Tesla Coil - Attempt II

This weekend I wound (by hand, of course) my second Tesla Coil secondary. This new secondary is wound with 0.28mm enamelled copper wire onto a 40mmx250mm PVC pipe and is much neater than my first attempt, a ~600 turn coil I wound on a multivitamin tube back in February.

The new coil has approximately 850 turns, possibly closer to 800 (I foolishly forgot to account for the enamelling). I originally wanted 1200 turns, but was agonising between wire thickness and available tube lengths. I settled for 0.28mm, I figured this would be easier to wind than the 0.2mm wire I used for my first coil.

Well, it was not easy, I reckon I spent in total about 7 hours of winding over two nights. I do have to say, the wire I used was from a reputable UK-based supplier on eBay, but was disappointed with finding several kinks as I was winding, resulting in small, but noticeable spaces between some of the windings, in contrast, the 0.2mm Chinese wire I got for literally pennies off eBay had no kinks at all! I really do not subscribe to the idea that British is necessarily always best or that anything from China is inherently bad, it’s just not true. And that’s why Maplin is no more.

Anyway, the new coil looks nice and neat, despite the hiccups. It’s currently taped up, I think I should give it a coat of (not oil-based!) varnish when I can.

Of course, I next had to test if the coil actually worked, so I hooked it up to my Slayer Exciter circuit running off a 19v power supply. I was somewhat disheartened to see the silent arc/corona coming from the end of the coil was not particularly bigger or more impressive than my previous coil! I measured the resonant frequency at about 1Mhz, I succeeded at least in making a lower frequency coil than my previous 4Mhz one. Argh, the Slayer Exciter just doesn’t seem good enough for what I want - I need to make a proper driver circuit!

I describe the Slayer Exciter circuit in my blog post ‘My first home-made ‘Slayer Exciter’ Tesla Coil!’, it’s incredibly simple, just an NPN transistor, a fast diode and a resistor - that’s it!

Slayer not very exciting…

Having fiddled with different transistors/diode combinations without much improvement, I’ve come to the conclusion that I need to make a new, better driver circuit, so to kick off I am thinking of building Electroboom’s Schmitt Trigger oscillator-based Tesla Coil circuit. I’ll either be making it on a protoboard, or even better, designing a PCB on EasyEDA.com/JLCPCB.

Schmitt Trigger oscillator for SSTC

I think I understand how it works - U2.1 is one half of the dual comparator, it creates a sawtooth waveform, the frequency of which can be adjusted by the potentiometer VR1 and the value of C4. This sawtooth signal is then fed into U2.2 and compared to the DC value of the potentiometer VR2 resulting in a varying pulse-width (duty cycle) square wave output to the MOSFET driver IC. Genius!

The circuit uses four N-Channel MOSFETs connected in parallel which drive the primary. I’m wondering if I could get away with using two or even just one, as Mehdi was using rectified mains voltage (170v) to his primary (which subsequently blew up his MOSFETs). I’ll be happy running mine off 24v as long as I get some half-decent arcs.

Before I can build this circuit, I will need a few components:

  • High speed dual comparator (MUST be fast)
  • MOSFET driver IC (DIP-8) - Must be fast and capable of high current
  • SOIC-DIP8 adapter board - Same footprint as a DIP-8 or not?
  • Multi-turn trim potentiometers
  • Solder
  • Electrolytic caps
  • Suitable heatsink(s), depending on which MOSFETs I use

The comparator Mehdi suggested, the MCP6562-E is only available as a SOIC package, I could use a SMD to DIP-8 adapter, but I’ve never soldered an SMD component before, plus these chips are about £3-4 each! As for the MIC4452YN MOSFET driver, they are about £4.99 each. Messing this up could be costly. I will try to find cheaper, suitable alternatives to these components if I can.

First prototype - first amazing arcs!

After the first turn-on, I quickly recorded a 9 second video on my tablet which I tweeted to Mehdi - and he retweeted me! I kind of wish I recorded a longer, better video now, also wish I demonstrated the CFL light trick too, d’oh! The coil interfered massively with the touchscreen of my tablet, so I had to keep my distance when recording it. I need a better camera/phone to record this stuff.

Something isn’t quite right!

Something weird is happening - the circuit seems to be self-oscillating without the oscillator part of the circuit even powered! I think my grounding isn’t right, should the ground of the MOSFET driver be going to the supply ground or the 5v circuit ground?

Next Objective - design a PCB!

Having had success with designing my first PCBs with EasyEDA/JLCPCB, I am now looking for a second PCB project and I think designing this Tesla Coil driver into a PCB is a great idea! Before that however, I need to perfect the breadboard prototype and work out this annoying self-oscillating issue.

Further info, watching & reading