Impact to lifespan of an electrolytic capacitor run near rated voltage?

[globoy]

I have a circuit that uses a pair of 470 uF capacitors as input bulk capacitance to deal with the inductance of a long DC power lead between the device and power supply.  The power supply outputs a regulated 24VDC (~2% regulation -> ~23.5 - 24.5 volts).  The design originally specified 35VDC working voltage caps but they are proving to be unobtanium (mainly because I'm using 10mm diameter caps).  But 25VDC parts are available.  I am under the belief that theoretically this should be ok but I am curious if there are any other issues I should be aware of running a capacitor so near its rated voltage.

[wraper]

Mouser has plenty of 470uF 35V 10mm diameter caps https://eu.mouser.com/c/passive-components/capacitors/aluminum-electrolytic-capacitors/?capacitance=470%20uF&diameter=10%20mm&voltage%20rating%20dc=35%20VDC&instock=y&sort=pricing

[Conrad Hoffman]

Nothing scientific to base it on, but I think heat is a bigger factor. Old radios and amps ran voltages very near the cap maximums and seemed to last decades. All the major cap manufacturers have general application guides that nobody reads. Download some of those and see if they mention derating the voltage at all. Example- https://www.nichicon.co.jp/english/products/pdfs/e-p_gui.pdf

[TimFox]

Related issue:  I learned (50 years ago) that one should not run an aluminum electrolytic capacitor at a voltage "far below" its rated voltage.
Can anyone cite a reference to quantify that recommendation?

[james_s]

Related issue:  I learned (50 years ago) that one should not run an aluminum electrolytic capacitor at a voltage "far below" its rated voltage.
Can anyone cite a reference to quantify that recommendation?

The ESR is typically higher, sometimes quite a bit higher on high voltage electrolytics. I'm not aware of any other issues besides size and cost but if there is something more to it I'd be curious too.

[Conrad Hoffman]

In that app note I referenced they say not to operate polymer caps at a very low voltage. I think over a long time the oxide film tends to adjust itself to the applied voltage but I've also heard this is far less of an issue with modern aluminum electrolytics. In my experience caps seem to have a longer life in working electronics than they do on the shelf with no voltage applied. I rarely use caps rated at less than 25 VDC because the dissipation factor tends to be higher. I had some old Sprague tests that also showed shorter life for the very low voltage parts.

[TimFox]

That's what I was taught, but I never found a good rule for how much is "a very low voltage".
In the vacuum tube days, it was normal to use a 15 to 25 V rated electrolytic to bypass the cathode resistor, but I was told not to use a 250 V capacitor in that location.

[Conrad Hoffman]

IMHO, I doubt capacitor manufacturers worry much about what their parts will be doing 20+ years down the road. Well, maybe for the military or the phone company. I use a lot of test equipment that's from the '60s or '70s and I bet the designers would have been flabbergasted if somebody said their creations would still be in use 60 years later.

[james_s]

I dunno, 20 years goes by pretty fast. I have a large amount of equipment that is older than that, when I'm designing something I certainly try to design it to last longer.

[bdunham7]

Related issue:  I learned (50 years ago) that one should not run an aluminum electrolytic capacitor at a voltage "far below" its rated voltage.
Can anyone cite a reference to quantify that recommendation?

That would generally be a problematic concept because many circuit designs use electrolytic capacitors at widely varying voltages.  Take a variable PSU that puts out 0-30V.  If you run a 3.3V device on it for a year, is that harmful?  Can a low applied voltage somehow be worse than storing it with no voltage?  Hmmmm.

[TimFox]

That's my question.  Obviously, if a capacitor is to be run over a range from, say, 5 to 100 V (in a variable power supply, for example), it needs to be rated for the maximum voltage.

[Haenk]

IMHO the "use the correct voltage"-thing may be a related to those very old electrolytic caps from the 50s/60s. I would not consider this to be an issue with modern electrolytic formulas.
As mentioned, better go with a higher temperature rating, most importantly go with a major brand and if it fits, a somewhat higher voltage. Most of the time, some higher capacity won't hurt, either.

[David Hess]

Related issue:  I learned (50 years ago) that one should not run an aluminum electrolytic capacitor at a voltage "far below" its rated voltage.
Can anyone cite a reference to quantify that recommendation?

I think that is only a concern with high voltage aluminum electrolytic capacitors because the leakage current while reforming at high voltage will overheat the capacitor.

The ESR is typically higher, sometimes quite a bit higher on high voltage electrolytics. I'm not aware of any other issues besides size and cost but if there is something more to it I'd be curious too.

Electrolytic capacitor ESR and dissipation drop up to 100 volts, and then jump up and increase at 160 volts and higher.  I suspect the jump and increase is caused by using a different electrolyte to support a higher voltage.

[RoGeorge]

Related issue:  I learned (50 years ago) that one should not run an aluminum electrolytic capacitor at a voltage "far below" its rated voltage.

Apart from what was already pointed, it might be also related with forming the Al electrolytic capacitors (slowly growing the oxide layers by applying a very small current up to the desired nominal voltage of the final capacitor), then the oxide layers slowly dissolve when no voltage (or lower voltage) is applied.  Therefore, a capacitor will de-rate itself with time if used at a lower voltage (or stored discharged) for very long periods of times (years/decades).

No idea how much of this was true, or if it is still applicable to the current technology of electrolytic capacitors.

[TimFox]

I looked in vain to a quantitative answer to my question.
A 22-page application note from Cornell-Dubilier  https://www.cde.com/resources/technical-papers/AEappGuide.pdf  discussed almost every other question, including definition of maximum voltage ratings, temperature derating, construction techniques, re-forming, and anti-series non-polarized connections, but not my question.

