My First Tesla Coil – MOSFET Version

In my previous post I discussed my build of the slayer exciter Tesla coil and provided a few tips to make the circuit a little more robust. This post is a follow-up with a more advanced and much more powerful version of the slayer exciter.

This circuit attempts to address a couple of the issues with the slayer exciter:

1.) The transistor in the slayer exciter dissipates a lot more heat than necessary due to the fact it spends a lot of time in the active mode as opposed to the saturated mode.

2.) It can be difficult to get enough (or the right) current gain for the slayer exciter to work correctly when using higher current bi-junction transistors (Such as the 2N3055)

It’s possible to address these issues by changing the slayer from a BJT transistor to a MOSFET and introducing a MOSFET driver.  The MOSFET driver rapidly transitions the  transistor from the off state, to the saturated state, without spending any time in the middle. This solves problem number one in the slayer circuit, and means we can deliver much more power to the coil with a lot less heating.

Schematic

 

As in the previous post, D3 and D2 need to be Schottky diodes, otherwise they may not be fast enough to prevent high voltages from damaging  U2. It’s also very important that D3, D2 and C3 and C4 are located as physically close to U2 as possible. Don’t skip on these capacitors! The driver will run very hot if you do!

U1 provides 9V to power the gate driver and may require a small heat-sink. Nine volts is chosen here as  Q1 will be saturated when its gate-to-source voltage is around 9V.  Charging the gate to a higher voltage will reduce the switching speed, and increasing the heat dissipated by the MOSFET driver. Charging the gate to a lower voltage on the other hand increases the turn-on resistance of the MOSFET and causes dissipates more heat there instead.

Q1 requires a honking great heat-sink, the bigger, the better. An old CPU heat-sink can work well here.

When selecting Q1 the main attributes to look for in the datasheet are:

1. Low Rds(on) resistance.

2. A maximum drain-to-source voltage (Vdss) of about 3 to 4 times the supply voltage

The IRF540n satisfies both of these criteria nicely:

1. Rds(on): 0.044 ohm
2. Vdss: 100V

If you don’t have a suitable MOSFET you can probably salvage a suitable one from an old PC (or other) switching power supply.

When selecting U2 the main attributes to look for are:

1. Switching speed, ideally 40ns or lower
2. High current, ideally 3A or more.

I’ve found in practice it’s possible to run MOSFET drivers in parallel, this seems to help them switch faster and run cooler. To do this you simply need to tie their outputs and inputs together. However ensure that each driver chip gets its own set of 0.1uF and 4.7uF capacitors.  Scaling up the circuit to a much more powerful version should be possible by utilizing parallel MOSFETS and gate drivers together.

An important thing to keep in mind with this circuit is it will only operate if the current is passing through the primary in the right direction, if your circuit fails to oscillate, first try reversing the connections on the primary winding.

Completed

Here I used dual-mosfet drivers in parallel as I only had 1.5A drivers available and they ran hot  on their own:

 

Videos of this circuit in action: