Lately, I've been trying to build a ZVS circuit that follows the schematic shown below

Schematic of the circuit

However, when I actually attach a DC power supply (eTM-305A) that can supply up to 5A @ 30VDC, and crank is past a certain limit, power supply indicator switches from constant voltage to constant current and the voltage drops.
At first, after only around 8V, the power supply changes to constant current and the input voltage drops to 3V and the current draw becomes 153mA
After removing the transformer, the limit changes to around 15V
After removing the capacitor, the power supply no longer switches to constant current at all.

The thing is, I've done this exact circuit before and it worked perfectly. My question is, what is causing such a problem?

Attached below is the list of components
Schottky diodes: UF4007
Capacitor: WIMA MKP 0.1uF rated at 250VDC
Resistors: 100Ω 5W cement resistors and 10kΩ through-hole resistors

  • \$\begingroup\$ Time to bring out your oscilloscope and start probing! \$\endgroup\$
    – winny
    Dec 20, 2021 at 13:03
  • \$\begingroup\$ How come the original circuit shows a different MOSFET if it's meant to be identical? \$\endgroup\$
    – Andy aka
    Dec 20, 2021 at 13:15
  • \$\begingroup\$ @Andy aka I couldn't find a suitable LTSpice model for the components I was using \$\endgroup\$ Dec 20, 2021 at 13:22

2 Answers 2


I've built quite a few of these circuits when I was young. They often have trouble starting up; if they don't manage to get a clean start-up, they lock up with both FETs turned on. Try setting the power supply to 12V with a 2A limit, then plug the circuit into the PSU when it's already on. If it goes into current limit, pull the plug back out and try again.

These circuits need the supply voltage to ramp up rapidly to start oscillating. The voltage needs to ramp faster than about one quarter period of the LC resonant frequency, which you can achieve by connecting the circuit to an already-charged capacitor (1000µF or so).

An alternative method is to connect power to L3 first, followed by the gate supply voltage (R3/R4). The circuit will start oscillating as long as the gate supply ramps fast enough.

Even better would be a circuit that monitors the ZVS oscillator and cuts the gate supply voltage when the oscillation stops. Then you can just retry starting the oscillator with a button and without risk of damage. Such a circuit can be found here. This also solves the problem that these ZVS circuits tend to stop oscillating when the load on them becomes too high - and as soon as the oscillation stops, you blow up your FETs.

  • \$\begingroup\$ I set the power supply to 12V with a 2A limit and it went into constant current with the voltage dropping to 8V. There was a huge whining sound and now im facing the same problem \$\endgroup\$ Dec 21, 2021 at 0:50
  • \$\begingroup\$ Was that with or without the resonant capacitor? Also, was there any load on the transformer's secondary? Does the transformer have a ferrite core? If so, are you sure it doesn't saturate? \$\endgroup\$ Dec 21, 2021 at 12:53

It turns out that the problem was both a combination of loose connections and my inability to properly use the power supply. Previously, the power supply output current was only set to 600mA. After resoldering the primary side of the transformer AND setting the power supply to 12V@2A, it seems to work. I tested it up to an input voltage of 20V and it seems to have a static current draw of 1.1A


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.