Caesum wrote on 2026-05-05, 16:22:

Pardon me for asking one more time, but I was checking PSUs because the last one I wanted to buy was from some not-so-reliable v […]
Show full quote

Pardon me for asking one more time, but I was checking PSUs because the last one I wanted to buy was from some not-so-reliable vendor. Instead I found someone else and he has a Fortron PSU FSP 300W FSP ATX-300PNF from 2003. It says it is "W/PFC" but I assume it means Passive PFC not Active. There is one question about voltages though, it says it has:

+3.3.V == 20.0A, +5V == 20.0A, +12V1 = 16.0A, +12V2 == 16.0A, +5Vsb == 2.5A, -12V == 0,8A
(+3.3.V & +5V = 120W Max) (+12V1& +12V2 = 264W Max)

And now my question is, since it is (I assume) a Passive PFC PSU from 2003 with higher voltages on 3.3v line and 5v line then why max W is higher on 12V lines and does it mean this PSU is suitable for older eras computers or actually not?

A passive PFC was a temporary workaround that allowed any existing power supply to improve its power factor without any modifications in PSU themself. Its presence or absence has no effect whatsoever (from a home user’s perspective)
The high power output at +12V is because, at that time, it was a power supply unit of the latest ATX 2.3 standard, which marked the start of the focus on this bus.

The PNF series consists of robust PSUs based on tried-and-tested principles, so you’re unlikely to encounter any issues, though the voltage might read +12.3V instead of +12.0V.
This doesn’t affect anything, apart from the lack of round figures on the monitoring (provided the motherboard’s monitoring isn’t giving the wrong reading).

Caesum wrote on 2026-05-05, 16:22:

he has a Fortron PSU FSP 300W FSP ATX-300PNF from 2003

+3.3.V == 20.0A, +5V == 20.0A, +12V1 = 16.0A, +12V2 == 16.0A, +5Vsb == 2.5A, -12V == 0,8A
(+3.3.V & +5V = 120W Max) (+12V1& +12V2 = 264W Max)

Instead of PSU like that you can just go and buy a modern one with strong +3.3.V & +5V power load. They also have 120W+ max load. I would just avoid Standard/Bronze PSUs.

Mike_ wrote on 2026-04-27, 16:31:

I opened up my Enermax 300W PSU out of curiosity. Capacitors seem to be fine, but I don't recognize their brand. […]
Show full quote

I opened up my Enermax 300W PSU out of curiosity. Capacitors seem to be fine, but I don't recognize their brand.

The attachment enermax_1.jpg is no longer available

The attachment enermax_2.jpg is no longer available

The attachment enermax_3.jpg is no longer available

The attachment enermax_4.jpg is no longer available

Speaking of PFC, is my Enermax active or passive PFC? It doesn't have a huge inductor like Fortron does, and it's only for 230V, with no selector switch.

AlexZ wrote on 2026-05-05, 16:43:

Caesum wrote on 2026-05-05, 16:22:

he has a Fortron PSU FSP 300W FSP ATX-300PNF from 2003

+3.3.V == 20.0A, +5V == 20.0A, +12V1 = 16.0A, +12V2 == 16.0A, +5Vsb == 2.5A, -12V == 0,8A
(+3.3.V & +5V = 120W Max) (+12V1& +12V2 = 264W Max)

Instead of PSU like that you can just go and buy a modern one with strong +3.3.V & +5V power load. They also have 120W+ max load. I would just avoid Standard/Bronze PSUs.

I thought so too at first but I was told that those might not provide a stable 3v/5v current especially the active PFC ones.

Caesum wrote on 2026-05-05, 16:22:

Pardon me for asking one more time, but I was checking PSUs because the last one I wanted to buy was from some not-so-reliable v […]
Show full quote

Pardon me for asking one more time, but I was checking PSUs because the last one I wanted to buy was from some not-so-reliable vendor. Instead I found someone else and he has a Fortron PSU FSP 300W FSP ATX-300PNF from 2003. It says it is "W/PFC" but I assume it means Passive PFC not Active. There is one question about voltages though, it says it has:

+3.3.V == 20.0A, +5V == 20.0A, +12V1 = 16.0A, +12V2 == 16.0A, +5Vsb == 2.5A, -12V == 0,8A
(+3.3.V & +5V = 120W Max) (+12V1& +12V2 = 264W Max)

And now my question is, since it is (I assume) a Passive PFC PSU from 2003 with higher voltages on 3.3v line and 5v line then why max W is higher on 12V lines and does it mean this PSU is suitable for older eras computers or actually not?

