re-capping a motherboard

[lilshawn]

anybody know why a motherboard i re-capped 2 months ago would be bulging caps again? it had a couple up next the the CPU that where bulged so i replaced them and now they are popped again. 

power supply looks good.

[BobA]

Might be a bad MOSFET overdriving the caps

[daywane]

Might be a bad MOSFET overdriving the caps

more input please.
I had a motherboard with the same problem. I also had a damaged motherboard and noticed same caps.
I de-soldered them and replaced the bad with used good ones.
not been a problem for almost a year.

what is a MOSFET.? (never mind googled it)
http://en.wikipedia.org/wiki/MOSFET
so thats what those things are called.

[BobA]

The MOSFETS are driven by a pulse width modulator and this controls the required voltages on the MB.   If a MOSFET shorts it puts too much voltage on the cap which can cause failure.   If the pwm malfunctions the cap may be over driven and be on longer than it's proper phase and this can also cause failue.

[fatfingers]

Wait, wait...

So a MOSFET is really just a transistor?

I wish I understood electronics.

[RandyT]

It could also be that the caps aren't of the quality necessary for use on a motherboard.  You didn't really say whether you got the good, Japanese capacitors meant for the job.  If you didn't, make sure you do next time .  I did, and my re-capped motherboard has been going strong for over year (even though the voltage seems a tiny bit higher than it should be, but it's still in spec.)

RandyT

[ubiquityman]

anybody know why a motherboard i re-capped 2 months ago would be bulging caps again? it had a couple up next the the CPU that where bulged so i replaced them and now they are popped again. 

power supply looks good.

I've re-capped several motherboards and power supplies and have not had a problem thus far.

Tell us what the original caps were and then tell us what you replaced them with.
If you don't know what the original caps were, tell us the motherboard model number.
For starters, I'd like to ensure you purchased the correct replacement caps before heading down further troubleshooting.

[lilshawn]

it's actually a motherboard out of a Big Buck Hunter Pro computer, it wasn't posting... brought it back to the shop and replaced all the caps (even ones without pregnant tops on them.) now it's about 2 months later and it's doing it again. (again not posting again pregnant caps)

EDIT: it's not the red "biostar" board it is a replacement green series board. unsure of manufacturer...

I'm unsure of the manufacturer of the caps, i picked them up at our local electronics supplier...they are an oddball size (2200uf 16 volt) where on most boards, the CPU caps are 1000 6.3 volt. Iv'e done several others with no (ha) problems. nothing like this anyways. Due to the area in which the caps a crammed in, i did not up-cap them i left them the same rating and voltage.

i'm leaning towards receiving a regular ESR cap as opposed to the low ESR type. They where the narrow skinny type that are normally found crammed on motherboards, so i'm not sure. I would think not, but you never know.......

is there a marking or identification on the cap that would let me know if it was low ESR or not?

@fatfingers, yes and no, the difference being that a transistor  when being switched on has a range where in between it's "kind of" on but not quite on yet...current begins to flow through at a reduced rate, a MOSFET is super narrow...where the difference of a few tenths of a volt can potentially turn it from off to on....wikipedia has a good animated GIF on the MOSFET page that demonstrates this quite well.
 

you will notice that the "connection" doesn't actually form inside this FET until about .45 volts... and is established by .5 volts

[northerngames]

are you overclocked any?

I recently took my 4000+ to 2.7gig and I was reading that if you crank your cpu, ram, voltage up to much it is possible it will start popping caps also and why alot of overclockers go with the jap cap's as mentioned prior above already becuase they can handle it where most other's cannot.

check in your bios and see what your feeding the ram and cpu voltage wise.

also if you have a power hungry video card and cpu but a crappy psu that can cause alot of stress, freezing, slow down, heat issue's etc. also.

[lilshawn]

No, no overclocking. Big buck hunter isn't huge on system resources. The only reason I think the computer is as fast as it is, is because they don't make anything slower.

well tomorrow ill pull the caps and see what the deal is... I'm on a 4 hour drive to go fix some stuff out of town.

[northerngames]

I would still check your bios voltage settings just to be sure.

[lilshawn]

10-4... Ill let everyone know what the deal is...

Anybody know yet if low/normal ESR caps are marked as such?

[MonMotha]

Low ESR caps are sometimes, but somewhat infrequently, marked exactly as such.

In addition to low ESR, high ripple current ratings are especially important in SMPS applications, especially at the input (not so much output) of buck (step down) supplies.  Low ESR does tend to be associated with high ripple tolerance, however.  Also, get the highest temp rating you can and up the rated working voltage if possible.  They'll last a whole lot longer.

ESR is more critical on buck outputs since it is where most of the ripple comes from once there is sufficient bulk capacitance.

Brand doesn't matter as much as people seem to say.  I've had decent luck with the off-brand stuff, but I still go with the quality Japanese brands (Nichicon, Panasonic ECG, Nippon Chemi-Con) when possible.  Sometimes they just don't have what you need, though.  Sometimes, the darned cheap caps can be smaller (since they won't actually stand up to the ratings they claim for very long) than the quality stuff which makes replacement hard.

If these are on the input of a buck SMPS, you might try some Sanyo OSCON caps.  These suckers have VERY high ripple current ratings, but they're not cheap.

When spec'ing electrolytics, I tend to do the following:
*Determine the capacitance range I need
*Determine the highest voltage the cap will see even as a transient then double it to get the cap working voltage.  At higher voltages (>30V), I'll use a smaller factor like 1.25-1.5
*Determine the ripple current the cap will see and similarly multiply
*Determine ESR range necessary
*Determine minimum temperature I expect the device to operate at then spec at least one temp range higher
*Fudge things if I absolutely can't find something that will fit

My caps don't normally fail unless I get one of the calcs wrong.

I find that, frequently, caps that fail quickly aren't poorly manufactured; they're just spec'd wrong.  Even the cheap crap tends to live up to the datasheet claims.  The claims just aren't very extraordinary   Designers commonly neglect secondary characteristics like rated ripple current entirely, and they almost never actually pay attention to the rated lifetimes at the given spec (often only 4000 hours or less!).  Neglecting to consider transient conditions in voltage or ripple current also seems common.  When performing replacements, overspec'ing will compensate for this and often get you something much more reliable.

[lilshawn]

Thanks for the tips.

Yeh, the ususal 6.3 volt caps I normally step up to a 10 volt cap, but for some reason tis board uses caps that are 16 volt and due to the crowded nature of the board I had to stick with 16 I couldn't find a 20 or 25 that would come close to fitting.

Ill be pulling the caps today and checking things out a little more here.

[lilshawn]

Seems the manufacurer of the failed caps is the same as ones I've had problems in the past with "OST"

I never even thought to check. Ill be replacing them with som Nichicon if I can find them

[Derrick Renaud]

I always use Panasonic FM series caps from DigiKey.  FC caps are decent too, but FM are better.

[ubiquityman]

I always use Panasonic FM series caps from DigiKey.  FC caps are decent too, but FM are better.

Ditto here. Panasonic FMs.  That's what I used to recap and haven't had a problem since.