It's an Everpower unit - a.k.a. a close sibling of Topower, which were overall pretty decent PSUs back in the day merely plagued by bad caps.
This one actually looks pretty well-built. Maybe it won't do 500W if we have to be realistic, but 350-400 shouldn't be unreasonable.
As for what the label claims in terms of available current for the 5V and 3.3V rails, that may or may not be strictly true. Just check what rectifier parts are attached on the secondary heatsink. If for example, the 5V rail has only one 30 or 40 Amp rectifier, then that's all it can do. Same goes for 3.3V and 12V rails. For identical parallel rectifiers, the current is usually what's printed on one of the parts + approx. 1/3 more. So for two 30 Amp rectifiers in parallel, that would be 30+1/3 (30) = 30 + 10 = 40 Amps.
Yeah, I would suggest to recap it. Fuhjyyu capacitors in Topower/Everpower PSUs pretty much always go bad, just like in old Antecs and old CWTs. At this age, they may be on borrowed time already.
It's an older half-bridge -based unit, so don't go with very low ESR caps (except the 5VSB circuit, as it doesn't care.)
Rubycon YXJ and YXF or Nichicon PW/PS/PJ/PM or United Chemicon LXZ/LXV/LXY/LXF or Panasonic FC or HFQ (obsolete series, but still be can be found online NOS) would work best.
A notch better (lower ESR) would probably be OK too, but it would actually probably degrade the output filtering performance slightly - e.g. the likes of Rubycon YXG, Nichicon HE, UCC KY & KYB, or Panasonic FK <- use only if you have to.
The big 200V 1000 uF caps can stay. They are under low stress and rarely go bad, even from the worst of brands.
As @annalog_programmer mentioned, you should also try to remove as much as possible from the brown glue - at the very least on the primary side. Many of its variations tend to go slightly conductive over time (with exposure to heat and/or moisture), and mess up signaling to high-voltage parts... sometimes enough for things to go banga-ranga. 😁
analog_programmer wrote on Yesterday, 09:24:
If the output voltages are within limits, do not bother to replace any electrolytic capacitors even if they are Fujih...whatever chineese cr*pacitors, until they play healthy enough.
IME, when it comes to the cheaper PSUs that aren't from one of the bigger OEMs like Delta, HiPro/Chicony, LITE-ON, and AcBel, then it's better to recap them. The bigger OEMs listed above know their designs pretty well and so know how to "get away" with using 2nd tier electrolytic caps by not over-working them... and usually they don't use anything that terrible either (e.g. Teapo, Taicon, Ltec, and OST.) This is why they often provide a decent useful life (usually 8-10 years). And for this reason, I don't always recap them, or at least not right away. But eventually they too will develop bad caps. From my own stash of "big OEM" PSUs, very few of them are left with running with the original capacitors. Most of them had to be recapped due to failing already. Anything with CapXon tends to go first, with Teapo and Ltec being a tie for 2nd place.
With the cheaper / less famous PSU manufacturers, the above cycle tends to happen even faster. So when I encounter PSUs like the one in this thread, I'd be more inclined to recap it right away than not to. Same goes towards CWT PSUs. And FSP, I consider it kind of in "the middle".
That said, it looks like the O/P's PSU may be in for a recap already. Going by what I see, the 2nd large cap from right to left in this image...
download/file.php?id=251645&mode=view
... appears to be bulging very very slightly.
*EDIT*
I see this is *not* actually a half-bridge PSU, but a forward topology one. As such, it should be OK to use the slightly lower ESR electrolytic caps I mentioned - e.g. Rubycon YXG, Nichicon HE, UCC KY and KYB, and Panasonic FK.
In a pinch, probably the following can be used too:
Rubycon ZLQ and ZLH*, Nichicon HW*, UCC KZE and KZH*, and Panasonic FM*/FR*/FS*.
The ones marked with a * should be fine for the 3.3V rail, at least.
Being forward topology, it's very likely the 5V rail will have either two 30 Amp or two 40 Amp rectifiers in parallel... in which case, the total available output on that rail should be very roughly 28-35 Amps, at least by my more conservative calculations.
