PIO Capacitor Calamity

[Thermionic Idler]

Posting to Beginners as it was my first implementation using PIO capacitors.

Today I had one of the 300B amps on the bench to install the soft-start PCB. When I took the bottom off, I was greeted by this sight:



It looks like one of the Jensen paper-in-oil capacitors has leaked oil everywhere. There was no deterioration in the audio to indicate this had happened, so I don't know how long it's been like this. However, I'm obviously not putting it back in the system until I've repaired it.

The amps have been in regular service (used for a good 5 or 6 hours each night) since January 2021. So this cap has lasted about 18 months.

So a question - assuming I can still get a replacement part (I think Jensen have stopped production of these and I've not checked HFC yet for stock) should I just replace it with same, or is this something that comes with the territory with PIO caps - so can I expect one of the others to let go soon? I originally fitted PIO coupling caps as Lynn Olson had specified them for the circuit I based this on, but I'm wondering if I should consider alternatives such as ClarityCap, for better reliability?

[IslandPink]

That is bad news, and it isn't that long.
What's the value on that one - it looks like quite a whopper.
I've been lucky with a pair of 0.47u Jensen coppers, had them on the breadboard for years, but I expect that they don't see much heat, on there.

I don't think this will be any surprise to a few people here, but I would rate the next-best caps ( in a power amp ) to be the Mundorf 'ZN' Tin /polyprop ones.

[Thermionic Idler]

That is bad news, and it isn't that long.
What's the value on that one - it looks like quite a whopper.
I've been lucky with a pair of 0.47u Jensen coppers, had them on the breadboard for years, but I expect that they don't see much heat, on there.

I don't think this will be any surprise to a few people here, but I would rate the next-best caps ( in a power amp ) to be the Mundorf 'ZN' Tin /polyprop ones.

It's 0.47uF - the circuit originally called for 0.1uF, the Spice model suggested I'd get better bass and phase response if I raised it a bit. That's the coupling cap between the 6V6 drivers and 300B's - there's also 0.33uF PIO between 12SN7 -> 6V6. It looks like HFC have spares still in stock.

Edit - change of plan. For now I think I'll order one replacement from HFC just to get the thing back up and running, as I don't have the budget right now for a full re-cap. Later on in the year I'll likely re-cap that stage with some of the nice Mundorfs. Again, for me reliability is key, I don't want to put the output transformers at risk (for one thing, because of the way it's been constructed they would be a bit of a nightmare to replace). I'm not sure if a coupling cap short would be something the secondary power transformer fusing would protect against. (C2, C9).

PDF of latest schematic

[Greg]

Just in case you’ve overlooked it, if PIO caps are what you specifically prefer, HFC stock plenty of alternative brands to Jensen.

[Thermionic Idler]

Just in case you’ve overlooked it, if PIO caps are what you specifically prefer, HFC stock plenty of alternative brands to Jensen.

Indeed, I'm told Miflex are also supposed to be very good and HFC carry those, too.

I'm actually completely open to trying other, potentially more reliable constructions. I read through the ClarityCap white paper research document for their MR caps (linked to from HFC) and the upshot, as I understood it, was that the most measurable performance parameter between different caps is the degree of mechanical resonance.

This led me to wonder whether PIO caps only sound the way they do because the oil acts to damp those resonances - there would be no difference if you were to encase the thing in a large brick of resin. Given the MR's design brief was to minimise those resonances as far as possible, logically they should come pretty close in terms of performance, whilst more readily coping with valve amp thermal cycling.

I've used MR's before (they are in the Transcendent Son of Beast which I sold to Nick and which was at a recent Owston), I'm beginning to think I should have just specced them again in the first place.

EDIT - all that said, when I built this, I incorporated fuses in both legs of each power transformer secondary in line with this article. I just pulled up the Spice sim of the circuit and simulated a coupling capacitor short. The circuit went from pulling 129mA from the power transformer at idle, to pulling 287mA (the transformer is rated at 230mA).

So that tells me that if a cap did go, the fuses ought to protect against any expensive consequences - which is my primary worry. I'm not too concerned about replacing caps occasionally, it's the potential collateral damage that worries me.

[Paul Barker]

A common tail over the years.

[izzy wizzy]

A small tale of woe. I was going to fuse the secondaries of my power amp as that article you posted coz it made sense. Note the "was". My worry was loss of negative bias and so the output stage going full conduction and who knows what happening. Was going to rebuild soon and so they were on the plan even though I should have had them in there. Even had them in my Farnell basket to do. Also had a program ready to go for an Arduino to monitor bias.

And then a wire on the bias transformer itself fell off but I was out of the room for a time. Came back to a bad smell, some very hot iron; mains tx, choke and OPT but no fire thankfully. I was very lucky that the OPT survived, I think. It sems to be performing OK in the other amp when tried but I'm still prepared for replacement - £500. The Mains tx and choke didn't survive so that was £500 shipped here. Ouch!

[Thermionic Idler]

A small tale of woe. I was going to fuse the secondaries of my power amp as that article you posted coz it made sense. Note the "was". My worry was loss of negative bias and so the output stage going full conduction and who knows what happening. Was going to rebuild soon and so they were on the plan even though I should have had them in there. Even had them in my Farnell basket to do. Also had a program ready to go for an Arduino to monitor bias.

And then a wire on the bias transformer itself fell off but I was out of the room for a time. Came back to a bad smell, some very hot iron; mains tx, choke and OPT but no fire thankfully. I was very lucky that the OPT survived, I think. It sems to be performing OK in the other amp when tried but I'm still prepared for replacement - £500. The Mains tx and choke didn't survive so that was £500 shipped here. Ouch!

Yikes! I feel your pain. Yeah, in my case it's not only the cost, because of the way I've constructed these things with the transformers under wood covers, and soldered-in circuitry sitting over some of the screws holding everything in, they would be an absolute sod to replace. So I figured anything I could do to reduce the risk of transformer burn-out was worth doing. It was fortunate actually because at first I'd cocked up the rectifier implementation (not enough power supply resistance) and I had a few GZ34's arc over before I sorted it out (it's now running Mullard GZ32 / Svetlana 5U4 with additional diode protection). The fuses did their job perfectly.