PSU worth replacing caps?

[stj]

Re: PSU worth replacing caps?

they usually touch in modern supplies even when they are the right size.

if you can fit 10mm then look at the list: D&H can use the same part number as E
C would be EEUFR1C332L IF you can fit 12.5mm in.

[mikay786]

Re: PSU worth replacing caps?

Thanks again stj

what would be best these Panasonics or the stuff behemoth posted?

Farnels is only 20mins away so useful.

[stj]

Re: PSU worth replacing caps?

i dont know anything about the ones behemoth listed,
i'm only really used to panasonic's range and about 20% of rubycon's range.

he was listng nippon-chemicon part numbers.

[Behemot]

Re: PSU worth replacing caps?

also that 10mm 3300uf cap would normally be 12.5mm

In like 60 % cases there are D10 caps used. Than >35 % D8, <5 % D12,5. So I guess "normal" (by definition) is D10. The other thing of course is that besides two special types - Samxon RS and Chemi-Con KZN both which I posses - you won't get that from any quality manufacturer.

Same with D8 longer than 20 mm. They now make pen caps in D10 so you can get caps as long as 50 mm but we will have to wait till they come out with brand new D8 pencaps nobody thought about before (besides chinese brands which have them for 10 years already and I am suggesting this can for over a year now).

And no, he won't fit D12.5 caps instead of D10 and D8 instead of D10 simultaneously. I already asked about KYA 3300/6,3s instead of those custom D8x30 and it would at least not be easy (or maybe impossible) to fit them, no way all of them. Because as you can see, it is 1,1kW unit in quite small chasis, that means it is frelling crowded inside already.

And personally, I would suggest you to get some of those 3300/16 caps I have as Panasonic will not make such capacitance in D10x30 for at least another decade. They just don't do custom stuff and don't care about ATX power supplies at all. How often do you see Panasonic caps on PSU secondary? One of 20 units (lets think quality caps used only)? It is 80 % Chemi-Con, 18 % Rubycon and maybe 2 % the rest. And while you will be at it, you can take some of those KYB 2200/16 I'll get in about 2 months as they are equal to your belowed Panasonic caps but I have everything much cheaper than those thieves from Farnell, Digikey, Distrelec, TME and all the others.

You know what? One such smaller local reseller, the GES Electronics, resell the same Chemi-Con KXJ 150/400 caps as I do. We both take them from the same distributor in same numbers and for the same price. I am 43 % cheaper even adding VAT (which I will soon start applying, so for non-VAT payer, I am now almost 60 % cheaper). And believe me, I already have very nice margin. To give you some idea, my expectation based on this experience is all these big-reseller thieves usually start their prices on three times (and more) the buying price, regarding these capacitors we talk about.

[stj]

Re: PSU worth replacing caps?

In like 60 % cases there are D10 caps used. Than >35 % D8, <5 % D12,5. So I guess "normal" (by definition) is D10. .

when i said "normally" i meant from parts suppliers.
we both know these bastards making psu's have a habit of ordering custom size parts or using caps that probably arent close to what's printed on them.

[stj]

Re: PSU worth replacing caps?

How often do you see Panasonic caps on PSU secondary?

i see lots of panasonic FC in psu's from SGI systems.
they are made by sony and delta mostly.

[Behemot]

Re: PSU worth replacing caps?

I think I never seen SGI. And nothing in Delta so far, nope. NCC and Rubycon.

[stj]

Re: PSU worth replacing caps?

SGI supplies are huge, 3times the size of ATX and usually about 800w continuous output without being stressed.

they look *very* nice inside!

[CapSwapper]

Re: PSU worth replacing caps?

not worth it. it has multiple cap failures which stresses other components. a cap can either fail open or become a huge resistor. those other components might be borderline on the verge of blowing. so you can spend money and caps and something else will fail down the line. once I see smoke the PSU goes in the dumpster

[c_hegge]

Re: PSU worth replacing caps?

^
Disagreed. Most of the time, the caps will be the only things that have failed, and the other components do NOT fail as a result. I've recapped hundreds of PSUs over the years and only had 1 or 2 fail shortly afterwards.

[kaboom]

Re: PSU worth replacing caps?

not worth it. it has multiple cap failures which stresses other components. a cap can either fail open or become a huge resistor. those other components might be borderline on the verge of blowing. so you can spend money and caps and something else will fail down the line. once I see smoke the PSU goes in the dumpster

Surprise?

Once that happens, it's too late for just a cap job.

The entire point of a cap job is to prevent a catastrophic/cascading failure.

A recapped PS will be more reliable than most new ones of questionable quality.

I've recapped Deltas, Hipros, AcBels, Bestecs, even CWT/Antecs for myself and others; all have had less ripple than before, plus they've been gone thru and heatsink screws were (re)tightened, and bad solder touched up, if required.

I take pride in my work; the last thing I'd suggest for someone with, say, an HP 3K SFF, would be to get one of those fake PSUs from ebay, when new caps for the (superior) original supply cost less than that junker!

Nope- I don't by "newer is better" in that case at all. Rip one of those "replace power" POSes open and you'll see- mssing filter sections, both on line and load side, small caps with "questionable" or outright fake values, diodes/transistors too small for (fake) wattage, absolute garbage passing for output leads (26-28AWG "disguised" by large insulation), fuse is in the neutral, etc.

As for repairability of legitimate PSes after a blow up? That depends on if there are SMTs and how far one wants to go...

[CapSwapper]

Re: PSU worth replacing caps?

there also environmental factors. did the psu take a huge surge from the power company or a lightening strike? was a surge protector ever used? el cheapo surge protector? one good surge to a cheap psu will cook it

[c_hegge]

Re: PSU worth replacing caps?

