Simple Micro SSTC

[Atomfusion]

Hello,

I am hoping you might be able to help me a little,
I have made a very simple SSTC based off Steve Wards Micro SSTC

I got rid of the antenna and the LM7812, and made the 555 timer variable, so that i can control its duty cycle.

I am blowing the FGA60N65 after only a few uses at less then 30 seconds a use, my problem is the Mosfet is not Hot or anything so I have no idea what is actually going wrong or how to add components / change to get it to last longer.

My question is can you give me any advice on what to look for in the oscilloscope or change to reduce the mosfet from blowing out ? Is there a more robust Mosfet to change to ? it runs at around 415 kHz

I have enclose just a picture of my current layout

A video of me blowing out my Mosfet with Mario audio in less then 2 seconds (this was only the second time that this Mosfet was turned on before that it was on for maybe 20-30 seconds with a 5 min wait (it was cold))

Any help / guidance you could offer would be great

[Mads Barnkob]

The antenna feedback is a very vital part of Steve Wards SSTC drivers, it ensures that the inverter is switching at the resonant frequency of the secondary LC circuit (coil inductance + topload capacitance), without this feedback to make the inverter feed current into the primary circuit at a timing/frequency where it can also be transferred to the secondary circuit, you will have conditions where no power is transferred and the current in the primary coil is so high that you blow up your MOSFET.

But since the MOSFET is a 650V / 60A and there are no explosions, but merely what I would call a silent death of the MOSFET, you must be killing it from overvoltage on the gate, at drain-source or maybe some MHz parasitic oscillations in your primary circuit.

Maybe you could try a RC snubber with diode instead of just using a single snubber capacitor across the MOSFET?

Anyway, you should redesign the circuit to use antenna or secondary base feedback and implement the duty cycle limitation in another way.

[Atomfusion]

Hello Madz and thanks for the quick reply,

I can add the antenna back and use 1N4148 diodes but i would like to understand how to protect the FGA60N65 better


E-Schmok has this around his Mosfet


What type/size Cap Resistor should C23 and R13 Be ?
does the resistor need to be a high voltage one ? Can i still use my .001 uf 2 kV
I have a couple different diodes can i just use the 1n4148 as D5,D6 ?


The LM4420 should only be putting out 9V as thats what i'm powering it with, I have been testing my Tesla Coil 4 inches from the electrical circuitry without any shielding could this be where im picking up the extra voltage from ?

I am using an digilent analog discovery for my Oscilloscope, but it has this annoying habbit of disconnecting from the Computer the second you switch on the Tesla Coil lol

Thanks again for your time

[E-Schmok]

Sers! The best way to find a good value for your snubber circuit is to try it out. try different rc-combinations and check them during operation on your oscilloscope. iam using around 56ohm/220pF for my single transistor coil. important: your snubber components has to withstand certain power losses. a better way could be a recovery circuit like the skory coil: http://skory.gylcomp.hu/tesla/mini_tesla2_sch.html

The best way is to optimise the design of your final stage to reduce leakage inductances. the final version of my single transistor coil don't even need a snubber anymore.

[Mads Barnkob]

digi-key has a small guide on RC snubbers, thats fairly simple and has a good we-do-not-care-about-RC-snubber-losses approach.

source: https://www.digikey.com/en/articles/techzone/2014/aug/resistor-capacitor-rc-snubber-design-for-power-switches



Where power dissipation is not critical, there is a quick design approach for the RC snubber. Empirically, choose the snubber capacitor Csnub equal to twice the sum of the switch output capacitance and the estimated mounting capacitance.

The snubber resistor Rsnub is selected so that

The power dissipation on Rsnub at a given switching frequency (fs) can be estimated as:

[Atomfusion]

I have changed the initial schematic to the way I am now setup.
My coil has about 20+ hours on the IGBT, Adding just the two Zener 18V diodes fixed my issue.
But then i updated to the UCC37322 IGBT driver that has an enable pin.

Voltage doubling works well but my Transformer seems to overheat (i have smelled it) can tuning help ? it seems to work at 83V with 1 to 3 wraps without a noticeable difference but at 43V it only worked well with 1 primary wrap


I have been working heavily on an Arduino interrupter that im going to start a new thread for, that can take a MIDI song and convert it to be played using the Arduino Uno or strongly recommended Arduino Mega