Because Sandia and Los Alamos national labs are nearby, Albuquerque had some really good surplus and electronic stores, metal suppliers and machinist related stores. Back in the 1980s, Sandia Labs had some really good auctions where you could buy pallets of electronic components, etc. That ended in the early 1990s, but you can still find some interesting and useful gear at the remaining surplus stores.
For coupling caps this is an issue, but for filter caps or energy discharge, this is usually not a huge issue.
Lots of the caps in that picture look like motor caps. 95% chance they're fine. Quite honestly, I don't see much in that picture that looks particularly ancient.
If you have a use for them, they're probably fine, but usually with audio electrolytics are the way to go for filter caps.
PIO caps can fail in a few different ways. Firstly, the dielectric can break down over time. Usually this shows up in the form of arcing internally. They can also leak, and once the oil leaks out of them, they can't withstand high voltages anymore.
Energy discharge caps (a specific type of PIO caps) can be a little more complicated. When they are charged up and left that way, the oil in them for some reason expands. They will puff up and leak. Usually they still work (unless it makes them blow apart).
The failure due to over-current is more interesting and dangerous. Due to the low ESR and ESL, these caps can charge and discharge extremely high currents, on the order of kiloamps. Some energy discharge caps can do over 100kA. When the maximum current rating is exceeded, the magnetic forces can tear the capacitor apart internally. This can create an EXTREMELY dangerous situation where the capacitor is torn internally and disconnected from the terminals, but still charged to several kV. This is just about the most dangerous failure for a big capacitor like this.
Also, it is very important to store capacitors like this with a shorting bar of some sort. It could be as simple as some copper wire, but without it they are notorious for recovering charge, sometimes several kV.
Most of what I discussed here is in regards to big energy discharge caps, but it's all stuff to be aware of.
Also yes, laboratory surplus is the best. A lot of times smaller, lower-budget labs will buy up pallets of this stuff to use in their own experiments at "free for the price of shipping" and eventually it trickles down to the rest of us.
Surplus auctions tend to have much better prices than eBay.
Surplus auctions tend to have much better prices than eBay.
The PCB containing oil caps are the best sounding ones. Of course if you drink the electrolyte you may get sick so I just advise against it. Also don't plug the speaker's wire into the walls mains sockets like those Euro morons notoriously do, which forced manufacturers to plug the amps receptacles.
I'm already doing that for someone here who PMed me. I'm going there Saturday to get the sizes and such. I'll let you know.
Much lower than electrolytic caps. It's a different technology. They also don't dry out. The purpose of the oil is to increase the dielectric strength, which is why some PIO caps have insane voltage ratings, far in excess of what you'd see in audio.
The low ESR is one of the things that makes PIO caps dangerous, as they can discharge enormous currents.
Some of the PIO energy discharge caps can do 100kA. Of course, you need to have extremely low ESR and ESL for that to happen. Fairly unique thing that tends to be limited to the lab.
Really, aside from maybe using motor run capacitors as filter caps, I don't see what people want to use these for. They can be very nice for transmitters and linear amplifiers due to their very high voltage ratings, and they're useful in the laboratory as well, but audio?
Even an 845 amplifier doesn't need that much voltage on the plate. I don't see where you're using them for coupling caps either, since coupling between tube stages is almost never more than 1 uF. I suppose you may have a few big caps coupling cathode follower or a parafeed configuration, but even then I struggle to see the point due to the cost and physical size of these things.
Big poly caps are reasonably inexpensive, far more practical and will sound exactly the same in these situations.
Also, for things that have to dissipate a lot of heat, that kind of heatsink can be a challenge to find. If you can buy a bunch of them, I think you'll have no problem selling as many as you can get your hands on to people on this forum. I certainly haven't had that kind of luck with heatsinks.
What I did for my 0-550V @500mA linear power supply (using MOSFETS) was to mount them to a 1/2" aluminum plate using a clamping bar. I then mounted the heatsink from a PowerMac G5 on the opposite side. The flatter the aluminum the better, of course. I happened to have a piece that was blanchard ground on both sides. I wound up lapping the surface of the PowerMac G5 heatsink, but it was pretty close to start with.
This sort of technique works great if you have to dissipate a lot of power. In my case, I knew I was pushing the FETs pretty hard, so I worked very hard to get the thermal interface as low as I possibly could. The G5 heatsink has an absolutely massive amount of surface area, so a very slow (and quiet!) fan is more than adequate even under maximum load. This power supply can dissipate up to about 175W at full load with an output voltage of 270 volts. Before I added tap switching on the transformer, it would dissipate twice that. If I were going to run that config permanently, I would liquid cool it.
Part of me wants to build a big Class-A liquid-cooled amplifier with an external cooling unit. Maybe if we ever send one of the big magnet power supplies at work to surplus I'll buy it and use the parts to build a monster 500W+ class A amp. I can dream...
What I did for my 0-550V @500mA linear power supply (using MOSFETS) was to mount them to a 1/2" aluminum plate using a clamping bar. I then mounted the heatsink from a PowerMac G5 on the opposite side. The flatter the aluminum the better, of course. I happened to have a piece that was blanchard ground on both sides. I wound up lapping the surface of the PowerMac G5 heatsink, but it was pretty close to start with.
This sort of technique works great if you have to dissipate a lot of power. In my case, I knew I was pushing the FETs pretty hard, so I worked very hard to get the thermal interface as low as I possibly could. The G5 heatsink has an absolutely massive amount of surface area, so a very slow (and quiet!) fan is more than adequate even under maximum load. This power supply can dissipate up to about 175W at full load with an output voltage of 270 volts. Before I added tap switching on the transformer, it would dissipate twice that. If I were going to run that config permanently, I would liquid cool it.
Part of me wants to build a big Class-A liquid-cooled amplifier with an external cooling unit. Maybe if we ever send one of the big magnet power supplies at work to surplus I'll buy it and use the parts to build a monster 500W+ class A amp. I can dream...
Last edited:
Take visual inspection of eventual leaking oil. And corrosion around the edges and pins. Also do not by any distorted shape or even slight damaged global shape. Check also the pin contacts, should be very strong and without any mechanical drift.
.
These visually and mech. good should be tested off-course, value and other parameters (Q, ESR...) before put on the high tension. Do not rapidly discharge Cs, but rather slow use higher value power resistor... I will put them first in lower voltage value without/with some load to check temperature and behaviour under tension? Latter IF everything is OK put them into high voltage?
cheers
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.