Mike_ wrote on 2026-04-28, 05:52:
Hmm, would you think it'd be worth the effort to recap that Enermax, even though output seems to be fine or just wait and see? I'd actually like to have that in working condition as it's a pretty nice PSU otherwise, so it would be a lot more likely to see actual use than the Fortron.
It all depends on you / what you like to do with your time and this hobby.
If you intend to keep the PSU long term and use it regularly, I'd say YES, just so that you can have peace of mind that it will be working fine.
But if you do elect to recap it, make sure to use good quality Japanese capacitors. Otherwise, any cheap 2nd tier brands could be just as bad as leaving the original caps in... at which point, recapping it would be pretty pointless.
Also, if you are not keen on spending any money on the Fortron or just want to get it "up and going", you could "recycle" the caps from the Enermax into the Fortron. I will admit, it's probably not the best thing to do... yet I do it all the time 😁 since I have/get too many PSUs, and sometimes the cheaper / gutless wonder stuff is not really worth buying high quality caps for (in fact, it's probably not worth bothering to recap/fix them at all... but when I can do it quickly, at least I can verify everything works in them in case I ever need to pull them apart for parts.) That said, I wouldn't grade the Fortron quite as a cheap / gutless wonder PSU. It's actually a decent PSU too. And I can see it's an APFC-less version, making it all the better (IMO).
Mike_ wrote on 2026-04-28, 05:52:
Ehh, it looks like there's lots of that yellow goo and even a couple of daughterboards with a bunch of small caps in them... If it's just a so-so PSU, I'd probably prefer not to spend the time for it.
You can replace just the big caps on the output / secondary side, as CharlieFoxTrot suggested, if that will make it easier for you. The smaller caps on the daughterboards indeed very rarely fail, so can usually be left alone. I only replace them on PSUs I really care about and/or any PSU that has a circuit(s) that may critically depends on such small caps (e.g. some 2-transistor 5VSB circuits, among others.)
shevalier wrote on 2026-04-27, 19:42:In the FSP on this platform, the supervisor is located on a small vertical board positioned perpendicular to the vertical board […]
Show full quote
In the FSP on this platform, the supervisor is located on a small vertical board positioned perpendicular to the vertical board housing the PWM controller.
Under the fan control board
If the board with the supervisor is present, it’s an excellent PSU.
If not (no voltage monitoring, only general short-circuit protection based on maximum power), then it’s just so-so.
A good source of spare parts.
Not quite.
That missing extra daughterboard labeled "OCP Control Board" does exactly that: OCP (over-current protection). If it's missing, the PSU just won't have over-current protection on the output rails. Voltage "monitoring" (i.e. under and over-voltage protection, aka UVP and OVP) is present and built into the PWM controller. Heck, even the old and very crappy PSUs with only a TL494 typically have a mostly working OVP and UVP. I say mostly, because sometimes the 12V rail and the negative rails lack either OVP or UVP (I forget which of the two was the usual one.) It all depends on how the OVP and UVP is implemented, though. For PSUs with just a TL494 / KA7500 PWM IC and discrete TO-92 transistors handling the PSU protections, OVP and UVP are typically both inaccurate and won't shut down the PSU. Instead, just the PG (power good) line gets pulled low to indicate to the PC and connected hardware to go into soft-off / power-safe mode. This is not a really good way to do it, of course.
But for the FSP / Fortron units, I believe the custom FSP PWM IC is a little more "advanced" in that it's a PWM + supervisor IC, so it has OVP and UVP built-in similar to how AT2005b, ATX2005, SG6105, and SDC2921 ICs do.
As for OPP (over-power protection), that should also be present on the Fortron above. Alas, like most half-bridge PSUs, OPP just may not work/trigger that well or be very exact where it triggers. The reason why is because on half-bridge topologies, the output power is "sensed" remotely by measuring current going into/out of the primary BJTs either via the BJT driver transformer or a smaller 4th transformer. A voltage is then generated on the secondary side of this transformer, which is rectified and picked up either by the PWM IC or a or a sub-circuit signaling to the PWM IC fault pin. Unfortunately, the rectification sort of "averages" the current pulses. So as is often the case, OPP on half-bridge PSUs is usually not very quick-acting nor very precise. In fact, it requires a careful setup to even be properly operative - something that is almost never done right on cheapo PSUs, which is why it's so easy to blow their primary BJTs when they get overloaded.