Yes, that line of Fortron / FSP (ATX-xxxPNF) is with passive PFC, not active.

Whether the PSU is suitable for older PCs or not is always a tricky one to define.
Generally speaking, if the power rating (Watts) on the 12V rail grossly exceeds that of the 3.3V and 5V rails *and* the PSU is not known to have DC-DC converters (like some / most? modern PSUs do), then it's probably *NOT* suitable for a 5V-heavy PC.

With that said, I can actually answer this question about the above FSP model directly from experience: it is not suitable for a 5V-heavy PC. In fact, it does rather poorly with a 5V-heavy PC. If you feel like taking a very long read, you can do so at the thread below, there user Analog Programmer has a higher-power -rated FSP PSU of the same model line and wanted to see if it would be possible to convert the PSU to make it more 5V-centric. As I suspected (and towards the end of that thread), it is revealed that this is not possible to do on this PSU - not without some serious modifications (most of which would result in diminishing the PSU's efficiency in an effort to balance the voltages between the 5V and 12V rails.) Here's the thread:
[SOLVED] PSU Fortron/FSP ATX-350PNF with problematic +5 & +12V voltages

AlexZ wrote on 2026-05-02, 07:41:

A more modern Athlon XP board with P4/V12 connector would be the best choice, 2nd option is socket 754 (but you would also need CPU, probably a Turion 64) as it also has P4 connector. You would use a modern PSU.

I would not bet on period correct PSU lasting long, it will have to be checked periodically and probably recapped as they are too old. If you prefer trouble free solution then avoid this route.

I would sell the board + CPU to someone who doesn't mind fixing period correct PSUs and for the same price buy board + CPU with P4 connector.

Problem with buying an AXP board with a P4/12V connector is that there aren't that many choices and they are all likely to be more expensive, since they will be at the tail end of AXP era, where 333 and 400 MHz bus was more prevalent.

Socket 754 could be cheaper and probably eliminate many of the problems... but then maybe that does not bring the same nostalgia level to the O/P as the old s462 one? I know for myself that I certainly don't give s754 stuff the same nostalgic value as s462 stuff - especially if the old s462 was some Duron CPU where you could do the L-bridge mods to unlock extra cash or do other cool stuff.

As for period correct PSU not lasting long - again, I think that really depends on what one gets... and also a bit of luck, of course.
A few months back, for example, I got a 250 Watt OEM PSU likely either from a Dell or HP PC (not sure which) made by LiteON. It's a model that I already have a few of. I opened it just to see if there was anything obviously wrong, but there wasn't. Not even that much dust inside, which means it probably sat on a shelf for a very long time. Needless to say, I powered it up and after verifying the output voltages were OK (at least with a multimeter... my scope is out of commission at the moment), I put it on a spare junky PC and let it rip away. It's been fine so far.

shevalier wrote on 2026-05-02, 08:33:

@Caesum from Poland. The main problem in Eastern Europe is that local flea markets are being cleaned out by professional dealers […]
Show full quote

@Caesum from Poland.
The main problem in Eastern Europe is that local flea markets are being cleaned out by professional dealers, and all the vintage stuff ends up being resold on eBay.
https://www.ebay.com/shop/asus-tusl2?_nkw=asus+tusl2
Finding something decent for a reasonable price isn’t exactly quick and takes a lot of luck.

Yeah, I see this is also starting to happen in Bulgaria now too. Though on my local flee market, some stuff somehow still falls between all of the dealers' and scrapers' fingers and then finds its way to me. For example, last week I was able to rescue a GeForce 4200 for a little under 5 Euro. As soon as did that, one of the known scrappers there went to the same guy I got the 4200 from and bought the rest of his GPUs and PC add-on cards. None of them were worth anything anyone would miss though, so I don't feel bad. The 4200 was a good save, though, and just at the right time.

Caesum wrote on 2026-05-01, 22:11:

Thank you guys for all the answers! While I didn't know buying a new PSU would be such a hassle, I learned quite a lot thanks to you.

At first I wanted to buy sth quicky so I could take it to my hometown for holidays, but it turned out to be so much more nuanced I missed my deadline so I'm taking a step back and will now look for offers more slowly lmao. I will update this thread if I find something worth mentioning.

Is the old refurbished Seasonic PSU you talked about a few pages ago not available anymore?
To me, that one seemed like a fairly good chance at being an OK option. (Unless it was just too expensive?)

Caesum wrote on 2026-05-05, 20:02:

I thought so too at first but I was told that those might not provide a stable 3v/5v current especially the active PFC ones.

