[Prev][Next][Index][Thread]
Oil Cap + Paper
Subject: Oil Cap + Paper
Date: Fri, 2 May 1997 19:49:12 -0400
From: "Thomas McGahee" <tom_mcgahee-at-sigmais-dot-com>
To: <tesla-at-pupman-dot-com>
CC: "Tesla-2" <tesla-2-at-emachine-dot-com>
Fellow Coilers,
A little while ago there was some discussion as to whether or not adding
paper to a Tesla coil capacitor would be useful. I have just finished a
few
experiments along this line, and would like to share with all of you my
observations.
I just finished building a flat plate capacitor using TWO sheets of .06
poypropylene between each metal plate. By the way, the poly rectangles
were 24 inches by 18 inches, and the metal plates 20 inches by 16
inches.
Border was 1" along each long side, and 2 inches at one end, the other
end
being the end the metal plate stuck out.
To reduce the problem of air bubbles I placed a sheet of paper between
each
surface. The resulting sandwich of materials then looked like this:
metal flashing
paper
poly
paper
poly
paper
metal flashing
paper
poly
paper
poly
paper
metal flashing
... repeated for entire capacitor.
I first did a visualization experiment so that I could see what kind of
effect the oil soaked paper would have on bubble removal. To this end I
substituted sheets of clear mylar for the plates and clear plexiglass
for
the poly. I used thicknesses that closely matched the metal plates and
poly
sheets. I used the paper as itself. I built up the equivalent of a unit
with three metal plates and used 1/4 inch plexiglass to hold the
assembly
together inside the casing that would eventually hold the actual flat
plate
capacitor. The casing was clear plexiglass. I then oriented the assembly
vertically and introduced oil. I backlit the assembly with an
incandescent
light unit so I could see through the paper. The "fake" capacitor
assembly
was under slight pressure, so that it had conditions as identical to the
final actual capacitor as I could get.
When the oil was initially introduced (slowly), I could see the oil
wicking
into the paper from the edges and the bottom. I decided to introduce the
oil rather slowly so that the wicking took place evenly. As I suspected,
this wicking action GREATLY reduced the incidence of air bubbles. Note
that
some small pockets of air bubbles were noted, but it was found that the
wicking paper helped to keep bubble formation down.
I then did an experiment to see what would happen to bubbles in a
regular
capacitor and in one with the added paper. To this end I built a
two-plate
version using a 4 inch square of poly and metal plates 3 inches by five
inches. This capacitor was built into a small plexiglass container
arranged
so that I could see sideways into the sandwich. I purposefully made sure
there were air bubbles between plates and poly with NO paper first. I
connected this little baby capacitor in parallel with a .01 mfd unit
connected to a Tesla coil. I removed the secondary as a safety
precaution
and fired up the machine. I had several heavy blankets covering the
little
capacitor, and I put my head under the blankets to observe what was
happening INSIDE the capacitor. The only light was that produced by High
Voltage activity within the capacitor. With NO paper I observed corona
activity where the bubbles existed. The bubbles would GROW as time
progressed, and it was evident that a fair amount of heat was generated.
To
help me see better what was happening, I used a few thin strips of mylar
to
create a LARGE bubble that I could view sideways. The capacitor plates
vibrate, and this agitates the bubble somewhat, but the bubble has a lot
of
activity going on within it, and it GLOWS quite brightly at all times. I
did experiments that verified that the air bubbles in a normal capacitor
create heat and promote dielectric breakdown. I did this by pushing the
capacitor to breakdown (It was rated at 1/2 the voltage of the big
capacitor it was paralled with, so it was the weakest part and failed
before anything else had a chance to.
By the way, when the poly gave way there was an internal arc followed by
copious gas production. Since I had this thing in a fairly controlled
environment, I let it cook for a few seconds before turning the machine
off. The oil was discolored and gas pressure in the chamber had risen,
but
there was no explosion or anything of that kind... probably because the
chamber had a large area for gas expansion. I threw out the bad oil and
cleaned the container.
I repeated the experiment, but this time with paper between each active
surface. There were fewer bubbles formed than without paper, so I
created a
few fair sized bubbles ON PURPOSE. Imagine my surprise when I fired the
machine and saw the bubbles diminishing in size! Here is what I found
out
was happening: when there is a layer of oil between the plates, air
bubbles
are broken up into smaller air bubbles and appear to be absorbed by the
oil.
