Index Home About Blog
From: "Barry L. Ornitz" <>
Subject: Re: Electrolytic capacitor question...please help.
Date: 05 Jan 1999

John Lundgren wrote:

>I don't think either one is 'absolutely WRONG', because as you said
>there is an electrolyte liquid and over time it will have some chemical
>action with the oxide, or else it will dry out.  But I believe that the
>purity of chemicals has come a long way and the manufacturing processes
>have improved so that the effects of this may be minimal over a long
>time, my guess would be a decade or more.  So I don't think it is, or
>should be considered, a concern.  Many have recommended that
>electrolytics be replaced periodically, usually 5 or 10 years.  It is
>probably prudent.

The aluminum purity issue was a problem in the 1940's.

The electrolyte will not dry out under normal conditions of temperature and
humidity.  The sugars, glycol, etc. are hydroscopic and will pull water out
of the surrounding air if the humidity is high enough.  Electrolytics are
probably the only electronic component that are damaged by storage under
very dry conditions.  Electrolytics are also affected by low temperatures.
If the electrolyte freezes (it must be pretty cold since the glycols and
such depress the freezing point), the capacitor quits working.  The reason
that electrolytics are worthless at high frequencies is mainly due to the
ion mobility of the electrolyte solution.  By a MHz or so, the ions cannot
move fast enough for the solution to be conductive.

> But whatever happened to the old Western Electric
>caps, the ones in the gray cases?  They used to be considered the best,
>probably because the telcos needed to have equipment that would last a
>couple decades.  The audiophiles used to claim the WE caps were the only
>way to go.  Well, what did they do to get them to last that long?

I am not aware that WE used many electrolytic caps in regular exchanges.
Electrolytic capacitors have no place in the high fidelity audio chain
anyway.  They are nonlinear and create distortion.  Their only real purpose
is as energy storage devices in power supplies.  They can also be used in
bypass applications were the effect of their nonlinearity is unimportant.

>If you ask someone that worked in the central office about their huge
>batteries, they would tell you that these lasted 30 years.  I guess the
>chemicals were very pure and the mfg'g process was well done.  Compared
>to the usual car battery, they were a lot longer lasting.  I think the
>same is true of today's electrolytics.  They last a lot longer and don't
>have problems as much as older ones did.

Many of the exceptionally long-life batteries were nickel-iron batteries
rather than lead-lead oxide batteries.  However getting long life from a
lead-lead oxide battery is not difficult.  Automobile batteries have a
limited life because of the size and weight requirements, and because of
the vibration they are subjected to.  Telco batteries also tend to be very
well maintained, and are much bigger than automobile batteries for the same
amp-hour rating.

>Any one remember one of the Reagan speeches where an electrolytic went
>bad in the microphone preamp/distribution amp?  They had no audio on
>_any_ network for several minutes before someone found and fixed the
>problem.  Shows what can happen when a simple part like a capacitor

Reagan should have taken this as an omen.  Remember when a science fiction
writer in the 1940's predicted that we would eventually elect an actor to
the presidency and was completely shunned because the public felt science
fiction should at least have some basis in fact or might possibly come
true.  The poor writer lived to see Reagan as governor but died before he
became president.  I wish I could remember the writer's name.

        Dr. Barry L. Ornitz     WA4VZQ

From: "Barry L. Ornitz" <>
Subject: Re: Electrolytic capacitor question...please help.
Date: 04 Jan 1999

Bill wrote in message <>...
>Rich Nelson wrote:
>> This is absolutely WRONG.  An electrolytic cap works because the the
>> voltage applied across the electrolyte forms the dielectric film.
>> Using a cap much below its rated voltage will not maintain the
>> dielectric, and the value of the cap will decrease over time.  And the
>> puncture voltage of the dielectric will also decrease - making the cap
>> more likely to fail!
>This is absolutely WRONG. It used to be true, but operating modern
>electrolytics at below rated voltages will not decrease the value of the
>capacitor over time.
>To quote, from an article in Radio-Electronics for February
>1970, by Richard R. Marsh of Cornell-Dubilier Electronics:
{Text deleted.}

This exact same thing was said by them in 1952 and they were lying then
too... or perhaps "stretching" the truth might be the polite way of saying

The effect is slow, but it does exist - and basically there is nothing that
can be done about it.  It occurs because of the basic chemistry of an
electrolytic capacitor in the formation of an anodic oxide film.
The negative electrode of an electrolytic capacitor is the electrolyte -
which is NOT dry.  It must remain moist for the capacitor to operate.
Without the polarizing voltage on the capacitor, the electrolyte will
slowly hydrolyze the aluminum oxide film on the positive terminal or anode.
It will do so until the voltage rating of the capacitor eventually reaches
the potential at which the capacitor operates.  The process is slow, but
measurable.  Various hydroscopic materials such as sugars, glycols, and
inorganic salts are added to the electrolyte paste in the so-called "dry"
electrolytics to maintain the moisture content.  As long as a polarizing
voltage is applied, the anodic oxidation reaction maintains the oxide film.
Even in new capacitors that have been stored for several months, when a
voltage is first applied the leakage current through the capacitor is much
higher than normal.  This quickly diminishes as the oxide film reforms.

