[TenTec] Electrolytic Capacitors

[Dr. Gerald N. Johnson]

On Wed, 2007-11-21 at 07:30 -1000, Ken Brown wrote:
> This post is quite off topic. It only relates to Ten-Tec radio equipment 
> vaguely, since there are electrolytic capacitors in them. I ask the 
> question here, because I know there are knowledgeable people here who 
> may be able to provide some insight.
> 
> Electrolytic capacitors require a voltage across them to maintain the 
> oxide layer which is the dielectric between the plates, right?

Yes, the applied voltage maintains the very thin aluminum oxide layer
formed on the foil by applied voltage when they were made.

> If they 
> have no voltage across them for long periods, then they are likely to 
> fail when full rated voltage is applied, right?

The best grades of capacitors conservatively have been figured to
maintain original specifications for about 2 years. After that they
sometimes recover by reforming. Some don't do as well, the purity of the
foil has an effect on that shelf life.

> What about an 
> electrolytic capacitor which has a much higher voltage rating than that 
> which appears across it in the circuit?

In the old days (before 1962 about) makers wanted the DC working voltage
to be 70 to 85% of the voltage rating, though long about then GE brought
out capacitors that could be used at much lower than rated voltage
without as much deterioration. That way they didn't have to make or
stock so many voltage ratings in the distributorships and radio/TV
repair shops.

> Would the capacitor ESR increase 
> over time until it fails, in an open (or real high ESR) mode?

The failure mode from age where the electrolyte eats the oxide is one of
increased leakage current. The thinner oxide stands voltage less well so
leakage current is higher.

But there is another failure mode where the electrolyte leaks or is
boiled out by high operating temperature and that raises the series
resistance. 

There can be confusion in that capacitor bridges tend to measure both
shunt and series resistance as Equivalent Series Resistance or ESR
because of the way the bridge is set up. Either shows up as a lowered Q
of the capacitor, but series resistance affects the ripple filtering
qualities of the capacitor in a power supply more than shunt leakage so
its considered the more critical element.

> For 
> instance what would you expect from a 25 VDC rated electrolytic 
> capacitor which never has more than 5 volts applied to it?

Over a long time I would expect the capacitance to be higher and the
leakage reasonable at 5 volts but to rise rapidly with a higher test
voltage. But that can depend on the operating history and the
environmental history. If run hot I'd expect higher ESR that didn't
respond to reforming. The higher leakage sometimes will respond to
reforming and sometimes the capacitor explodes in the attempt or soon
after. I don't like cleaning the remains of exploded capacitor so I much
prefer to purchase new electrolytics from distributors with a high turn
over so their parts are fresh. Many times I've rejected miniature
electrolytic capacitors (the poorest made, especially when cheap) direct
from the local distributor's drawers for old date codes. That comes from
testing on my capacitor bridge and finding them showing a high power
factor or high ESR. I prefer my capacitors to have a date code within 12
months of the time I wire it into the circuit. I've failed NEW miniature
electrolytics because of high power factor even with recent date codes.

Miniature electrolytics thees days are a major contributor to equipment
malfunction, sometimes as bad as runs of them failing the first time
power is applied. I have little confidence that Ying Yong's back yard
capacitor factory in the wilds of central Asia has access to good
materials, clean processing, and proper forming. When replacing
electrolytic capacitors I try to purchase from known mainline
manufacturers and choose their highest temperature rated, lowest ESR
rated products. When there is room I like the Vishay-Sprague Atoms that
have a significantly larger case for better dissipation of heat from
internal leakage.


> 
> Thanks in advance.
> 
> DE N6KB
> 

-- 
73, Jerry, K0CQ, Technical Advisor to the CRA
All content copyright Dr. Gerald N. Johnson, electrical engineer