Now, I have not tried abusing one of these half-bridge FSP/Fortrons to see what they do. But I presume their OPP should be working more or less correctly.
As for any newer FSP / Fortron based on a forward-converter topology, the OPP should definitely be working and on-point, as is almost always the case with forward-converter topologies (since the PWM IC is on the primary, allowing it to do pulse-by-pulse current sensing and limiting, which is way more accurate.)
tehsiggi wrote on 2026-04-28, 09:33:
My go it often: If I'm re-capping anyways, I go all in.
Same.
Though I should note that on PSUs without APFC (i.e. PPFC or no PFC), I almost always leave the two big 200V primary caps alone, since they almost never fail, even if they are a really crappy brand. And these big caps tend to be expensive. So this is one reason I prefer PSUs without APFC - just one less thing to fail / worry about.
On PSUs with APFC, if the capacitor is not a Japanese brand or not rated for at least 420V, I pull it out and measure it. If still in spec, I might put it back in, or I might replace it with something better if I have anything in stock (I rarely do, however.) Generally, I like to see the cap be rated for at least 420V (450V better) and be a Japanese brand and rated for 105C as well. Anything less tends to get cooked / overworked over time. On that note, this is why I also absolutely hate PSUs with "eco" fan modes - it absolutely and unnecessarily shortens the life of the primary cap (my theory is the manufacturers did the calculations and probably did it so the caps will last past the warranty, but maybe not much more than that, seeing how many PSUs with failed primary caps I get over the years.)
Caesum wrote on 2026-04-28, 21:00:
Okay one last question then before I decide to buy. I found this cheap FSP Fortron that has pretty good 5V and 3.3V, but it says that it's an Active PFC one. Do you think it could work or should I ignore any active PFC PSU entirely?
I am not 100% positive, but I think that FSP might be from the newer ATX version 2.x era, and so the 5V rail (despite its high-ish rating) may still sag quite a bit without a proper load on the 12V rail.
FYI, there was a thread on Vogons about this particular problem with a similar (or same) model FSP.
So my hunch is to avoid this one, at least for now.
Caesum wrote on 2026-04-28, 21:00:
Also another silly question, how do you find out if a PSU is non-active PFC? I wouldn't have guessed this one was an active one if it wasn't for this huge sticker on it lmao. I know good rule for me is to look for PSUs with beefy 3V and 5V but some modern ones have them good too and yet they are not recommended for retro hardware.
Voltage selector switch = never APFC
That's about the only 100% guaranteed way to tell.
Some other ways to tell if there is APFC or no APFC:
1) if label says something along the lines of "full range input" or "100-240V AC in", then it very likely is a PSU with APFC. Only exception would be for some very old AT PSUs rated under 100 Watts, as these might use flyback topology and thus can be full range input without actually having APFC.
2) On some really cheap PSUs, the voltage selector switch may be missing, simply because the manufacturer was too cheap to install one (in order to save a few cents). On these, the PSU may often be labeled as either "115/120V only" or "230/240V only". It's not a 100% proof way to tell for sure, but in probably over 90% of the cases, the PSU will be without APFC.
shevalier wrote on 2026-04-29, 04:37:
Whilst a unit without an APFC has most likely been sitting on a shelf for the last 15 years, one with an APFC has most likely been in operation somewhere in an office for the entire 20 years since its release.
Well, not necessarily.
Granted, OEM PC builders started switching to PSUs with APFC around 2nd and 3rd gen Intel i-series... so yes, roughly 15 years ago. But that does not mean older office PCs with PSUs without APFC were not used longer. I actually have seen quite a few older PCs running to as late as 2019 / pre-covid times. Windows 7 EOL was really the "last nail in the coffin" to most old office PCs that were still kept around otherwise.
Also, I'd venture and say that an older PSU without APFC sitting on the shelf is probably a lot better than a newer PSU with APFC that's been hammered ever since it was bought... but it all depends on the specific PSU we are talking about, so I probably shouldn't make sure generalizations (and probably neither should you 😉 )
Caesum wrote on 2026-04-29, 09:37:
Frankly I am surprised as I thought nobody makes passive PFC units anymore. Although they say it should be used with "current and next generation of platforms".
