Small Tesla coil

The goal was to make a very cheap but nevertheless agreeably well performing Tesla coil for home operation. The coil is run from a 8kV 50mA (400VA) neon sign transformer from NeonComp , multiple of which were gained for free at a local neon sign construction yard.

Most of the materials are low-cost, household materials like cardboard, old wood panels, chicken fence, used transparency sheets, and household aluminium foil. Secondary wire is ripped from an old transformer, primary is made of bathroom copper tubing.

The only things paid for were: roll of aluminium duct tape and air duct (for the toroid), cables and mains sockets, RF filter, switch board components, boat varnish (urethane), a bunch of screws.

Total cost around $50. Low-power (<1/2) performance is 4"-7" streamers, full power gives 12" streamers.

Here are the specs:

mains side:  
controller: Simple triac switch with an opto-coupler, switched by  9V input pulse. (Now replaced by solid-state relais). 
A simple 18VDC relais switch could do, too.
Includes internal mains RF filter and fuse.
PSU: 220V neon transformer, rated 400VA
gives 50mA current at 8kV

PFC cap for ~100% correction installed

commercial (cheap) RF filter put in reverse on the mains side

grounding: casing of the switch device / controller is grounded,  and the center tap of the NST is also grounded.


primary side:

spark gap: Bad design.

Two concentric metal cylinders i.e. tin cans from ananas containers that have the walls spaced about 6mm apart. Wires are connected with screws .

The SG is placed in a plastic marmelade container, bottom removed, and is blown by a 25W 220V microwave oven fan 

The fan is placed in parallel with the NST mains connectors and of course slows down when the NST charges the HV capacitor... not good.

Now replaced by a RQ-style gap - multiple copper tube pieces in a ventilated plastic cylinder.


tank cap: Self made 8nF capacitor (10nF was used for calculations), using old and used A4 sized transparency sheets (free, university) and household aluminium foil (1" shorter than A4 paper on all four sides) stacked in the manner of a MMC cap - 1 foil layer out to one HV connector then 10 alu-foil & transp. sheet layers in between, and on top comes another alu foil layer out on the opposite side to other HV connector.

During building, each transp. sheet was thinly(!) covered with cold-pressed vegetable raps oil, then one alu sheet was put on top, it also covered with oil, then next transp. sheet placed on top and air and left over oil squeezed out by hand.

This should give about 0.66nF for each single stack.

The resulting cap is very compact, for 8nF it is just 3cm thick, and A4 sized.

It IS tedious to make, but if you really use all 10  intermediate sheet layers, and 1" clearances along the edges, you'll have no corna worries even if there are tiny air bubbles trapped between some (or most) foil layers. No vacuum techniques necessary.

The voltage accross each single sheet in a stack is about 2kVDC at peak, 800VAC normally. The sheets did stand 8kVDC and pulsed 180kV tiny tesla coil output very well, but the lower the voltage over each sheet is preferrable - less corona problems for one, and the risk for HV punch-through is greatly reduced (transp. sheets generally are not high quality and high purity plastics...)

The container in which the cap was built inside from start on is made from plexiglass. It is easiest to assemble the cap right into the container.

After finishing, pour out excess oil and place a heavy weight evenly over the cap, let it stand a few days, draining oil out in between. In the finished cap there will be no liquid oil left that could make a mess of the cap. 
All oil is as a thin "capillary" layer between the transparency sheets, making them almost impossible to separate. Additionally, vegetable raps oil and also lin oil dry out in a few days, glueing together the sheets on their edges.

With the sheet material I got and because of only short TC operation times, there was no heating due to RF losses inside the cap, so excess oil for cooling was unnecessary.


primary coil:  
connectors: to cap and to SG: standard HV wire from the same neon sign construction yard as the NST was from

other: no broad copper or aluminium strips yet, instead HV wire and partly also thick non-HV litz microphone or audio wire (cap to primary tap connector).

clips: standard big rounded end clips.


grounding: none, aside from NST center tap going to mains ground


secondary side:

coil form: - 1.4" diameter cardboard roll (left over from the household alu sheet rolls)
- covered with additional 0.4" of cardboard layers, sometimes a layer of (poly)urethane varnish and wax paper (brown baking paper for the oven).
- on outside coated a few times with urethane varnish, the inside is also coated
- center filled up with candle wax (~3 candles)
- NO wires led inside the secondary


wire: AWG is close to #27, it is 0.35mm wire diameter


winding: ~650 turns by hand, no extra spacing between turns
top: About 0.75" clearance between uppermost turn and a small round metal plate, which is mounted on top.
The plate has a screw sticking up through it, inteded for mounting a toroid or other terminal.


topload: various toploads for experimenting:
1 - round lid of a tin can, small pieces of cardboard roll sticked around it so it looks toroid-like, then covered with aluminium ducting tape.
2 - "oversized" toroid: air duct hose with 8cm diameter, wrapped to toroid with inner dia of 8cm, covered with alu duct tape. Tin can center for mounting the toroid on the secondary.