I verified this by creating a capacitor with ONLY oil as the dielectric,
and found that when the bubbles are surrounded by oil that the
application
of high frequency AC does indeed make the bubbles break up and dissipate
throughout the oil! This effect is only noticed when high frequency AC
is
applied. It was NOT observed with HV DC or 12KV 60HZ AC. It seems that
the
oil wicked up in the paper does something similar, and does indeed help
considerably in dissipating bubbles when the machine is in operation. I
sort of expected something like that to happen because Tesla had written
about how oil breaks up little bubbles in a high frequency coil that is
oil-immersed... but I was not aware of how quickly this can happen in a
capacitor with paper added. Without the paper there is not enough oil in
contact with most bubbles to break them down. So instead the bubbles get
hot and expand and make the problem worse. The paper helps to ensure
that
every bubble has oil surrounding at least 50% of its surface. This make
a
BIG difference!
***
Once I had done these experiments, I decided to do a few additional
experiments as part of my construction of the *full-sized* oil-immersed
paper and poly capacitor.
Since I was interested in whether or not there was any reduction in
capacitance when adding the paper separators, I decided to first
assemble
the capacitor DRY and with NO PAPER. The measured capacitance was .0101
mfd. I then rebuilt the capacitor with the paper in between as shown.
The
measured capacitance BEFORE adding UNIVOLT 61 Transformer Oil was .0098
mfd. The measured capacitance AFTER adding oil was .0098 mfd (no
change).
As you can see, the difference in capacitance value when adding a paper
separator is minimal. In this case, about 3%.
The paper I used was not Kraft paper, just a good 100% cotton paper that
I
got from our print shop. I cut it into sheets the same size as the poly.
The only reason I didn't use Kraft paper was that this other stuff was
handy, and it looked to me to be very strong and absorbent. You want the
paper to be smooth and absorbent, and thick enough that there is a
decent
amount of oil there to help break down the bubbles during operation. Do
NOT
use coated stock!! You want it smooth, thin, absorbent.
****
When I make my next capacitor of this kind I am going to make the
capacitor
plates about half the size I did on this unit. I had to make a custom
plexiglass case for it because I wanted it to be a sealed unit with the
lowest inductance I could get. But the size and weight makes it a bit
inconvenient to move around. My suggestion to others who want to build
flat
plate capacitors is to first find a good sturdy container and then
determine the actual size of the plates from those dimensions. I decided
to
go for minimum waste of poly, and that meant large plates. It also meant
the expense and time of building a custom enclosure. Unless you have a
really CHEAP source of plexiglass, buy a plastic cooler or rectangular
garbage can or whatever for your enclosure. Save oil by making the
actual
capacitor a tight fit (but with at least 1/4" all around for allowing
the
oil to move freely). I use small pieces of 1/4" plexiglass scraps glued
to
the enclosure to make sure the capacitor stays in one place. I also
assembled the capacitor on a piece of plexiglass that serves as a
platform
that holds the thing together. That way none of the plates or poly move
around relative to one another.
In keeping with my desire to make the capacitor high quality, I used
three
1/4" x 2 1/2" bolts to hold each set of plates together. Large nuts and
wide washers were built up between each set of metal plates so that they
were held at the proper distance. The wide contact surface area allows
huge
currents to flow from each set of plates to the appropriate large flat
strap aluminum connector that connects from the set of plates to a large
1/2" bolt. It is this large bolt that connects the capacitor plate to
the
rest of the Tesla tank circuit. This may appear like overkill to some of
you, but recall that the total resistance in the tank circuit is the sum
of
all the individual resistance losses, and that ONE bad connection can
GREATLY affect the MAXIMUM current that will flow in the tank circuit.
When I built the actual capacitor, I waited 24 hours to let bubbles
escape,
with several inspections and shakings to help find and dislodge the big
bubbles that like to hang around. I connected it to a Tesla coil and
slowly
brought it up to full power and then ran it for 15 minutes continuous.
It
was cool as a cucumber and not a hint of corona. I only tested it at
12KV
using a 30 ma neon so far, but with the equivalent of 120 mils of poly
this
baby should be good for quite a bit of voltage. Does anyone out there
know
what the maximum AC RMS voltage rating would be? I don't have a list
with
polypropylene on it handy. By the way, I carefully inspected each piece
of
poly for defects, and ended up rejecting two because they had deep
surface
scratches from when the stuff was handled in the warehouse where I
bought
it. Looks like someone used one of those razor blade knives to cut the
sheet loose from its packaging. I'm glad I found that out in time to
reject
that section from the capacitor. I will use the reject pieces for other
HV
things that are not as critical.
I hope these experiments are useful to some of you. If I find any
deterioration of the paper or any other items of note in the future, I
will
send a post about it to the Tesla list.
Fr. Tom McGahee