[Lowfer] Question on building homebrew variable caps...

[Ralph Hartwell]

...combining replies to two messages at once here...

> > > consider a set of fixed cap plates separated by glass plates
> > > now sliding the glass in and out between the plates will make
> > > the cap vary. Advantage would be a cap with good mechanical
> >
> >Glass is a very lossy dielectric at RF.  Ask the guys who build Tesla
> >Coils.  I have popped more glass plate caps than I want to think
about.
> >Much better is PVC plastic.

...and...

> Are you sure that glass is lossy at 185KHz? Or is it just the way that
these
> glass plate capacitors are constructed - allowing uneven heat
generated from
> the metal plates to crack the glass. PVC isn't that great either. Ask
the
> guys that build the really high-Q loading coils.

Most Tesla Coils run in the frequency range of 75 to 500 KC, and the
experience of many people, including my own, has shown that a glass
plate cap will fracture if run at any sort of high power level.
Replacing the glass with polyethylene plastic will often increase the
coil's output by as much as 50% with no other changes.

The loss tangent of most glasses at RF is high enough so that at high
power levels the RF loss heating causes the glass to become partially
conductive if it gets hot enough, at which point the dielectric fails.

Usually, however, the glass simply either punctures or cracks due to the
localized heating occurring at places where there are microscopic
imperfections in the glass.  Ordinary window glass is particularly bad
in this respect.  As for various types of glass, I don't know of any
that are much better than others in this respect, although I am sure
there are some.  I would expect that the cost would be higher than
readily available window glass.  Sheet Quartz should be better, though
expensive.  But plastic is much better.  If you look up the loss
tangents of various materials, you'll see they pretty much go from super
good to poor as you progress down the line of:

Vacuum
TFE
PTFE
Paraffin wax
Polystyrene
Polyethylene
.....
.....
Glass
etc.

The difference between TFE and Polyethylene is vanishingly small for use
at these low frequencies.

Standard capacitor construction methods used by many serious Tesla Coil
builders for making effective HV RF rated caps use multiple layers of
thin poly sheet placed together so that any imperfections in one sheet
will not line up with imperfections in the adjacent sheets.  Several
thin sheets of dielectric are much better than one thick sheet.  This is
due to the fact that commonly available poly sheet will unavoidably have
some imperfections or inclusions in it which will decrease the
dielectric strength at the point of the defect.  These imperfections
will increase the concentration of the electric field at that point,
leading to the possibility of failure of the dielectric at lower than
expected voltage. Usually three sheets of poly are used.  Do not try to
exceed 3 KV per cap if you wish to eliminate corona failure.  Many have
tried; many have failed. <G>

A quick trip over to one of the more serious Tesla Coil lists will
provide you with a wealth of HV RF capacitor construction info.  Most of
it, of course, is aimed at building plastic insulated, oil filled caps
of up to 0.05 Uf rated to withstand 30 KV RF pulses without expiring.
Nevertheless, the concepts used are applicable to LF uses as well.

73,

Ralph   W5JGV / WC2XSR / 13

The current On-The-Air status of WC2XSR/13 is available at:
http://home.att.net/~shmrg/wc2xsr-13_log.htm

My 600 Meter Research Group files are at:
http://home.att.net/~shmrg

My Home Page is at:
http://home.att.net/~ralph.hartwell
Home of the Rife Biomedical Research Web Ring