## DL4YHF's Homebrew Homepage: High-Voltage Nonsense

Main site: http://www.qsl.net/dl4yhf/high_voltage/index.html
Backup: http://www.mydarc.de/dl4yhf/high_voltage/index.html

Last updated: December 2020

## High Voltage nonsense (experiments with no serious purpose)

Having build several Tesla transformers (spark gap and "solid state") and oscillators with TV flyback transformers with 2N3055 transistors some decades ago, I wondered about the efficieny obtainable with modern MOSFETs in a zero-voltage switching configuration (ZVS). A very simple but efficient configuration is the Baxandall oscillator (which LOOKS similar like a Royer oscillator, but the principle is different. A Royer oscillator's frequency depends on core saturation, a Baxandall oscillator uses a resonant tank as the primary).
I didn't want to build yet another directly-mains-powered "QRO" generator but something adjustable, fed by a current- and voltage limited switching mode lab suppy (1-32 V DC, adjustable from almost zero to 30 Amps). When powered with only 20 V DC, current limit 7 A, the ancient TV flyback transformer was good for an impressive 'Jacobs Ladder' display:

If the browser refuses to show the above video,

The DC current drawn by the ZVS resonant oscillator (again, Baxandall principle but using MOSFETs in a 'Mazzilli Converter') can be seen on the lab power supply in the background.

Baxandall oscillator for testing several old TV flyback transformers
for their 'power handling capabilities'. Requires a current-regulated
power supply with a fast voltage rise on power-on, otherwise this
simple circuit may not start to oscillate and fry the MOSFETs instead !

The DC input current depends on the length of the arc: The longer the arc, the more voltage and thus power is consumed. The AC current through the arc is fairly constant. With the transformer's secondary shorted, the DC input current was 1 Amp (at 20 Volts), without load on the secondary, it was also just 1 Amp, and just before the arc broke (at the top of the Jacob's Ladder) the input current rose to 7 Amps, or 140 Watts DC input power. The power supply's digital current display (seen in the background of the video) was too slow to catch the current peaks, only the red LED indicating 'constant current' instead of 'constant voltage' reacts fast enough.
For such a simple circuit as the Mazzilli Converter, the power efficiency is suprisingly good. The MOSFETs remained cold, the only losses occurred in the gate pull-up resistors (56 Ohms in my case, since I wanted a fast voltage rise on the transistor gates) and in the tank capacitors (3 * WIMA-FKP 150 nF in parallel). Also (but this was expected) the old TV transformer's high voltage secondary coil (without internal diodes as seen in "modern" TV flyback transformer) got a little warm after a 10-minute run driving the Jacob's ladder.
The secondary coil had approximately 2600 turns (calculated from the measured voltage ratio between primary and secondary), and a DC resistance of 1 kOhm. You don't want to push 0.1 amps (for a really hot arc) through such a coil, so how to turn this beast into 'Frankenstein Monster Lab' equipment ?
A different secondary with a similar number of turns but a significantly lower loss resistance was needed.

A nice effect can be achieved by covering the electrodes of the Jacobs Ladder with a solution of salt - the sodium chloride gets atomized by the arc, and the sodium gives the flame an intense amber colour:

Jacobs Ladder as 'electronic firework' for the window sill.

To achieve the amber-coloured arc, the steel electrodes were painted with a saturated salt solution (in water). Wait for the 'paint' to dry before turning on the arc. The video was taken at 32 Volts * 6 Amps = about 200 Watts DC input power.
Other flame colours can be achieved by covering the electrodes with Strontium chloride (intense red) or copper chloride (green), but this is definitely not recommended for indoor operation.

The current in the 'TV flyback' secondary caused it to heat up quickly, so this isn't good for a 'permanent' installation.
To make the transformer suitable for unattended operation with 100 percent duty cycle, without fearing a 'secondary melt-down', a homemade secondary was planned.

... to be continued - or not :o) ...