Oh yes, they are still made.
It's just about all of them fall in the cheapo "gutless wonder" territory. Most serious (or even semi-serious) manufacturers have moved onto APFC, simply because it allows them to pack more power with smaller parts / into smaller space. And it all has to do with energy storage.
For those curious, it particularly has to do with the size of big electrolytic capacitor on the primary size. The energy, E, stored in a capacitor can be expressed as follows:
E = 0.5 * C * (V)^2
where C is the capacitance in Farads and V is the voltage in Volts.
What the math tells here is that the energy stored in the capacitor increases only linearly with an increase in capacitance. But an increase in the voltage increases the stored energy exponentially (by a square factor.) As such, the manufacturer can use capacitors with lower capacitance, so long as the voltage on the capacitor is increased to make up for that stored energy. In essence, it allows manufacturers to squeeze more performance from smaller parts (primary caps). And as a bonus, it makes the PSU full-range (typically 100-240V AC) -capable, which is generally another pro for the user, as now he/she won't have to deal with voltage selector switches and the possibility of messing something up with that. However, it's not all without a trade off. Firstly, the APFC circuit is essentially a voltage-boosting circuit (well, very roughly said), so it will have losses... and thus actually decreases PSU efficiency in almost all cases. Also, because manufacturers are always looking for ways to save on costs, they regularly use only 400V -rated primary caps when 420V or (better yet) 450V should have been used. Most APFC circuits run north of 375-380V DC... which cuts it really close to the primary capacitor's voltage rating... which is why the primary reason primary caps in PSUs with APFC don't last as long, along with that from the high-frequency ripple / harmonics from the APFC switching.
Caesum wrote on 2026-04-29, 09:37:
Akyga is a Polish brand if anyone wonders.
That may be, but the PSU is more than likely something Chinese inside.
In other words, Akyga is probably just the importer company.
Mike_ wrote on 2026-04-29, 19:18:
On second thought, I guess I could recap that Fortron PSU.
I happen to have 3300µF 6.3V chemi-con KY series caps around that have been left unused, I bought them several years ago for recapping a motherboard but ended up not using them. So could I use them for replacing those 10V 3300µF caps? It sounds like it should be fine, as those caps must be for 5V rail. Or is there any reason why this would be a problem?
Should be fine.
Chemicon KY is a usually a very fit replacement for Teapo SC and SY series... or for PSU use, in general.
Mike_ wrote on 2026-04-30, 07:18:
Yes, I'd also say that if the voltage ever gets up to 7V, there are worse problems... 😁
Indeed.
On some older motherboards with TTL ICs that are rated for maximum of 5.5V and directly connected to the 5V or 5VSB lines, anything past 5.5V for more than a few moments could technically fry them.
appiah4 wrote on 2026-04-30, 11:38:
I bought around 400 caps today. Low ESR Samyoung and Samwha mostly for 1000uf and above, some Koshin for 470uf and below.
...
I had no issue with these brands so far..
Then this one goes out to all of the TV techs that have worked on Samsung and LG TVs and cried silent tears. 🤣
Now I indeed wouldn't go out and say these are all terrible brands (well, for Koshin, I still can't make up my mind.) In many cases, Samwha and Samyoung caps did last well past the warranty of many of the stuff/devices they were put in. So I won't down-talk them (too much). But just like Teapo, OST, Ltec, and all these other 2nd tier cap brands - once you go past the 7-10 years of use, the failure rate really increases and can be "all over the place". With Rubycon, Panasonic, Nichicon, and the rest, you pay more, but you don't have to worry for easily at least twice that time. So you get what you pay for them.
But I get it, not every piece of hardware really needs to be made to last forever (or as long as possible.) On that note, I probably shouldn't preach here, since I regularly recycle caps of all brands in all kinds of hardware.
That said, I do want to make note here that you can get (some) high quality Japanese brands of capacitors for not that much more than the other lesser brands - usually NOS on Ebay. Just have to be careful, since there are a lot of fakes / counterfeits. But I have repeat-shopped from a few sellers and have gotten genuine parts each time. Admittedly, some of the caps were a little old. But I will take an old(er) Japanese cap almost any day over a 2nd tier brand.
