Replacing Electrolytic Caps - Putting Two Caps in Series?

[Stovemananon]

Hi all,

I have a very simple question.

I need to replace a 100v 10uf capacitor.

Can I safely use two 22uf 50 v capacitors in series?

Will this total 11uf at 100v?

Thank you!

Edit: I did this on a repair recently and the amp seems to function great. Will there be issues down the road?

[martin manning]

Yes, correct, but you should add balance resistors across the caps to insure that they share voltage equally, like this: https://www.vishay.com/docs/48296/_did- ... 5-1709.pdf
From that article Rmax = (2*Vmax - V)/Ileak. If you assume that the caps are run at rated voltage (V = Vmax), and Ileak is 0.04CV (typical data sheet value for 50V cap), then Rmax = 1/(0.04C), or 1.14 Meg. A 1M resistor across each cap should do it.

[Stovemananon]

Yes, correct, but you should add balance resistors across the caps to insure that they share voltage equally, like this: https://www.vishay.com/docs/48296/_did- ... 5-1709.pdf
From that article Rmax = (2*Vmax - V)/Ileak. If you assume that the caps are run at rated voltage (V = Vmax), and Ileak is 0.04CV (typical data sheet value for 50V cap), then Rmax = 1/(0.04C), or 1.14 Meg. A 1M resistor across each cap should do it.

Thank you I appreciate this!

Naïve question here: Why is there a leak value? Is this for DC? Aren't caps supposed to "stop" DC from passing?

[pdf64]

…Why is there a leak value? Is this for DC? Aren't caps supposed to "stop" DC from passing?

You need to clarify the application you’re asking about.
Normally ecaps are used for DC power supply smoothing, but we shouldn’t assume that.
eg 10uF could be a coupling cap.

[R.G.]

Naïve question here: Why is there a leak value? Is this for DC? Aren't caps supposed to "stop" DC from passing?

This is one of those things where the real world is filled with asterisks and footnotes. Ideal caps don't leak, but real-world caps always do.

It's all about the kind of insulation inside the caps. Electrolytic (polarized, + and - ) capacitors have high capacitance values and low voltage ratings because of their unique construction. The insulator in aluminum electrolytic caps, for instance, is a layer of aluminum oxide grown on the surface of the aluminum plates. This layer is - literally! - molecules thin. Since the thinner the insulator, the higher the capacitance, the electrolytically-grown oxide layers make for high capacitance.
The problem is that they're not nearly as good an insulator as a thicker layer of plastic film or ceramic. They leak.

All caps leak some, but for most kinds of cap and types of circuits, it's so small you can ignore it. This is not true for aluminum electrolytic caps specifically. They leak. You get high capacitance, but you have to deal with the leakage.

[Stevem]

R G could you also comment on why it’s best for a electrolytic cap to be run up to within. 20% of its rated voltage?

[roberto]

I usually consider 10% voltage margin, not 20%, due to voltage supply variations.
20% is indeed more conservative for places at the end of a supply line in rural areas.

[R.G.]

R G could you also comment on why it’s best for a electrolytic cap to be run up to within. 20% of its rated voltage?

Sure.
Running an electro at a whole lot below its rating makes it prone to thinner points in the insulating oxide failing, as well as capacitance drift. Keeping it up near the specified voltage maintains the insulating foil layer better.

The insulating oxide is grown on the aluminum foil plates by passing a current through the plate and a water based bath. This is often in the form of a voltage source a bit higher than the desired voltage rating run through a current limiting resistance or current limiter. At low currents, the oxide grows in thickness. At high currents, you get hot spots and damage to oxide already formed. The degree of forming is watched by looking at how much voltage is dropped across the current limiter versus the oxide. Once the voltage across the limiter drops enough, the voltage is being resisted by the oxide layer, and the layer can now withstand the desired voltage. Caps are generally formed to a bit higher than rated voltage for a bit of surge protection and some shelf life.
When unused on a shelf or run at a lower voltage, the oxide slowly un-grows, dissolving out into the electrolyte until the applied voltage re-grows it back into place. Running it at a lower voltage lets the oxide un-grow down to that maintained by the lower voltage. Shelf life has no re-grow ability at all. The initial oxide forming is done under conditions where thinner spots are subject to more growth to help prevent thin spots. Un-growth doesn't necessarily do this, so you can have spot failures when a higher voltage pushes damaging currents into a thin spot.

Using the cap at 80% or more of its rated voltage is a way to keep it near its nominal rated voltage and capacitance as long as possible.

The normal variation of line voltage makes this harder to do. So using the cap near but not exactly at its rating is a good compromise. It's a juggling act.

[martin manning]

Using the cap at 80% or more of its rated voltage is a way to keep it near its nominal rated voltage and capacitance as long as possible.

