"Titch" Solid State Tesla Coil

Current best performance is a 2.5" streamer to ground running from a 30v bench supply.

Pictures

Coil still under construction

Movies

Coil still under construction

Secondary

Based on a small screw top white tub that used to contain some AVX (I think) capacitor samples. Wound with 521 turns of 0.063mm diameter Grade 1 ECW on the transformer winding machine at work.

Using the transformer winder was a revalation - a task that would have taken several hours and lots of broken wire was shortenend to a lunch hour. Seven secondaries were wound in that time as adjusting the wire diameter to get the feed rate correct was quite tricky. Overlaps resulted if this wasn't correct. The turns counter was also very useful as it allowed me to see how close I was to a 100% packing density on the winding.

The base of the secondary was at the screw top end of the tub, with the ground connection was made by a piece of copper tape on the screw threads. The secondary was then screwed into the cap, making contact with another piece of copper tape on the inside of the cap, which was subsequently connected to ground.

The secondary received two coats of polyurethane varnish to keep the fine wires from being damaged. It sustained some damage due to the antenna falling onto the secondary but this was cleaned and repaired.

Topload

A Dunlop single yellow spot squash ball wrapped in copper tape and hot-melted to the top of the secondary. Breakout was a small screw wrapped in copper tape pointing vertically from the spherical topload.

The copper tape didn't provide a particularly good finish, but it was satisfactory.

Primary

The primary was wrapped around the screw cap of the coil and consisted of two turns of thick wire. Secured in place with hot-melt, the bodgers friend :-)

Primary Driver

The power switching was provided by a full bridge of IRF540 MOSFETs (100V, 33A) mounted on individual heatsinks. These were connected to a piece of copper clad board which formed the connections for the full bridge. Anti-parallel diodes were MUR420s.

Power supply was made by modifying an AT power supply to give a variable output voltage up to 50V. Input decoupling was provided by two Nichicon 820uF, 50V low ESR electrolytics and a Wima 1uF PP film cap across the supply.

The MOSFETs heatsinks were attached to each other in pairs with M3 nylon bolts to maintain isolation between the drain pins. The top and bottom of the heatsink were tapped with some M3 threads to allow mechanical fixing to the lid of the PC power supply containing the control circuitry for the coil.

The screw cap lid of the secondary former was bolted onto the other end of the heatinks through some spacers.

CONSTRUCTION PHOTOS HERE

Controller + Gate Driver

This coil, like Stubby, is based on Steve's PLL control circuit and uses a pair of heatsunk Ixys IXDD414s to drive the GDT primary, with the outputs clamped to the rails with pairs of 1N5819s. AC coupling was a 330nF PP film cap.

A regulated 9V supply was used to power the circuitry. This was preferred to a 12V supply as the supply current is determined by 0.5 * C * V^2 * freq. Essentially, the higher the supply voltage, the higher the currents.

To get inverted but synchronised gate drive signals to the Ixys drivers from a single square wave could have been done using an inverter. However, that introduces a delay into one of the drive signals which can lead to cross conduction.

Instead, two gates from a CD4070 XOR logic IC (can also use CD4030) were used as below.

Delay through the CMOS gate is 22ns from input to output, but the outputs are matched to within 6ns on the rising edge and 3ns on the falling edge (measured at 500kHz).

GDT consisted of a Ferroxcube 3E6 25/15/10 toroid wound with 7 turns of penta-filar 0.2mm diameter Grade 1 ECW. The screenshot from my design spreadsheet is here.

An LM78L05 was used to provide a regulated 5V power supply for the PLL and the interrupter circuits. This is because the line regulation of the 9V supply, working over such a wide range of input voltage, cannot be guaranteed. Since the PLL oscillator uses the supply voltage as a reference, variations could throw the oscillator off enough to cause the bridge to be driven at the wrong frequency.

The control circuitry was mounted on the underside of the lid of the modified power supply. When running it on the bench open frame during intial testing it was found to be slightly susceptible to noise from the resonator causing the driver to cut out. Running whilst attached to the grounded metal case removes these problems.

Interrupter

Comparator based interrupter, 0-100% duty cycle, 50Hz approx interruption frequency, duty cycle adjustable using dial on outside of casing. Some experiments with interrupter frequency were made, with the selected frequency resulting in longer straighter streamers.