Again, no. It does still run despite the caps, which indicates that it did not fail from a surge, and that there is nothing else wrong. It's just a case of low quality caps doing what they do best.

[kaboom]

Re: PSU worth replacing caps?

there also environmental factors. did the psu take a huge surge from the power company or a lightening strike? was a surge protector ever used? el cheapo surge protector? one good surge to a cheap psu will cook it

First of all, at the end of a long branch ckt, mutual coupling between the three conductors in typical NM cable (Romex) will result in common-mode voltage between a remote ground point (See below, 1). So the MOVs will not conduct in such cases, since all lines will be equal in potential.

As far as the connected device (computer, whatever) is concerned, everything could be 5KV above that remote ground, but as long as L-N is ~120V and L-EGC/N-EGC are ~120/"lowish" respectively, it'll run just fine.

The real hazard is when this "floating" computer dumps potential into true grounds originating someplace else.

Point 1: No CATV/telco/waterpipe bond. Not ground in this case, bond- attempting to keep all conductors as near as possible in potential, in an attempt to minimize destructive current flow during lightning strikes or "line hits."

Bonding the CATV to the service entrance causes direct or induced currents on the coax shield to return [i]at the service.[i] W/O a CATV (or telco) bond, those currents will instead travel up the coax to your cable (DSL/dialup) modem and blow thru the isolation transformers, returning to your electric service/waterpipe/rod return, via your expensive stuff.

In the case of silly USB cable modem interfaces, that USB cable will become the return path for the current "delivered" to it; by way of the modem, energized by the unbonded coax. In that case, the motherboard is damaged if the USB-chassis bond is poor. It's often damaged anyway, since the current coming up the shield of the USB cable induces current into the signal wires inside, which often go straight into the bridge chip.

Similar events happen with long audio cables, from PCs to other equipment, especially if said equipment is also grounded.

Bonding is far more imporant. Considering how cheep some "surge directors" (poor crimps, tiny brass strips, little contact area) are made, I'd just as soon build an old fashioned plugboard with metal electrical boxes...


Now, for those power supplies that still work but need a recap, they're obviously not degraded in any way. The only component that may be an issue is the first across-the-line capacitor in the unit. They're the first thing to swallow any differential transient coming in; that's not their main purpose, which is forming a section of the line filter, which prevents noise from the power supply from getting out.

While there may be other caps across the line, after common mode chokes, remember that no such inductor is perfect. There's leakage inductance too, even though this is minimized in their design. So that leakage inductance is in series with any downstream X-caps after the first one, limiting their admittance to any incoming spikes.

Sustained overvoltage, though, will kill 'em all.

Speaking of which, anyone care to bring up the proverbial lost neutral?

Like when people think their hacked-in portable generator "creates surges," but are using a ragged dryer cord (three-wire! ) to backfeed their system without checking how the genny's wired.

The slightest imbalance, especially when that fridge, oil burner, or washing machine start, kills all electronics on the opposing leg. Especially those "surge strips," which, if it's from before UL required thermal fuses, smoke and burn as they attempt to shunt and "balance" the overvoltage.

I'm not having this "It's junk, you need a new one after the caps blow" nonsense again.

When myself and others on here do recaps, the result is something better than the original unit. Plus, in having someone go thru the unit, things like long leads and bad solder can be corrected.

I take pride in my work; the last thing I'd suggest for someone with, say, an HP 3K SFF, would be to get one of those fake PSUs from ebay, when new caps for the (superior) original supply cost less than that junker!

Nope- I don't by "newer is better" in that case at all. Rip one of those "replace power" POSes open and you'll see- mssing filter sections, both on line and load side, small caps with "questionable" or outright fake values, diodes/transistors too small for (fake) wattage, absolute garbage passing for output leads (26-28AWG "disguised" by large insulation), fuse is in the neutral, etc.

Capisce?
(pronounced "ka-peesh!")

[stj]

Re: PSU worth replacing caps?

Speaking of which, anyone care to bring up the proverbial lost neutral?

Like when people think their hacked-in portable generator "creates surges," but are using a ragged dryer cord (three-wire! ) to backfeed their system without checking how the genny's wired.

The slightest imbalance, especially when that fridge, oil burner, or washing machine start, kills all electronics on the opposing leg. Especially those "surge strips," which, if it's from before UL required thermal fuses, smoke and burn as they attempt to shunt and "balance" the overvoltage.

sounds like how Haliburton was killing people in the showers in Iraq!!

[CapSwapper]

Re: PSU worth replacing caps?

my rule of thumb is I only recap a psu that has never failed whether it be new or used. I've experienced the fizzle pops and fireworks of a psu with cheap junk caps that took out a motherboard with it. every psu I ever recapped has never failed so far

[c_hegge]

Re: PSU worth replacing caps?

^
The pop and smoke was probably just one of the caps exploding, not a transistor. Otherwise, the thing probably wouldn't still run.

[mikay786]

Re: PSU worth replacing caps?

Again, no. It does still run despite the caps, which indicates that it did not fail from a surge, and that there is nothing else wrong. It's just a case of low quality caps doing what they do best.

The PC the PSU came in hadn't been used for 3 years and had been in storage, when the owner powered it on for the first time in 3 years did it fail.

The unit still functioned after this once I got it and even had PC running for a few hours, although did make some buzzing sounds at first.

Anyhow just waiting on caps to be delivered from behemoth, hopefully should be fine. Fingers crossed

[Pentium4]

Re: PSU worth replacing caps?

Nice. The PSU will be fine once you have the new caps in it. And it will be fine for as long as you want to use it, if it in fact has a San Ace fan too

[Behemot]

Re: PSU worth replacing caps?

Still have not arrived? Strange. Cannot check the tracking now though, have the papers in my workplace. I am sitting on my ass at home writing review now…