If a cap is installed in a device and it operates at say 50% of its rated voltage, it won't ever need to operate at its rated voltage. Is there any problem with that, other than it might drift a bit in value, and be larger and more expensive than a cap that is designed for the lower voltage?

[Stevem]

Thanks R G!

[R.G.]

Using the cap at 80% or more of its rated voltage is a way to keep it near its nominal rated voltage and capacitance as long as possible.

If a cap is installed in a device and it operates at say 50% of its rated voltage, it won't ever need to operate at its rated voltage. Is there any problem with that, other than it might drift a bit in value, and be larger and more expensive than a cap that is designed for the lower voltage?

Good question, and one I've pondered several times. I've looked and never come up with a definitive answer from a cap manufacturer or industrial chemist, etc. I have several theories, but no clear "here's the answer".
The obvious ones are what you note: using a cap at much less than its rated voltage means you are paying more and the cap is bigger; and the capacitance drifts with thinning. As a green power supply engineerling I was pounded upon soundly for specifying too much capacitor, so I still twitch when that's mentioned, but I try to control it.

I have found hints, but no definite facts about differences in the electrochemistry of forming the oxides for new capacitors and regrowth in existing caps. I have found descriptions of electro caps made by winding up the aluminum foils and the porous spacer, inserting into the can, and then filling with electrolyte. After that, the terminals were fitted and the can put on a forming rig to form up the oxides. But I think that these days the foils are oxide coated in a forming bath as a continuous roll to roll process, and then wound with plain foil and spacer to be inserted into cans. That makes some sense, as it's possible that an optimized bath for continuous wet forming might not have the same chemistry as you want for long-term stability and performance inside a sealed capacitor can. You'd want the forming bath to be more fluid, for one thing; high viscosity in the final can is a good thing. Maybe a different combination of sodiumhydraperfluoroborates in the can would give better heat resistance or longer life. It makes sense to me, but no solid data.

The import is that the stuff in the can may not be the best stuff for repair and regrowth.

However, I have noticed that the MBAs have been after the capacitor makers, too. I could not find any 1uF 25V electros or 1uF NP 25V electros last time I tried. They were only stocked in 50V. Also, it's common enough for tube power supplies to use two 350V caps stacked for the B+. That gets you to a nominal 350 used at about 225-250V. Seems to work. I think that for voltages much over 500, the yield for aluminum oxide electros goes to cr@p. There's some kind of limit there, but I don't know if it's fundamental in aluminum oxide or just an MBA difference.

I wish I had a more definitive answer for you. I personally fall back on the engineer's mantra: get something working with the parts you have or can get, and do the perfect thing when you can.

[martin manning]

Thanks. It seems to happen with some regularity, as in the examples you gave. Just looking to get some reassurance that there aren't any serious consequences.

[roberto]

@R.G.
do you have some literature about this phenomenon?

I'm interested into it. I remember something like 25 years ago a discussion on a forum about it, but never some papers.
Thanks in advance.

Roberto

[R.G.]

@Roberto
A fair amount of it I picked up pre-internet, as I was designing power supplies for a living.

I did a quick search and came up with https://www.nmr.mgh.harvard.edu/~reese/ ... s/tec1.pdf, which adds some modern info from Nichicon. Looks like they're firmly in the camp of forming the oxide layer first (section 1-5) and the forming bath is water and borate based. The anode and cathode foils are slit to width, wound up with a porous separator, put in the can and electrolyte added.

Nippon Chem-Con https://www.chemi-con.co.jp/en/faq/deta ... id=29BDBH5 says that using them at less than rated voltage has very little effect on lifetime.
Here's a (dense!) overview: https://www.ijrte.org/wp-content/upload ... 078219.pdf.
I found this fascinating, but it's very much old news, as it's from 1962: https://conradhoffman.com/papers_lib/Sprague_62_7.pdf
This one had a bit on the different forming and functioning electrolytes.https://conradhoffman.com/papers_lib/Sp ... _64_11.pdf It reinforces the two-different-electrolytes theory between forming and use. Again, old news since it's from 1964.

So it looks like there's not much if any lifetime penalty for using electros below their rated voltage. Lifetime is hugely affected by the temperature of the inside of the cap. High temps force electrolyte vapor out the vents and capacitance goes down while ESR goes up. There didn't seem to be much info in this recent pass on the topic of geriatric capacitors. Caps that have been just sitting without being powered do go bad; I saw a lot of references to shelf life, but these seemed to be hugely short - they were talking about a couple of years, and 1000 hours (that's 41 days!!!) in one case, but surely that's not a general guideline.

[roberto]

@R.G.
Thank you very much!