Hi all,

a W&G switching power supply in my SNA1 spectrum analyzer had a large 15mF filter cap with a total short. The screws showed some brown-black goo, otherwise no sign of failure like overheating. Another capacitor of the same type is perfectly in spec. Once the capacitor replaced the supply and the rest seems to work fine.

Up to now I thought these caps gradually loose capacity and go open. Here is the first time I've seen one shorted. The supply worked for 20 minutes or so before it failed. Before that it was unused for years, and when switched on it was stressed by a shorted through capacitor in the -12V supply. The 15mF capacitor belongs to the 5V supply, probably unrelated.

Any opinion on this?

cheers
Martin

开云体育

Yep, I too have found low voltage "large value" electrolytic cap's self-depolarise if left unused for some years (10+!)

If you suddenly hit them with the full operating voltage (and ripple current) the resulting heat generated internally can result in the infamous "smelly snow storm" as the insides are ejected via the vent, or explosively by firing the entire end cap off the can!

Or, the internal distort and go short cct.? Also sometimes creating a lot of heat and pressure, but usually something else up-stream opens up, hopefully a fuse of circuit breaker first.

I have though had fair success in recovering them, but you need to check them "BEFORE" applying main power.

The last I recovered (some years ago, based on previous experience with HV electrolytics) were a bank of three 31000uF/40V caps, from a very old computer PSU.

When checked with a meter, they were virtually short cct.? (All three, individually!)

As I wished to use them as a bank of three again, I set up a lab PSU (with diode & fuse in series on the output) for some 41V, and less than 25mA current limit.

It took several months (4+!) before the current limit light went out, but they came back to life (as indicated by slowly rising terminal voltage) and are happy again with low leakage running at some 28V average (37V peak) running stone cold.? Also their "capacity" is not far above 100,000uF total again.?? Didn't measure ESR, but the way things were working, and cool with it, I don't think there is an issue there.? (The application was an 18V raw DC PSU, upstream of several DC/DC buck converters.)

The disadvantage with this, is the time it takes.? The slower the better.? (If the terminal voltage does not start to rise after some days, or stays at a low level without moving (or even fall's back!) Then it's probably not going to recover.)

Small wire ended electrolytic cap's also suffer, but the useful recovery rate is abysmal.? There again, the replacements are much more economic than trying to source those "bin sized" monsters.

This is based on my own experience over the years.

Remember too, that a large value cap with even a moderate voltage, stores a considerable amount of energy, that at best can cause damage to other things,? At worst, well...? Take care.

Dave G8KBV



--
Created on and sent from a Unix like PC running and using free and open source software:

Hi,

thanks for all these stories.

Although the PSU was running for 20 minutes or so, I did not observe any signs of heat or smell. The only thing was the dark goo around the screws.

Could it have vented during this period? There is a little white deposit on the blank soldered parts around, but not on the wires or elsewhere. Power consumption of the PSU wasn't higher than expected, and the 5V rail must have been OK otherwise the analyser would have stalled.

When the cap shorted the pre-regulator shut down and periodically tried to start up. No fuse was blown. Not a bad design from that point of view.

I'll try to reform the cap with Dave's receipt, although it shouldn't work...

cheers
Martin

A couple of things:

The dark goo around the screws is usually simply tar. Older
caps, particularly the FP style used tar as a sealant, and
to keep the guts from banging around in the can.

Also, the electrolyte is very, very, very low resistance stuff.
If it leaks and bridges between the electrodes, it will present
a dead short at voltage.

The more usual case where this will be seen is when a radially
leaded capacitor is mounted tight to a PCB. If a leak happens,
the electrolyte will pool between the leads, and on traces (if
there are any on the component side), and burning can commence.

-Chuck Harris


On Tue, 21 Mar 2023 13:48:37 +0100 "Martin via groups.io"
<musaeum@...> wrote:

Hi,

thanks for all these stories.

Although the PSU was running for 20 minutes or so, I did not observe
any signs of heat or smell. The only thing was the dark goo around
the screws.

Could it have vented during this period? There is a little white
deposit on the blank soldered parts around, but not on the wires or
elsewhere. Power consumption of the PSU wasn't higher than expected,
and the 5V rail must have been OK otherwise the analyser would have
stalled.

When the cap shorted the pre-regulator shut down and periodically
tried to start up. No fuse was blown. Not a bad design from that
point of view.

I'll try to reform the cap with Dave's receipt, although it shouldn't
work...

cheers
Martin

Greg