When you take away the electrolyte, the film does not hydrolyze.  But
likewise it can no longer repair itself.  Completely dry electrolytic
capacitors have been made, where the anode foil is removed from the
electrolyte bath after the oxide film is formed.  The oxide layer is then
sputtered with aluminum to produce the other electrode.  However, in the
remainder of the manufacturing process, usually during the rolling, the
oxide film is damaged and the capacitor fails prematurely.  Such capacitors
have never been marketed successfully.

There is usually no problem when using a capacitor originally rated for a
higher voltage in a low voltage circuit.  However, you do pay a penalty
both in price and in internal resistance when you do this.  The effect on
price versus voltage is high, but the added internal impedance of high
voltage electrolytics is generally not a problem except at high frequencies
where the internal inductance of the electrolytic capacitor is already
making the impedance increase.

For those that want to learn more about the chemistry of electrolytic
capacitors, I suggest the book below.  I will warn the readers that this is
a book written more for metallurgists and material scientists than for
electrical engineers (much less those who do electronics as a hobby).

By the way, the theory predicts that an electrolytic capacitor, once formed
at a particular voltage, will increase in value as its anodic film
hydrolyzes.  For bypass applications this is fine, but this effect is
partially why electrolytic capacitors are never used in critical frequency
determining or timing applications.  In practice, the change in capacitance
is not easy to predict since only the most reactive portions of the oxide
film degrade.  What many people attribute to loss of capacitance is really
a large increase in internal impedance.

        Dr. Barry L. Ornitz     WA4VZQ

AUTHOR: Young, Lawrence.
TITLE: Anodic oxide films.
PUBLISHED: London, New York, Academic Press
DESCRIPTION: 377 p. illus. 24 cm.
SUBJECT: Metallic films.
NOTE: Includes bibliography.

The following book has not been released yet, but it should have more
recent information:

 AUTHOR: Conway, B. E.
 TITLE: Electrochemical supercapacitors : scientific fundamentals and
technological applications
PUBLISHED: New York : Plenum Press, 1999.

SUBJECT: Storage batteries.
SUBJECT: Electrolytic capacitors.
SUBJECT: Electric double layer.

NOTE: Includes bibliographical references and index.

From: "Barry L. Ornitz" <>
Subject: Re: Finding right voltage capacitor
Date: 23 May 1999

Bill Smith wrote in message
>Voltage tolerances for most electrolytics are +/- 20 %.   1.2 x 450 VDC =
>540 volts.
>I don't think you will have any difficulty using the 450 volt capacitor
>directly in your circuit.

This is very bad advice.  In fact, it is downright dangerous.  The surge
voltage for a 450 volt capacitor is generally only 500 volts and the surges
must only be applied intermittently with a series current limiting

Normally you use a capacitor at LESS than its voltage rating.  In fact, it
is impossible to make aluminum electrolytics over approximately 500 volts
(just as it is impossible to make tantalum electrolytics over about 75
volts).  Commercial electrolytics of up to 800 volts were once sold, but
there were two series units internally.

It is possible to reform a capacitor at a slightly higher voltage rating,
but you are asking for trouble.  Not only will you have to run the voltage
up slowly, you will use up some of the internal electrolyte in the
reforming process.  The capacitance will also drop as the dielectric layer

You can run an electrolytic capacitor at any voltage below its rating
safely, but if you go significantly below its rating and operate at this
voltage for a long time (year or more), the capacitor will gradually lose
its original rating and approach the minimum voltage rating needed for
operation is the circuit.  This is because of hydrolysis of the anodic
oxide film.  Contrary to a few advertising claims by Mallory and others in
the early 1950's, this process is fundamental to the chemistry of
electrolytic capacitors.  Most capacitor manufacturers suggest that new
capacitors have to sit for six months in storage, they be reformed before
being installed in a circuit.

People should quit thinking of electrolytic capacitors as just polarized
capacitors.  Actually they are electrochemical cells somewhat akin to
batteries.  Once viewed this way, the negative aspects of these capacitors
is much easier to explain.

