Generic question - power supply failure

[darkstar49]

I know there are numerous threads on the subject, as well as zillions of tutorials, etc...

But I'm busy with a power supply (Cherokee/Lineage Power) SP651 that works well for a few minutes, then randomly shuts down...
I'm definitely NOT a PSU expert, but I did quite some research on the Net and found nothing helpful... I did not find any hints on what could be wrong with a working PSU that shuts off randomly, not after a few seconds, or during startup, but after several minutes... and it doesn't get worse... there's just something that makes it think after a while that it's time to sign off...

I'd be very thankful for any ideas/hints as to what component(s) could produce this kind of failure and/or how to troubleshoot such failure ?

[wn1fju]

Could it be a thermal problem?  Are any parts getting real hot?  After it shuts off, can you immediately restart it or do you have to wait a while (until it cools off)?  Also, some better power supplies have thermal cutoff switches.

[darkstar49]

Thermal issue is indeed my best guess as well... and restarting it immediately after a shutdown indeed leads to a pretty quick shutdown.

But the question is, given that the fan is working OK, what components could be involved in such a failure ? 
Are thermistors ageing ?? Is it a good idea to replace those anyway ?? (assuming the labels are readable and I can find a suitable replacement, because as usual, there are no schematics nor parts lists available)

[Grandchuck]

Some components partly fail.  That could be what is going on.  A switching transistor might work well enough to produce output but run extra hot due to leakage and thus eventually cause shut down.  Does your nose detect any odor that might indicate over heating?  (don't use a nose meter up close, especially under power

[Nusa]

After anything found with visual inspection, the usual thing when dealing with strange symptoms on old stuff is to replace the electrolytic caps. Or at least test them, but that takes more time than replacing them.

Pricing a replacement unit is also a useful exercise, as that tells you whether the repair is economical in terms of time and money spent fixing it.

[tunk]

Bad or cold solder joints?
(And if it's the same unit as in your other thread, clean it first.)

[darkstar49]

Pricing a replacement unit is also a useful exercise, as that tells you whether the repair is economical in terms of time and money spent fixing it.

Oh yes !!!! I bought a few spares from Lecroy a while ago, but those PSU's are no longer available, and Lecroy is not selling any spares anymore (tried 2-3 months ago), I'm not even sure they have some... those suit-case models (Xs, Xi and derivates) are now out-of-support, or under 'best effort' support (I know for sure that those motherboards with the PCI/104 slot are now unobtainium, Lecroy has 0 stock, and they're out of production since many, many years).

So, even at 4-500 USD a piece, I wouldn't bother... but unfortunately, there's no other way... I also tried to get in touch with Cherokee / Lineage Power (now GE Power), with no luck... (this was a custom design for Lecroy anyway, so I'm pretty sure they wouldn't nor sell anything, nor transmit any technical details/schematics/...).

Regarding the caps, yes, that's a standard action on old PSU's... but I was wondering wether the symptoms here are 'compatible' with 'semi-dead' caps... (they all look 100% fine, btw).
My assumption was until now that those symptoms were not compatible with problems with the electrolytics, but maybe I'm wrong ??

[darkstar49]

Bad or cold solder joints?
(And if it's the same unit as in your other thread, clean it first.)

yep, same unit, and of course cleaned up !!  ;-)
Still same symptoms... I could have used the other thread, but thought that the question was in fact more 'generic'... as stated, I found lots of tutorials for PSU repair, but none that mentions a PSU that fails after several minutes... so I created a new thread, as I also hoped to have feedback from people that wouldn't be interested by a Lecroy specific topic.

As for the solder joints, nothing obvious, to me at least...

[darkstar49]

Some components partly fail.  That could be what is going on.  A switching transistor might work well enough to produce output but run extra hot due to leakage and thus eventually cause shut down.  Does your nose detect any odor that might indicate over heating?  (don't use a nose meter up close, especially under power

No risk with the odor meter, even at close distance, as we all have to wear a mask !!!  ;-)

But no, couldn't spot anything special... (without mask !). I'll get a thermal cam this week-end, could be interesting as well...??

[Squoip]

I've been successful with a can of freeze spray when there are hot shutdown problems. I have been able to isolate to a single component by spraying smaller and smaller areas until a short burst on a specific component does the trick.

[darkstar49]

Some components partly fail.  That could be what is going on.  A switching transistor might work well enough to produce output but run extra hot due to leakage and thus eventually cause shut down.  Does your nose detect any odor that might indicate over heating?  (don't use a nose meter up close, especially under power

Stop !!! I've got it now with the PSU outside the casing, and geee... yes, it smells !!!!!!!!!
But I couldn't say what it smells like... as stated, I'm no PSU expert.

There's no smoke, just the smell, but quite strong...   I'd say it smells like 'electronics stuff', just much stronger...  although I know this is pretty childish and doesn't help much...
(and yes, I'm running it with a big fan blowing at it !!)

   

[Grandchuck]

Thermal camera time.

[darkstar49]

Still don't have the 'real' cam... but some Flir One gave me maybe some hints...

And strangely enough, the smell is almost gone today (I started it yesterday for the first time since quite a while...).
Anyone knows what temps are 'normal' for what kind of component ?

On 2 relatively big resistors near line entry (just after what looks like a relay), I had 130°C (the little Flir One might not be the most precise tool, but still quite a lot ), connected to the mains, but idle...
Also had the same kind of temp on a resistor near what I believe to be the rectifier bridge ?? For the remaining part, nothing went upside from 65°C... (although the scope was not pumping a lot, as it has a boot disk failure, but it had it yesterday, too.

Any comments welcome !

