Using "new" old electrolytic caps in new circuits - Yay or nay?

[WyverntekGameRepairs]

Hello, I was working on a circuit at college a couple weeks ago using old Rubycon electrolytic capacitors (from around 1980, 1982... Pretty old), and it only just hit me now - How would age affect these capacitors?

So I have decided to ask all of you. When I received my capacitor design kit from a supplier a week ago, I noticed that it indicated that the components "expire" over time. I can see why - Age and the ambient environment temp, moisture levels, etc. can cause imperfections and decrease reliability in any component, especially electrolytic capacitors.
Knowing that age affects the performance and reliability of these capacitors, how good of an idea was it for me to use old electrolytics (even if they are Rubys) in a medium-voltage circuit? (The voltage being held by the caps was around 10VDC, and the caps were rated 1,000uF @ 24V). Another example is that I did a cap replacement on my Atari 2600 at my college campus as a small project. I replaced the main big cap on the switchboard (the one that filters the DC voltage in before it reaches the regulator) with an equivalent capacitor that was from around 1980 - it was in the original unopened Radioshack box packaging and everything - as well as a signal filtering cap on the switchboard with a axial leaded cap from around the late 1970's. Thinking about how even unused components are affected by time, I'm wondering if it is safe for me to be using these capacitors on my console or any other personal project I work on in the future? (Of course, I don't use New Old Stock parts for repairing other people's things, I only use new components.)

What do you think?

[Nusa]

Hey, if it looks good and tests good and its for a personal project, nothing terrible about reusing parts. Especially if it's an expensive part to replace.

[WyverntekGameRepairs]

Hey, if it looks good and tests good and its for a personal project, nothing terrible about reusing parts. Especially if it's an expensive part to replace.

The Atari 2600's TIA chip comes to mind quite easily...
Anyway, I can see your point. I guess I'm just worried about old parts looking OK from the outside, but when they are put into working circuitry for the first time in decades (or ever, in terms of their life span) they end their short functioning lives with a loud *bang* and puff of smoke, and leave me dumbfounded as to what the hell just happened. Not that that has happened to me before, but it is a possibility of happening, and it is an eerie, disconcerting feeling. Like, I'm okay with using old components if I don't have new ones to replace it with (in personal projects only, of course!), but yet I sometimes have that eerie feeling that the life of the component could end loudly and unexpectedly after spending most of it's "life" in a parts bin for a few decades.

[mcovington]

I looked into this recently, reviewed the literature, made some tests, and came up with this:
http://www.covingtoninnovations.com/michael/blog/1910/index.html#x191006

I think 1980 is a bit old, but with Rubycon (high quality) it might be OK.  The guideline I ended up with is:

"My current practice is to freely use capacitors up to 10 years old, use them with caution up to 20 or 25 years old, and distrust those that are older. There is certainly no need to toss them when they're just 2 or 3 years old. ...

I expect people to write in and tell me I'm wrong. But I expect to be told I'm wrong in both directions (too cautious or not cautious enough), so it will balance out."

Capacitor manufacturing has improved recently.  Panasonic capacitors no longer have date codes.

[james_s]

Capacitors (electrolytic) don't normally go bang, they just tend to increase their ESR and drift out of tolerance. They can also physically leak, so inspect the legs for signs of leakage and corrosion.

[MagicSmoker]

This is one of those questions where the answer is always, "it depends."

Loss of electrolyte over time due to imperfect sealing around the terminals is a major cause of "death by senility" in aluminum electrolytic capacitors. This failure mechanism causes ESR to rise - sometimes dramatically - and there is no curing it. So, a good first pass test of any NOS (New Old Stock) capacitor is to measure its ESR. And how do you know if the ESR is too high or still ok? Well, that very much depends on the capacitor, but a general purpose elko will have a dissipation factor, DF, which is the ratio of ESR to Xc, of around 10-20% (0.1-0.2). An elko is usually considered EOL (end of life) when DF has doubled.

For example, a 100uF capacitor will have a capacitive reactance, Xc, at 1kHz (a common frequency used to test ESR) of 1.59 Ohms, so a good capacitor should have an ESR of between 0.159 and 0.318 Ohms (ie - 10% to 20% of Xc), while an ESR of >0.6 Ohms would be suspect.

