FYI, cheap and big film caps

[JoshK]

I came across this epay listing when searching for largish high voltage caps for tube preamp building. Thought I'd point it out to others since some of you might like to use these for crossovers, especially ones like the Modula MTMs that use huge caps in the tweeter network.

60uf 400Vdc Film Cap on ebay for <$10

[servicetech]

Looks like it's designed for industrial use. Oil filled air conditioner/motor caps are also a good way to get high voltages for relatively low prices. I don;t know how they sound but they are designed to have 240V A/C for hours on end.

[JoshK]

Yeah, we use motor run caps for tube amps, but not sure if they are always appropriate for crossover duties as they can have dc leakage. I'll defer to someone with more experience. They could sound awful, but for the price it is worth trying.

[servicetech]

I'd 2nd guess that DC leakage, they can hold a charge for MONTHS. Some manufactures put a 100K resistor across the terminals to bleed off voltage when the equipment shuts down.

[Deward Hastings]

Making up 60uf using the 10uf poly caps at Madisound is still less than half that price . . .

http://www.madisound.com/catalog/product_info.php?cPath=356_5_12&products_id=1463

[Amphiprion]

If you just need some bulk HV caps for the power supply, I don't think it would much matter if they were meant for audio or not.

Now if someone can find me some 40kV ceramic door knobs at a good price let me know

[JoshK]

Making up 60uf using the 10uf poly caps at Madisound is still less than half that price . . .

http://www.madisound.com/catalog/pro...oducts_id=1463

True, but they're a 1/4 of the voltage spec too. I thought people typically used caps over 200V spec for crossovers.

[Amphiprion]

They do, but I haven't looked in to understanding why. It may just be that using thicker film is cheaper to manufacture, and we don't really care about size in passive DIY crossovers anyway. A lot of the cheap ones from Digikey are meant for board-level installation (the radial caps usually) and size matters a lot then. Lower voltage = smaller cap (generally).

[Deward Hastings]

Under what circumstance can you envision voltages of over 100V (in all probability a conservative rating anyway) in a crossover (not counting multikilowatt 16 ohm pro speakers, of course)?

Has anyone ever demonstrated an arcover failure with those particular caps (in home audio use)?

A higher voltage spec than you actually need buys you exactly *nothing* in capacitor performance.

[ThomasW]

Under what circumstance can you envision voltages of over 100V..... A higher voltage spec than you actually need buys you exactly *nothing* in capacitor performance.

FWIW, JonMarsh typically recommends using 250V or higher rated caps. If the opportunity presents itself I'll get his explanation for that recommendation.

[Dennis H]

Under what circumstance can you envision voltages of over 100V (in all probability a conservative rating anyway) in a crossover (not counting multikilowatt 16 ohm pro speakers, of course)?

Looking at the schematic, the QSC RMX2450, which is pretty similar to the medium-power pro amps used by many home users, has 110V rails.

http://www.qscaudio.com/support/library/schems/Current/RMX%20Series/RMX2450.pdf

[Deward Hastings]

It seemed so obvious and simple to test . . . so I did.

Madisound "surplus" 10uf polypro cap (measured 10.24uf) in series with a 8 ohm dummy load . . . sort of the basic first order tweeter protection cap.

Plugged straight into the basic 120V 60Hz power amplifier in the sky. Measure 116VAC across the cap. That would be your basic 1.5 kilowatt per channel AV receiver running about max output.

After a couple minutes (and no fireworks) remove the cap and re-test. 10.24uf.

Didn't think twice about throwing it back in the parts box to use in a crossover someday . . .

[Amphiprion]

I killed a 3KV metalized film polypropylene capacitor by hooking it up in parallel with the output of a 7500VRMS neon sign transformer (and running the output through a spark gap full of diesel fuel - this was at my previous job). Anyway, I knew the cap was going to die but I didn't want to wait another day for a digikey order, so I hooked it up to see if it would even last a few seconds. The waveform on the scope showed occasional drops from a few kV down to 0V several times, which I assume was dielectric failure. After 20 seconds it stopped working completely and I pulled it from the circuit. It was noticeably warm. I remember the warmth because I thought it really odd that a film cap would do that. I guess when you put 7.5kV across it, it can happen. Total power of the setup was only about 40W being pulled from the wall by the transformer; I don't know how much was going through the cap.

[JonP]

I killed a 3KV metalized film polypropylene capacitor by hooking it up in parallel with the output of a 7500VRMS neon sign transformer..

It almost sounds like you WEREN'T expecting it to go, when overvoltaging it like that? You put 10.6KV peak voltage across a 3KV rated cap... must have been pretty good that it didn't short instantly. Heating was probably from multiple internal arcs. Smoke, then fire would have been appearing shortly, or maybe internal burnt areas brought it to full current limit.

Luckily, neon sign transformers are purposefully lousy, designed to current limit and voltage drop to 0 for safety and prevention of accidental full arcs in the neon tubes. You REALLY would have seen fireworks if the transformer was good for a kilowatt...

[benchtester]

Madisound "surplus" 10uf polypro cap (measured 10.24uf) in series with a 8 ohm dummy load . . . sort of the basic first order tweeter protection cap.

Plugged straight into the basic 120V 60Hz power amplifier in the sky. Measure 116VAC across the cap. That would be your basic 1.5 kilowatt per channel AV receiver running about max output.

I think you are a bit lucky. At 116 VAC, the peak voltage would be in the 160 volts range (164 volts with a clean sine wave). The caps shown in your Madisound link are rated at 100 volts which is usually a DC rating; so, as I understand it, you were about 60% over spec.

I do like small caps to fit into small speakers. I just bought caps with a 75 volt rating, and for better or worse, I am confident that driver will fry long before the cap is stressed.

[Deward Hastings]

Well, you're correct about the peak voltage (at which the tested sample *did not* fail), and a sample of one hardly makes for good statistics. If I were going to use them at 100V continuous I'd probably test a couple of dozen (out of the hundreds I have on hand, love that 60 cent quantity price), and put a couple on a HV supply and test to actual breakdown. Just to know, you know.

But for crossover applications it's pointless. As you say, the "driver will fry long before the cap is stressed". Like you, I don't use any drivers that would not fail at 100V, or, for that matter, have any amps with 100V rails. I'd go to active crossovers (just as the pros do) before that kind of power level anyway. The voltage capability of the Madisound caps is more than enough for loudspeaker crossovers . . . it's just not an issue worth worrying about. There isn't any measurable "performance" difference (relative to other poly caps), either. They work just as well as "beautique" components, for a *lot* less money.