Electrolytic capacitor life ratings, KS/KA series

Started by thelonious, December 22, 2014, 10:16:03 AM

Previous topic - Next topic

thelonious

I love the Panasonic KA/KS series electrolytics because they are inexpensive, small, and so far have seemed bulletproof (although I haven't asked haveyouseenhim Mike to field test them). But according to the datasheet it looks like the tradeoff for that small size is shorter life, lower ripple handling, etc. The KA/KS series life rating of 1,000 hours doesn't even make sense to me. If someone leaves their pedalboard plugged in all the time, they'd hit 1,000 hours in 42 days. Now, obviously, the caps aren't going to just up and die at their rated life, but... 42 days??

So, two questions:
1) Can I expect that if I buy the EE or EB series, they will actually last much longer than the KA or KS?

2) What are your experiences with cap rated life vs. actual life, especially with miniaturized caps?

R.G.

I have not read those specific datasheets, but the hours of life are normally under full max conditions; that is, conducting full rated ripple current at max temperature and possibly humidity.

Electro caps in general last longest if stored at not-hot temperatures and left with some polarizing voltage on them. The internal chemistry aging effects are worst at high temps, and leaving a voltage on them without being at high temps actually helps reform any weak spots in the oxide layer *at least with older-chemistry electros*.

So no, I would not expect them to start dying off after 42 days under bias. On the other hand, they *are* telling you that they won't honor warranty after 1000 POH.  :icon_wink:
R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.

PRR

> life rating of 1,000 hours doesn't even make sense to me

A sample WWI Liberty aircraft engine was tested for 50 hours at 400HP before being approved for military purchasing. (Production engines failed nearer 30 hours, but that's the start of another story.)

Liberties in tank and mail-plane service sometimes turned hundreds of hours at 200HP-250HP. Liberties in rum-runner boat service might run many hundred hours putting 50+HP, peaking to 350HP in a chase, before flipping or getting caught.

A Liberty powering a generator could probably light several houses (40HP) for a thousand hours between overhaul.

Wear-out failures are often a trade-off between how hard it works and how long it lasts.

Extreme example: a tire-maker gave a 50K mile free-replacement warranty. A guy put it on a V-8 power motorcycle and wore it out in a weekend (not even 5 miles, 1/4 mile at a time). Monday he brought the sad skin in for a free replacement, to burn-up the next weekend.

Is your pedal burning-up?

Are you running near boiling point?

Are you near the rated ripple current?

Can you finger-hold the cap after heavy abuse?

In HIGH power abuse, like DC motor and welder supplies, the caps DO cook pretty quick. The designer uses the de-rating factors to stretch expected working life to significantly longer than his warranty period. 2,080 hours will cover 8 hours a day 5 days a week 52 weeks a year, and is nominally "safe" for a 1-Year warranty. 2X the life may just be running 10 deg C cooler than rated max temperature. In a hard-worked Power supply, this may need an over-size box or a fan, but these are cheaper than shipping welders back for warranty work in the 11th month.

In small audio, e-caps typically last longer than we care about. Exceptions include "beloved" vintage gear, like 1963 Fenders, which now have 10,000+ hours (at high temps). These have been failing since 1973, and most older Fenders now in service have had cap-jobs. But in younger less-intense service, e-caps last plenty long, being run not-so-hot, some below voltage rating, and far below current ratings.

Another factor is that it is easy to screw-up e-caps. It is a big messy job, but parts per million of chlorine, sodium, or other contaminants cause early failures. No shop has this completely under control, but they do get better every decade.

FWIW: around 1975 I built a box with a lot of e-caps, notably a couple large "35V" power caps worked at 37V actual. It ran barely-warm about 10,000 hours over 30 years. When I left it, there had been no e-cap failures.
  • SUPPORTER

thelonious

Quote from: R.G. on December 22, 2014, 12:13:11 PM
the hours of life are normally under full max conditions

Quote from: PRR on December 22, 2014, 06:05:55 PM
In small audio, e-caps typically last longer than we care about. Exceptions include "beloved" vintage gear, like 1963 Fenders, which now have 10,000+ hours (at high temps). These have been failing since 1973, and most older Fenders now in service have had cap-jobs. But in younger less-intense service, e-caps last plenty long, being run not-so-hot, some below voltage rating, and far below current ratings.

That makes much more sense. Thanks guys!

PRR

Here is some -old- statistics which give some perspective:



Obviously any part has a short life if you abuse it.

This data was observed from huge numbers of parts in "average abuse".

Note that the data is in percent. So "0.2" is really two per thousand. And that is in 1,000 hours. So the extrapolated life for a "0.2" part is two deaths in a *million* hours. (However other things happen in that length of time; none of the data is from million-hour parts!)

The real use is like the example. A pocket radio is not un-like a complex transistor pedal. They get a MTBF of 48,000 hours or 5.5 years steady. Detailing, the worst part is the coils, and most pedals have zero to one coil, not six. Without coils we are near 100,000 hours or 10 years steady. (1969 was still a maturing-period for transistors, so I suspect with post-1975 transistors we would be over 15 years steady. Or at 7 hours a day, hypothetical 50 years, and you will be how old and feeble-fingered?) (Apologies to Les Paul.)

Some data is obsolete. Modern carbon-film resistors are much more reliable than any carbon-compositions resistors were. Paper caps are gone (or cost a mint).

The take-away thought is that your power transformer and your POTS will give ten times more troubles than your e-caps.
  • SUPPORTER

bool

WRT lytics, I had good experence with various Panasonic, Nichicon and Philips (now Vishay iirc) caps. Some I have in my DIY-ed gear, in daily service since early 90s.

All three brands are not really "expensive" if you shop around. Nichicon have "automotive" caps now that are rated up to 150 deg C so should be fairly sturdy at lower temps.

bool

Another low-cost brand I had good experience with is Yageo. Small footprint and up to 105 deg C.