Well, with FSP / Fortron, there's no definitely separation like that.
They did have some really old models with APFC that were actually pretty good for a 5V-heavy PC.
They also have much newer models without PFC or with just passive PFC that were horrible for 5V-heavy PC.
And then there's the reverse of both of these too: old models without APFC that were good for 5V-heavy PCs and new models with APFC that are akin to any modern group-regulated PSU with APFC.
So in other words, you still have to look at the label and use some "intuition" to more or less guess how the PSU would behave.

But again, the ATX-xxxPNF series from FSP / Fortron are not a good fit for 5V-heavy PCs, as I mentioned above.

By the way, I just got one of those ATX-xxxPNF series PSUs (an ATX-350PNF, I think??) last weekend with a PC from the local flea market. It is currently in pieces and drying outside on my balcony (needed a wash, since it was extremely dusty/dirty). Worth noting is that every single capacitor on the output is bulged and needs to be replaced. I didn't even think to plug it in as-is. Will be recapping it when it's dry.

So yes, with old PSUs, it is very possible to get one that won't be any good / will be in need of a full repair / recap.
Thus, I am certainly not claiming that all of the warnings that other members mentioned here about used PSUs are wrong in any way. In fact, they are pretty sound.
In my case, I just buy old PSUs, PCs, and other parts in larger quantities and cheaper... so if I get one PSU that is no good, I move onto the next.
With that said, if you can find a cheap source of used PSUs, just buy a bunch and see what sticks. Again, if you try to stick to the better OEMs, you have a higher chance of getting something decent that also probably works too. In particular, look for old Dell, HP, and Gateway PSUs that may have the following text in the model numbers:
any model beginning with "DPS" or "NPS" - these will be either Delta or Newton PSUs.
any model number beginning with "HP" - these will be HiPro/Chicony.
any model number beginning with "PS-" - these will be LiteON
Anything that says "Astec"

Mike_ wrote on 2026-05-05, 18:26:

Speaking of PFC, is my Enermax active or passive PFC? It doesn't have a huge inductor like Fortron does, and it's only for 230V, with no selector switch.

Active PFC for sure.
Single 400V cap with a large toroidal inductor close by is the tell tale sign, usually.

momaka wrote on 2026-05-05, 21:18:

Is the old refurbished Seasonic PSU you talked about a few pages ago not available anymore?
To me, that one seemed like a fairly good chance at being an OK option. (Unless it was just too expensive?)

It's still available but the vendor seems massively untrustworthy as he has a lot of opinions about lying about their hardware + it costs about 23 EUR / 28USD which I feel is way too much to pay for a sketchy vendor. I will be looking for other PSUs.
Worst case scenario I'll just give up, strip the PC of the more interesting parts and throw the rest to the trash.

I don't think anyone will be selling trully refurbished PSUs for 23 EUR. Probably just visually inspected + cleaned with a shower.

Some of the good brands suggested like Seasonic, Chieftec, Delta, HP may also have been used in servers, some sellers openly admit it locally. It may have been in continous operation for a very long time, like 10 years. Similarly to that is PSUs used in schools, libraries etc. Beware if anyone is selling more than 1-2.

AlexZ wrote on 2026-05-06, 13:29:

I don't think anyone will be selling trully refurbished PSUs for 23 EUR. Probably just visually inspected + cleaned with a shower.

I agree with you because I gather it takes time and resources to fully refurbish a PSU. Unfortunately I don't see any vendors in Poland who actually do that. Mostly they just say it's "refurbished" with no details and when asked directly they either not respond or say they just cleaned it. They also have really bad opinions and some vendors even sell chinease non brands with Chieftec stickers or so. So I feel like paying 23EUR is a bit much for something that could just as well be a big lie.

It's still an option of course and I might buy it if nothing else shows up, but until then I'd just like to search a bit more (and ask for y'alls opinions if you don't mind).

AlexZ wrote on 2026-05-06, 13:29:

Some of the good brands suggested like Seasonic, Chieftec, Delta, HP may also have been used in servers, some sellers openly admit it locally. It may have been in continous operation for a very long time, like 10 years. Similarly to that is PSUs used in schools, libraries etc. Beware if anyone is selling more than 1-2.

Regarding that, I was also asking about PSUs on a Polish board and they recommended another option: buying a 2008-2012 era PSU with at least 130W on the 3x5v line combined. They also gave me a link to a particular PSU as an example. The question is, wouldn't that PSU be still too modern? It's a Chieftec 350W GPS-350EB-101 A and it looks really promising. It's not refurbished, but it's clean and tested using voltage tester, so the offer looks much more solid than anything I've seen yet + it's dead cheap too.

