QCW Tesla Coil Gun

Why?

I started this project at the end of High school as a sort of personal capstone project. While I am sure there is no functional use for a tesla coil mounted into a backpack, I have learned so much while undertaking this project and I hope you will agree the finished project is suitably cool.

Tesla coils fascinate me because they combine so many electronics disciplines and also physical construction challenges. From RF to Power electronics this project certainly tested me and (even though it has taken me nearly two years) I believe it was fully worth it.

Some Process Photos

I took so many photos while this project developed, I have attached a few of my favorites below

The finished Hand Piece

After 3 different revisions I landed on this combo of laser cut, 3d printed, pieces. The handle contains the control electronics and hopefully the careful grounding makes this as safe as it could be

The State of the Lab

Towards the end of the project the "lab" underneath the bed in my dorm got a little out of hand. Testing was not easy in the space because I could reach both walls while standing in the middle of my room.

Early Testing

Before anything even got put into the pelican case there was so much troubleshooting. I like this photo because at this point it felt impossible that it would ever be finished

Taking Form

At this point I was feeling more confident and the project began to look like I had imagined it would for so long.

Resonator

4" by 5.5" coil form. 28awg magnet wire. Internal Capacitor string made with series mlcc capacitors each has a capacitance of 10pf, eventually replaced with a ferrite core. Topload originally made with ware and laser cut wood but replaced with a 3d printed for wrapped in aluminum foil

Battery

I'm using two 9600mah 11.1v lipo batteries to power the entire thing. They are directly soldered in parallel to a BMS system and charger. This assembly feds through a switch and fuse (maybe a relay as well in the future) then is put into an off the shelf flyback converter to feed the inverter the nominal 400vac the inverter will run at.

Ramp Generator

I used an Teensy 3.2 to control a simple ramp generator. The ramp generator must create a pwm ramp, around 20ms in duration, to control the inverter board. I designed it all to fit within a cavity in the handle, it uses a button a the trigger and transmits via fiber optic to the inverter board (the signal has to run parallel to the inverter output cables so regular wires would not work). The pulse amplitude is adjustable via a potentiometer. The entire thing runs off of some AAA batteries and is coded in C++. Eventually I added a 9v battery plug and a voltage regulator, as well as a power on light and a pulse indicator light

Inverter

3x 2200uf 450v DC bus capacitors. Input from flyback converter for 400v nominal bus. New generation SIC MOSFETs by ONSEMI. PCB designed in kicad and produced by JLC PCB. Designed for minimal parasitics and adequate high frequency operation. I had issues (noted in the troubleshooting section) with some gate traces running too close to the inverter output, raising the gate drive board and running wires (I removed the pcb traces for gate drive) fixed the issue

Trouble Shooting

And Lots of it

Current Inducing Fet Turn off

Capacitor String Problems

I had so many capacitor strings issues that eventually I decided that to scrap the idea all together and replace them with ferrite cores within the secondary to increase coupling. Every few pulses a trace on the capacitor string board would explode and the resonant frequency would shoot up due to the drop in capacitance. There were other issues with standoff distance. I think that while capacitors could be used effectively my implementation was not going to work.

This problem was almost the end of me. The magnetic field of the output of the inverter was enough to pull the gate of the upper MOSFET low for a fraction of a second whenever the inverter switched on. The actual problem was traced to a PCB trace that I had haphazardly run too close to the RF output of the inverter. Shielding this trace fixed the issue. But it took hours and hours to discover what the issue actually was

Finished Product (coming soon)

While the cool is not yet done click the video link below for low power testing