"non-polar" capacitor?

Chuck Guzis cclist at sydex.com
Mon Aug 1 11:56:23 CDT 2016


On 08/01/2016 04:08 AM, Dale H. Cook wrote:

> 1) When there is no constant DC offset to the applied voltage - if
> the voltage applied to an electrolytic capacitor reverses the
> capacitor can be damaged. When there is no constant DC offset the
> second reason can come into play:
> 
> 2) When a large value non-polarized electrolytic can, as previously
> noted, cost less than a large film, oil-filled, or other non-polar
> capacitor.
> 
> Some of us still deal daily with analog circuitry at work.

There's one other aspect that I believe is germane to this particular
application:  Linearity.

Aluminum electrolytic capacitors are very curious beasts--they're not
all capacitance, but rather operate as several different "pure"
properties hooked in series.  Most people here know about ESR, but
there's another one--wet aluminum dielectrics also function as if they
had a leaky diode in series with the series resistance and capacitance.

This is a result of the way they're constructed--basically two aluminum
foil strips, separated by a membrane (often paper) saturated with an
electrolyte (commonly borax).  A DC current is applied to "form" the
capacitor and establish its polarity.

Probably nobody here is old enough to remember liquid rectifiers, but my
own father related them to me (he grew up in a poor family during the
teens and 20s).  The idea was that you took a plate of aluminum and a
plate of tin, lead or other metal and immersed them in a solution of
sodium bicarbonate, borax or some other electrolyte (dad used ammonium
sulfate, a waste product of the nearby steel mills' coke ovens, and ran
an AC current through them.  Eventually, the device polarized and formed
a hydroxide/oxide layer on the aluminum and developed a preference to
current direction.  The sparkbangbuzz website has some interesting
observations on this animal.

For use in audio circuits, this "series diode" effect can lead to
distortion, hence the use of a non-polar device.  Take an aluminum
electrolytic and put it on your curve tracer and the effect is quite
visible.  For what it's worth, I don't know if tantalum caps exhibit the
same behavior, but I suspect that they do.

--Chuck



More information about the cctalk mailing list