Mr Diodey went short circuit and the capacitor let the smoke out

[lscw]

It went right for me eye too! Lucky I had my specs on to get a good look

The half wave rectification diode barely visible behind it went short and that happened. That was actually the replacement capacitor, the first one just puffed up and smelt funny. There was quite a bit of spikey noise on the scope on the original cap, it wasn't any better with the replacement.

The diode is a SR260. 2A, 60V schottky. Is there any reason why it may have failed? or just bad luck

This secondary output of the transformer just powers the cooling fan.


Search terms: Hyelec DC Power Supply HY3005ET, 30V 5A

[Gyro]

In the absence of any other clues, you probably fitted the new cap reverse polarity (impossible to prove one way or the other now the can has blown off), but it takes a major insult for them to blow like that rather than vent.

From what you say, it sounds as if the old cap was just failing in the usual way, drying out -> causing high esr -> causing high internal pressure. You say that the waveform was spikey with both caps. That implies that the rail was still outputting DC at that point (you would have noticed AC), so the diode probably failed after the measurement due to the reversed capacitor being a virtual short.

Sorry, that's the best diagnosis I can make from the available information. It sounds as if your eyesight had a lucky escape.

[lscw]

That's an interesting hypothesis. I figured the AC coming out the secondary was the cause but, as you say, spikey DC doesn't correlate. Would it look DC-ish with a reversed cap?

I replaced the diode for a salvaged SR260 and ended up putting the original cap back in as nothing else I had was a suitable fit and all is well, for now.

[Gyro]

Would it look DC-ish with a reversed cap?

Hard to say, it might possibly hold up as a very leaky capacitor for a short time before becoming a more or less total short. You mentioned that it is the fan supply, so a small auxiliary winding compared to the main secondary, that probably means that a short on it wouldn't be picked up by the main overcurrent protection. Yes, the original cap will have gone high ESR if anything so as long as it isn't actively spilling its guts and corroding things it should hold for a short while. Something higher voltage rated or +100%/- 50% capacitance ought to be fine as a medium term replacement if you have one though.

You might want to consider putting a small wire-ended fuse or PTC thermistor in series with the diode. If the overcurrent protection doesn't notice a short on that output then a prolonged short would probably burn out the winding and cause shorted turns or melted insulation tape. That would effectively write off the (economically irreplaceable) transformer.

[lscw]

Good thought. I failed to mention that the power supply would cut out after a couple of seconds over ~1A when the fault was present.
I'll bare that in mind though when I come to replace the cap (inevitably not until it fails again).

I guessed it must be now high ESR, the voltage waveform across the capacitor has quite a large envelope while there is enough voltage to spin the fan.

I did try a few cheap capacitors I had in stock in the 200-300uF 16V range from an assorted aliexpress kit. I thought they were insufficient but I may have been chasing my tail with a dry solder joint. The original cap is 47uF 50V so good excuse as any to expand my stock collection.

Thanks for your help

[Gyro]

You're welcome.

Yes, there will considerable ripple across an already high-ESR failing capacitor. As it's on an SMPS transformer, use it as an opportunity to add some useful low-ESR caps to your stock (the cheap ones probably weren't Low ESR). 


Edit: You could probably up the capacitor value to 100uF. It sounds as if the original was a bit too stressed anyway.