How Long?

[Structo]

I was looking at some new electrolytic capacitor specs today
and it seems that many high quality electrolytic capacitors
have a life expectancy of 3,000 hours or 125 years or more.

Does that shoot a hole into the 10 year life span theory of electrolytic capacitors in tube amps?

I'm sure that heat and other variables play a part in the life expectancy of
the capacitors.

But still......

[Firestorm]

Obviously, the 10 year lifespan has to be taken with a grain of salt. We've all seen some Tweed Fenders that still performed brilliantly with original caps. And others of the same vintage that would not work at all without an overhaul. Then, there is the contentious issue of sonic impact of cap construction. Modern caps are highly etched to increase the surface area of the foil. Old caps, not so much (why they were bigger). Some "authorities" claim it makes a difference.

[Gaz]

125 years, huh?

[vibratoking]

Looks like a math problem?

[sonicmojo]

I know math. 3,000 hours is only 125 days if you play 24/7!

[Gaz]

I know math. 3,000 hours is only 125 days if you play 24/7!

^^^ That's what I was thinkin'.

[Structo]

Doh!

You're right.

[Malcolm Irving]

... high quality electrolytic capacitors have a life expectancy of 3,000 hours ...

It depends what you do with them. Connect a 25uF 25V electrolytic across the mains and its life expectancy drops to about 3 seconds!

[gingertube]

When you design high reliability stuff in the day job you learn the tricks.

The killer of electrolytics is heat. There is radiated heat from transformers and tubes etc but usually the culprit is self heating inside the electrolytic. This is calculated in the usual current squared times resistance where the current is the ripple current and the resistance is the ESR (equivalent series resistance).

Then there is a calc. based upon the Thermal Resistance to give you the expected internal temperature rise.

Finally there is a lifetime calc based on the inverse of the square of the internal temperature rise.

In summary there are 2 squared terms in the calc , one which determines internal temperature rise
and then a second squared term to give expected lifetime.

When given the opportunity choose your cap for higher ripple current rating and lower ESR. You get quadrupple benefits with respect to lifetime.

THen there is the fact that modern electrolytics are a far cry from what was around 20 or more years ago.

You can improve lifetime of some electros by going knaked. If they are the chassis mount types in a metal can with a printed plastic jacket, peel of that plastic jacket.

Also mount away from heat sources and in an area with some ventilation.

Output Tube Cathode bypasses too. Often a 63V or 100V cap will have much higher ripple current rating and lower ESR than a 25V or 35V part.

Cheers,
Ian

[rp]

We've all seen some Tweed Fenders that still performed brilliantly with original caps.

Been there, very true, not just Fenders, no hum and sweet sound, but when you finally do the recap the amp really perks up and then you notice what was missing. I've had old old e-caps both filter and bypass look and test perfect (at low voltage) but always when replaced the amp wakes up. BF Fender's can seem to benefit from old but still sound caps if you want a soft sweet sound. I always feared doing a recap on gigging friends' amps for just this reason. I really warn them ahead of time.

Hifi it really is 10 years. You can hear it with the extra resolution and (hopefully) quiet background your stereo has. You'll start hearing the extra distortion creep up which you never would in an instrument amp, as said you might prefer it. Even with ss which is much lower V and heat, my Naim gear turned to poop after 10 years. I have a very nice Onix ('80s Brit integrated) as a second unit that I never even noticed was off, did a recap with Nichicon Gold Tunes and it just took off.

If top fuel high performance is a goal I'd say replace every 10 years.

[Phil_S]

I don't know much. Some real gurus here have commented. I'll put in my 2¢. I'm perfectly willing to accept 3,000 hours as the rating. If you keep the heat under control, maybe you get a little more. If you run the amp 20 hours a week, that's about 3 years. I'm thinking that you can get more like 5 years, maybe 10 with less frequent use or if the thing just happens to last. I'm not seeing 125 years. That seems ridiculous. In any case, rated life gives a reasonable service interval.

[Leo_Gnardo]

I don't know much. Some real gurus here have commented. I'll put in my 2¢. I'm perfectly willing to accept 3,000 hours as the rating. If you keep the heat under control, maybe you get a little more. If you run the amp 20 hours a week, that's about 3 years. I'm thinking that you can get more like 5 years, maybe 10 with less frequent use or if the thing just happens to last. I'm not seeing 125 years. That seems ridiculous. In any case, rated life gives a reasonable service interval.

No guru here, can't afford a magic carpet or any of the other stuff. But I'm thinkin' Tom made a simple decimal-place error which brought on the original post. You'll get a lot more than a 'little' extra life by keeping caps away from the heat. See what St. Leo (Fender) did when he put caps in the bathtub/doghouse, veddy smart he was! OTOH when I see a Bogner modded amp with hi voltage caps parked right atop the power transformer, I gotta say "what. was. he. thinking?"

FWIW some 20-25 years back, studio gear gooroo David Manley was suggesting a 7 year replacement schedule for power supply electrolytics. Guaranteed employment plan for studio techs if they follow that advice.

Or you can skip all that, use film filter caps and let your great-grandchildren's great-grandchildren worry about when to change 'em out. Every time I look at a catalog I see more and cheaper hi capacity hi voltage fillums. Carr's been doing it all along, for what, some 20 years now? Whether you like the amps or not, Carr filters ain't gonna fail. Old fashion' engineering "build it so it won't break" as opposed to modern engineering "build it so it lasts one day past the warranty, then too bad Charlie."

[Structo]

Yeah I didn't do the conversion right from hours to days to years.

[Firestorm]

Yeah I didn't do the conversion right from hours to days to years.

Try projecting the moon's orbital increase from the current 3.8cm/year back over 4.5 or 4.6B years, accounting for solar gravitation. There are some real errors to be made on that scale.

I was just revisiting an old hobby as a planetary dilettante because of a Nova show.

[Leo_Gnardo]

Yeah I didn't do the conversion right from hours to days to years.

Try projecting the moon's orbital increase from the current 3.8cm/year back over 4.5 or 4.6B years, accounting for solar gravitation. There are some real errors to be made on that scale.

Oh noes! Earthlings of the future will have funny looking eclipses, diminished tides, no need to sing "when the moon hits your eye like a big pizza pie that's amore' ". An inch and a half a year, that adds up over time don't it? How many furlongs per fortnight...

I expect the moon's rate of retreat isn't constant, has been either accelerating or decelerating over time. Then there's the matter of mass gain - both bodies gain mass over time from whatever "space junk" runs into them, meteoroids from sand-grain size on up constantly crashing into both. Could this eventually balance the gravity field so there's no more runaway moon, or (ulp!) it starts heading thisaway?