How To Build A Simple But Powerful Flyback Driver

Picture of How To Build A Simple But Powerful Flyback Driver
Tired of little purple sparks? Want bigger hotter sparks? Then try: 

The ZVS Flyback Driver

It is probably the most powerful and efficient flyback transformer driver that was fairly recently invented by Vladmiro Mazilli. It uses resonant zero voltage switching (also know as ZVS) to drive the flyback transformer. This means the MOSFET's are designed to switch (on or off) when the voltage across them becomes zero.

Because the MOSFET's switches when there is no voltage across them, it will generate very little heat, the only source of heat is caused by the MOSFET's internal resistance. Unlike the simple 555 timer flyback drivers, The ZVS flyback drivers will allow you to run your flyback transformers for much longer periods of time before the MOSFET's overheat. If you get really good MOSFET's, it might be even possible to run your ZVS flyback driver infinitely! (Or until the circuit is interrupted)

Video of it working! 

Remove these adsRemove these ads by Signing Up

Step 1: The dangers of the ZVS flyback drivers

Picture of The dangers of the ZVS flyback drivers
Not only the ZVS flyback driver is powerful, but it is very dangerous. You can easily pump several hundreds of watts into the flyback transformer and the output current would be around 50mA to 200mA (or even more), which is way above the lethal rate which is 10mA.

Do NOT attempt to do this as your first flyback transformer driver project, I recommend you to start with using simple 555 timer flyback drivers before thinking about building an ZVS driver.

And finally, you are solely responsible for any harm to others or damage or any other problems that a ZVS driver may cause. The ZVS driver should be used for educational and research purpose only.

That is the end of my health and safety rant. :-)
vknezevic2 months ago

I have wondered how to plug in rectified 220V without burning out gates of mosfet. Then I found solution, wich is described on

somewhere on middle of page. People complaining that can't use more than 70v, and with solution on upper page cann be used any voltage

Dillon123 3 months ago

use these Infineon 17n80c3

hohenheime3 months ago
How can I tweak the circuit to avoid putting too much amps on the MOSFETs, I couldn't find better than irfp450 (14A)
I don't mean to be jumping your case, but I must advise all people who try this schematic. Any of you recall the color code provided for a 10k ohm resistor? It is listed as orange/brown/black. This is not a 10k ohm color code, as I checked it on a calculator, and looked at the color/digit significance. This color code would actually be for a 31 ohm resistor. DO NOT USE this resistor if you are seeking a 10k ohm.

The actual color code for 10k ohms is of the same colors, but it is actually supposed to be in the order of BROWN/BLACK/ORANGE!

Thankfully, I recognized this prior to taking out the resistors needed from a TV motherboard. Do not worry about the 470 ohm color code provided, as the publisher has it listed correctly.

You're welcome.

MadDoctor Whitson
I recall how you were wanting to know why the negative terminal gets hotter than the positive when arcing. I don't know if anyone has answered this for you yet, but I will provide my input according to my comprehension of current direction.

First, you must understand that all the electrons come from the negative terminal of the circuit, and the high voltage positive terminal acts much like an electron vacuum. This is because all positive wires lack electrons, and when they come in contact with a negative, or ground wire, the electrons are sucked out from its resource.

Negative/Ground wires can't shock you, as the electrons can't flow into any conductor unless it lacks electrons - like a live wire.

Due to the negative/ground terminal possessing so many electrons, and the live terminal having virtually none, the terminal with the most electrons will get the hottest, as there is more electron friction in it. The live terminal doesn't get very hot, because any electrons to make their way into it are quickly "sucked" away, thus the live terminal still lacks electrons, and this lack in electrons results in very little friction.

Seemingly, everybody thinks the current flows from the live wire, when it is actually in reverse. It somewhat bothers me why schematics aren't designed in reverse as well. Due to this, I once drew a Tesla coil schematic in reverse.
Cormaxzyz4 months ago
What is the amp hour rating on the batteries you use to power this? Also do the batteries have to match (that is, same manufacturer and same ratings)? I'm asking because I've already burnt out a 5 amp hour battery with a few high-current experiments and I don't want to burn out any more!
Techno12138 months ago
Hello! I recently took apart a microwave and i didnt get a transformer, i got this. It came with no capacitor. How do i power it?
13, 1:55 PM.jpg13, 1:55 PM.jpg13, 1:55 PM.jpg13, 1:55 PM.jpg
This Is a flyback MOT. you can use the zvs on this kind
There is a transformer on there, it's the thing with the curved black piece going over the coil. You have to take it out.

This driver will not work for your transformer, it's not a flyback and isn't designed for high frequencies.

