Hardware Insights Forum

Believe it or not, I found this thing BRAND NEW in the box from 2007. The two 1000µF 10V Asia X on the 5VSB bulged just sitting in the box. They read 1226µF 0.68Ω, and 968µF 1.15Ω. It looks almost identical to the Allied that JG reviewed, however, this one has a 20 pin 2008 IC instead of the usual 16 pin 2005. I want to recap it, as these are decent PSU's for basic systems, with that 40A schottky on the 12V. My question is about those extra caps for the controller. They're three 0.1µF 50V Asia X TNX. I know it's a small value, but should these still be replaced? I don't like the combination that these have been sitting for 8 years, being a crappy brand, and being GP. I don't have any suitable replacements. Lowest value I have is 0.47µF 50V nichicon PW. I could maybe replace them all with ceramic caps, I'd have to find some though.

ALSO, I think I discovered that Globe Fan may have manufacturing ties with Solytech? I've never seen Yang Chun caps outside of Solytech, and the Globe Fans in this unit, if you look carefully, I could see Yang Chun caps installed on the board of the fan.

Typical Fuhjyyu caps bulging just from sitting...

If I'm replacing an electrolytic below 1uF, then I usually use a ceramic cap to replace it. The PSU looks decent enough. I'm not sure if it would do 460W, though. 400W might be OK if the silicon is up to the job.

I guess that manufacturing ties with YC and Globe fan are possible (although you do occasionally see Solytech PSUs with other brands of caps and fans.)

I know. A lot of the other caps were almost out of spec that weren't bulging.

So, would you use 1kv or 2kv ceramic caps? Don't their voltage capabilities drop with temperature increase?

Yeah, it does look good. The switchers are 2SC3320's. The two fans move a lot of air, and they're both ball bearing. Main toroid is nicely wound and of good size.

guess that manufacturing ties with YC and Globe fan are possible (although you do occasionally see Solytech PSUs with other brands of caps and fans.)

Yeah, I know they used caps such as G-Luxon, Elite, Su'scon, JunFu, DON before but those were mostly in their old ones. They're purely Yang Chun except for the input. They're all ANODIA now on the high voltage caps. Same strange color though.

With 2SC3320s and 820uF primary caps, 400W would probably be OK. I just use the same voltage rating as the original caps (usually 50V) when replacing with ceramics, although there's probably no reason why a 100V or a bit higher rated cap wouldn't work. It's actually the capacitance that changes with temperature, not the maximum voltage. Exactly how much it changes depends on the dielectric used. X7R is a fairly common dielectric (and I'm pretty sure it's what I've been using to replace the 0.1uF electrolytic in Antec SP-xxx power supplies. The total capacitance change is rated at 15% over the -55*C - +125*C temperature range.

see http://en.wikipedia.org/wiki/Ceramic_ca ... apacitance

Oh gotcha, it's the capacitance that changes. Okay, and it looks like where those are located they won't get very hot. Thanks for your input. Do you buy new ceramic caps or do you just pull them from old psu's?

I just get new ones. They are only a few cents

Got some new ones and installed them. Looks great now. Runs well. Thanks for your input Not sure what I'll do with it yet but it will probably go into the typical Core 2 Duo system with a 9500GT I expect it to last forever now Since it has good caps now and ball bearing fans.

c_hegge wrote:Typical Fuhjyyu caps bulging just from sitting...

I'm surprised they weren't more out of spec given the fact that they had bloated. That just goes to vindicate that the degree of "swelling" in an electrolytic capacitor has little correlation with its actual internal condition.

I think they bulged because they just sat there unused for so many years without any voltage across their plates. The oxide layer just thinned and thinned until some weak point broke down in the electrolyte and hydrogen gas formed. I imagine they don't smell too pleasant, either. However, this is not just characteristic of Fuhjyyu capacitors or NCC KZGs and KZJs. The same thing can happen to any capacitor with bad aluminum foil, bad electrolyte, and bad seals (bad bungs - the reason why early yellow Fujitsu polymers were sometimes found bulged with little stress - moisture ingress - bad seals entrapped moisture inside and ESR increased until they vented).

