zvs with dual mosfets

depends on how good those resistors and zeners and capacitors are. the top side gate resistors are already burning two watts of heat while idle. 10V is really too low for a gate voltage upper limit, and 40V supply is really pushing the upper limits of the mazzilli design. when switching, those drains are going to see about two or three times the supply voltage, so the opposing side gate diode has to be fully conducting very early to ensure none of this hits the gate by capacitive coupling across the junction before the zener can get into reverse conduction. esr on the capacitor has to be very low, as it heats up it will increase, increasing reactance and thus increasing the drain voltage when the fet switches off. irfp260 has a Vgs turn on between 3 to 3.6V cold, but drops down to 2V when hot. the miller capacitance between drain and gate can send voltage spikes back onto the gate terminal, turning it back on slightly, making it heat up, and because the gate is clamped to source there's no way for this circuit to pull the gate lower than the source to prevent accidental turn on. the mazzilli is a fun circuit to play with but once your supply gets higher than your Vgs max voltage and you start getting into silly power levels it gets very destructive very quickly; and this is all before you've even started making pretty sparks and arcs with it, which makes a whole new set of stresses on all of the components.

This circuit is fun untill the fets self destruct my last destruction was violent enough to let out the magic smoke and sparks Im using a 12v battery jump starter transformer it will do about 38v open circuit and supply 200 amps but im sure the voltage will drop down to about 10 volts since it is for starting vehicles I put a 20amp breaker on the output before the rectifier I tried a 30amp breaker but thats when a blow my zvs module for the second time its the module you can get off ebay for 33$ not a bad deal I was wondering if more mosfets would mean it could survive better if they lock on all 4 of the fets should knock out the breaker before they explode I know makerj101 on Youtube did a similar design Thanks for the input

fets in parallel is good in theory but zvs again does not have the drive current to switch them all on quick enough to prevent a failure cascade

your fets aren't blowing from too much current

I made the circuit but with 4 fets per pair it works fine on low power about 12 V but as soon as I turn up the voltage the wave form on the gates gets very jagged and rough so I need to redesign a better gate drive circuit One of my fets has a blow hole in it now so hopefully the others are okay

yes, you have discovered the limitations of the mazzilli zvs driver; it's a fun toy but not a good design for these power levels.

http://everycircuit.com/circuit/4919635115245568

see my example circuit above; even though gate supply is 10V, the Vgs voltage only reaches 6V. look at the current load on the gate supply too. The mazzilli circuit does not provide this gate power. it charges the gate passively from the load supply, then uses a diode to discharge it and sink the gate drive current from the supply through the bias resistor and down the opposite gate. if the load supply voltage dips, the gate won't get high enough to fully turn the fet on. if the fet doesn't turn on fully, the fet will not reach a low enough Rds-on resistance. if the fet has high resistance it won't pull the voltage down low enough to drain the opposite gate. the net result is fets on both sides of the circuit end up with gates just sitting right near their Vgs turn on threshold, not switching, their Rds sitting around 1Ω and they start burning off 20V * 20A as heat. couple this problem with all the other issues of high voltage spikes spraying across the gates as well and it just turns to disaster.

mazzilli circuit is still a great little circuit for learning about zvs though. just not a good high power design as its simplicity is also it's limitation. it doesn't teach you about the importance of maintaining supply drain and source resistance, managing fet capacitance and inductance, or why gates need isolated power supplies, or why gates need craploads of power; which all become relevant as you ramp up power from desktop toy to Fun Levels

Would it be possible to use a comparator circuit to tell when one side of the cap bank is positive and use a fet to switch the gates on and off so that way the gates get turned on and off fully

Don't worry faceblast, the mazilli zvs driver is just the most efficient flyback driver at the moment, and if well done even for other tipes of transformers