This is a showcase of the solid-state tesla coil project I’ve built, and yes, IT PLAYS MUSIC ☺ (WIP)
Ever since I visited the Nikola Tesla Museum in Belgrade in high school, and saw the tesla coil in person, I wanted to build my own version for myself. After all, it is pretty hard to make one and it is also a dangerous device.
I've always tinkered with electronics but never had a chance to build one until now. Although this one and the one in Belgrade differ quite a bit (besides the size 😄), the principle of the operation is the same.
Unlike the one in the museum, this one uses semiconductors instead of the spark gap.
I started researching what would be the most cost-effective way to build the tesla coil and a decision was made for it to be an SSTC.
SSTC uses semiconductors instead of the spark gap to switch the primary coil. The most significant difference is the absence of resonant capacitors (primary capacitors). Because of this, the SSTC cannot generate huge voltages at the output; however, it is much cheaper to build.
My plan was to look at the builds other coilers made and try something similar. I went for the proven design of the Steve Ward with a few modifications I did regarding the logic board.
After that, I started to do calculations using an awesome tool called JavaTC.
My plan was to build a platform that houses all the electronics and above the platform would sit the secondary coil. The secondary coil resonant frequency I was aiming for was 300kHz max (preferably below that).
The first thing I did was visualized everything in my head, how I should place the components and ECBs.
The second thing was to design the PCBs and have them manufactured. JLCPCB was chosen for board manufacturing.
I started with the logic board PCB and that was the first ever PCB I designed and manufactured. The logic board uses the UCC27425 driver to drive the FETs. The feedback signal is picked up by the antenna and together with the optical receiver, it triggers the driver which drives the FETs through the gate drive transformer.
I used the gate driver transformer to provide galvanic isolation between the logic and switching boards. The output of the GDT is around 18V. GDT is constructed using a ferrite toroid made of N30 material and a Cat5 cable. I twisted the wires together to lower the parasitic inductance which needs to be kept at a minimum at high switching frequencies.
The second board I designed was the driver/switching board. It consists of the gate circuitry and two FETs in a half-bridge configuration. I decided on the half-bridge because it is cheaper (it uses only two transistors) but it has lower performance. The bus capacitor is a large electrolytic capacitor with screw terminals (1000µF/400V) which is directly fed with full-wave rectified 230VAC (Red and blue wires).
One of the most painstaking things when building the tesla coil is the secondary winding. The problem with the secondary winding is that it needs to be made with hundreds of windings of very thin copper wire, in my case that is around 1030 turns of 28 AWG wire.
I made a little contraption to ease the winding process. It consists of a simple holder which holds the motor of the cordless drill and it is controlled by the PWM controller. I made my secondary in about 3 to 4 hours with this device.
After the secondary was wound, I applied two coats of polyurethane varnish to the secondary windings to prevent corona formation and make the secondary sturdier. To install topload on top of the secondary I designed two parts in fusion 360 which were later 3D printed.
I chose toroid for the shape of the topload. On most coils today, toroid is the dominant shape for the topload instead of the sphere. To achieve desired resonant frequency of the secondary I needed to build the topload with specific dimensions.
Topload was built using wood, cardboard, glue, PU foam, and aluminium tape. First I made a circle out of wood, 300mm in diameter. The next thing I did was printed the toroid shape using the printer, glued it to the cardboard, and carved the shape with scissors.
 |
 |
I glued all the pieces to the wood and filled empty space with PU foam. When the foam hardened I shaped the toroid and taped it with aluminium tape.
After gathering all the components, I assembled the coil and it was time to test it. For the testing I made simple voltage divider with the incandesent light bulbs which could be switched to voltage divider/series resistor (to limit the current).
First test went okay but discharges were pretty low, my guess is that the coupling between the coils was too low. (Unfortunately I didn't took any photos of the first test).
Second test took place the next day.I raised the primary in an attempt to improve the performance and see what would happen 😄. Performance was a little bit better but still far from what I expected.
I wanted to push the coil a little bit harder so I played some random music from my phone and one of the transistors popped.
Dead fet
Dissasembled fet, die area (could be counterfeit 🤔)
Picture of the coil and it's electronics
Note the raised primary coil (with the help of the bricks 😂), feedback antenna to the left, and piece of wire on top of the topload which helps with the breakout.
After the coil was out of function because of the blown fet, from the tests, I knew I needed to increase the coupling of the coils. Characteristics of the new primary:
- 8 turns
- ~4.78uH of inductance. Much lower than the previous one, allows more current through the primary.
- Coupling coefficient increased to around 0.4.
One more change I did was to change the IRFP460 I started with with the K60H603 IGBTs.
NOTE: The frequency of this coil is pretty high for an IGBT. This IGBT has a much higher gate charge than the IRFP460 so my driver is working harder to drive it efficiently, IGBT heats up much more than the MOSFET. So this is not the ideal configuration but I chose the IGBT because I wanted to push more current through the primary coil which IRFP460 probably couldn't do for a very long time. The primary peak current should be below 30A with the new configuration
Side-by-side comparison of the primary coils.
After everything was assembled once again, I started testing the coil.
Performance was much better, even better than I expected.
I took a couple of photos and most importantly, a couple of videos 😁
Probably the most interesting thing of this coil is the fact that it can produce music.
Note the primary wires in the first picture below. They are pretty long in my case and that should be avoided to reduce parasitic inductance.
Checkout my channel to see the videos of the coil in action. https://www.youtube.com/@bojandolic4820/shorts