Those chemical reactions are things that only heat accelerates and only the thickening of the dielectric prolongs. They (the chemical reactions in question) will eventually reach fruition whether it takes two years or twenty years. With aluminum that's only of 97% purity and without the proper neutralizers, additives, and oxidizers in the electrolyte (in order to dull or keep at bay the aluminum foil and the water base electrolyte's tendency to attack the aluminum oxide layer), these capacitors have nothing to preclude the formation of pressure and hydrogen so they can bulge on the shelf with no encouragement across the leads. I think it was Chemi-con that said in a certain document that not only is water-base electrolyte, at some point, no longer effective at reducing the impedance in a capacitor, but that lytics that use that technology have a poor aging factor (very interesting!). That said, even good capacitors will eventually fail as it is natural for the liquid electrolyte inside to eventually evaporate, but with the 99%+ purity of the aluminum and the proper inhibitors and passivating mechanisms in the electrolyte, those capacitors will far outlast the useful life of the equipment they're assembled into.

The 9500GT IIRC is somewhere in between the 8600 GT and 8600 GTS in terms of performance, though both of those cards were considered disappointments for the time (2007) given that they were deemed budget cards. But the alternative at the timeframe was an 8800 GTS and those beasts not only often ran way too hot for their own good but weren't too light in way of vacating space. GPUs have come a long way since then, though. That PSU does look like it would be decent for a basic system at least.

Wester547 wrote:That said, even good capacitors will eventually fail as it is natural for the liquid electrolyte inside to eventually evaporate, but with the 99%+ purity of the aluminum and the proper inhibitors and passivating mechanisms in the electrolyte, those capacitors will far outlast the useful life of the equipment they're assembled into.

"Useful life" depends on the equipment in question. (How long does it take for an audio amplifier to become obsolete, for example? Even in the case of PCs, you don't need anything new for e-mail or word processing.)

I noticed that both Chemi-con and Nichicon say that:

• The lowest ambient temperature used for lifespan estimation should be 40°C
• The maximum expected lifespan should be 15 years (131,490 hours), even if a higher value is calculated

This includes conductive polymer capacitors, by the way.

That first point makes you wonder why they even bother making 85°C 1,000-hour capacitors…

Of course there are other things that can be expected to fail sooner (HDDs, BGAs, etc.). Ceramic capacitors can last almost forever — or at least, they could until recently…

I noticed that both Chemi-con and Nichicon say that:

• The lowest ambient temperature used for lifespan estimation should be 40°C
• The maximum expected lifespan should be 15 years (131,490 hours), even if a higher value is calculated

This includes conductive polymer capacitors, by the way.

I think that includes shelf life as well, not just power on hours. That said, solid polymers should also far outlast lytics because again, no liquid electrolyte to dry out.

That first point makes you wonder why they even bother making 85°C 1,000-hour capacitors…

Ha! You think that's bad? Remember the Aerogel "super capacitors" clock capacitors that were notorious for leaking in the original Xbox? They were only rated for 1,000 hours at 70*C with a maximum temperature of 70*C. Yes, 70*C. Look at the datasheet. Coupled with those crappy circular bungs they used, no wonder they failed prematurely... (at least inconsistently - the ones I pulled had bullseye bungs)

Of course there are other things that can be expected to fail sooner (HDDs, BGAs, etc.). Ceramic capacitors can last almost forever — or at least, they could until recently…

Ceramic capacitors, like tantalum capacitors, are perfectly capable of burning a hole right through your motherboard or HDD PCB, or even GPU PCB, without any discernible reason. Interesting how Intel can voice concern over that but not over using Lelon lytics on their motherboards, for an example...

As to "useful life", I'm talking around 30 years or so. Of course some stuff is nice to keep just for the sake of boasting or checking out vintage electronics of old. And that chip of the year may have a 2008 date code but those transformers still have a 2005 date code.