same sentence, but what do I know. Hope to get a chance to put a load on it later this week and also grab another one of these out of storage to see if it also has the "extra wire". As always appreciate any wisdom y'all have to impart, this is only the 2nd psu I'v ever opened so I definitely don't know what I'm on about.

I have one of these, I also got it NOS. They are well built (manufacturer is CWT as mentioned above), but you definitely want to replace ALL the capacitors as they used Fuhjyyu, which are of abysmal quality.

Other than that they should be more than enough for any AT build you have.

As for the extra wire, if it goes to a connector you should solder it to the PCB, probably a manufacturing defect.

Well made AT unit. 2004- 2005 maybe. As TheMobRules said those Fuhjyyu capacitors need to go. The loose black '"extra wire" should be soldered to COM (common ground).

DeathRabbit679 wrote on 2024-10-28, 04:06:

As always appreciate any wisdom y'all have to impart, this is only the 2nd psu I'v ever opened so I definitely don't know what I'm on about.

There used to be a pretty good PSU write-up on Hardware Secrets in regards to what does what and the different topologies. But I believe that's gone now, as are many old the old PSU reviews. Might be up on Waybackmachine, though (haven't checked.)

In any case, here's how it breaks down:
The big yellow transformer in the middle is what separates the primary and the secondary sides.
The primary side is the one that is connected to the AC line in your home and therefore at (relatively) high-voltage (read: DANGER of shock or electrocution when the power supply is plugged into the wall.) You can actually see on this power supply, the manufacturer even put a yellow sticker on the primary heatsink to warn of the live / high / non-isolated line voltages.
The secondary side (a.k.a. the "low-voltage" or "isolated" side) is what the PSU output voltage rails are connected to (i.e. 5V rail, 12V rail, Power Good signal, and etc.)

DeathRabbit679 wrote on 2024-10-28, 04:06:

So I finally got time to take one of these apart.

Yup, that's a legit Antec PSU.

Old half-bridge design, but pretty dependable. And looks well-built too. You can see all of the primary-side EMI/RFI components are present (two yellow X2 caps, a common mode choke, and a power receptacle with built-in filter with 2x Y2 caps.) There's also a bridge rectifier (with what looks like two MOVs in front of it for AC surge protection), and appropriately-sized primary electrolytic filter capacitors (the two large 200V ones.) The main switching transistors are also nice and large (TO-247 case size rather than skimpy TO-220 seen in cheapo units.)

Moving onto the secondary... again, everything looks good. The output rectifiers (attached to the secondary heatsink) are appropriately-sized (one is in a TO-247 case, so probably a 30-Amp -rated and likely for the 5V rail, while the other is TO-220 case and probably a 6 or 10 Amp for the 12V rail.) The output toroid inductor looks large enough for a 150-200 Watt unit. And there's even a fan controller board for the PSU's fan - something rather rare for AT units. Most older AT units just ran their fan at full blast / 12V all the time and were rather loud.

Lastly, the output capacitors - these also look like they were sized well for the unit. Only issue is they indeed look like Fuhjyyu brand, judging by the "Y" style vent on top. Since it's a NOS unit, you could just use it as-is... at least for a little while. But for any long-term use, you may want to change out the Fuhjyyu capacitors on the secondary side, as they have a tendency to fail in CWT PSUs (and in general too.) While at it, maybe get some replacements for the smaller capacitors too. Only the large 200V caps don't need to be changed, as they almost never fail, even after many years of use (or lack thereof.)

For replacements, you can use something like Rubycon YXJ or YXM or YXS or YXF series, Nichicon PW or PS or PM or PJ series, Panasonic FC series, and United Chemicon LXZ or LXV or LXY series. If you don't have any high quality Japanese capacitors like this available where you live, they you might better off leaving the original Fuhjyyu caps than changing them with new no-name Ali-express or Amazon c[r]aps (worth noting here is that most local shops / independent electronics stores usually won't carry any good brands either.) If you're in the US, finding the above suggested replacements on Digikey or Mouser should be relatively easy. For the rest of the world, it will vary. Farnel and RS Components are two other legitimate online parts distributors. eBay can be hit or miss - there are a few sellers that sell NOS Japanese capacitors suitable for PSU use. But the majority of stores/sellers are just slinging garbage straight out of the China malls. If the description and the pictures don't specifically show the brand and series of capacitor(s) you're buying, run away.

momaka wrote on 2024-11-01, 12:39:

There used to be a pretty good PSU write-up on Hardware Secrets in regards to what does what and the different topologies. But I […]
Show full quote

DeathRabbit679 wrote on 2024-10-28, 04:06:

As always appreciate any wisdom y'all have to impart, this is only the 2nd psu I'v ever opened so I definitely don't know what I'm on about.

There used to be a pretty good PSU write-up on Hardware Secrets in regards to what does what and the different topologies. But I believe that's gone now, as are many old the old PSU reviews. Might be up on Waybackmachine, though (haven't checked.)