In the application discussed, two series connected 300 volt capacitors
would be a good choice.

        73,  Barry L. Ornitz     WA4VZQ

From: "Barry L. Ornitz" <>
Subject: Re: Help: Strange "electrolytic"???? **SOLVED** [Non-Polar 
Date: 10 Dec 1998

Jeroen Stessen wrote in message
>If I were you then I would make sure that it is permitted to put
>a constant DC voltage over these bipolar electrolytic caps,
>because they are made for applications in which there is usually
>no DC voltage component, particularly in loudspeaker crossover
>filters. A constant DC might ruin your caps in short time !!

Actually non-polar electrolytic capacitors are designed for circuitry where
the polarity might _occasionally_ be switched.  They handle DC just fine
when kept within their ratings.  Unlike regular electrolytic capacitors,
however, the DC field does not keep the anodic film "formed".  Their
application in AC circuits is not considered good design, especially in
circuits like crossovers.  They are used in these circuits only for their
small size and low cost.

     73,  Barry L. Ornitz     WA4VZQ

From: "Barry L. Ornitz" <>
Subject: Re: What is the best capacitor?
Date: 17 Jun 1999

Anthony New wrote in message <>...
>> Christopher R. Carlen wrote in message
>> <>...
>> >I have learned, thanks to this group, that Al electrolytic caps don't
>> >have low impedance at high frequencies, due to ESR.

>It's not the ESR that's the problem, it's the inductance. That's why you
>need to parallel different values of capacitance for decoupling. In fact
>in some applications some ESR is desirable - high-valued ceramic
>decoupling capacitors with very low esr can form a tuned circuit with
>the inductance of parallel electrolytic capacitors, even tantalum, and
>resonate with hf supply current ripple.

It is possible to wind the foil in an electrolytic capacitor such that the
inductance is minimized.  But the capacitor will still perform poorly at
higher frequencies (above 100 kHz).  This is because an electrolytic
capacitor has the electrolyte as its cathode.  The anode is the oxide
coated foil.  The foil connected to the negative lead serves only to
provide an ohmic connection to the electrolyte.

At high frequencies, the diffusion of ions in the electrolyte is too slow
to travel the small distances in the exceptionally short times involved.
The result is an increase in impedance as the frequency increases.  This is
normally considered ESR when describing the electrical analogy.  This
effect is generally larger than inductance effects, so manufacturers rarely
try to wind the low inductance units.

    Barry L. Ornitz     WA4VZQ     ornitz@dpnet./net

From: "Barry Ornitz" <>
Subject: ??? Re: Non-Polarized Electrolytic Capacitors
Date: Sun, 19 Nov 2000 22:02:41 -0500

"Stonecatv" <> wrote in message
> Does anyone know of a good source for non-polarized capacitors?  I am looking
> to replace a resistive line cord.

While the poster is actually asking about AC capacitors, his subject line
is deceptive with the word "electrolytic" in it.  I am not sure anyone else
has caught this slip-up.

Non-polar electrolytics are _NOT_ suitable for use in this application.
Non-polar electrolytic capacitors are only suitable for use in DC circuits
where the polarity might change occasionally.  They may be used in AC
circuits for very short durations - but at the expense of a shortened life.

Electrically non-polar electrolytics are two conventional electrolytic
capacitors in series with the cathodes connected together.  In actual
construction, they are composed of two pre-anodized aluminum foils wound
together.  Unlike a conventional electrolytic capacitor, the applied
voltage does not serve to keep the anodic film properly formed.
So their life is automatically shorter than conventional electrolytics.

To quote from CDE's (Cornell Dubilier) application note (with my emphasis
added in capital letters):

    "While non-polar aluminum electrolytics are available for
MOMENTARY-duty AC applications like motor-starting and voltage-reversing
applications, the high dissipation factor of aluminum electrolytic
capacitors - from 2 percent to 150 percent - causes excess heating and
short life in most AC applications."

In the application where you use a capacitor for its AC reactance to
replace a resistive voltage dropping line cord, what you need are modern
dry-film AC capacitors, or oil-filled motor RUN capacitors.  These are
designed for continuous duty AC service.

Note that motor START capacitors are only for intermittent service since
they are non-polar electrolytics.  Occasionally you will find non-polar
electrolytic capacitors used in some small bi-phase motor applications like
antenna rotators and garage door openers.
These are used because of cost since only brief usage occurs in normal

Motor run capacitors can usually be purchased at any large electrical
distributor, motor shop, or air conditioning supply.  Grainger is a common

         Dr. Barry L. Ornitz     WA4VZQ

From: "Barry Ornitz" <>
Subject: Re: ??? Re: Non-Polarized Electrolytic Capacitors
Date: Mon, 20 Nov 2000 22:39:25 -0500

"Sylvain" <> wrote in message
> Errr... sorry Doc, what have I been using for the last 20 years in those
> speakers cross-overs, and countless other speakers makers too ?
> I think they are non-polar electrolytics...