The other age-related failure mechanism is a loss of the dielectric layer over time which causes the DC leakage current to increase, perhaps to the point of causing the capacitor to overheat and possibly even explode. However, this failure mechanism can be reversed by merely applying a slowly increasing voltage to the capacitor over time which re-forms the dielectric layer. How slowly? No prizes if you guessed the answer is, "it depends...," but I've seen recommendations as wide as going from 0V to rated voltage at the rate of 1 minute per month of storage to 1 hour per year of storage; in my experience it doesn't really matter, just keep an eye on the actual current flowing through the capacitor after each increase in voltage and wait until it has dropped below a few hundred microamps before increasing the voltage again in steps/amounts no more than 1% of the capacitor's rated voltage.

[WyverntekGameRepairs]

Wow, this is a lot of great information. Thanks! I think I can safely go over a few points while I'm here...

I looked into this recently, reviewed the literature, made some tests, and came up with this:
http://www.covingtoninnovations.com/michael/blog/1910/index.html#x191006...

I read up on that little article, and it is actually quite helpful. Thanks! Also, I appreciate you highlighting the important notes out.

Capacitors (electrolytic) don't normally go bang, they just tend to increase their ESR and drift out of tolerance. They can also physically leak, so inspect the legs for signs of leakage and corrosion.

Oh, I thought they would just go *bang* instead of waver off like that. Interesting. I should have predicted it would taper off instead of just explode, but hey - Now I know
And it indeed makes more sense that it would just fade out like that, as I don't think aged capacitors lose any voltage tolerance and therefore can't explode from overvoltage.
(Not counting shitty capacitors of course, we all know those pretty much explode regardless of what you do!)

This is one of those questions where the answer is always, "it depends."

Loss of electrolyte over time due to imperfect sealing around the terminals is a major cause of "death by senility" in aluminum electrolytic capacitors. This failure mechanism causes ESR to rise - sometimes dramatically - and there is no curing it. So, a good first pass test of any NOS (New Old Stock) capacitor is to measure its ESR. And how do you know if the ESR is too high or still ok? Well, that very much depends on the capacitor, but a general purpose elko will have a dissipation factor, DF, which is the ratio of ESR to Xc, of around 10-20% (0.1-0.2). An elko is usually considered EOL (end of life) when DF has doubled...

Thanks for this, I'll be sure to use this testing method before I use any NOS caps in the future.

Any additional votes / comments are welcome!

[Nusa]

Installing with reverse polarity sometimes gets you a bang, depending on the amount of energy involved. But usually they don't explode unless you're exceeding their specifications by a LOT.

[WyverntekGameRepairs]

Installing with reverse polarity sometimes gets you a bang, depending on the amount of energy involved. But usually they don't explode unless you're exceeding their specifications by a LOT.

Or, if I'm not mistaken, you show them AC. Electrolytic polarized caps don't really like having AC put through them. If, for example, a bridge rectifier in a switch mode power supply were to fail and the DC smoothing capacitor was to start seeing the AC mains, it would probably last only a few seconds before it blew up and left big skid marks all over the power supply, and then causing a chain reaction as each part of the power supply failed and it snowballed until eventually it spelt "disaster" in great, big, flaming red letters.

[Nusa]

A bridge rectifiers is four diodes, and generally only one diode will fail at a time. They don't fail often, but when they do they generally fail open. What do you think happens to the rectifier output in that event? But it's possible for them to fail closed as well, so also think about that rectifier output case as well. Now think about the less likely case of multiple diode failures. How do they have to fail to actually achieve a scenario that passes direct AC? Not so easy, is it? And in any case, you're wrong....DC smoothing capacitors aren't going to go boom from a little 50/60 Hz AC in the short term...remember half the cycles are in the correct direction.

[WyverntekGameRepairs]

A bridge rectifiers is four diodes, and generally only one diode will fail at a time. They don't fail often, but when they do they generally fail open. What do you think happens to the rectifier output in that event? But it's possible for them to fail closed as well, so also think about that rectifier output case as well. Now think about the less likely case of multiple diode failures. How do they have to fail to actually achieve a scenario that passes direct AC? Not so easy, is it? And in any case, you're wrong....DC smoothing capacitors aren't going to go boom from a little 50/60 Hz AC in the short term...remember half the cycles are in the correct direction.

Hm. You know, you're right. I guess I was thinking of a different scenario? Or maybe that was just a bad example. My point was that generally capacitors tend to not accept AC voltage all that well (though now that I think about it, it might be dependant on the period that the voltage goes negative?)

