Don't forget about "soak", also known as dielectric absorption or
dielectric hysteresis. In some capacitors, particularly electrolytics,
charge "soaks" into the dielectric and can take from seconds to days to
work its way out. The result is a "discharged" capacitor that seems to
spontaneously recharge itself. (One story I heard during my broadcast
days was someone who put a screwdriver across a "discharged" capacitor.
After he picked himself up off the floor, he found just the handle of
the screwdriver. He was lucky his eyes weren't hit with molten metal.
You'll often find large capacitors stored with a shorting wire across
them -- a good idea.) I recommend leaving a heavy cliplead in place
across the high voltage capacitors for the entire time you're working on
the unit. Put some tape over the on/off switch and/or a tag on the plug
to remind you to remove it before turning the power back on.

As a side note, the aquadag coating of a CRT is notorious for this. I've
gotten a healthy bite off a CRT the day after it was discharged for an
hour or more with a cliplead.

Roy Lewallen, W7EL

On Thu, 14 Jul 2005, Roy Lewallen wrote:

Date: Thu, 14 Jul 2005 13:27:57 -0700
From: Roy Lewallen
Newsgroups: rec.radio.amateur.homebrew
Subject: QUESTION: Fun with Svetlanas or Staying alive with kV power
supplies

Don't forget about "soak", also known as dielectric absorption or dielectric
hysteresis. In some capacitors, particularly electrolytics, charge "soaks"
into the dielectric and can take from seconds to days to work its way out.
The result is a "discharged" capacitor that seems to spontaneously recharge
itself. (One story I heard during my broadcast days was someone who put a
screwdriver across a "discharged" capacitor.

Yep, I can vouch for this effect. I've seen it, too. Short the cap with
alligator clip-tipped wire for a few seconds if you want that voltage down
to microvolts. I think it is not "soak" but simple RC time constant decay.
Just like radioactive half-life. Residual voltage on a cap _never_ goes to
true zero, only according to the decay equation. However, there is a
so-called "electret" effect which really sounds like your "soak" effect.

After he picked himself up off
the floor, he found just the handle of the screwdriver. He was lucky his eyes
weren't hit with molten metal. You'll often find large capacitors stored with
a shorting wire across them -- a good idea.)

D'arsonval meters, especially sensitive ones, are also best stored with a
shorting wire accross the terminals.

I recommend leaving a heavy
cliplead in place across the high voltage capacitors for the entire time
you're working on the unit. Put some tape over the on/off switch and/or a tag
on the plug to remind you to remove it before turning the power back on.

As a side note, the aquadag coating of a CRT is notorious for this. I've
gotten a healthy bite off a CRT the day after it was discharged for an hour
or more with a cliplead.

Roy Lewallen, W7EL

In article g,
straydog wrote:

Don't forget about "soak", also known as dielectric absorption or dielectric
hysteresis. In some capacitors, particularly electrolytics, charge "soaks"
into the dielectric and can take from seconds to days to work its way out.
The result is a "discharged" capacitor that seems to spontaneously recharge
itself. (One story I heard during my broadcast days was someone who put a
screwdriver across a "discharged" capacitor.

Yep, I can vouch for this effect. I've seen it, too. Short the cap with
alligator clip-tipped wire for a few seconds if you want that voltage down
to microvolts. I think it is not "soak" but simple RC time constant decay.
Just like radioactive half-life. Residual voltage on a cap _never_ goes to
true zero, only according to the decay equation. However, there is a
so-called "electret" effect which really sounds like your "soak" effect.

Yup. Formally "dielectric absorbtion". In a lot of types of
capacitor, enough stored charge can come back out, over the course of
a few minutes, to raise the open-circuit terminal voltage of the cap
to a significant percentage of its fully-charged level, even if you
had previously discharged the cap all the way to a zero reading.

It can be enough to give you a nasty bite.

--
Dave Platt AE6EO
Hosting the Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!

straydog wrote:

Yep, I can vouch for this effect. I've seen it, too. Short the cap with
alligator clip-tipped wire for a few seconds if you want that voltage
down to microvolts. I think it is not "soak" but simple RC time constant
decay. Just like radioactive half-life. Residual voltage on a cap
_never_ goes to true zero, only according to the decay equation.
However, there is a so-called "electret" effect which really sounds like
your "soak" effect.

No, it's not a simple time constant -- the slope is much different. A
reasonably good model of it is a series R, shunt C, series R, shunt C,
etc. A large number of Rs and Cs with various time constants have to be
used to imitate it reasonably well.

I don't know about an "electret effect". An electret is a dielectric
with a permanently trapped charge, so it produces a static electric
field. It's the analog of a magnet, with permanently trapped flux and
producing a permanent magnetic field. A magnet that's not moving can't
produce a current (much to the dismay of the perpetual motion crowd),
and an electret that's not moving can't produce a voltage.


D'arsonval meters, especially sensitive ones, are also best stored with
a shorting wire accross the terminals.

That's for an entirely different reason -- to protect the meter from
damage, rather than the user.

Roy Lewallen, W7EL

On Thu, 14 Jul 2005, Roy Lewallen wrote:

Date: Thu, 14 Jul 2005 19:47:23 -0700
From: Roy Lewallen
Newsgroups: rec.radio.amateur.homebrew
Subject: QUESTION: Fun with Svetlanas or Staying alive with kV power
supplies

straydog wrote:

Yep, I can vouch for this effect. I've seen it, too. Short the cap with
alligator clip-tipped wire for a few seconds if you want that voltage down
to microvolts. I think it is not "soak" but simple RC time constant decay.
Just like radioactive half-life. Residual voltage on a cap _never_ goes to
true zero, only according to the decay equation. However, there is a
so-called "electret" effect which really sounds like your "soak" effect.

No, it's not a simple time constant -- the slope is much different. A
reasonably good model of it is a series R, shunt C, series R, shunt C, etc. A
large number of Rs and Cs with various time constants have to be used to
imitate it reasonably well.

Well, if you want to get technical about it, yes. All those layers of foil
all, individually, add up to all of those Cs and Rs.

I don't know about an "electret effect". An electret is a dielectric with a
permanently trapped charge, so it produces a static electric field.

I doubt if there is very much in the universe with a real, practical
_permanence_. The second law of thermodynamics says that trapped charge is
going to try hard to become untrapped.

It's the
analog of a magnet, with permanently trapped flux and producing a permanent
magnetic field.

Most magnets that I've heard about will slowly lose their magnetism via
one of many mechanisms.

A magnet that's not moving can't produce a current (much to
the dismay of the perpetual motion crowd), and an electret that's not moving
can't produce a voltage.

In a laboratory, once, I made an electret out of, basically, frozen water.
The various pieces of these electrets, bathed in liquid nitrogen, behaved,
qualitatively, like little magnets. Bring one near another, and whatever
"pole" was at some point on one piece would spontaneously jump,
appropriately, to the piece I was holding.

An electret that is not moving can't produce a voltage? Semantics. There
_will_ be an electric field between one pole and the other and it _will_
be measureable and calculable and the results will be in volts regardless
of whether the electret is "moving". The units can be expressed in more
than one way depending on reference systems, definitions, and goal of the
measurement or calculation.


D'arsonval meters, especially sensitive ones, are also best stored with a
shorting wire accross the terminals.

That's for an entirely different reason -- to protect the meter from damage,
rather than the user.

For many people (since, in my life, it came up many times in
conversation), protecting the meter was more important.

Roy Lewallen, W7EL


Art, W4PON