There are plenty of MOT tutorials out there.
Liam.great989 months ago
(removed by author or community request)
Yes, it can be, i have a .56µF and a .22µF in parallel combining, .78 µF.
And it works fine.
jasshopper7 months ago
could i use irf460 instead of irf250?
astrikos1 year ago
this is a graph from the same circuit but now the 1microfarad MKP capacitor is replaced by a ceramic 10nanofarad trying to achieve higher frequencies

as you can see the half cycle is 1microsecond which means 500khz frequency

the problem now is that I do not get high voltage on the flyback

as far as I know the ferrite core goes up to 2MHz, may the problem be the capacitor which can not handle big currents so I will try a more powerfull capacitor

the diode may be another problem because it's not ultrafast

the FET is ok according to the datasheet for this frequency
XTL astrikos7 months ago
Its a resonant tank you are making with the coil inductance and capacitor. You are searching for the resonant frequency so just driving it faster doesn't mean it wil be more efficient. You need to balance the two out and drive it a frequency where the coil becomes resonant. Efficiency means the power will get through.
astrikos XTL7 months ago
I wanted to get high frequency and high voltage, not necessarily high power. I found out that this magnetic core works up to 160khz which is the frequency of the modern flybacks for big screens.
Mattef7 months ago
I think the reason why the negative terminal gets so hot, is that positive charged ions (Nitrogene+ and Oxygen+) are accelerated to it. The positive terminal gets bombed with electrons, but because they are less heavier, they don't generate much heat. Nitrogene and Oxygen are much heavier, so the "friction" (cross section) between the metal is even greater.
XTL7 months ago
The same driver can be used to heat any metal you put in the red wire loop (replacing the ferrite core flyback transformer). The cap value and number of turns (less) will need tuning to get the metal to heat up . The metal acts as a poor transformer core and resists the change in energy flow thereby disspating the wasted energy as heat. Generally a bad idea unless you want to heat metal up in an induction furnace.
XTL7 months ago
IRFP250 = 250V, 23Amps, RDS=140mOhm, Gatecharge(GC)=140nC
IRFP254 = 200V, 30Amps, RDS=85mOhm, GC=140nC
IRF540=100V, 33 Amps, RDS=44mOhm, GC=71nC
The 540 has a much lower voltage and might have had a problem switching, but its also likely the the Gatecharge (which is much lower) means you need a different tuned network. Try diff cap values.
arai1010 months ago
Hey! thanks for this tutorial.i have almost completed the flyback circuit.But i don't have a suitable 15V,10A PSU.Can i get 15V,10A from the mains only by using a transformer ,bridge rectifier and smoothing capacitor,please suggest me the circuit.Your help would be highly appreciated.
Use a computer power supply. They have over current protection, short-circuit protection & various other things. they can be easy to come by. don't bother with transformers for your power supply... cause you need at least 10 amps for this circuit. I have yet to come across a wall transformer that's rated at 12V @ 10amps. PC power supplies easily have at least 20 amps... not to mention various voltages... ( +3.3, +5, +12, -12V )
qwertyboy3 years ago
You can get the MOSFET's from Digikey for $2 each. Doesn't sound like that much money.
You could also consider going to texas instruments for samples. Don't order a crazy load of them tho. I've ordered many samples from TI & they're awesome. Priority shipping. Best of all, it's free!
Plasmana (author)  qwertyboy3 years ago
If you are going to make them go pop a lot, then it can get pricey.. :)
LOL, yeah, just be careful not to let the magic smoke out
Plasmana (author)  qwertyboy3 years ago
Yes, it is hard not to.. :)
a2262631 year ago
How do I find which pins are the inputs to the primary coil on the flyback?
This site ( says "With a simple ohmeter this is very simple, you will find easily the primary inputs because the coil resistance is about 1 ohm"

I bought a new flyback a couple of weeks ago. My flyback has 10 pins. I got these results:

Pins ..... Resistance
1 + 5 .... 2 ohms
1 + 9 .... 1.8 ohms
5 + 9 .... 0.9 ohms

2 + 8 .... 0.9 ohms

3 + 4 .... 0.8 ohms
3 + 6 .... 0.8 ohms
4 + 6 .... 0.6 ohms
All the other combinations (i.e; pins 1 + 3) produced no resistance, so that means that pins 1 + 3 are not connected. Conversely, this means that pins 1+5+9 are connected; pins 2+8 are connected; pins 3+4+6 are connected; and pin 7 and pin 10 are not connected to any other pins.

The first problem is that many of these gave me a resistance close to 1 ohm, so I still don't know which pins to use for the primary coil.
The second problem is that a flyback should have two sets of connected pins; pins that connect to the primary coil, and pins the connect to the secondary coil. However, I seem to have 3 sets of connected pins (or five sets, depending on how you count). Does this mean something is disconnected inside my flyback?