[Conrad Hoffman]

Well, how old are you? Do you have time for a 50 year duration experiment? 

[globoy]

Thanks for the input everyone!  There is one note in the guide Conrad linked indicating to make sure ripple + constant DC doesn't exceed the max working voltage.  Doesn't appear so in my application but that makes sense.

Regarding part type I should have specified the parts are SMT.  Definitely more TH parts available.

[jonpaul]

Operating temp far more important, the cap rating is 85C (worst) 105 C better, 135 C best,


This means at those ambient tempos and at rated ripple current, the cap looses 50% or capacity and esr may increase 20..50%after 2000 hrs.

In hot industrial we used 105 and 135C caps.

Jon

[Conrad Hoffman]

That reminds me, watch out for very low temperatures because the capacitance goes down and might not be enough for the circuit to function correctly and/or cause something else to fail.

[james_s]

Operating temp far more important, the cap rating is 85C (worst) 105 C better, 135 C best,


This means at those ambient tempos and at rated ripple current, the cap looses 50% or capacity and esr may increase 20..50%after 2000 hrs.

In hot industrial we used 105 and 135C caps.

Jon

I typically use 105C capacitors when I repair things.

That makes me wonder though, is there any disadvantage to high temperature capacitors other than cost?

[TimNJ]

I like the attached PDF published by Jianghai. Also check this: https://www.chemi-con.co.jp/en/faq/detail.php?id=29BDBH5#Anchor-3

2　Applying Voltage Effect on Lifetime

Where a capacitor is used at lower than the rated voltage, the lifetime may not be adversely affected, which means that the effect of the applying voltage is negligibly small, while the effect of the ambient temperature and heat generation due to ripple current is significant.

Looks like a 100V cap was tested at 35V, 50V, 60V, 80V, and 100V...and the loss tangent and capacitance change over the test time is essentially identical. (They do not actually note if it's a 100V rated cap, just guessing.)


Standard lifetime model:

L = L0 * Kt * Kr * Kv

where L = predicted lifetime, L0 = datasheet lifetime @ Irms/Tmax, Kt = temperature factor, Kr = ripple current factor, and Kv = voltage factor

More or less, the attached PDF agrees with the above sentiment that applied voltage is not so important (for small sized radial caps). For small sized caps, they suggest Kv = 1.0, suggesting there is no improvement to lifetime with significant voltage derating.

So, why derate at all? Well, I'd say the above assumes you are using a reputable manufacturer with good process control and test methods. Maybe for a lower tier manufacturer, the above does not hold.

But more generally, derating is just a nice fudge factor to make sure that your design doesn't blow up when it inevitably runs into some corner case that you didn't think to test. A little peace of mind?

In a super well controlled electrical environment, not subject to transients, no changing load conditions, etc. maybe you're fine running it at 100%.

[coppice]

Nothing scientific to base it on, but I think heat is a bigger factor.

Huh? There is plenty of science to base it on. The specs for capacitors include a derating curve for lifetime vs temperature, and it follows well established equations. The only thing you really need the curves for is to instantiate the values in the equation.

The derating as you operate a capacitor near its rated voltage is vaguer. Most sets of design rules for high reliability call for capacitors to be used well below their rated voltage. However, this is a pretty lose way of setting a rule, as the way different manufacturers specify their capacitor voltage rating, especially for high voltage capacitors, varies a lot.

[Conrad Hoffman]

Read what I was thinking, not what I wrote! I meant that I've no references about running at maximum voltage, but am guessing heat is a (well known) and likely larger factor.

[jonpaul]

James S: See The Arrhenius equation

https://en.wikipedia.org/wiki/Arrhenius_equation

It  applies to Lytic capacitors, as chemical reactions double in speed with every 10 deg C increase in temp.
The temp rating depends on: ESR, ripple current rating, seals construction, seal materials, vent effectiveness, etc.

Definition of the temp rating is at that ambient temp (on capacitor surface) and rated ripple current, the internal temp rise will cause a 50% degradation in C or ESR (forget exactly) after 2000 hrs at that temp and Iripple.

Disadvantages of 105 and 135 C caps are indeed cost and availability as 99% of consumer/cheap/Chinese are 85 C or less.
The small volume and market for Industrial and Ballast grade means tiny volume of prod and long lead time.
Ballast caps are made very limited sizes and capacity/V as intended for 120/240 V A lighting eg HID/LED/CFL ballasts.

One example we used was in our 500 kHz resonant 70W HID ballast of 1990, a 68 uF 450V cap Nichicon, dim 25 mm dia 25 mm H
Refer to the spec sheets and app notes of Rubycon, Nichicon, and other high quality electrolytic cap manufacturers for more info.
As for voltage rating, indeed lyitcs used at far below rating may have degraded capacity over time. (forming)
In the example, the 450V cap rating was on a 360V rectified line bus, so derated ~ 60%.

Bon Soiree,


Jon

[wraper]

Disadvantages of 105 and 135 C caps are indeed cost and availability as 99% of consumer/cheap/Chinese are 85 C or less.

What? 85^oC caps are becoming extinct in small sizes, very small amount is used similar to dip IC packages. Only large can and high voltage capacitors are still dominant in 85^oC flavor. Not to say temperature rating in not a direct indication of quality, there are plenty of super cheap 105^oC capacitors of questionable origin. 125^oC are usually specialty capacitors for bulbs. And in most cases they are cheap caps from 2nd-3rd tier manufacturers.

85 C or less.

Only a large can specialty capacitors are available in lower temperature rating. And they are not cheap at all. Like these https://eu.mouser.com/ProductDetail/EPCOS-TDK/B43415C9108A000?qs=Aj7me2ELROJfDmc0Z1koLg%3D%3D