My new Seasonic has questionable caps.

[PCBONEZ]

Re: My new Seasonic has questionable caps.

Who worked for NASA?

[PCBONEZ]

Re: My new Seasonic has questionable caps.

acstech,
Did you work for NASA?

[acstech]

Re: My new Seasonic has questionable caps.

Oh heck no.

[PCBONEZ]

Re: My new Seasonic has questionable caps.

I don't see how a list of uF values and voltages with no other information is going to give away anything to anyone.
-

Lacking that information you claim as to the difference in cost has no credibility.

.

[Spectre]

Re: My new Seasonic has questionable caps.

So your generalization about all PSU's that are recapped is based on a very small sample size of a problem model of PSU.

How is it that you determined the caps failing caused the failure rather than some other defect causing the caps to fail.
-
Caps are very unlikely to case a failure on a load tester.
If it failed on a load tester with new caps then it's not the caps causing the problem.
More likely under rated MOSFETs or tranformers which you didn't change.
.

I am getting really tired of repeating myself because you want something to be there that isn't. The short version one more time. When load testing units very few units with questionable caps survive load testing, as such there are very few cases when the unit would be recommended based on my testing so that the capacitors would come into play as a reason for the unit to fail long term sicne the unit has already failed thereby nullifying any capacitor issues. The best examples of this not occuring are with Turbo-cool's and their Teapo's. Which do fine in the field.

So you see if they get elminated before the unit is ever opened to inspect the capacitors and component selection questionable capacitors are a non-starter because the unit is already dead and gone. Do you get it now?

I don't see how a list of uF values and voltages with no other information is going to give away anything to anyone.
-

Lacking that information you claim as to the difference in cost has no credibility.

.

Its an unannounced unreleased product so it is still covered. That is fine you will find out once the unit becomes available or if someone else wants to divulge the information. And again for the last time the wholesale price of the capacitor is not the end cost in changing them in a power supply.

Lacking inforamtion though is a funny thing.

Who worked for NASA?

It is called sarcasm.

[PCBONEZ]

Re: My new Seasonic has questionable caps.

>>>
And again for the last time the wholesale price of the capacitor is not the end cost in changing them in a power supply.
<<<

BULL!

You aren't pulling out installed caps and putting in better ones at a factory.
-
There is no difference in the cost of *installing* good vs bad caps in a new unit.
-
The only difference is the cost of the caps.


~~~~

>>>
It is called sarcasm.
<<<

Good.
Because while I might ride a dragon I won't ride a space ship.
They are way too dangerous.

.

[PCBONEZ]

Re: My new Seasonic has questionable caps.

You are repeating yourself because we are not talking about the same thing.

You made a blanket statement that recapped PSUs are not reliable.
-
Come to find what you are really saying is new PSUs that fail load tests tend to have bad caps and so aren't worth recapping.
- That is probably true but that's not where I/we/here are coming from.

A PSU that's been in service for a year (or 2, or 3+) and then fails due to the effects of the capacitors aging (which is accelerated with crap caps) performs perfectly well when re-capped. (Or at least as well as it originally did.)
- That's what the population here deals with every day and we KNOW it's true.

We weren't really talking about the same thing.

...

[acstech]

Re: My new Seasonic has questionable caps.

A PSU that's been in service for a year (or 2, or 3+) and then fails due to the effects of the capacitors aging (which is accelerated with crap caps) performs perfectly well when re-capped. (Or at least as well as it originally did.)
- That's what the population here deals with every day and we KNOW it's true.

QFT!!!

[PCBONEZ]

Re: My new Seasonic has questionable caps.

Dispelling the fan myth:

Hmmm,

A 500 watt with ONE TINY lil 40mm fan.
-
Sparkle/FSP , FSP500-80BU , 500w , single 40mm Fan , -> MTBF 100,000hr

http://www.fsp-group.com/english/1_p...d=61&proid=101

A 1200 WATT with only ONE 80mm Fan.
-
Zippy/EMACS , PSL-6C00V , 1200w , single 80mm Fan

http://www.zippy.com.tw/P_product_de...code=PSL-6C00V


OH SHIT! In the learned wisdom of PSU reviewers those are gonna melt down into lumps of useless metal!!

OH SHIT! OH SHIT! - They'll catch fire!!

Spectre!
You better go call the Engineers at FSP and EMACS and tell'em they don't know what they are doing!
I'm sure they'll listened to you.
You obviously know more about what it takes to cool a PSU than those Engineers do.

