Capacitors---changing to higher voltage

[Elf Six]

Hi all.

So, you can swap caps with higher voltage ones(not thinking about ESR etc.)

But, are there any advantages to swapping to a higher voltage cap? Is it maybe better produced,
will it live longer, will it stay in spec for an extended time compared to its lower-spec counterpart?

Thanks!

[uwezi]

All of your questions depend on which type of capacitor we are talking about...

Higher voltage electrolytic capacitors for example will show significantly higher ESR - just because this parameter is not as critical for a high-voltage cap as for a low voltage one. Also the parasitic inductance of the high-voltage part might be worse.

Other than that it is - in my view and experience - always a good idea to be on the safe side, if board size, budget and availability allow.

[Elf Six]

Hi

Just your standard electrolytic capacitor. I mainly use Rubycon, Panasonic and Nichicon.

I have plenty space on my boards, and I have the caps already. So you say that I should use them

Thanks!

[T3sl4co1l]

Electrolytics are generally chosen for the next voltage rating higher than operating.  A 12V supply using 16, maybe 25V caps; a rectified 120VAC line (= ~160VDC) using 200V caps; etc.

It's not generally recommended to use vastly higher ratings (e.g., 100V rating for 25VDC) because the capacitor degrades over time, resulting in off-spec value, ESR and such.

Electrolytics degrade over time because the dielectric very slowly breaks down in the electrolyte; effectively, its voltage rating decays over time (years to decades).  This process is reversed by leakage current, which causes an electrochemical reaction that reforms the dielectric.  After enough time, the effective voltage rating drops to the voltage it's being operated at.  If this voltage is a lot lower than normal, the thickness of the dielectric, electrolyte and electrodes will all be different from spec, giving different C and ESR values, probably out of spec.

In contrast, you generally choose tantalum capacitors for several times nominal operating voltage, for safety reasons; the same goes for ceramic caps, but for stability reasons; aluminum polymer and film capacitors are fine at ratings, and also generally expensive or bulky enough that you don't want to get vastly overrated parts anyway.

Tim

[TheRuler8510]

Higher voltage electrolytic capacitors for example will show significantly higher ESR - just because this parameter is not as critical for a high-voltage cap as for a low voltage one.

Correction - Higher voltage electrolytic caps have LOWER ESR.

[TheRuler8510]

It's not generally recommended to use vastly higher ratings (e.g., 100V rating for 25VDC) because the capacitor degrades over time, resulting in off-spec value, ESR and such.

Tim

My understanding of this situation is that, higher voltage caps will have lower ESR, and hence lead to potentially higher ripple current, which if too high will degrade your cap life. However, if you analyze it and use the right components, that are in-spec, you should be OK.

Also, caps that are used verses caps that are not used (i.e. in storage) will have a higher life, analogous to the low voltage use question--however this is over DECADES of time that you will notice it. I have some unused caps that are 30 years old and measure fine, although I would not use them for anything new.

Of course if you use el-cheapo caps, all bets and rules do not apply.

[rexxar]

In contrast, you generally choose tantalum capacitors for several times nominal operating voltage, for safety reasons; the same goes for ceramic caps, but for stability reasons; aluminum polymer and film capacitors are fine at ratings, and also generally expensive or bulky enough that you don't want to get vastly overrated parts anyway.

I've never understood this. Why are tantalums (for example) rated at 25V when they're only safe to use at 6V or what have you? Surely it would make more sense for the rating to reflect safe operating range, instead of engineers having to derate them so much.

[mariush]

Why do we have cars capable of 200 miles per hour when it's only safe to drive at 80-120 miles an hour?

[rexxar]

Why do we have cars capable of 200 miles per hour when it's only safe to drive at 80-120 miles an hour?

Well, your car doesn't explode when it gets up to 200MPH, it just hits something and launches you out the windshield.

and then it explodes

[mariush]

Your car doesn't, but your tires can. Some are only rated for 120 miles per hour, some even lower speeds.

But we're getting really off topic now.

[T3sl4co1l]

The answer however is the same in both cases.  You can... that doesn't mean you want to

In fact, they are manufactured with ratings 2-3 times higher still: enough damage occurs during manufacture to derate them that much already.  Right out of the box, they're probably okay at that rating, but handling and soldering causes more stress, plus the rigors of operation, environment and age.  Also, peak currents are a no-no, especially first time application of an incrementally higher voltage (they reform / self-heal to some extent, like electrolytic and film types do, the problem is if too much tries to "self heal" at once, it erupts).

Tim

[CaptnYellowShirt]

Why do we have cars capable of 200 miles per hour when it's only safe to drive at 80-120 miles an hour?

Erectile dysfunction.

[TheRuler8510]

Electrolytics are generally chosen for the next voltage rating higher than operating.  A 12V supply using 16, maybe 25V caps; a rectified 120VAC line (= ~160VDC) using 200V caps; etc.

It's not generally recommended to use vastly higher ratings (e.g., 100V rating for 25VDC) because the capacitor degrades over time, resulting in off-spec value, ESR and such.

Electrolytics degrade over time because the dielectric very slowly breaks down in the electrolyte; effectively, its voltage rating decays over time (years to decades).  This process is reversed by leakage current, which causes an electrochemical reaction that reforms the dielectric.  After enough time, the effective voltage rating drops to the voltage it's being operated at.  If this voltage is a lot lower than normal, the thickness of the dielectric, electrolyte and electrodes will all be different from spec, giving different C and ESR values, probably out of spec.

Tim

Tim/T3sl4co1l,

I defer to your expertise on this.  A power supply guy told me that Electrolytics have "memory," that is, for instance a 100V cap operated for a LONG time at 10V would be a problem if you moved to 50V, due to this "memory" effect. He could not quantify the "time" or voltages limits, or what would happen when operated higher.

It sounds believable due the chemistry, but why can't I find anything like that in the datasheets or an App Note that attempts to quantify this? Do you know of any references?

Regards,

[vk6zgo]

A lot of the Electrolytics used in various equipment are not highly critical---a particular type  & voltage rating may be used
because "there were a lot of them in the store"!

I've never,over many years seen any problem with using capacitors of higher voltage rating as replacements.
If it made that much difference,repair companies would have needed a much larger inventory of such parts.

High ESR in critical positions will cook in short order,but again,most applications are not critical.

Circuits which rely upon the use of a capacitor with poor ESR characteristics are examples of poor design,in my opinion.

[u666sa]

I repair power supplies often, computer, TV, and I often have to swap a cap from let's say 16 volts to 24 volts, usually with capacity around 480 uf.  Also, often we don't have needed capacitance in our local store, so I go for higher volts and somewhere around the needed capacitance.  Works.   

[T3sl4co1l]

Oddly, I didn't see any manufacturers' articles on a real quick search.

This looks handy,
http://digital.library.unt.edu/ark:/67531/metadc3104/m2/1/high_res_d/thesis.pdf
Skimming, I didn't notice a section on leakage in the 'reforming' voltage range.  It should look sort of like avalanche breakdown with a lot of series resistance, so that leakage goes up sharply at some voltage, and that voltage threshold rises as charge is delivered, so that after reforming, it does not break down at voltages lower than the reforming voltage, and the leakage current is the standard leakage current alone.

It's entirely possible, though I think unlikely, that newer formulations don't age much, or at all.  But they haven't been around for the decades that it can take to become apparent, so...

Tim