Caesum wrote on 2026-05-07, 17:41:

It's a Chieftec 350W GPS-350EB-101 A and it looks really promising. It's not refurbished, but it's clean and tested using voltage tester, so the offer looks much more solid than anything I've seen yet + it's dead cheap too.

Its a cool power supply for an ATX whith a flyback design.
https://sector.biz.ua/docs/power_supply_schem … -350EB-101A.pdf
It’s a very rare topology.
As a collector’s item for enthusiasts – absolutely, but as a main power supply – I wouldn’t touch it.
It delivers 300 watts out of 350 via the +12V rail.
In other words, it has no advantages over a modern PSU (which comes with a warranty and a money-back guarantee).

Caesum wrote on 2026-05-07, 17:41:

It's a Chieftec 350W GPS-350EB-101 A and it looks really promising. It's not refurbished, but it's clean and tested using voltage tester, so the offer looks much more solid than anything I've seen yet + it's dead cheap too.

That's a PSU made by Delta Electronics - one of the best PSU manufacturers out there.
You can tell by the UL certificate number on the label (the backwards "UR" logo), which is E131881.

It is also a newer 12V-centric design... but I'd trust it would stay in spec/regulation a lot better than most other group-regulated PSUs.
So it may actually end up working OK with your PC, despite being a newer 12V-design.

And if it's dirt cheap, it's definitely worth a try, IMO. (BTW, how cheap are we talking about? If under 10 Euro, might even be worthwhile to grab two of them as a "warranty". 😉 )

Capacitor-wise, it's probably going to be all (or mostly) 2nd tier Taiwanese brands inside, but nothing downright terrible. Just the usual Ltec, and/or maybe CapXon.
I'm not sure what Delta does with their designs, buy these caps usually tend to last in Delta PSUs a lot longer than with anyone else - at least, that's been my observation over the years on many units.

shevalier wrote on 2026-05-07, 18:19:

Its a cool power supply for an ATX whith a flyback design.
https://sector.biz.ua/docs/power_supply_schem … -350EB-101A.pdf
It’s a very rare topology.

It's not flyback, it's single transistor forward (typicaly called STF for short). Technically, they look the same on the primary side. But the transformers and output side is different. Flyback designs are discontinuous in nature, whereas conventional forward-converter designs are continuous. The difference is, the former does not have an output (coupled) inductor/toroid, as the transformer is gapped and thus acts like an open-core inductor. In contrast, the transformer used for continuous designs is not gapped and the output inductor after the transformer serves that function.

As for the Delta above, the only "unique" thing about it is that it actually uses two parallel MOSFETs for its STF topology.

In any case, STF is not a rare topology at all. Delta used it on quite a few of their PSUs starting from the very late '90s / early 2000's up until maybe 2009 or 2010, when they started also producing more two transistor forward PSUs. Two transistor forward is essentially almost the same thing as STF, but here the 2nd transistor (MOSFET) is added to cut-off power "above" the transformer, and then two diodes are used to re-capture the transformers back-EMF and dump it back into the primary bulk cap. This allows the primary-side snubber to be much much smaller, therefore greatly improving the efficiency.

shevalier wrote on 2026-05-07, 18:19:

As a collector’s item for enthusiasts – absolutely, but as a main power supply – I wouldn’t touch it.
It delivers 300 watts out of 350 via the +12V rail.

Collectors item? Why? 🤣
You do realize there are still PSU designs still made with this topology. It's still a very widely used topology. Just not so much with ATX PSUs, mostly because STF typically tops out at about 350-400 Watts. Beyond that, the snubber on the primary really needs to grow in size to keep up with the higher power, and at that point it just doesn't make sense to use this topology.

shevalier wrote on 2026-05-07, 18:19:

In other words, it has no advantages over a modern PSU (which comes with a warranty and a money-back guarantee).

Maybe so. But what about price?
A modern new PSU sure isn't going to cost peanuts. Whereas some of these older used PSUs can be found very cheap and still provide years and years of useful life without breaking the bank - in many cases, longer than the intended purpose of the PC hardware they are intended for. Again, this is a Delta PSU we are talking about here. While not impervious from getting bad capacitors, it is a lot more rare for them.
I run such PSUs almost exclusively around here... and it's not like a have much of a choice either - I have probably around 50 or so retro PCs. Imagine if I had to buy a new PSU for each one of them - I'd be broke already!
Also, you don't hear Xbox 360 owners massively dumping their old power bricks (PSUs) and getting new ones (that is, for those who still have a working 360, which not many are remaining - at least from the older models). Yet their power bricks rarely fail. Most of them are made either by... well... guess! - Delta, LiteON, and Chicony/HiPro.
And laptop owners? It's the same exact scenario.
Somehow people don't frown upon plugging in a 10+ year old laptop power adapter to their retro laptop. But with an ATX PSU, they are treated like it's some kind of a bomb about to go off.
I just don't get why people make this distinction / discrimination when it comes to power sources.