[darkstar49]

Damn !! this repair is not going the way it should... was probably way too tired yesterday night, and was then wondering why the scope had a boot disk failure (on top of smelling like hell... I guess the PSU had nothing to do with the burning smell  )

Now I've got it...    had mis-inserted the IDE cable into the 2.5" disk 

I'll be very lucky if the mobo is not fried... these are 100% unobtainium !!!!!!

[tautech]

My gut feeling is dead or dying mainboard caps.
The PSU current sense circuitry is shutting the PSU down.
If you can get it to run for a couple of minutes and there is no HV present run your fingers over a few caps and you should find some getting hot.

[darkstar49]

My gut feeling is dead or dying mainboard caps.
The PSU current sense circuitry is shutting the PSU down.
If you can get it to run for a couple of minutes and there is no HV present run your fingers over a few caps and you should find some getting hot.

I've got another PSU (not the same revision, but same type), and everything works fine with that one, thus I assume that the problem is with the PSU.
I couldn't get the thermal cam yet, so I've put this aside for the moment. Many components are glued (incl. caps) and not so easy to access, so I'm a bit reluctant to start swapping caps just to be sure... I'll first try to spot any abnormally hot components under load, and then refine the diagnostic.

[Nusa]

When you get it, take some images of the working supply in operation, so you have a baseline on what's normal.

[tautech]

My gut feeling is dead or dying mainboard caps.
The PSU current sense circuitry is shutting the PSU down.
If you can get it to run for a couple of minutes and there is no HV present run your fingers over a few caps and you should find some getting hot.

I've got another PSU (not the same revision, but same type), and everything works fine with that one, thus I assume that the problem is with the PSU.
I couldn't get the thermal cam yet, so I've put this aside for the moment. Many components are glued (incl. caps) and not so easy to access, so I'm a bit reluctant to start swapping caps just to be sure... I'll first try to spot any abnormally hot components under load, and then refine the diagnostic.

Oh well, if nothing else that narrows the field some. 

Very likely it's caps then or a faulty joint that's breaking down with heat.
For SMPS I generally focus around the controller IC particularly its VCC storage cap. They seem to drift way outta spec with age and go high ESR and/or low capacitance. Pull and measure is all you can do with these as they never show signs of degradation.

[Miti]

My bet is on the caps. Their ESR went up and they heat up from the ripples, further increasing the ESR. Touch the top of the low voltage caps with the finger to check their temperature.

[Rasz]

>2 relatively big resistors

they are big for a reason, but I would replace capacitors directly near them, living next to source of heat tends to dry caps pretty fast

> works well for a few minutes, then randomly shuts down

what shuts down? have you monitored its rails when it does that (ideally using another scope)? does PSU stop producing prgood signal before going down (im guessing its ATX like )?

[darkstar49]

>2 relatively big resistors

they are big for a reason, but I would replace capacitors directly near them, living next to source of heat tends to dry caps pretty fast

> works well for a few minutes, then randomly shuts down

what shuts down? have you monitored its rails when it does that (ideally using another scope)? does PSU stop producing prgood signal before going down (im guessing its ATX like )?

any idea what temp to expect on those resistors, as stated, the FLIR One might not be the most precise, but 140+ °C seems pretty much... is this rather indicating a problem somewhere else, or could the resistors themselves also age and dissipate more heat than they're supposed to ??

I'm still waiting to redo the exercise with a 'real' thermal cam, I'm willing to change some caps, but the PSU is relatively complex (see pic above), and components are not easily accessible, so unless the only answer is 'change them all', I'd prefer to be more specific...
And again, I'm a PSU rookie (with no intention to become an expert !!).

As for the shutdown, no, I didn't monitor the power-good signal. But don't forget there are 4 ATX-like connectors, and the driver is actually the scope's acquisition board, not the PC board (the PC mobo's power-on switch is connected to the acq board), and I've no clue how the startup sequence looks like, nor if and how the PSU and the acq board interact with each other... apart from the board with the switching IC, there are 5 other (vertical) mini-boards with different ICs in the PSU...

[darkstar49]

My bet is on the caps. Their ESR went up and they heat up from the ripples, further increasing the ESR. Touch the top of the low voltage caps with the finger to check their temperature.

Yep, will do that, but once I get my hands on a 'real' thermal cam, not with the fingers... 

[tautech]

Rasz's comment on the PWRgood rail is something that need not be overlooked as that's another little independant SMPS AFAIK. Trace the PWRgood rail backwards and you'll find where it originates.

[darkstar49]

Very likely it's caps then or a faulty joint that's breaking down with heat.
For SMPS I generally focus around the controller IC particularly its VCC storage cap. They seem to drift way outta spec with age and go high ESR and/or low capacitance. Pull and measure is all you can do with these as they never show signs of degradation.

I'm really not an expert, but there are many IC's in this PSU...  as stated, it's not an ATX PSU, it's a custom design, that includes an ATX-compliant output, but there are 4 ATX-like connectors (1 ATX 20-pin, 1 22-pin, 1 18-pin, 1 16-pin).
So you're saying that regardless of their aspect, caps can be seriously out-of-specs ?? Not good... is there really no way to locate bad caps in a non-intrusive way ?? This PSU is far from simple and physically quite dense, desoldering all the electrolytics is gonna require a serious amount of time ! Is it possible from the picture above to identify 'more critical/suspect' caps to begin with ?

[darkstar49]

Rasz's comment on the PWRgood rail is something that need not be overlooked as that's another little independant SMPS AFAIK. Trace the PWRgood rail backwards and you'll find where it originates.

it's not very clear to me what you want me to learn from the power-good line ?? Are you trying to detect whether the PSU shutdowns on its own, or on request of the scope ?