While I loathe audiophools who search for every possible non-scientific
tweak for the ultimate sound, using proper capacitors in cross-over
networks will make a measurable difference.

Electrolytic capacitors are chemically more akin to batteries than to true
capacitors.  Their response is quite nonlinear.  This is easily measured
and many articles have been published on the subject.  Adding a large DC
biasing voltage to a regular electrolytic will improve its AC linearity.

Try using film or oil-filled paper capacitors in the cross-overs.  They
will sound much better.

By the way, electrolytic capacitors act as a conducting diode in reverse
polarity.  The old electrolytic detectors used by early radio experimenters
operated much like electrolytic capacitors.  In one direction they
conducted; in the other direction they anodized the aluminum electrode.

        Dr. Barry L. Ornitz     WA4VZQ

I'll send you some additional information on the chemistry of electrolytics
in a separate message.

From: "Barry Ornitz" <>
Subject: Re: Faking Non-Polarized Electrolytics
Date: Mon, 20 Nov 2000 22:50:46 -0500

"Bill J" <> wrote in message
> > There's a "trick" which I have always considered a bit dodgy:
> To simulate, say, a 16 uF 250 V non-polarised, get two of 32 uF 350 V
> ordinary, and connect them back-to-back. Also connect 100 kohm to 1
> megohm, sufficiently high wattage, across each.
> The theory is that at the worst, each C will take it in turns to look like
> a "short", whilst the V across it is the wrong way. And the R across it
> will prevent the natural non-exact equality from allowing one C to do all
> the work.

This works, but you do not need to double the capacitance.

When an electrolytic capacitor is reverse-biased, it conducts if the
applied voltage is greater than approximately 1.5 volts.  So when one
capacitor is reverse-biased, the other capacitor is charged.  As the
polarity reverses, the roles reverse.

If you measure the capacitance with a very low voltage bridge, the
capacitance will measure half the capacitance of each unit.  But if you
increase the voltage the capacitance measured will increase as you get over
the 1.5 volt excitation level.  This diode action is why non-polar
electrolytics generate so much distortion in audio circuits.

        Dr. Barry L. Ornitz     WA4VZQ

From: "Barry Ornitz" <>
Subject: Re: How to form an old electrolytic ?
Date: Wed, 13 Dec 2000 02:01:08 -0500

"Douglas Snowden" <> wrote in message
> What is the best way to form an old elcctrolytic,
> say 30ufd at 250vdc that has been sitting for many
> years?

If the electrolytic is a wet, or liquid filled electrolytic from before
1940, you may have to add distilled water.  Most electrolytics made after
WWII are the "dry" type which are actually made with a paste electrolyte.
The paste contains materials such as propylene glycol, sorbitol, etc. that
absorb moisture from the air to keep the electrolyte moist.  So if the unit
has been stored in an extremely dry place, let it set a few days in a
typical home environment to re-moisten.  [Most electronic components store
best when kept extremely dry.  Electrolytic capacitors are the exception.]

The aluminum oxide film that forms on the capacitor anode is the
dielectric.  The anode is the aluminum foil upon which this film forms.
The cathode is the electrolyte.  The other foil serves merely to provide
electrical contact to the electrolyte.  With age, this oxide film
hydrolyzes and dissolves, lowering the voltage rating of the capacitor.

To reform, several approaches may be used.  The traditional method is to
use a variable power supply, or a variable transformer feeding a
conventional supply, to produce a voltage across the capacitor.  The
voltage is started quite low and is increased slowly while watching the
current drawn by the capacitor.  If the current rapidly increases, back off
on the voltage and let the capacitor remain at the lower voltage for a
while before increasing again.  A forming current of 100 microamps to 1
milliamp is a typical maximum allowable current.

Another approach is to hook the capacitor in series with a current limiting
resistor and connect the combination to a power supply of the desired
voltage.  Size the resistor for the initial desired charging current.(the
starting current).  The capacitor voltage will initially follow a typical
RC time constant charging curve until the anodic film starts forming.  At
this point the voltage will rise almost linearly until the desired voltage
rating is met.

To save some time over the second approach, I use a constant current
source.  This is a little faster but it is hardly necessary.

        73,  Barry     WA4VZQ

Index Home About Blog