Haunted Capacitors

Hi Everyone,

I have just been reading an interesting discussion on rec.pyrotechnics where
they have been talking about high voltage capacitors that seem to recharge
themselves after being discharged. Has anyone ever had this happen to them
using Tesla coil capacitors?
I remember once seeing a demonstration of a dissectible Leyden jar- the
lecturer charged up the jar and then pulled the inner and outer plates out
and shorted them together- there was a spark when they touched. He then held
the jar and he didnt get a shock. He then reassembled the capacitor and
shorted the two metal plates together with a wire- there was a great big
spark this time - even though the capacitor had apparently been discharged!
Apparently this demonstration showed that most of the charge in a Leyden jar
capacitor is stored in the dielectric. Anyway I have reproduced the
discussion from rec.pyrotechnics below just in case anyone is interested:

       Friday, 14 July 1995 20:08:40
          rec.pyrotechnics Item
  From:           David Brooks,daveb-at-perth.DIALix.oz.au,News Only
  Subject:        Re: TC Transformer Phasing (was RE: tesla coils(help))
  To:             rec.pyrotechnics
Steve Roys <sroys-at-umabnet.ab.umd.edu> writes:

>> 	Also, if you do use the capacitors, you should be aware that they 
>> store the electricity, and even if you shut off the input power to it, 
>> they MAY STILL BE CHARGED.  You must discharge the capacitors to make 
>> sure that they are safe.

>Another thing is that high voltage capacitors can recharge themselves 
>after being discharged, so It's a good idea to short out the caps 
>immediately before making any adjustments.  It's also good practice to 
>store HV caps with the terminals wired together so you don't get any 

>> Also, before playing with the stuff and getting your hands dirty, grab 
>> a voltmeter, and check for any residual voltages.  It could save your

>The might not work too well for a cap charged to 10kV, your voltmeter would 
>probably complain. :-)  For large HV caps it's safer to assume they're 
>charged and always short them out.

  Yet another good precaution is to connect a bleeder resistor 
_permanently_ across that nasty HV capacitor. The value will naturally 
depend on the application, but see descriptions of high power amateur 
radio transmitters for some examples.
 Another is to provide a HV relay, which shorts the relevant points to 
ground. Said relay being pulled open-circuit by the DC supply power. This 
ensures that everything fails safe when power is off. BTW, put a 
low-value _high wattage_ resistor in series with those relay contacts, so as 
not to burn them when they do close on a fully charged HV cap.


          Monday, 17 July 1995 4:03:10
          rec.pyrotechnics Item
  From:           Tim429SCJ,tim429scj-at-aol-dot-com,News Only
  Subject:        Re: TC Transformer Phasing (was RE: tesla coils(help))
  To:             rec.pyrotechnics
>BTW, I wasn't aware that a HV capacitor could recharge itself...how does 
>it do it?  I work with some capacitors that are about 2 feet cubed, but 
>run on about 150vdc, so I'm not as familiar with high voltage caps...and 
>we have bleeding resistors in our gear, so they discharge the caps for 
>us, if we leave them for a couple of minutes.  (BTW, this beast in 
>particular uses 6 400 Amp fuses, and I've seen them blow before.)


Brad, the phenomenon you are asking about is a result of dielectric
absorption.  When a capacitor is charged, the charge is actually stored
in the dielectric.  When the cap is discharged, a portion of the charge 
remains trapped in the dielectric.  Over time, this trapped (absorbed) 
charge can escape to build up again on the plates.  The amount of 
dielectric absorption displayed by a given cap varies with the dielectric
material used.  Hope this helps.  You might check with some of the high-
end audio news groups for more on dielectric absorption.


          Sunday, 16 July 1995 17:21:25
          rec.pyrotechnics Item
  From:           DLSwank,dlswank-at-aol-dot-com,News Only
  Subject:        Subject: Re: TC Transformer Phasing
  To:             rec.pyrotechnics

: Brad K. Browne (aa957-at-ccn.cs.dal.ca) wrote:
: : BTW, I wasn't aware that a HV capacitor could recharge itself...how
: : it do it?

A note from the electronics world:

There is a well-known phenomenon with "plastic" (polystyrene
and polyethylene) capacitors called "voltage soak". If held at
voltages over about 100VDC. for a few hours and then discharged,
in a few hours (or days) a terminal voltage perhaps as much as
1/3 to 1/2 of original will reappear.

Have never seen any rigorous investigations of it, but most
opinions seem to be roughly that:

(1) The original charge causes some mechanical stress to the
dielectric material.  (This part is firm, and a widely known loss
mechanism for capacitors of any kind.)

(2) The plastic (dielectric) cold-flows and/or absorbs the stress
in its molecular lattice directly. (The energy is STORED via one
or several mechanical mechanisms.)

(3)  After stress is relieved the stored (mechanical) energy is 
changed back to electrical charge, probably via piezoelectricic
means. Most large plastics molecules "should" show some
piezoelectric effect. Some have large amounts and are used
in "electret microphones".  Common capacitor dielectrics 
definitely do have small but predictable amounts. Some forms
of contamination during manufacture might cause large
increases in the effect.

Electrolytic capacitors usually have much higher internal leakage
and this effect is hard to find although theoretically it 'should
be there'  for them as well.

Since the effect reduces performance of the capacitors, makers work
hard to eliminate it. Users grumble but tolerate it. Nobody doubts
that the effect is quite real.

There is also some evidence that rapid temperature changes on the
capacitor can cause some charge to appear, probably from the
same mechanisms.

Anyone using HI-V capacitors should definitely keep some
loading resistance on them during periods when they wish them
to be "discharged". Even 100K ohms via a toggle switch will keep
them 'cooled off' . Note that the load should be DIRECT and not
through some semiconductor junction.

-Doug Swank-

Best Regards,

-- Mark
       _/_/_/   _/_/_/_/       Mark Conway
      _/    _/    _/          Deep Thought BBS, Auckland, New Zealand
     _/    _/    _/          A FirstClass(tm) Macintosh GUI BBS
    _/_/_/      _/          Internet: mconway-at-deepthnk.kiwi.gen.nz