Old Capacitors.

[jmcilkley]

I just started recapping a 1970s Akai AS980 quadraphonic receiver. There are 4 caps on the amplifier output rated at 2200uF 35V which I have replaced. I then tested the old ones and all 4 test at only between 1000 and 1150 uF. Is this normal to lose half their capacitance after 45 years? All 4 by about the same amount? Other caps seem ok so far. It just seems odd to me.
i'm attaching the circuit diagram to show C8, the cap concerned.

[Radio Wrangler]

It's not unusual.

In an electrolytic capacitor, one electrode is an aluminium foil whose surface has been matted to increase the surface area. This surface has a thin layer of aluminium oxide built up on it to form the dielectric. The problem then is to have a second electrode shaped to intimately follow the surface of the oxide. Anything less than perfectly coating the oxide with conductor will leave air gaps and high capacitance won't be achieved. The answer to the problem is to use a conductive liquid as the electrode. The liquids chosen are viscous enough not to go walkies, and are soaked into a paper strip with a second aluminium foil on the other side. This second foil is naked: No oxide on its surface, so that it connects well to the conductive liquid.

The really clever part is that the capacitor can be made with identical foils, and that by applying DC voltage after assembly, the capacitor will take current and electrolytic action will break down and remove the oxide on one foil, while building it up on the other. So assembled capacitors-to-be have to be formed to become capacitors. This forming can undo itself over a long period and it's not unusual to find old capacitors which need re-forming. Also reverse biasing one undoes the forming rather quickly and leads to failure as it goes conductive.

There is another, and usually bigger ageing effect. That conductive liquid is water based. Water tries to evaporate and vapour pressure builds up inside the capacitor's housing can. The pressure depends on temperature, and you get pressure related leakage of water vapour past the seals where the rubber/plastic bung end meets the can and the conections throug it. The hotter the capacitor, the faster the water loss. It may be heat from things in the vicinity, or heat made internally due to the amount of AC current the capacitor is passing. Water vapour pressure goes up, life expectancy goes down. Better sealing helps.

As water is lost, the conductive liquid dries out and voids form, reducing the capacitance as you've found.

Also, the net resistance presented by the fluid part of the current path goes up.

Measuring the ESR of the capacitor is a better way of checking for dying/drying capacitors as the Equivalent Series Resistance starts to rise before there is much reduction in capacitance. So a lot of people on here have built or bought ESR meters.

A great many people think electrolytic capacitors operate by some chemical process, like little batteries, but the dielectric is a solid, and good old aluminium oxide. Nothing fancy is going on when you use them. The fancy stuff is when they are formed in manufacture. While they don't work chemically, they are made that way. Impressively clever.

Progress in foil etching and the formulation of the conductive liquid (well, jelly) have shrunk the physical size of modern capacitors, and reduced ESRs.

There is a tale that industrial spies stole the secret formula of an advanced electrolyte from a Jamanese firm, and passed it to China where it rapidly spread around their capacitor makers. It's also said that the formula stolen was not the final one and it lacked additives needed to control corrosion. Like all such tales of espionage, who knows the truth? There is certainly a lot of money at stake, but we did see a plague of unreliable electrolytic capacitors and a surge in failures of consumer goods.

So, in buying new capacitors, look for their rated life expectancy and their max temperature. Oh and their ripple current rating. You may also want to go for for reputable Japanese manufacturers.

David

['LIVEWIRE?']

A long time ago during the CB boom, I bought an American textbook on the subject, which, IIRC, included a recommendation that Electrolytic Capacitors should be replaced after five years, no doubt for reasons basically stated by David (RW) in the post above, but I don't recall, even when my full-time job was repairing car radios, ever doing that as a matter of course. It was more a case of 'if faulty or maybe suspect replace it, otherwise leave well alone!

