[sdiy] Correction of two electrolytic cap related thing

[Harry Bissell Jr]

Bob wrote:

> However the points I made about the nature of
> dielectric materials in general and now no matter
> how fast you alter the field on them within the
> rated voltage limits will you 'hurt' the dielectric
> material still applies I believe. 

Actually... capacitors are specified with a dV/dT
(usually spec'd in V/us) rating of how fast of a
change they can tolerate...

So a 450V cap might fail if the change in voltage over
time were too fast.

You have probably not encountered this before,
because...

1) Electrolytic capacitors usually are not specified
or intended for pulse current applications.  Usually
the only concern in filter applications (like the tube
amp example) is the 'ripple current' ... an AC signal
superimposed on the DC level of the cap. This causes
heating and can shorten the life of the cap... but
this  is not what would happen in the 'arc' condition

2) The internal inductance of the electrolytic cap may
actually HELP you by limiting the current in a pulse
application

Film capacitors for pulse applications do have dV/dT
ratings published.

Now... how fast does the discharge actually happen ?

(the maximum slope is at the beginning of the
discharge)

...and does it exceed the rating of the cap ?

(answer : who knows ?)

If you do draw an arc on the cap, it may not be
constant in discharge. It may make and break several
times (in tens of microseconds). The internal
inductance which now has a high current flowing
through it may 'ring up' ie. resonate  (its a series
RC circuit damped only by the ESR). This could produce
a voltage MUCH higher than the applied voltage, and
could punch through the dielectric.

As Bob has noted, this does not happen in most cases,
but it is not impossible.

If the 'foil leads' to the terminals were to break
open
from heating, the dielectric could be damaged as well.
Again, not too common.

The danger here is that you are using the electrolytic
cap in a way that it is not specified or intended for.
The resistor discharge is a much safer way to go.

H^) harry  (when caps fail I have to do the
post-mortems ;^)