Attention Gurus- New Unused Caps- Shelf Life and Should I Charge Them

[g1]

Been re-formed occasionally and still tests good? I'd use them.

Which get's us to the argument about whether to automatically replace electrolytics which have reached a certain age. In this case, you could install those caps, the amp could go somewhere else for an unrelated problem where someone may insist on replacing those "25 yr. old" caps., which are in fact unused, and tested good.

[R.G.]

Yep. Nothing's perfect. What's a mother to do?

[The Dude]

My own personal view is that if a cap tests good, it's good- plain and simple. I don't care how old it is (within reason). I've probably replaced as many or more defective caps that were made in the 80's as those made in the 70's. I do not care for the term "re-capping", and with modern QC standards and manufacturing processes, you might be installing a part that will die before the existing part would. And, how long has that "new" cap been sitting on a warehouse shelf somewhere?

However, I do understand the "logic" behind the "re-capping" thing so usually I leave it to the customer. I explain both points of view and let them decide if they want to fork over the extra bucks for parts and labor. Some do and some don't. I do a lot of contract work and these guys just want stuff to work so that they can sell it on. They are not interested in spending the additional cash to completely rebuild an amp. Others, are looking for reliability and will spend the money- mostly because they have lots of money.

I'm not saying I'm right and you're wrong- just my take on the subject. It reminds me of the "French Model" commercial. "Why are you re-capping that amp"............"I read about it on the internet".

[Leo_Gnardo]

Electrolytics that heat themselves up when charged are fit for the recycle bin. Same for ones that emit spuzz* of any sort - not even worth trying to charge. Ten this week so far, all from 70's Fenders.

The most recently made caps I've seen fail were 18 years old, in a Mesa Boogie, don't remember which model.

And I've seen e-caps made in the 1940's that are still running just fine. If I let an amp run for say 6 or 8 hours, and no heat-up or failure, they stay in unless amp owner requests a change.

* liquid, solid, gas or any combination.

[The Dude]

^^^^^ Agreed. And, just to clarify, I did not say and do not believe that you shouldn't check caps. Anything leaking, split, high ESR, etc. should definitely be replaced. What I am referring to is wholesale replacement of parts simply because they are old.

[Chuck H]

^^^^^ Agreed. And, just to clarify, I did not say and do not believe that you shouldn't check caps. Anything leaking, split, high ESR, etc. should definitely be replaced. What I am referring to is wholesale replacement of parts simply because they are old.

I admit to advising this on occasion. But I do have a good reason. Often there is a poster with an old amp that's having problems and wouldn't know better one way or the other. Replacing filter caps removes doubt. I of course always recommend that the bias supply caps be included. Many bias supplies have failed because of old caps. This alone is worth the price of admission.

In my own space and considerations I only replace caps when one or more is showing symptoms of failure. But then I do replace them all. And I have seen 40+yo caps working perfectly and left them in place. I've also seen cheap, modern electrolytics that fail inside of ten years. It's absolutely true that you can buy (and most manufacturers do) modern caps that don't stack up compared to what is found in vintage amps. But, because of modern technology, materials and manufacturing tolerances you can also buy superior caps to what was used in vintage amps and they don't cost much more than the cheap ones, relatively speaking. A bean counter at Fender who is only looking at the bottom line for 20,000 units see's considerable savings at a dollar per unit. Too bad though. A extra ten bucks for an amp doesn't amount to a pile of grunt. But once a concession for quality is made where does it stop? Well, the better transformers are only $XX more. And the better speaker is only $XX more... Pretty soon the product can't be marketed competitively anymore.

[The Dude]

^^^Yes, and I've begrudgingly advised it myself solely for the reason that someone doesn't have a cap tester and, generally, a handfull of 'lytics is cheaper than a good cap tester. If you're going to do this for a living or even a steady hobby, you ought to own a good cap tester.

[Alan0354]

I don't believe there is age of those ceramic, and other small non polar caps.

I have been using electrolytic caps I got 15 years or more ago.

I was very into photography and I have many studio flash that are very high power. I was told before that I need to power the power pack on once a while to keep the caps alive. Don't ask me why, I have no idea. But caps use in flash has to work so so much harder than in amps. You are talking discharge everything in mS which is 100+A current. I don't know this even applies to filter caps that never see large current surge. I do make it a point to at least power them up once a year and let them sit for 10 minutes and flash a few times.

[Joe L]

^^^Yes, and I've begrudgingly advised it myself solely for the reason that someone doesn't have a cap tester and, generally, a handfull of 'lytics is cheaper than a good cap tester. If you're going to do this for a living or even a steady hobby, you ought to own a good cap tester.

Dude!