~~~~~~

Those are power supplies where the priority is reliability in a brutal environment (servers) and not in impressing ignorant buyers (and magazine editors).

As I said before:
It does NOT take two fans to cool a PSU.
That's load of advertising HYPE.

The hype as existed so long and been repeated so frequently that people actually believe it and call it 'common knowledge'.
-
Once upon a time common knowledge said the world is flat.

.

[PCBONEZ]

Re: My new Seasonic has questionable caps.

Well if all you want is a circulation of a couple hundred readers go for it. But people do not read the parts that I get the liberty to spread in there let alone get more detailed. So sure go read Gabbe's stuff he does a great job but it doesn't draw the readers neccesary to pay the bills around here.

I actually in reviewing the Alexa ratings within a month of him starting in 2006 he had 1/3 of the readers you did.

Since then his readership has been stable or slightly rising, yet yours has rapidly and steadily declined.

At present there is no significant difference in your page views and his and he's only been around 2 years to your 10? or something...

.

[acstech]

Re: My new Seasonic has questionable caps.

Well I finally got around to "load" testing it. I put "load" in quotes because I dont have the fancy test equipment the big sites have, so I do the ghetto method of running 3dmark06 while it's hooked to a multimeter. So of course I'm not going to have ripple specs or anything like that.

With the system in my sig (see page 1) the voltages are as follows:
Idle:
12v: 12.05v
5v: 4.96v

Load:
12v: 12.05v
5v: 4.97v

Under this same load my old Scythe 550w would drop about .15v on the 12v.

Apparently this thing is well built other than the caps. I think I'll look into recapping it.

It would be interesting to see how this unit does under a full test.

[acstech]

Re: My new Seasonic has questionable caps.

So far, for the OST replacements, I have RLP --> Panasonic FC, and RLX --> Panasonic FM. Does that sound about right?

The JPCE-TUR and Teapo are yet to be determined.

[PCBONEZ]

Re: My new Seasonic has questionable caps.

Office with PC with caps data on it is apart for carpet cleaning.
Can't look up OST # right now but FM in the outputs sounds more than adequate and FC for the others is fine. Physical sizes on FC and FM is sometimes a problem so watch that.

[acstech]

Re: My new Seasonic has questionable caps.

Well, I just had it apart, and here's what I found:

Teapo SY 1000uF 16v 8mm (the pad was for a 10mm and the cap was raised off the board)
Jpce-tur 2200uF 16v 10mm x
OST RLX 3300uF 16v 12.5mm
OST RLP 2200uF 6.3v 10mm
OST RLP 2200uF 10v 10mm
OST RLX 2200uF 6.3v 10mm
OST RLX 3300uF 10v 10mm
2x OST RLS 220uF 16v, probably 6.3mm

I'll have problems finding replacements for the Jpce-tur and the 3300uF 10v OST.

I like to match the cap series to their replacements, I think it makes for a better recap. Or maybe I'm paranoid again.

[PCBONEZ]

Re: My new Seasonic has questionable caps.

The 3300uF 10v.
Trace the line.
If it's on 5v or 3.3v output you can use a 6.3v cap.
PSU's tend to use 10v where 6.3v could go because 10v usually has better ESR/Ripple.
A better quality cap using 6.3v will get to the same place.
-
If it's not on 5v or 3.3v output then a 16v would be an acceptable replacement.

.

[PCBONEZ]

Re: My new Seasonic has questionable caps.

>>>>>
I like to match the cap series to their replacements, I think it makes for a better recap.
<<<<<

Me too. I totally agree.
Sometimes there just isn't a perfect replacement so you have to compromise.

.

[PCBONEZ]

Re: My new Seasonic has questionable caps.

Break from the carpet again! Woo Hoo!

Jpce-tur 2200uF 16v 10mm
If you can't fit a 12.5mm then you'll need to talk to Topcat or resort to eBay.
I only found one listed anywhere I usually look and they are out of stock until April.
Best thing would be to squeeze in a 12.5mm because those are easy to find in that size.