I didn't know what to do, so I just went ahead and connected it to the CFL circuit. I tried pins 5+9, 2+8, and 3+4 as the pins to the primary coil, along with every combination of the 4 pins on the CFL circuit, but did not get a single arc for any of them. Does this mean that my brand-new flyback is bad?
I had the same issue... IMO. don't bother with the flyback primaries. Just wind your own primary coils onto the ferrite core. once you have your ZVS circuit build & connected to your newly wound primaries, very carefully sweep your HV lead by all the other pins... you'll know when you found the pin for the secondary. You'll see a nice stream of HV. I like to solder a wire onto the secondary pin & do what ever then... Just be sure you discharge the coils!! being zapped by HV isn't fun lmao
build an easy circuit, such as mine and find the pins by arcing.
aadz9311 months ago
Hey I have some stw88n65m5 mosfets which are rated at 650v at 84 A max will these work?
Lol yes, those are more then enough, as long as they are mounted on a heat sink & are N-channel fets as the schematic calls for, it will work lol. I had used many different kinds of fets before. The worst that will happen, is they'll pop into smoke... ( witch in that case, you'd better have a fuse, or a PC power supply. or you're asking for an electrical fire lmao )
fatevian11 months ago
Hi everybody!
Someone can clear me a doubt?
It works if I put a HV diode to rectify? Do you think it's possible with this circuit, to supply high voltage to magnetron oven? What do you think?The current is strong enough to break down the barrier of diode?Thanks.
I doubt it. The voltage is too high and current is too low. MOTs output much more current than this. Also, most microwaves are around 1000 watts, and this ZVS only has 400 watt input.
Liam.great9810 months ago
Will these work? They're a dollar and eighty cents each!
darktwilite12 months ago

built this circuit and when i first plugged in a 9v battery, i could hear the high pitch from the flyback.. consequently though I didnt see any arcs and the high pitch tone changed when i tried touching the ground output from the flyback to the positive output. Randomly though the circuit suddenly worked and i saw some nice arcs, but a bit later the circuit was doing the same thing as before. does someone know whats happening?
Try a rechargeable drill battery, if tou don't want to buy lead acids
you don't have enough current from a 9v to turn the driver on, you need at least 10 Ampere to satisfy the driver. hope this helps.
astrikos1 year ago
A great caution to be taken.

When you start up the oscillator and before arching you should hear the 10khz sound.

If not then the oscillator has not started up and this means that on of the FETs is constantly ON and it will heat up to destruction within seconds.

This type of oscillator for some reason does not always starts up so you must stop the power supply immediately and start it again. This happens more often when you give low voltages, for example if you power it with 5 volts it's very difficult to get it oscillating but it is possible.

I managed to melt the case of the FET although I used heat sinks within seconds but the FET is OK.

I will device some method to be sure that the circuit will oscillate 100% if it is powered up.
astrikos1 year ago
this is the arching between the ground of the flyback and the high voltage output

the distance at which an arc can start-up strongly depends on the shape and the capacity of the electrodes

as you can see I have used two wood screws for not damaging the pins when they reach incadescence. between the tips of screws the arc starts nearly in contact only. between the bodies of the screws the arc starts at about 1cm.

I will use various other geometries as well as the Wimshurst machine electrodes which are spheres.

Conserning the sound you may hear when not arching, this happens beacause as you can see from the graphs the frequency reduces to one-fifth of the one when arching, which means that goes down to 10KHz, this frequency goes back to the power supply making the power supply to vibrate in this frequency which is an acoustic one.

I suggest that you don't run the ZVS without arching or anyway loading it some way.
a small correction, the 10KHz vibration and sound comes from the primary of the flyback, that is the two 5-turn coils and not from the power sypply, this is good beacause it means the choke toroid coil blocks the AC signal from going back to the voltage source.
astrikos1 year ago
yesterday I completed the permanent implementation of the Vladimir Mazili ZVS.

I attach two photos and two oscilloscope graphs of the two FETs. I used both channels of the oscilloscope, one for each drain of the FETs.

The low frequency and higher voltage graph is valid when there is no load/no arching on the flyback

When the flyback arches then the frequency increases and the voltage drops as you can see on the respective graph.

The destortion while the FETs go on and of may have to do with the diodes, I will try faster ones than the FR207.

Get More Out of Instructables

Already have an Account?


PDF Downloads
As a Pro member, you will gain access to download any Instructable in the PDF format. You also have the ability to customize your PDF download.

Upgrade to Pro today!