Just scanning the 'Bay for a few minutes I find a bunch of cap testers for under $50. You want to at least be able to check value and ESR. ESR can even be tested in circuit. Leakage can be tested using the HV supply on your amp with a 10 ohm current sense resitor and your DVM.

From the Wiki on Dialectric Absorption:

"Dielectric absorption is a property which has been long known. The value of it can be measured in accordance with IEC/EN 60384-1 standard. The capacitor shall be charged at the dc voltage rating for 60 minutes. Then the capacitor shall be disconnected from the power source and shall be discharged for 10 s. The voltage regained on the capacitor terminals (recovery voltage) within 15 minutes is the dielectric absorption voltage. The size of the dielectric absorption voltage is specified in relation to the applied voltage in percent and depends on the dielectric material used. It is specified by many manufacturers in the data sheets."

As you can see, it isn't very expensive to setup to tests caps. If you start doing it with new caps and recording the values on the cap itself, you will quickly gain experience on what a bad cap looks like.

[g1]

From the Wiki on Dialectric Absorption:

"Dielectric absorption is a property which has been long known. The value of it can be measured in accordance with IEC/EN 60384-1 standard. The capacitor shall be charged at the dc voltage rating for 60 minutes. Then the capacitor shall be disconnected from the power source and shall be discharged for 10 s. The voltage regained on the capacitor terminals (recovery voltage) within 15 minutes is the dielectric absorption voltage. The size of the dielectric absorption voltage is specified in relation to the applied voltage in percent and depends on the dielectric material used. It is specified by many manufacturers in the data sheets."
Dialectric absorbtion can be approximated by charging the cap to full voltage, safely discharging it quickly but completely, disconnect and wait for about ten seconds, then check the residual voltage."

As you can see, it isn't very expensive to setup to tests caps. If you start doing it with new caps and recording the values on the cap itself, you will quickly gain experience on what a bad cap looks like.

I'm probably missing something here, does dielectric absorption relate to ESR or leakage? The wiki article goes on to state that "For most electronic circuits, particularly filtering applications, the small dielectric absorption voltage has no influence on the proper electrical function of the circuit." It lists 10 to 15% as acceptable absorption for electrolytics. Is it beyond this where there is a relation to leakage or ESR?
Could you be more specific about your leakage test with 10R resistor?

[Joe L]

I'm probably missing something here, does dielectric absorption relate to ESR or leakage? The wiki article goes on to state that "For most electronic circuits, particularly filtering applications, the small dielectric absorption voltage has no influence on the proper electrical function of the circuit." It lists 10 to 15% as acceptable absorption for electrolytics. Is it beyond this where there is a relation to leakage or ESR?
Could you be more specific about your leakage test with 10R resistor?

Dielectric absorption is not related to ESR or leakage. It is the ability for the dielectric to hold a charge even after you discharge the cap down to zero volts. You can charge up an electrolytic, discharge it, then with your voltmeter, see the voltage rise up from 0 once the short is removed. It's usually one of the first signs of a cap going south.

The leakage test is as simple as hooking up a cap with a 10 - 1K ohm resistor in series to a supply that is a high as possible without going over the rating. You measure the voltage drop across the resistor and use ohms law to calculate the current flowing. Make sure and wait until the current stabilizes at the lowest reading which depending on the time the cap has spent on the shelf or out of service, can take hours. In fact, if you don't know the cap's status, start with a 10K-100K resistor at first, just to safely reform the electrolyte.

[R.G.]

I read back through this and it's clear that there's not a general understanding of what electrolytic caps are, why they go bad, and what/when to do about that.

First, no electronic designer *wants* to use electrolytic caps instead of other cap types. They are inferior in most ways, except the biggie: they can have very, very thin insulating dielectric, and so they can put a huge amount of capacitance in a small volume. For a long time, electros were the only way to get multi-uF caps in sizes that were practical to put in portable equipment. Designers are willing to live with the high inductance, high esr, short life, built in failure time, polarization needs and other oddities to get the high capacitance per unit volume and the mostly unlimited amount of capacitance possible.

The reason aluminum electros have such high capacitance per volume is that aluminum (and a few other metals, like tantalum and niobium) can be arranged to grow an insulating layer of the metal oxide over the outside of a film of the metal, and by growing this layer electrochemically, it can be very thin - a few atoms thin, if you want to mess with trying to do that. By growing the oxide layer just barely thick enough to withstand some voltage target, you get the maximum possible capacitance by using the thinnest possible insulator. It is also good that aluminum oxide grown this way has a high-ish dielectric constant, which means that the capacitance you get is even higher than you'd have if something like vacuum spacing was the insulator.