Same goes for the 3300uF. If 12.5mm will fit will be easier to find.

If either of those are output caps don't be afraid (like a lot of people are) to jump up to much much better ESR and ripple ratings.
(It might be -necessary- to get them in 10mm.)
If you think about it the designer of the PSU has no idea what motherboard it is going to be connected to and the output caps in a PSU are in parallel with all the caps on that rail on the motherboard. The specs/requirements are pretty loose there. The minimum ripple/ESR they will choose will be worst case (a crappy motherboard) and the maximum will be a cost consideration.
-
What I'm getting to. (By way of two trips around the world)... Is that any effect of changing the output caps will be reduced by all the caps in parallel with it on the motherboard and there are usually only 1 or 2 output caps in the PSU to xxx many in parallel with them on the board.

.

[starfury1]

Re: My new Seasonic has questionable caps.

well bonez Knows more them me on cap series etc but using 12.5 mm
(if you can get them in there)
would probably add a bit to the life anyway

The only thing that seems to be the consensus of opinion round here is its preferred to use Non aqueous type caps for psu.
So panasonic FC is preferred to FM for psu use

ESR is not overly critical (like VRM) and yeah technically it makes sense that the reason that 10 volt caps are used is cause they have lower ESR in el-cheapo caps then the 6V3 part.

Anyone know of any other reason why 10 Volt might be used...?
(maybe sloppy supply regulation at some point...???)

OK could have this a bit wrong but my understanding is that the lower you go ESR wise the shorter (Generally speaking) will be the life of the cap
(I know the endurance thing life expectancy and when it goes out of spec is is a valid point but I am talking in a more general sense here, so I suppose any of the above apply depending on what you want to ref to and at what point of expected failure)

BTW that doesn't mean they wont work and have a reasonably long life either

Just a side point really
but yeah maybe the best thing you can do for any PSU is make sure it has good ventilation and gets rid of hot air quick.
Cooler it runs the longer life its liable to have over all.

anyway just my thoughts, no expert

Cheers

[PCBONEZ]

Re: My new Seasonic has questionable caps.

using 12.5 mm .... would probably add a bit to the life anyway

It should yes.

The only thing that seems to be the consensus of opinion round here is its preferred to use Non aqueous type caps for psu.
So panasonic FC is preferred to FM for psu use

They way people think of aqueous and non here FM is aqueous and FC isn't.

But I don't buy the whole aqueous / non aqueous thing as being as important as some people think anyway. Water isn't the whole story. Other additives play a big part and manufacturers tend to keep those details secret.
Also all low ESR caps are aqueous. The only difference is the degree of their aqueous-ness(?). You can tell that by their low temp rating. The lower it is the less water they have.

FC is -55
FM is -40 (More water than FC)

Chemicon advertises KZE as designed for PSU's.
Rubycon advertises ZL as for PSU's.
- I think Chemicon and Rubycon know what they are talking about.
Both KZE and ZL are 'aqueous' and have a -40 low temperature rating.
Additives my friends matter as much as water content.

OK could have this a bit wrong but my understanding is that the lower you go ESR wise the shorter (Generally speaking) will be the life of the cap

This (Generally speaking)[hehehe] is because to get ESR way down low they have to use caps with more water in them.
Caps with more water have a lower endurance rating.

BUT!
The endurance rating is NOT a reliable or accurate indicator of expected lifetime as people seem to think it is.
-
The endurance test is not a test to failure. It's an "at least this long" kind of thing and so is not a solid definite number.
-
The endurance test is accelerated, uses approximations as inputs, and completely ignores the effects of time on materials. (For example, if there is an impurity that leaks out of the Aluminum at xxx parts/million/month and that breaks down the oxide layer or electrolyte an accelerated test will miss those effects completely. Same same if an additive to the electrolyte absorbs into the Aluminum (or something else) and weakens the electrolyte by being absent.)
-
You could easily have two caps (series) with the same endurance rating and one's actual lifetime consistently be twice as long as the other.

.

[acstech]

Re: My new Seasonic has questionable caps.

I thought it was possible for ESR to be too low in some instances. Do I not have to worry about that on a PSU, or with electrolytic caps?