[Craig Sawyers]

Originally Posted by Radio Wrangler

There is a tale that industrial spies stole the secret formula of an advanced electrolyte from a Jamanese firm, and passed it to China where it rapidly spread around their capacitor makers. It's also said that the formula stolen was not the final one and it lacked additives needed to control corrosion. Like all such tales of espionage, who knows the truth? There is certainly a lot of money at stake, but we did see a plague of unreliable electrolytic capacitors and a surge in failures of consumer goods.
David

A computer I bought at around that time (20-ish years ago) by a long defunct company called TIME. The main board used these Chinese capacitors of death, and when the computer stopped working, the capacitors had voided their contents onto the board. The ones that hadn't were bulging badly.

I managed to resurrect it by removing all the capacitors, cleaning the board, and using new and decent electrolytics. That lasted long enough for me to archive the contents onto an external HDD, and I then scrapped the thing.

Craig

[Boulevardier]

David (Radio Wrangler), thanks for that fascinating account of the manufacturing of electrolytics. The depth and breadth of knowledge/expertise on this Forum takes my breath away sometimes - quite often actually!

Mike

[GrimJosef]

Originally Posted by 'LIVEWIRE?'

A long time ago during the CB boom, I bought an American textbook on the subject, which, IIRC, included a recommendation that Electrolytic Capacitors should be replaced after five years ...

The insides of electronic equipment built into cars do seem to run very hot - I guess because they've been squashed into the limited space and also because there's little ventilation behind the console. As David has said, heat can be the death of electrolytics. I recall a working rule that each 10C temperature reduction, at least for a few tens of degrees below the maximum rating, doubles the capacitor's lifetime.

In the end though most electrolytics from reputable manufacturers have their nominal lifetime at their maximum temperature listed in the datasheet. The more you pay, the more life you get. The cheapest decent ones tend to be rated for 2000hrs at 85C. Inside warm valve equipment I try to fit ones rated for ~5000hrs at 105C. A few are rated for as long as 15000-20000hrs and others (I think for use in car engine bays) at 125C.

5 years seems a bit short. But if they're running for long hours in a hot environment and you want to keep the failure rate below, say, 0.1% per annum then that might be reasonable.

Cheers,

GJ

[Ted Kendall]

The "capacitor of death" syndrome isn't confined to Chinese components - many Japanese DAT transports and camcorders made extensive use of surface-mount electrolytics, and it is not uncommon to find a mass evacuation of fishy-smelling corrosive on the boards from duff capacitors. Worse, the action of said on solder renders affected joints very difficult to clean up.

[Maarten]

There's a few separate problems.

- The big scandal from around the turn of the century concerned, in its least inaccurate retelling, a thief who sold a formula without the additives that were needed to prevent electrolysis to various Taiwanese component manufacturers. For some years, Taiwanese motherboard manufacturers explicitly mentioned "Japanese capacitors" to regain their customers trust.

- The Chinese repetition of the Taiwanese scandal, early and often, any time when cost is to be saved and the quality control engineers are asleep.

- Various capacitors with bad seals, spilling their contents under certain relatively reasonable conditions (SMD and various through hole Elna types probably did this even when used normally, but temperature cycling while not in use might speed things up).

- Various capacitors with bad seals, spilling their contents after some years of relatively heavy usage (often found in VCR power supplies from Grundig, Sony (Grundig production, but leaky Elna in their own products), Philips (former Grundig production) in the late 1980's - early 1990's).

- Underdimensioned capacitors (either by the capacitor maker or by the equipment maker)

The borderline between the last two can be vague. Panasonic G-deck machines suffered from something like this. Probably no bad seals but using borderline suitable capacitors under harsh conditions. Also Samsung TV's and DVB receivers suffered from one or both during a few years, using Samxon and Samwha electrolytics. I think this happened just after the Taiwanese scandal.

[ricard]

I've read somewhere that the electrolyte in modern (as in anything made after the era of 'wet' electrolytics with replacable electrolyte, that is 1940s and on) use an electrolyte that is hygroscopic, meaning that it absorbs moisture, rather than the water evaporating over time. Is there any truth to this? It would seem to be the opposite of what David is saying above if so.

[Radio Wrangler]

If you take apart failed, high ESR type failure capacitors, you find that the insides are dry. It's an experiment that can be performed at home anytime you have a dud.

