Re: failed cap

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Dr. Resonance" <resonance-at-jvlnet-dot-com> 


Any home brew caps using a relatively thick dielectric (64 mil) is subject
to this type of failure.  You had some small voids in the dielectric.
Inside these voids the dielectric constant is different from the surrounding
material.  The result is corona inside the small void.  Corona produces heat
and eventual failure occurs.  Some of these voids are so small they almost
can not be seen with the human eye.  As the corona forms immediately,
breakdown begins.  It's a slow process but eventually the void produces
dielectric failure as you have experienced.

Professional manufacturers always use 0.5 (1/2) mil thick dielectric
material to avoid this problem.  This dielectric material is so thin it can
only be handled by expensive winding machines and can not be touched by
human hands.  The entire surface area of this dielectric material is checked
for voids prior to acceptance by the cap mfgr.

I toured Maxwell Capacitors years ago and they use a "clean room" for all
their special energy discharge cap winding processes.  Some of their other
processes such as soldering extended endfoils together were equally
interesting.

Stick with the MMC caps and a 2.5:1 ratio (DC to AC) and you will have
virtually no trouble in the future.

Dr. Resonance

Resonance Research Corporation
E11870 Shadylane Rd.
Baraboo   WI   53913
 >
 > Hi.
 >     i built a cap with 1.6 mm or 64 mil thick PE dielectric and put the
 > whole thing in canola oil. It worked great for a while but while i had
just
 > the cap across the nst ,,, with no tesla coil connected.. to test my new
 > mulitgap sparkgap... the PE between 2 of the plates broke down and now it
 > blows bubbles and hisses and only produces a spark in the gap if the gap
is
 > 1 mm or 30 mil . Its pretty much had it and must be breaking down at 1 or
2
 > thousand volts..IS it more stressfull on a cap with no tesla coil
connected
 > and is it suprising my cap did not last in my config. The plates are 1.75"
 > clearance from each end and 1" along the sides. !6 sheets of makes each
layer
 >
 > Phil