In any case, here's how it breaks down:
The big yellow transformer in the middle is what separates the primary and the secondary sides.
The primary side is the one that is connected to the AC line in your home and therefore at (relatively) high-voltage (read: DANGER of shock or electrocution when the power supply is plugged into the wall.) You can actually see on this power supply, the manufacturer even put a yellow sticker on the primary heatsink to warn of the live / high / non-isolated line voltages.
The secondary side (a.k.a. the "low-voltage" or "isolated" side) is what the PSU output voltage rails are connected to (i.e. 5V rail, 12V rail, Power Good signal, and etc.)

DeathRabbit679 wrote on 2024-10-28, 04:06:

So I finally got time to take one of these apart.

Yup, that's a legit Antec PSU.

Old half-bridge design, but pretty dependable. And looks well-built too. You can see all of the primary-side EMI/RFI components are present (two yellow X2 caps, a common mode choke, and a power receptacle with built-in filter with 2x Y2 caps.) There's also a bridge rectifier (with what looks like two MOVs in front of it for AC surge protection), and appropriately-sized primary electrolytic filter capacitors (the two large 200V ones.) The main switching transistors are also nice and large (TO-247 case size rather than skimpy TO-220 seen in cheapo units.)

Moving onto the secondary... again, everything looks good. The output rectifiers (attached to the secondary heatsink) are appropriately-sized (one is in a TO-247 case, so probably a 30-Amp -rated and likely for the 5V rail, while the other is TO-220 case and probably a 6 or 10 Amp for the 12V rail.) The output toroid inductor looks large enough for a 150-200 Watt unit. And there's even a fan controller board for the PSU's fan - something rather rare for AT units. Most older AT units just ran their fan at full blast / 12V all the time and were rather loud.

Lastly, the output capacitors - these also look like they were sized well for the unit. Only issue is they indeed look like Fuhjyyu brand, judging by the "Y" style vent on top. Since it's a NOS unit, you could just use it as-is... at least for a little while. But for any long-term use, you may want to change out the Fuhjyyu capacitors on the secondary side, as they have a tendency to fail in CWT PSUs (and in general too.) While at it, maybe get some replacements for the smaller capacitors too. Only the large 200V caps don't need to be changed, as they almost never fail, even after many years of use (or lack thereof.)

For replacements, you can use something like Rubycon YXJ or YXM or YXS or YXF series, Nichicon PW or PS or PM or PJ series, Panasonic FC series, and United Chemicon LXZ or LXV or LXY series. If you don't have any high quality Japanese capacitors like this available where you live, they you might better off leaving the original Fuhjyyu caps than changing them with new no-name Ali-express or Amazon c[r]aps (worth noting here is that most local shops / independent electronics stores usually won't carry any good brands either.) If you're in the US, finding the above suggested replacements on Digikey or Mouser should be relatively easy. For the rest of the world, it will vary. Farnel and RS Components are two other legitimate online parts distributors. eBay can be hit or miss - there are a few sellers that sell NOS Japanese capacitors suitable for PSU use. But the majority of stores/sellers are just slinging garbage straight out of the China malls. If the description and the pictures don't specifically show the brand and series of capacitor(s) you're buying, run away.

Wow, thanks for the extensive write up! The cap recommendations will greatly help me when I decide to rebuild this guy with fresh caps, so I'm not stumbling around digikey blindly. Thanks for pointing out the danger noodle areas too. I've got some lineman's gloves I can wear whenever the thing is powered. Bought them to take the edge off a bit if I caught a shock fiddling around inside a CRT television a while back.

Also, I opened all the rest of them up. To my surprise, there seemed to be two different hardware revs. Staring and comparing, most of the circuitry lokked the same, though the location and pinout of the p8/p9 and molex connectors was moved around a tad. The one I opened that was identical to one I posted above, it had the extra wire soldered to COM. Really just must have been a manufacturer defect or perhaps the way the cords were tightly wound pulled in it for 2 and a half decades till it let go. I did verify that at least all the large caps are fujhyyu trash on both BOMs, so I'll probably test them all, see what state they're in. If good, I'll recap one of each style and then measure again. Nice that I have the extras for reference so I don't have to take dozens of reference pics ahead of time. I think I have my project for the coming holidays 😀 Hopefully, my meager soldering skills will be adequate, at least it's old enough it should still be SnPb. The last thing I recapped was ROHS and it took me like 4 hrs to replace 6 caps.

DeathRabbit679 wrote on 2024-11-05, 03:16:

I think I have my project for the coming holidays 😀 Hopefully, my meager soldering skills will be adequate, at least it's old enough it should still be SnPb. The last thing I recapped was ROHS and it took me like 4 hrs to replace 6 caps.

It'll be fine, these units are from the pre-ROHS era, and being single sided boards it is very easy to remove the old caps and clear the holes with some braid dipped in flux. Of course a desoldering gun will make the job even easier, but it's not really necessary here. Also, these AT power supplies are not cramped with tons of components like later ATX units so you have ample space. Another nice thing is that you can remove the PCB without desoldering anything as Antec/CWT used connectors for the mains wires, fan and all that.

It's a nice learning experience I think! Just be careful when soldering and clean up well to make sure you don't create any accidental solder bridges as in a power supply this can be... problematic.

For what it's worth, I used Rubycon YXF capacitors on mine and they work perfectly. Any of the other series suggested by momaka should be fine as well.