5th April 2021
When i first thought of making a DRSSTC i wanted to do it big style, without really thinking i started winding the large secondary coil in the excitement. I wound around 2500 turns of 0.35mm double enameled copper wire on a 100x16 cm PVC pipe. Very soon i realized that it's not really a good idea to make such large coil as my very first Tesla coil. I wound the secondary in April of 2019 and after a month or so i put this project aside and started working on my DRSSTC I. When i got quite happy with the results of DRSSTC I, i resumed my work on this project.
The secondary coil
I started winding the coil at 1st of April in 2019 and it took me 3 days to finish (20hours of work). I tried my best but i still made many imperfections during the winding. Right now i'm hoping that the double enamel insulation on the wire will protect the windings from any flashovers.
In the summer of 2020 I decided to select a driver for the coil. By this time i just found out about a nice OCD feature called pulse-skipping. With pulse-skipping instead of turning the H-Bridge off when an OCD threshold is hit, it only skips primary cycles until the current drops below the threshold. This way the coil can continue to oscillate even when the current is high and it prevents the current from climbing even higher. It's a clever way to push more energy into the system and making the streamers longer and hotter. I searched the web and found Philip Slawinski's website who is offering his UD+ driver with pulse-skipping feature for very reasonable price. You can read more about UD+ on PhilsLab web. Soon I received the driver and it worked pretty much "plug-and-play".
Primary coil was wound with 10mm copper pipe, total of 8 turns were wounded on a custom 3D printed primary holders with the help of my brother David.
Primary coil platform
3D printed primary holders
My brother's children helping with winding of the primary :-)
For a topload I custom 3D printed 7 20mm rings and covered them with aluminum tape. With the secondary coil i got a resonant frequency of around ~64kHz
Secondary coil with printed topload sitting on a platform with primary coil, next to my DRSSTC I
For MMC I used 14S2P combination of MPAPB1K2W22J0M6 2.2uF 1.2kV snubber capacitors. They seem quite robust and their mounting with screws is perfect for high current connections. I also used 2 of them in parallel on the H-bridge as snubber caps. The holes for the screws were perfect fit for my IGBTs.
Finished MMC on a custom 3D printed holder
The same capacitors used as snubber caps on the H-bridge
I have bought Semikron SKM400GB12E4 IGBTs from eBay for around 30$ a piece. I was paying attention that the seller is a verified and trusted seller from anywhere else but China :). I have found these IGBTs from a German seller with great ratings, so they are hopefully not fake. I have measured their gate capacitance and they have around 27nF, which is pretty close to datasheet value. So far so good. As a DC bus capacitors i used 2S2P combination of 450V 10000uF capacitors for total of 10000uf at 900V. I used 22kOhm 50W resistor across the primary terminals for MMC charge draining. The gates of IGBTs are charged through 4.7ohm resistors and discharged via schottky diode and internal 1ohm resistance. I also placed 27V TVS diodes across the gates to suppress possible voltage transients.
Close up at the SKM400 terminals
IGBT with gate charging circuit
Heatsink with single halfbridge installed
Almost finished bridge
At these power levels we need better and more efficient inrush current handling than a simple NTC. I designed a simple circuit which uses 24 8.2kOhm 10W resistors in parallel for total of 340 Ohms 240W for charging DC bus capacitors. After around 20seconds, this circuit switches relay which bypasses the resistance with a short circuit. This is way more efficient than a simple NTC. But you have to pay attention, the coil cannot be run while it's powered through the resistors! That's why a 5mm green LED was added to the front panel which indicates that the relay has switched. Soft-start circuit together with the UD+ driver are powered with Meanwell 24V 5A SMPS.
Soft-start circuit using 2 relays in parallel with 3phase bridge rectifier mounted on a heatsink
Finished soft-start circuit mounted on custom 3D printed front panel
Inrush current into a single 10000uF 450V capacitor charging to only 120V. 20A/div, measured after full wave rectification
Schematic of the soft-start circuit (without the indication LED)
Gate Driving Transformers (GDTs) and primary current transformers (CTs)
Unlike GDT and CTs in my first DRSSTC, this time i gave it some thought instead of just randomly picking cores and guessing windings count. I did some calculations and ordered large ferrite cores with low enough permeability so they can provide a lot of current. I picked large cores so they do not reach saturation even if they ran at 30kHz (20kHz below planned operating frequency). GDTs are wound with twisted wires (2 in parallel for primaries), and they have 10 turns. Feedback primary CTs consists of two cascade transformers 1:33:30 (1:990 in total).
Ferrite cores for GDTs and CTs
1:33:30 cascade transformer
Unmounted CTs for feedback(black) and OCD(red)
Finished coil and first light
After a while I assembled everything together and started first testing. I captured some waveforms in the process with the primary tapped at 7.5 turns.
10th August 2021
Low side IGBT gates
Primary current @54kHz, 100A/div, 350us ontime
75cm discharge at 110VAC input voltage, 350us ontime, 350A OCD
~80cm "ground" strike
200VAC input, >3kVA and first explosion
After some time I finally ran the coil outside just to be left with very small <10cm sparks. Everything was set up the same way as during the inside testing where the coil was running fine, nevertheless outside the coil performed poor. I took it back inside and everything was working the same as before. I checked everything I could and made really solid ground connection for the secondary but still, outside the coil did not wanted to work well. It took me a week to realize that inside the capacitance of the topload is much higher, because it is closed room with ceilings full or mains wiring and earth ground which increases the capacitance by a great deal. I found out that outside I needed to decrease the windings of the primary coil from 7.5 to 6.1 turns.
After fixing the issue I finally got the coil to spit out over a meter long arcs with a modest 400A OCD setpoint at ~180VAC input.
Catching arcs at 350us ontime, 180VAC input, 400A OCD (>2kVA)
DRSSTC II and DRSSTC I duo
14th September 2021
During my duo Tesla coil run something went horribly wrong. A loud brief explosion can be seen at the end of the following video
Thankfully no transistors were damaged. What actually exploded were the input power wires!
While I was running the coil I totally forgot about the power rating of the input power cables. This wire was not rated nowhere near 15A which I was pushing in to the coil. The insulation on the wires melted and then they touched each other, causing a short circuit.
Coil survived without any harm. Thankfully I managed to capture the explosion on a video.
You can see it here frame by frame: