Re: My extended-foil cap failed

To: tesla-at-pupman-dot-com
Subject: Re: My extended-foil cap failed
From: "Malcolm Watts" <MALCOLM-at-directorate.wnp.ac.nz> (by way of Terry Fritz <terryf-at-verinet-dot-com>)
Date: Sun, 11 Oct 1998 21:49:30 -0600
Approved: terryf-at-verinet-dot-com

Hi Gary,

> Original Poster: Gary Lau  10-Oct-1998 1204 <lau-at-hdecad.ENET.dec-dot-com> 
> 
> Hello All,
> 
> I was eager to log some time on the extended-foil rolled-poly capacitors
> that I've documented on my new web site.  This capacitor is two units in
> series, each using .04" thick LDPE dielectric.  The caps were housed in
> 4" PVC SDR pipe and had no provisions for pressure relief.
> 
> Last night I ran the coil for 1 minute, checked for any signs of cap
> heating, ran for 2 minutes and checked.  The ends were just slightly
> warm.  I ran for 3 minutes and checked.  The ends were again just
> slightly warm.  I was about 2 minutes into the next run when the roar
> diminished, then POOOMMM!  One of the PVC end caps burst, leaving about a
> pint of mineral oil on my garage floor and egg on my face.
> 
> A post mortem disection revealed dielectric puncture under the long edge of
> the recessed foil, plus a slight discoloration of the poly elsewhere
> along the recessed edge of the foil, suggesting that the dielectric
> thickness may not have been adequate.  All poly and foil surfaces were
> evenly wet with oil.  Unclear is how long this cap would have survived
> had I limited run-time to a more conventional 30 seconds at a time.  Is
> running 7-8 near-continuous minutes something that is foolish even with
> the most robust capacitors?

That is exactly the discolouration I observe on unrolling a blown (or 
partially used) dry fired extended foil poly cap. It seems to me that
the real problem is not dielectric thickness (compare the DC handling 
capacity with the applied voltage in TC use) but the edge corona 
suppression. This suggests to me that the real limiting factor 
regardless of poly thickness has to be applied voltage per cap. I can 
get 1/2 hour+ from a dry fired cap which is fired for two or three 30-
40 second bursts and then left to cool internally completely. Were I 
to make a serious cap building effort now, I would build a large 
number of sections to run in series with no more than a couple of 
kV/section. It is a breeze to build a cap to that spec. Size is just 
not going to be competitive with the pros but....

> I will update my web site with this info, and in the future, recommend
> and use pressure release mechanisms on caps!

Hear hear!

Malcolm