|back to main page|
"Stubby" Solid State Tesla Coil
Current performance are continuous 7" arcs to ground with the occasional one out to 8". The below photos are 6" to a strike rod.
5" arcs to ground
The first Tesla coil I have completed to my satisfaction, total project time from start to this point is approx three years. In the meantime, I got married, bought and renovated a house and started a family. I think that lets me off the hook for taking so long ;-)
This coil started out with a poor bridge layout which had tons of stray inductance in the bridge. Despite this, it achieved a 4" arc to ground before deciding enough was enough. An all too brief appearance (fzzzzzt pop) at the '04 Derby Teslathon (see Adam Horden's Site for photos) convinced me I needed to sort it out. So I ditched my old layout and custom control circuit and built Steve's PLL driver instead, adding a much better, low inductance bridge layout in the process. And there it lay, untouched for a couple of years, watching, waiting for that one lunchtime...
Short piece of cable company green pipe of 100mm diameter and wound with 0.1mm Grade 1 Enamelled Copper Wire (ECW). White Perspex end caps on top and bottom, sealed with hot-melt glue. One coat of polyurethane varnish.
Two expanded polystyrene half-toroids purchased from local craft shop glued to a piece of card. Covered in self adhesive aluminium tape and glued to top. Note, heat generated from high power running is enough to deform this polystyrene and make it susceptible to knocks. Result: the toroid looks a little worse for wear.
Six turns of thick wire around base of secondary.
Half bridge made up of two ST STE70NM60 ISOTOP packaged N-channel MOSFETs (600V, 70A, 0.05 ohm) with built in back to back 30V Zeners on the gate. These were kindly donated by ST as sponsorship of this project. Across each MOSFET was fitted a Rectron HER508 high efficiency diode to catch any spikes.
Input decoupling provided by 3 x 330uf, 450V electrolytics from to ground with the half bridge coupling provided by a 68n and a 4n7 polypropylene foil in parallel on each leg. These were fitted with 100k resistors across each for stabilising the mid-point voltage. The mains input was passed through a 3A mains filter integrated into an IEC mains inlet, through a 6A MCB and then fed through a Fairchild GBPC3508 800V bridge rectifier.
Connections between capacitors and MOSFETs was all made on double sided copper clad board with the trace areas cut with a scalpel and then peeled off to form the tracks. The underside is the positive rectified mains and the top is the return. This was done to reduce the inductance between bridge and decoupling caps.
The MOSFET bases were mounted on a section of earthed 130mm wide aluminium U-section that provided heat-sinking and shielding for the control circuit within.
Looking at the electrical performance of the half bridge, the PLL phase was adjusted so that the bridge voltage (top trace) switches before the primary current (bottom trace) crosses 0A.
Controller + Gate Driver
I used Steve Conner’s DWSSTC PLL driver, modified for slightly higher running frequency by changing the 47k and the 470p in the PLL circuit to a 56k and a 270p. Other than this, the circuit was identical in nearly every respect. I can recommend this control circuit as being very easy to build and get operational. The phase adjust facility is particularly useful.
The MOSFET gates were driven as per Steve's circuit, with UCC37321/2 driving each leg of a GDT with 1N5819s clamping to the rails. The GDT has bi-filar wound secondaries enclosed in a metal braid primary for maximum coupling. I used 8 turns on an Epcos N27 toroid about 22mm diameter. The leakage inductance was good enough to not require any series gate resistance with acceptable leakage.
I used a 12V DC input from a bench supply initially, but moved on to a plug top power supply. I was initially concerned with the amount of power being taken - 350mA @ 12V when the coil was operating in CW mode = 4.2W - and that the UCC drivers would be getting hot. However, most of this current seems to be used for charging and discharging the gate capacitance, with the UCCs remaining fairly cool during CW operation.
Op amp based interrupter, 0-100% duty cycle, 80Hz approx interruption frequency, duty cycle adjustable using dial on outside of control housing. The circuit is based on this one.