UHVLab DRSSTC I

4th April 2021

DRSSTC I

Introduction

I started this project in august of 2019 and it was my first DRSSTC and not just that, it was also my very first Tesla coil. During the construction i learned so much about high power electronics and about electromagnetism in general. And since this was the very first coil i've ever made i've done so many mistakes and destroyed many IGBTs in the process.

The Struggle

In total i destroyed the bridge 4x before finally completing the coil. My first destruction came when i first tested the coil at higher power levels without any ground strike rail. A spark from the secondary arced to the primary coil and IGBTs got fried. This happened in the september of 2019. Another fail just 2 months after this came when i was pushing the IGBTs too hard. At this time i had really no clue about what is too much and what is okay for the transistors. I was using ontime of 200us without any overcurrent protection! My estimate is that those poor TO247 IGBTs were switching roughly 600A at 218kHz. No wonder they didn't survive for a long time.

And this time it really was a loud bang. At least i got my lesson, 200us is too much and OCD is needed. It didn't took long before another explosion occured. This time the coil operated really well, OCD was working and i used almost full mains voltage to power it. I kept ontime at about 40us and it seemed like the coil is finally ready for full power. At the end of february of 2020 i wanted to measure the voltage at the bridge and i used home made twisted wire. What i didn't realized in time was that the insulation on the twisted pair wasn't rated nowhere near 300V and it broke down which shorted the bridge, destroying another set of IGBTs. And we're still not finished... Yet another explosion occured in the summer of 2020. My first test at full mains voltage. I used 440V TVS diodes on the bridge to protect IGBTs from overvoltage transients. It turned out that the transients were so long that they fried one of the TVS diodes on the bridge and the diode went short circuit, which led to another fail. I ordered 1200V IGBTs and 880V TVS diodes to have large overvoltage headroom and it seems like everything is working now even at input voltage of 250V+ which is even above mains.

Construction

The first thing that i got ready was the driver. I ordered UD2.7 driver board and all the components and started soldering.

I ordered parts for 2 boards as it's cheaper to buy components in larger quantities and already knew that this will not be my last DRSSTC. There were couple of small mistakes on the PCB, the footprint of the optical receiver was incorrect and led to short circuit on the board, and one of the resistor dividers on the board was missing ground connection. After correcting these errors, the driver worked immediately. I got my secondary coil from eBay, it had roughly 1680turns of 0.15mm wire. Topload was 3D printed and covered with aluminum tape. The resonant frequency of the secondary is approximately 240kHz. I ordered Tesla TC343 capacitors to make 6kV 42nF MMC which resonates with primary coil at 218kHz. For the bridge i used the same Tesla TC343 capacitors as snubber capacitors and custom designed a PCB for it. As a power supply for the driver i used Meanwell 24V 3A SMPS which i covered in metal enclosure together with the driver to prevent EMI problems. For DC bus capacitance i used 3 400V 3800uF capacitors in parallel. After 3D printing the case and some assembling, it looked like this:

Soon i started scoping waveforms which as i found out is not quite as easy when there is so much EMI in the proximity of the coil.

Low side gates

Low side gate and primary current, low power

Closer look at the voltage on the bridge during the switching.

Primary voltage and current at higher power levels. At higher power levels the EMI makes the reading almost unreadable

After pretty long time (1st April of 2021) i got the coil finally running at mains voltage. The project is not finished though. At mains voltage, the secondary insulation broke down and a new secondary is needed. Also i'm going to add some NTCs to protect it against the large inrush current. I'm using over 11000uF of capacitance! Inrush current to these capacitors is too large for any breaker and also for the bridge rectifier to handle. Until now i was using a variable transformer for the input. But for now i'm quite happy and satisfied with the result. The streamers from the coil are in the 60cm domain which is 2.5x the height of the secondary winding.

29th April 2021

Currently I'm in the process of winding new secondary.

With the new secondary in place and NTC inrush protection,

the coil should be finally fully ready for mains voltage without any problems.

New secondary, first ground strikes

A few weeks ago I have wound a new secondary coil. Previous secondary was 24x6.5 cm, consisted of 1680 turns of 0.14mm wire. I bought the secondary from eBay long time ago. This time I have wound the coil by myself. I bought 0.2mm wire for this purpose to wind around 1300 turns on a slightly higher secondary. The idea behind thicker wire and less turns is that the resistance of the secondary will be way lower. And it is! Previous secondary had resistance of 400ohms, new secondary has only 160ohms. The lower the resistance, the higher the Q factor, which leads to higher voltages if properly tuned. The wire I used have a double layer of enamel insulation which should prevent any breakdowns.

New secondary winding

Replacement of the secondaries

New secondary in place

And the Q-factor trick works! The sparks are now pretty much the same length to secondary height ratio as before but this time the secondary is a bit higher, so the sparks are even bigger than before even with 400 less winding.

No more variac, NTC it is!

Since the coil is already ready to run at full mains voltage, it would be silly to carry the 30kg variable transformer just to prevent the inrush current. I made a very crude NTC inrush protection circuit. With 6 large NTC thermistors and capacitor dropper power supply for a bypass relay.