[Prev][Next][Index][Thread]

Re: Building your own HV transformer



Great to see someone interested in building HV transformers.

The Iron rods will surely cause eddy current losses.  Maybe he intends
to run the thing at high loss for short periods.  Like he says "the
design is untested."  I figure large number of typos in that article . .
. 500 feet of wire in the primary and 800 in the secondary, not too
likely!  I might believe 8,000 but just don't buy only 100 turns in the
primary because I did build a open core transformer.

I decided I needed a better transformer than the oil burner ignition
transformer I was running and didn't really have a good idea where to
find a suitable core.  Finding cores wasn't the problem, finding one
with enough space to accommodate the insulation required was a problem.

I have a book "The Tesla High Frequency Coil," by Haller & Cunningham
(originally 1910, but reprinted by Lindsay Publications in 1995). 
Misters Hallar and Cunningham outline building a Tesla coil from
scratch.  And I do mean scratch!  They give instructions on cutting the
laminations for the induction motor used to turn the rotary gap!

The HV transformer they outline is a sort of big induction coil.  Their
instructions call for a lot of stuff that isn't readily available
anymore, but are easily substituted.  

The core I used is a bundle of steel wire (about 11 awg) used for
hanging ceiling support.  It is one and a half inches in diameter and
twenty inches long.  The primary is wound in four layers of 350 turns
each (I brought out the start and finish ends of the wire and can
configure the primary for different turns ratio and primary impedance). 
With 120 volts, 60 cycles, I operate with two windings in parallel and
in series with the remaining two, for a total of three windings or 1050
turns.  With 140 volts in, I use all four windings, and with a
transistor chopped 100 volts DC, 200 cycles, I run all in parallel.  

The primary (excitation) current with no load, and three windings in
series, is on the order of one and a quarter amps.  Secondary output is
over 8,000 volts at a design nominal of 900 watts (easily pushed to 1.5
KW, due to the size, and thermal inertia of the transformer).

The secondary consists of four homemade bobbins wound with a total of
67,800 turns of 32 AWG magnet wire.  The inside bobbins have 55 layers
each and the outside 45 layers each.  I used one section of 2" PVC pipe
for the primary to secondary insulation, and the bobbins were made with
2-3/8" ID tubing (Plexiglas end plates) placed over the primary
insulator.

I rigged a vacuum impregnation system using a pressure cooker and vacuum
pump and impregnated each coil with liquid paraffin wax, assembled the
transformer in a oak box and back-filled the box with molten wax.  

If I use the transformer as an induction coil, the output voltage is 50
to 80 KV chopped at a 80 HZ rate, with a 12 mfd capacitor across the
'points.'  The tesla coil is ecstatic  with the increased voltage and
rep rate, but it is hell on primary capacitors.  The 12 mfd has to
endure a spike of ~600 volts on it.

To characterize the iron (they just don't include hysteresis curves with
ceiling wire) I wound a smaller test coil and scaled it up for the large
coil.  I arbitrarily set a limit of 1 amp, figuring that %10 was a
reasonable amount of power to waste to excite the iron.  The inductance
is directly proportional to the cross section of the iron in the core
(allow about one diameter to stick out either end past the primary). 
The inductance is proportional to the square of the turns (increase the
turns by a factor of two, increase the inductance by a factor of four).  

An open core will not saturate at any reasonable input voltage (the
primary would melt first).  An open core does tend to self limit current
to an extent.  I run un-balasted at all times. If I want to lower the
current I lower the input voltage, increase the primary turns, or
increase the speed of the rotary spark gap (unsynchronized, 6000 RPM,
eight breaks per revolution, with two fixed gaps in series)

My primary winding and iron can be slid out of the box because the PVC
tube protrudes from either end of the box and isn't encased in wax. 
I've experimented with finer and coarser wire with very little
difference in the performance at 60 cycles (at least in the 12-18 AWG
range).  The gauge of the core wire makes a profound difference in the
operation as an induction coil.  (I even tried running steel wool
through my garden shredder to make powdered iron - the permeability fell
faster than the eddy currents at 60 cycles - but allowed operation to
over 1 KHZ - and colored everything within about twenty feet of the
shredder a nice rust tint in a few days)

It only took seven days to wind the secondary, using a primitive motor
driven winding lathe.  There is about 13 miles of wire in the
secondary.  It took about six months to complete the entire project,
from planning to playing.  It wasn't particularly hard, and was very
rewarding.  I was sorry to finish it; building it was fun.

The three phase transformer sounds like it might be a winner if you have
enough room for the insulation.  Do you plan to cut part of the core
away, or wind the primary/secondary in pairs on the outer legs and phase
them in series?  

Don't be too tempted by high a secondary voltage, the insulation
requirements in the transformer get hairy over ~5,000 volts/winding, and
the TC's capacitor design gets bigger and more expensive.  

Consider the weight of the finished transformer, mine weighs in at over
fifty pounds and I neglected to allow for handles on the case.  stupid. 
If I do another one, I will increase the size of the core to allow a
more winder-friendly turns ratio.

To characterize your three phase core, you only need to wind a test coil
with a few hundred turns of wire and use a variac to bring up the
voltage on it.  If and when you reach saturation, the current really
takes off with a small increase in input voltage.  Bear in mind the
current will drop a little as the primary wire heats because of
resistance increases at higher temperatures.  The core must be closed to
see a sharp rise in current, even a small air gap will cause a sloppy
'knee' in the curve.

The existing primary turns and voltage rating, if you have them, may
give some information.

My copy of the ARRL Handbook, 1990, devotes about three pages to 60HZ
transformer design, but they are chock full of the basics.  If you want
to go with the transformer/induction coil route (I recommend it) get
Lindsay's reprint, make the cross section of the core larger than
1-1/2," and put handles on the box (or wheels).

Post your experiences! 
good luck.