Though that being said... If I take a 450V 150uF cap from a power supply and put it directly on mains voltage, will it explode? What about a 250V cap? I'm genuinely quite curious. I'm afraid to try it out myself in case it explodes with unimagineable force or excessive fire.

[David Hess]

Aluminum electrolytic capacitors can fail with low capacitance and high ESR after being stored for decades.  Even if they test out as acceptable, a large part of their remaining operating life may have expired.

This also applies to aluminum electrolytic capacitors which have been in service with essentially zero ripple current for decades.  I have refurbished old test gear where all of the small output bulk decoupling capacitors were only at 10 to 50 percent of their marked value with the larger ones doing better indicating that the failure was gradual evaporation of the electrolyte over time.

[helius]

The standard failure mode for old (or non-operated) electrolytic capacitors is "de-forming", thinning of the oxide dielectric layer through etching by the acid electrolyte. In this scenario, the measured capacitance increases (because the dielectric separation is thinner), but the charge leakage is very high. Capacitors in this condition can be restored to near-factory condition by re-forming.
Sometimes the electrolyte has evaporated (dried-out capacitors). This would lead to low measured capacitance (because less than the full foil area is acting as a capacitor). This condition cannot be remediated and the caps should be thrown away.

[james_s]

Hm. You know, you're right. I guess I was thinking of a different scenario? Or maybe that was just a bad example. My point was that generally capacitors tend to not accept AC voltage all that well (though now that I think about it, it might be dependant on the period that the voltage goes negative?)

Though that being said... If I take a 450V 150uF cap from a power supply and put it directly on mains voltage, will it explode? What about a 250V cap? I'm genuinely quite curious. I'm afraid to try it out myself in case it explodes with unimagineable force or excessive fire.

150uF at 60Hz is going to present an impedance of 17.7 Ohms which placed across 120VAC will draw 6.7A of reactive power. This is quite a lot of current and will result in heating the capacitor, whether it can tolerate this or not depends on the particular capacitor and how much ripple current it can tolerate.

If you are using a polarized electrolytic capacitor though it will be connected backwards 50% of the time and will not last long. I suspect it would blow the top but probably not immediately.

[WyverntekGameRepairs]

Hm. You know, you're right. I guess I was thinking of a different scenario? Or maybe that was just a bad example. My point was that generally capacitors tend to not accept AC voltage all that well (though now that I think about it, it might be dependant on the period that the voltage goes negative?)

Though that being said... If I take a 450V 150uF cap from a power supply and put it directly on mains voltage, will it explode? What about a 250V cap? I'm genuinely quite curious. I'm afraid to try it out myself in case it explodes with unimagineable force or excessive fire.

150uF at 60Hz is going to present an impedance of 17.7 Ohms which placed across 120VAC will draw 6.7A of reactive power. This is quite a lot of current and will result in heating the capacitor, whether it can tolerate this or not depends on the particular capacitor and how much ripple current it can tolerate.

If you are using a polarized electrolytic capacitor though it will be connected backwards 50% of the time and will not last long. I suspect it would blow the top but probably not immediately.

Actually, that clears up a lot! Thanks for that info, very helpful. I never thought about it that way. I guess I taught myself to not put capacitors in any AC situations because it could shorten their lifespan (as they do see negative voltage, basically incorrect polarity) or even cause them to explode depending on how badly they were constructed. But hey, guess I learnt something new today
Basically what you are saying is that it really depends on the construction quality and ESR ratings of the capacitor. It also is affected by the frequency and voltage, of course, and there are a lot more factors than I thought. Pretty cool info, even if it is off-track from what I posted (granted, I did ask in the first place, so it's my fault really).

The standard failure mode for old (or non-operated) electrolytic capacitors is "de-forming", thinning of the oxide dielectric layer through etching by the acid electrolyte. In this scenario, the measured capacitance increases (because the dielectric separation is thinner), but the charge leakage is very high. Capacitors in this condition can be restored to near-factory condition by re-forming.
Sometimes the electrolyte has evaporated (dried-out capacitors). This would lead to low measured capacitance (because less than the full foil area is acting as a capacitor). This condition cannot be remediated and the caps should be thrown away.

I saw BigCliveDotCom "activate" a fake electrolytic capacitor by adding electrolyte to it (it was a dry capacitor, like the kind you see in scam fake "oem" caps sold by shady bootleg vendors). In theory, would adding new electrolyte to a cap after it has dried out "reactivate" it in a sense?