There is a youtube video of a guy opening new capacitors and placing them in really nice looking wooden cases, for audiophile, reasons of some sort. I think his sealing may be somewhat degraded by his actions.

David

[Techman]

It's a very interesting topic.

I'm not one to go replacing original electrolytic capacitors unless the ones fitted have failed or show signs of failing.

An original electrolytic capacitor that has survived for sixty odd years and is still going strong, may well survive another sixty years, but most of us won't be around to see if that's the case.

I've often wondered about the life expectancy of modern replacement electrolytic capacitors, particularly because of all the known problems with modern electrolytics causing early failure of modern consumer products, hence, unless there's very good reason, I prefer to keep with the well proven 'old' original.

The question that I would always ask is will the replacement last another sixty years or will it need replacing again five years down the line?

Some will no doubt say that they've fitted replacements to vintage kit ten years ago and they're still fine. My answer would probably be that the kit no longer gets the 24/7 use that it probably would have had in it's previous life, so can't really be compared. This is a reason that I'm not entirely comfortable with 're-stuffing'. In a normal situation of heavy use as when the radio, or whatever, was new and in continuous use for many hours a day, the original capacitor would have been in 'relatively' free air, whereas a capacitor stuffed inside the container of an original capacitor has no cooling air flow around it and I tend to feel that the main reason that premature failure doesn't happen is that the kit doesn't get the amount of use as a restored vintage item that it would have got when originally first used.

For the record, I've never yet re-stuffed an electrolytic capacitor. I'm not saying that I would never consider re-stuffing one, and those that do, often make a very good job of doing so, it's just that personally I'm not entirely comfortable with it. Capacitors such as smoothing electrolytics were never designed to be effectively sealed inside another container. Also, how would you be able to tell years into the future whether a capacitor 'stuffed' in such a way was starting to bulge and fail? What would happen if failure went unnoticed and the capacitor exploded within the can? What would be the effect of this explosion on the possible shorting out of the non-original wiring used to connect the capacitors within the re-stuffed can? All things to think about when considering whether to re-stuff, or replace below the chassis out of sight, or perhaps just fit a close looking replacement and call it an honest historical repair.

As regards what has been said regarding the replacement of all the capacitors in such kit as vintage CB radios, I have noticed that there have been people selling complete capacitor kits for blanket replacement - good for them! I've never found the need to replace all these capacitors in these virtually forty year old sets, many of which still function a well as the day they were made.

[McMurdo]

Specifically those output capacitors lead a hard life in audio equipment, perhaps more so than other forms of coupling or decoupling capacitors. I see failed ones quite often where the others in the circuit are fine.

[Techman]

Kevin has nicely brought the thread back onto the actual topic that the OP has asked about with regards to his Akai receiver. However, since the actual thread title is just "Old Capacitors", I think we are justified in discussing the wider subject that this title would indicate.

[Radio Wrangler]

I bought a Revox B261 FM tuner. A very well reputed piece of kit. It was ex-BBC where it had been in either a rebroadcast or monitoring rack for a few tens of years, running day and night, with other hot equipment.

It worked. The dial bulbs were kaput and it had a definite hum problem.

LED festoon bulbs and a re-wire of the supply to a suitable voltage fixed the lighting. I suspected the reservoirs for the hum. One had been changed already, but it turned out to be a low-value electrolytic in the regulator of one rail. A value/rating that got used extensively through the rest of the set.

Enter a Peak ESR meter and a quick run through all the electrolytic capacitors. 70% of the small values had gone high enough to show that their reserves of electrolyte were gone. The reservoirs themselves were OK.

Seeing the lie of the land in a unit which had certainly spent years stewing 24/7, it made sense to just replace all the small ones. I'm not a fan of blind 're-capping' it seems to be done unnecessarily and stirs up all manner of subsequent problems when done with poor workmanship or by people who can't pick suitable replacements or get the polarity right. But when you find a large proportion of the same things all iffy and know that they have been exposed to life-accelerating factors, it is justified.

The hum has gone. Everything worked first time, and I have a fine-sounding tuner. It's the only time I've done a blanket re-cap.