As for warranty, I can wipe my rear end with that more than half the time.
Last year I got a very modern 500W Seasonic PSU from the flea market on the cheap. I knew there had to be something wrong with it, as the stuff that ends on this flea market almost never works right / is there for a reason. So I opened it, and sure enough it was full of bad caps - every single one on the output, that is. The PSU was made in 2022. It still had a 5 year warranty. Luckily, I didn't know this so I recapped it and all was well. But then I did some research on the matter about what I could have done had I wanted to use the warranty. Short answer: even if Seasonic would honor the warranty (despite the fact that I don't have a receipt for the PSU, since I am not the original owner), I'd still have to pay shipping to send it to one of their "repair centers" (aka exchange depots), where I'd likely just get another unit with the same crappy caps back, but just not failed yet. Or maybe, if I'm really lucky, I'd get a refurbished unit from them, where they had a newer revision of the PSU with better caps - but I seriously doubt that. These days, big manufacturers don't have the time and don't want to spend the cash to hire techs to fix warranty units. They just send you a new/replacement unit if the old one has failed under warranty, and that's all.
Well, I say 'no, thanks' to that. Not only is this a waste of time and money for me (to have the broken PSU sent to some warranty depot), but it also does not address the original problem. So with that said, I am glad I didn't know the PSU had a warranty and that I just elected to recap it myself.

momaka wrote on 2026-05-07, 20:01:

It's not flyback, it's single transistor forward (typicaly called STF for short).

You're probably right, judging by the phase configuration of the primary transformer.
About 15 years ago, I tinkered with the same OEM Hipro unit—without a clear schematic—and I really didn't like it. Overall, for this power range, I prefer the classic half-bridge topology.

Just not so much with ATX PSUs

this.
It produces the same high-voltage spikes as a flyback topology, but lacks “built-in short-circuit protection via power limiting.”
I’d buy the inventor of the Two-Switch Forward a beer.

Maybe so. But what about price?
A modern new PSU sure isn't going to cost peanuts.

If the new power supply burned out, you went and replaced it under warranty.
As for the old one—you wrote yourself: “Well, it’s not that complicated. Spot the differences between flyback and SFT topologies. Replace the switches and electrolytic capacitors, and lubricate the fans.”
Of course, any retro computer enthusiast will learn to solder sooner or later, unless he’s a Saudi prince.

Speaking of old supplies.. I just went through the bay and looked for some.

Look at this Codegen one.. the maths do not math on this one at all. I'm waiting on the 520W variant to arrive which has 42A on 5V - for some project. Re-cap worth for sure..

tehsiggi wrote on Yesterday, 15:28:

Look at this Codegen one.. the maths do not math on this one at all.

Oh, and when it arrives, could you post a photo of the +3.3V linear regulator on the MOSFET?
It was present in half of the Codegen power supplies 😀

shevalier wrote on Yesterday, 15:44:

tehsiggi wrote on Yesterday, 15:28:

Look at this Codegen one.. the maths do not math on this one at all.

Oh, and when it arrives, could you post a photo of the +3.3V linear regulator on the MOSFET?
It was present in half of the Codegen power supplies 😀

You mean like this?

https://www.eevblog.com/forum/general-computi … y-in-the-world/

I remember another EEVblog thread where someone got a power supply that was missing its heatsinks about 10-15 years ago.

tehsiggi wrote on Yesterday, 16:57:

shevalier wrote on Yesterday, 15:44:

tehsiggi wrote on Yesterday, 15:28:

Look at this Codegen one.. the maths do not math on this one at all.

Oh, and when it arrives, could you post a photo of the +3.3V linear regulator on the MOSFET?
It was present in half of the Codegen power supplies 😀

You mean like this?

https://www.eevblog.com/forum/general-computi … y-in-the-world/

yep.
I’ve had to deal with a Codegen like that before.
It all ended up with me having to wind an extra winding for +3.3V on the transformer, fit diodes instead of the MOSFET, and route part of the group stabilisation choke windings to the +3.3V channel.
It actually worked, and not badly at all. For a P4-based system where there is no direct use of the +3.3V voltage from the PSU.
Ever since then, whenever I hear “Codegen”, I get the shivers.