Had a very interesting fault in my Cambridge Audio music streamer which I use to play music from a NAS server.
The RCD in the house tripped without warning and plunged the house into darkness.

As usual I was not at home when this happened but with some instructions to my son via mobile phone, traced it the above device. Unplugging it enabled the RCD to be reset and power restored.

When I got the unit on the bench I confirmed the fault by plugging it in and immediately tripping the RCD again.

I opened it up and found a nicely made SMPS board mounted beside the main motherboard.

I removed the SMPS and set it up on the bench on it's own.

I plugged it in via it's IEC socket expecting the RCD to trip again but it didn't!! I then made some voltage measurements and found a + and - 15V rail and what was probably meant to be a 5V rail which measured about 4.5V.

I switched it off and measured across the main filter cap on the primary side of the supply as these can store enough voltage to give you a nasty belt if they are not fully discharged. It measured only a volt or so.

I then powered it up again and measured across the main filter cap expecting to see about 330v DC but only saw about 250V.

Powered it off again and decided to remove this cap and measure it. Well I didn't have to as when I tried to desolder it one leg came off and it was leaking electrolyte from it's base.

It was a 47MFD 450V cap which I fortunately had a suitably rated high ripple spare for. I fitted the new cap and loosely fitted the supply back into the unit and everything worked perfectly with no tripping of the RCD.

What confuses me is why it tripped the RCD when the main filter cap was open circuit?

The mains is rectified and then filtered and then using varying mark to space ratio switching into the primary of the SMPS transformer to regulate the output voltage in the usual way.

Why would the filter cap being OC cause the RCD to trip??

The unit is now back in service and working perfectly but the question still remains.

Any on got any ideas???

Why would the filter cap being OC cause the RCD to trip?

Are you sure it's OC? Maybe it's SC.

Was the RCD tripped because of current overload or active/neutral imbalance?

I am unsure of the exact reason a capacitor in any particular state would cause an RCD to trip but RCDs will trip because of spikes and surges. It's not commonplace but does happen.

Maybe the leaking electrolyte allowed enough current to leak from the mains to an earth ground or secondary (that leads to earth ground) of the SMPS to trip the RCD? Did you clean up the leaking electrolyte off the circuit board?

The cap was definitely open circuit. Otherwise the main fuse would have blown and the cap most likely would have exploded.

There was very little electrolyte leakage on the PCB. Certainly not enough to cause a current path.

I've seen filter caps in SMPS primaries o/c many times before but never had one trip an RCD.

I don't like fault mechanisms I cant explain. Hmmm.

One of the things I have discovered recently and mentioned, was with the combined RCD / circuit breaker. These are clearly heat sensitive.

Clearly the Pump Shed (now modified) exceeded 50 degrees & the breaker down rated. That meant no pump until things cooled down, hours later.

Simple. The electrolyte leaked across the PCB and across the hot - cold barrier.

Seen it happen before, no mystery there. Nice to know your RCD works, probably saved you from more damage.

It's probable that the electro was not fully O/C until you pulled it from the board, else the SMPS would have shut down thinking there was an over-current fault.

The electrolyte leakage was confined to just under the cap and certainly did not bridge the primary to secondary gap.

It cleaned away easily and left no sign of corrosion unlike other instances I've seen where tracks were dissolved etc.

Well you had primary to secondary leakage somehow and that electrolytte is the most likely cause.

You say it didn't trip once the SMPS was removed from the case. What about electrolyte on the inside of the case?? It wouldn't take much to trip the RCD.

Probably worth running an insulation test on that transformer.

If it is a combined RCD & circuit breaker, it just might have had enough current to trip it. It would be a good idea to have a look at the circuit. I have had instances of non approved oil filled caps found exploded on the mains side.

I have also found metal cased transformer equipment with no ground connected to chassis, no earth wire & a cap from mains to frame. A recipe for disaster. One would hope that there are RCD's & they trip if the thing is touched in fault mode.

If I have suspicions about a transformer, I will not just do the normal Tag & test thing. I will actually isolate its secondary / secondaries as well & check for leakage, not only to ground, but between their windings.

What I did do whilst I had it apart was to use an insulation tester to test between Active and Earth on the IEC connector.

It showed no leakage whatsoever.

This was done before I changed the main filter cap and with the supply disconnected fro the mother board.

As said before the leakage under the crook electro was very small and in a position which could not bridge the primary to secondary gap.

I guess it will remain a bit of a mystery but it's up and running perfectly now and I was very pleased that it wasn't something more complicated.

I do like to listen to music and was sort of going into withdrawal without

As noted, with the summer we are having. Have a look at your fuse box & see if its in direct sun & when.

RCD will not trip for earth leakage on a transformer that is not an auto-transfomer. It will not trip at 30mA if the voltage is under 130V, if it is an auto transformer and if the transformer secondary winding is isolated from the primary, RCD will also not trip for earth leakage.

Who was the "Einstein" who mandated whole house RCDs?