[Gyro]

I saw BigCliveDotCom "activate" a fake electrolytic capacitor by adding electrolyte to it (it was a dry capacitor, like the kind you see in scam fake "oem" caps sold by shady bootleg vendors). In theory, would adding new electrolyte to a cap after it has dried out "reactivate" it in a sense?

In theory, yes. The practical problems are that the electrolytes used are weird (usually secret*) cocktails which vary from manufacturer to manufacturer (and probably type to type too, eg long life vs Low ESR or temperature rating). The other one is how to introduce the correct amount of electrolyte without creating a new leakage path - I suppose something like a very fine hypodermic followed by some sort of sealing compound might work.

So yes, replacing what has evaporated off would theoretically do it, if the electrolyte has leaked past the seal though, you'd face serious issues with corrosion eating your PCB tracks.


A little OT, but at one time, people used to try the same sort of thing with dried-out NiCd cells, it mostly worked... Again the problems were sealing and existing leaks but at least in that case, there was no fancy recipe to worry about, just deionized water.


P.S. * A few years back, there was a huge issue with low ESR capacitors on PC motherboards etc. It turns out that it was mainly due to one manufacturer attempting to 'duplicate' another manufacturer's formula. Google 'Capacitor Plague for more details.

[tggzzz]

Knowing that age affects the performance and reliability of these capacitors, how good of an idea was it for me to use old electrolytics (even if they are Rubys) in a medium-voltage circuit?

Often answering a question is much easier than working out which is the right question to ask.

I suggest a useful question would be "what is the cost if one of the capacitors fails?".

[WyverntekGameRepairs]

Thanks everyone, all of this information is highly useful!
So in conclusion, it is indeed okay to use an old or NOS capacitor in new personal projects as long as the ratings are still accurate and the capacitor is showing no signs of leakage or damage. There is an easy test to determine if an old cap is still useable or not:

Loss of electrolyte over time due to imperfect sealing around the terminals is a major cause of "death by senility" in aluminum electrolytic capacitors. This failure mechanism causes ESR to rise - sometimes dramatically - and there is no curing it. So, a good first pass test of any NOS (New Old Stock) capacitor is to measure its ESR. And how do you know if the ESR is too high or still ok? Well, that very much depends on the capacitor, but a general purpose elko will have a dissipation factor, DF, which is the ratio of ESR to Xc, of around 10-20% (0.1-0.2). An elko is usually considered EOL (end of life) when DF has doubled.

For example, a 100uF capacitor will have a capacitive reactance, Xc, at 1kHz (a common frequency used to test ESR) of 1.59 Ohms, so a good capacitor should have an ESR of between 0.159 and 0.318 Ohms (ie - 10% to 20% of Xc), while an ESR of >0.6 Ohms would be suspect.

The other age-related failure mechanism is a loss of the dielectric layer over time which causes the DC leakage current to increase, perhaps to the point of causing the capacitor to overheat and possibly even explode. However, this failure mechanism can be reversed by merely applying a slowly increasing voltage to the capacitor over time which re-forms the dielectric layer. How slowly? No prizes if you guessed the answer is, "it depends...," but I've seen recommendations as wide as going from 0V to rated voltage at the rate of 1 minute per month of storage to 1 hour per year of storage; in my experience it doesn't really matter, just keep an eye on the actual current flowing through the capacitor after each increase in voltage and wait until it has dropped below a few hundred microamps before increasing the voltage again in steps/amounts no more than 1% of the capacitor's rated voltage.

It is also possible to "reactivate" an electrolytic cap by adding electrolytic solution, but it is definitely not recommended for use in projects, and is more for experimentation with capacitors and capacitor engineering - not general hobbyist cap-revival methods.
After doing some investigation into the posts above, it seems that capacitors that experience use, even if occasional, have a longer life than unused caps of the same age that have not been maintained for shelf life. This depends on the quality and materials of the capacitor.
And finally, it is best to just use common sense when attempting to use NOS caps in a project, of course. If it looks or measures suspicious, then it *is* suspicious.

[strawberry]

'used' capacitors are still fine (spot with the eagles eye) to use them, most cases 90% of juice is still there.
Old good quality capacitor is better than brand New dirt cheap one
Electrolytic capacitors placed near hot heat source is no point testing
Larger electrolytic capacitors are more durable than smaller or low profile
In last 20 y capacitors are made out of water based electrolyte with lower ESR values than older capacitors before, downside is if chemistry is wrong it will generate H and eventually went