Big caps failing, but small ones seem ok

[dreamlayers]

I would like to fix a 10 year old Soltek SL-65KVB motherboard because it has a P3-1000 and an ISA slot, making it a nice way to use some old hardware. It has a lot of GSC electrolytic capacitors.

Over a third of the 6.3v 1000uF and 1500uF capacitors with gold/brown lettering are at least bulging. I obviously have to replace all of those.

There are 27 or so smaller GSC electrolytic caps with silver/white lettering. Two are 470uF and the rest are 100uF or less. None of those show any obvious physical signs of failure. Do I have to replace all of those too?

It's interesting how the 1500uF caps have a different GSC logo and worse paint coverage in the markings. My first thought was that they're counterfeit, but the failing 1000uF caps have the same logo as the small caps and well painted markings. I'm attaching photos of two of the large caps and one small cap.

[Agent24]

Re: Big caps failing, but small ones seem ok

GSC are such rubbish anyone making counterfeit ones would be insane.

If it was me I'd probably replace the lot for peace of mind since they are a useless brand, although the 100uF etc might be OK to leave.


Make sure of course you replace them with something not crap.

[dreamlayers]

Re: Big caps failing, but small ones seem ok

I ended up building a simple ESR tester, using a 555 to drive step-down ferrite core transformer, and observing the output waveform using my oscilloscope. I had some difficulty with test lead inductance, but after some changes I got it to work well enough that I could test the small caps in circuit. It seems they're all good. Almost all of the large caps with gold/brown lettering are obviously bad, even including most of the ones that aren't bulging. Here's my blog post with a bit more info.

[PCBONEZ]

Re: Big caps failing, but small ones seem ok

There are 27 or so smaller GSC electrolytic caps with silver/white lettering. Two are 470uF and the rest are 100uF or less. None of those show any obvious physical signs of failure. Do I have to replace all of those too?

Standard practice is to replace anything 6mm diameter and larger - and everything over some uF - and any 85C caps.
"Some uF" depends on who you talk to.
For me is '220uF-and-up' on an extra important or stressed board, otherwise anything 'over-220uF'.
Other people say '470uF-and-up' or '330uF-and-up'. No one 'informed' leaves 470uF anymore.
Around 5-6 years ago most people [including me] were usually saying 8mm/820uF and up - but we learned better since.
Anyway...
On that one the 470uF should go. The 100uF and less would be optional.
Given they are GCS and the age of the board I'd at least consider replacing all of them if this board is used for something important like in your business or to run a machine or something like that.
.

It's interesting how the 1500uF caps have a different GSC logo and worse paint coverage in the markings. My first thought was that they're counterfeit

They are just a different series [model line].
I can see the green ones are RE series.
Unfortunately GSC didn't always mark the series.
.

[Agent24]

Re: Big caps failing, but small ones seem ok

Around 5-6 years ago most people [including me] were usually saying 8mm/820uF and up - but we learned better since.
Anyway...
On that one the 470uF should go. The 100uF and less would be optional.

But will you stick to that story in another couple of years?

I think given they are GSC it'd be best to get rid of the lot. GSC sucks as we all know, why even give them the chance?

Wasn't there a story on here of an on-board SCSI controller that went nuts because of a bad 100uF (or smaller!) ?

[PCBONEZ]

Re: Big caps failing, but small ones seem ok

Don't you know by now that I edit things like 6 times before I'm happy with it.. LOL

I was stating 'standard practice' and trying to be fair to those less picky that I am.
- You missed the last line where I recommended replacing them all if the board is important. If this is a hobby-play-toy or for occasional use to test drives or RAM or something then I really wouldn't worry about the small caps. If it's running a machine I would.

Yeah, I've seen 100uF and less failures but most all were hot spots next to something [like the Dell Optiplex 'hot boxes' that have some next to the HDD with no air flow] - or they were 85C caps.
That includes the 4&5mm GSC. Haven't seen any 105C go bad - yet - on a P-III.
P4 and up is a different story and that's probably heat.
I -have- seen OST 4&5mm 105C go bad on a P-III.

SCSI Controller - Hot spot.

My concern is that at this point they are getting kind of old so dry-out may not be too far away even if they are 105C.
Some 85C start going dry after 4-5 years if the case is on the warm side and it's 'on' most of the time.
Dunno how long it takes 105C to get to that point.
105C with the same Endurance rating -should- take 16-20 years to get to the same point if under the same conditions.
-
I have 15 year old 486 boards with 85C caps and they are still fine. Presumably that's because there's almost no heat compared to newer gear. I also don't know how long they were in actual use before they got 'parked'.
.

[Agent24]

Re: Big caps failing, but small ones seem ok

Yeah the older boards all still seem to work just fine.

Then again my 486s etc seem to mostly use just tantalums. And as you said, less heat etc.

I see you've just got 8888 posts! Very auspicious time to buy some Chinese Sacon FZ?