David

[Diabolical Artificer]

As to the OP's question, all caps aren't born equal, Akai for the most part used cheap a*se caps, the then equivalents of Teapco's, that and it depends where they are sited. the worst looking caps I've come across were in a DAB radio, a cheap brand siting right next to a voltage regulator with thin as tin foil heatsink. They looked liked they'd been on the cap equivalent of 24hr drinking binge.

I had a jar full of 2 yr old NOS Panasonic 47u 63v 105 deg C electrolytics, by general consensus not cheap shoddy jobbies ruined by several having vented electrolyte and also most showing high ESR but we've all come across old 1960's date stamped caps still within spec. From experience smaller caps (size and value) seem to fail more often whereas their bigger brothers are fine for the most part.

Andy.

[whiskas]

My simple advice is - don't buy Chinese electrolytic caps - AT ALL - ever.
(Nothing against Chinese people who are generally lovely folk.)

Good caps by Panasonic, Nichicon, Nippon Chemicon are available from RS, Farnell etc at reasonable cost.
Also it's fitting and appropriate to use a Japanese brand cap in Japanese equipment.
To install cheap caps in good equipment is a fallicy, a waste of time and money, and makes a mockery/mess of the equipment.

The Japanese have known how to make reliable capacitors for many years now.
(The Chinese cap companies are still working on it....)

[G6Tanuki]

I agree on the 'avoid unknown-brand Chinese electrolytics' - these are often the cause of failure in cheap CFL and LED lightbulbs.

It's not only the Chinese that have issues though - a ~professional~ SMPS I worked on recently had grey-sleeved electrolytics made by "Illinois Capacitor" - they must have been 20 years old but after testing a few and finding a few which had become not-capacitors I didn't bother to test the others and just replaced the lot.

As well as the brands mentioned upthread, Rubycon is one of my go-to brands for new electrolytics - but beware - there are Chinese ones called "Rulycon"!

Finding the right capacitance/temperature/voltage-rating for replacements isn't usually a problem; the bigger issue seems to be that manufacturers are reducing their range of axial/radial leaded electrolytics in favour of SMD - it's annoying to search the Internet on the basis of voltage/capacitance and find over 90% of the results are now for SMD parts.

[Diabolical Artificer]

"Rubycon is one of my go-to brands for new electrolytics - but beware - there are Chinese ones called "Rulycon"! " There are also lots of cheap caps that copy Rubycons brown colour and logo.

A.

[ColinTheAmpMan1]

I have come across a peculiarity in comparing some old electrolytics with some modern ones. Let me say first that the old capacitors are in a valve amplifier; a Vox AC30 to be exact. On checking with my Peak ESR70, I get the following figures:

TCC "Picopack" 8uF + 8uF @ 450 VDC:-
Red- 9.6 uF, ESR 0.74 ohm. Yellow- 10.2 uF, ESR 0.42 ohm.

RS 16uF + 32 uF @ 450VDC:-
Red- 15.8 uF, ESR 0.84 ohm. Yellow- 32.4 uF, ESR 0.82 ohm.

A couple of new dual electrolytics branded "MA" show:

8 uF + 8 uF @ 500 VDC:-
Red- 9.24 uF, ESR 4.8 ohm. Yellow- 9.39 uF, ESR 4.4 ohm.

16 uF + 16 uF @ 500 VDC:-
Red- 18.01 uF, ESR 3.9 ohm. Yellow 18.28 uF, ESR 3.7 ohm.

I am inclined to leave the "original" capacitors in place, as they seem to test just fine. What is puzzling me is that these old electrolytics have significantly lower ESR than the brand-new capacitors, which to me is counterintuitive . Does anyone have any comments on this oddity?

Colin.

P.S. I describe the old electrolytics as "original" but I have no proof of this. I actually suspect that the RS-branded 16 uF + 32 uF @ 450VDC is not original, as the parts-list describes it as 16 uF + 16 uF @ 450 VDC.

[MotorBikeLes]

Try putting the new ones in circuit for a few days then test again. i suspect the ESR may have fallen by then.
Les.