The oxide layer is grown by diddling with the aluminum foils' alloys and surfaces to get lots of surface area, then dunking the foils into a chemical bath and running electricity through it. The electricity is in a polarity that forces oxide to grow on the aluminum. The current flowing through the aluminum surface forces aluminum oxide to form over the surface, with most growth where there is most current, and the surface covers over as the current will dodge around to find open areas until those plug up with oxide. Even when the surface is covered, more voltage can force current to flow through thin oxide layers, which grows more oxide.

Eventually, the oxide layer gets so thick that for the applied voltage only a trivial amount of leakage current flows. The oxide is now "formed" up to this voltage. When it's up to 10V, 50V, 500V, whatever is needed, the finished foil is wound and stuck in a can, and electrolyte goop is soaked into the can, and the can is closed. Cap is done, and it's shipped to buyers after testing.

So far so good, but notice that the oxide was forced to form by pouring energy into it. Over time, it UN-forms. The electrolyte goop in the can has at least one of its aims to chemically prevent this slow un-forming, but can't stop it entirely. With time, some of the oxide film in isolated spots, at random, gets to where it would support less and less voltage. The oxide gets thinner, so capacitance goes up, but withstanding voltage goes down. Eventually, with time, if it simply sits on the shelf, some spot(s) will un-form to the point that putting the normal voltage on it will let massive currents flow, the heat burns spots, and the cap dies.

But if you catch the cap before it un-forms to the point that high leakages happens in spots destructively, the merely leaky spots can be re-formed by putting a voltage across the cap. This is how it was made in the first place. Leaky spots actually re-form back to higher voltages to some extent. It's not perfect, because the electro goop for use is not the optimized bath for originally forming the cap, but it can work. Caps that are in service at modest temperatures continuously do this random-un-form and voltage re-form, because the spots that leak most also re-form most. Keeping a cap at its working voltage all the time without also cooking it at high temperatures is the way to keep the cap in the best shape for the longest time.

Even periodically reapplying voltage lets the weaker spots re-form by the leakage current re-forming the weak spots.

So that's why there is a shelf life. If they sit on a shelf too long and you put normal voltage on them without some kind of current limit, the re-forming can't happen fast enough to make up for the local hot spots and distruction, so you can get catastropic failure. Note that other issues can happen, including chemical degradation of the electrolyte, which used to be drying out when it was an aqueous solution, but most of them are non-aqueous now. Still, they degrade with time.

But if you periodically, or once when you pull them off the shelf, put them on rated voltage with limited current so the re-forming leakage can re-form the voltage ability of the weak spots, they can be re-formed to low enough leakage to be put into service, and then they get on the curve of normal voltage reforming weak spots until something else kills them.

Re-forming consists of applying the desired voltage through some current limit. The military wants a tungsten bulb. A 100K resistor works too. The leakage current in the weak spots is limited to a value that re-forms the oxide without burning a failure spot there. A cap that won't re-form is dying from one of the other failure mechanisms.

Over time, the state of the art in terms of aluminum alloys and electrolyte chemistry progresses. You can't buy caps like they used to make 'em - thank God. Modern etched foil electros are a fraction of the size and much less prone to internal chemical failure as well as drying out and have lower ESR and ESL, as well as better heat dissipation.

[Rhodesplyr]

My grandfather was a professor of electrical engineering at Clemson and NCSU from the 1930s through the 1960s. He believed that electrolytic capacitors that hadn't had voltage applied in a long time needed to be reformed, something he taught me when I was ~ten years old.

I recently wanted to power up a 1957 Dumont amp someone gave me for initial testing purposes that probably hadn't been turned on in 30-40 years. I have a rebuilt Heathkit C1 Condenser Checker that I've found to be a very useful tool for this kind of thing. It will essentially auto-reform an electrolytic capacitor and has a leakage indicator built in. It took over three hours for the caps to reform, but they finally did, so I was able to test the amp to see what I might want to do with it. With the Heathkit tester, I can just set it up and leave it running while I'm doing something else (periodically monitoring). It also discharges them safely when you turn it off.

[tedmich]

^^ nice thorough answer RG! one question: how is it that many reforming protocols for stored EL caps simply have them heated in an oven with no applied voltage? It would simply be too expensive for Vishay to take back shelf stock of close to expiring EL caps and actually apply current limited voltage to them... I had always heard reforming involved putting the electrolyte back into solution after it had crystallized out, thus the cap reforming ovens in common use. I think Ted Weber made one and spoke about it on the interwebs a while back. So that would be thermo-chemical reforming as opposed to the electrochemical you described.

[Chuck H]

I don't recall ever reading about cap reforming being done with heat exclusively. I think that is just one step in a process. The heat would help re emulsify the electrolyte and would be followed by current limited application of voltage. At least that's what I've read. I've never tried it.