Looking for a good ZVS circuit

[FPS]

I require a ZVS circuit to use as a transformer driver, up to 1800W (60V 30A max).

I'd saved the attached schematics; it was from prominent members here, but I cannot remember where specifically they were saved from.

I realise I can buy one from Aliexpress, but I'd like a good base to play with and modify. What would be the most suitable circuit to fulfil this need? What changed could be made to achieve say 60V input - 20kHz output? Preferably something that'd work over a good voltage range.

I did try to make this one https://www.schematix.co.nz/forum/how-to-s/1-4kw-induction-heater but I failed, both MOSFETS just lock on and i have no idea why.

[FPS]

This is the failed one, the capacitors are mounted underneath. I used terminal blocks so I could swap out MOSFETS, inductors and capacitorsm while this isn't ideal it shouldn't affect its function to that extent, should it?

[davekni]

I'd saved the attached schematics; it was from prominent members here, but I cannot remember where specifically they were saved from.

That's one of my older ZVS oscillator versions.  Keeps gate pull-up resistor power dissipation low, but does not solve the startup issue that most ZVS oscillator (Mazzilli oscillator) circuits have.  Standard "solution" is to apply DC power rapidly, as in a DC switch on power supply output.  I have other past posts where I thought I'd solved that startup issue, but was just having luck, not a true reliable solution.

The one reliable (at least so far for multiple users) solution is to apply gate power prior to applying voltage to power inductors feeding output.  That way neither supply needs to ramp up quickly.

Other solution I've used is to turn on main inductor power first, then switch on gate power rapidly.  I use this method for switching on and off using logic control.  Not a universal solution, but works in at least some cases.

BTW, I have one other solution that I haven't tested yet.  That's to use depletion-mode FETs for feedback instead of diodes.

This is the failed one, the capacitors are mounted underneath. I used terminal blocks so I could swap out MOSFETS, inductors and capacitorsm while this isn't ideal it shouldn't affect its function to that extent, should it?

Image shows 9V and 14V on FET source leads.  Sources are connected to 0V.  Implies ECB and/or FET connector blocks are damaged.

[FPS]

I had two of them made via PCBWay, but they both have the same fault, just locks on 24V at both outputs. Circuit picture and schematic pdf is attached. The other is a bit more even at the MOSFETS (G/D/S - 12V/24V/10V) on both sides, but acts the same.

Yeh I have a 3 premade Aliexpress ZVS and they all have the problem of one side locking on if it's not switched via a mechanical jolt. The other attachments are ways I was looking to prevent that.

[petespaco]

Yeh I have a 3 premade Aliexpress ZVS and they all have the problem of one side locking on if it's not switched via a mechanical jolt.

What are you using for a power supply?  I don't know anything about the circuits in your first post, but I do know that you will need a power supply that can deliver at least 25 amperes or more for the 1000 watt "premade Aliexpress ZVS" that use 2 IRFP260N Mosfets.
Like this one:
https://www.aliexpress.us/item/2251832672195220.html

Also, make sure that your power supply is not set for current limiting below its max. capability.  If the power supply voltage drops much below about 12 volts, you will have problems.

[davekni]

Your circuit diagram clearly shows FET sources connected to ground (0V).  Your ECB must have an issue such as missing ground plane connections.

U1 regulator is likely too low current to power 100 ohm gate pull-up resistors.

FETs are obsolete 900V parts.  Not likely to work for this simple ZVS oscillator circuit due to high on resistance and likely high gate capacitance.

Not familiar with your PFET part either.  Can it handle sufficient current?  Can it handle high power dissipation during slow turn-on of your circuit?

[FPS]

Petespaco, the Aliexpress ZVS modules available all work perfectly well, running off a 48V battery through a 0-48V 24A buck converter. I just require a good "base circuit" to modify and play with.

Dave sorry for the confusion but that circuit isn't mine, rather it's just one that I'd saved for inspiration. I'm not sure if this circuit is a made circuit or just a proposed one.

 I guess IXTH140P10T/IXFH140N10T and the 2 x IRFP4668PBF would be suitable FET replacements.

[davekni]

Dave sorry for the confusion but that circuit isn't mine, rather it's just one that I'd saved for inspiration. I'm not sure if this circuit is a made circuit or just a proposed one.

No confusion.  You made it clear that you'd copied it from some previous post.
Yes, it is one I've made.  Does work and drastically reduces gate pull-up resistor power.  However, still requires rapid application of supply voltage as with simpler ZVS circuits.

[radbloke]

I made a ZVS which I’m directly driving the primary coil of a DRSSTC at resonance. Using 1/2 wave rectified mains, we get great ramped output. I have worked on this with Jerry Miller of ArcAngel Tesla Coil fame and we have put together a PCB for which I can give you a link to the Gerber files. The PCB has been tested and works great. The primary resonance starts when 12V gate voltage is applied. Then the ramps start when mains is added via a variac. Here is a link:
https://teslascience.wordpress.com/cw-tesla-coil-running-directly-off-a-zvs-driver/

[davekni]

I made a ZVS which I’m directly driving the primary coil of a DRSSTC at resonance. Using 1/2 wave rectified mains, we get great ramped output. I have worked on this with Jerry Miller of ArcAngel Tesla Coil fame and we have put together a PCB for which I can give you a link to the Gerber files. The PCB has been tested and works great. The primary resonance starts when 12V gate voltage is applied. Then the ramps start when mains is added via a variac.

Key point relevant for this topic is last two sentences above.  Applying gate power before main power is a reliable way to get these ZVS oscillator circuits to start oscillating.

radbloke:  Your coil would be good as its own topic under DRSSTC category.  I have other thoughts and questions, but don't want to clutter up this thread with an unrelated project.

[AstRii]

U1 regulator is likely too low current to power 100 ohm gate pull-up resistors.

Just to clarify, the circuit shown on the schematic is mine first try at high power ZVS driver. https://www.uhvlab.org/zvs-driver

The U1 regulator was sufficient to power the Gates, but from what I remember the circuit was running pretty low frequency of ~50kHz. These TO-220 package linear regulators can deliver plenty of current, but need to be cooled well. For Vcc>30V, I had to replace the LM317 with discrete regulator made with a TO-247 package transistor on a big heatsink.

FETs are obsolete 900V parts.  Not likely to work for this simple ZVS oscillator circuit due to high on resistance and likely high gate capacitance.

Indeed, I was working with what I had laying around, at input current of over 40A the circuit was latching up and didn't work.

Not familiar with your PFET part either.  Can it handle sufficient current?  Can it handle high power dissipation during slow turn-on of your circuit?

For input currents of 20-40A I ended up using 4 of those in parallel on a heatsink.

I need to find some time to put all of this information to the website..

Beside all of this, the circuit has never failed me to start-up reliably at various loads and input voltages.