Hi

A self-oscillator circuit based push and pull MOSFET and the concept is very similar to the Royer Converter as shown below. This circuit comes from MIT textbook where engineer experimented with wireless power transfer. In fact I have no interest in wireless power transfer but I am interested to understand the circuit but I am struggling at the moment. I would be grateful if you provide any explanation.

As far as I know that two MOSFETs need to generate square waveforms (PWM) which come from the current within the the centred tap transformer or two inductors when powering at DC voltage.

I am struggling to understand what is the use of the diodes connecting to the gate of MOSFET?

 

Looks like it operates in LC resonant mode.

The diodes allow the cross-coupled FET's to pull the opposing FET's gate voltage down to shut it off. The resistors bias BOTH the FETs in the ON state, which is an unstable condition- they cannot both be ON because this would turn them both OFF... it becomes bi-stable, the resonance of the LC tank sets the pace.

When the resonant cycle reaches a point where the drain voltage of the OFF state FET begins to drop, it sucks the bias voltage of the opposite FET away through the diode, turning it off.

A very interesting circuit!

 

I have tried to simulate the circuit using SIMETRIX and I could not get any useful waveform out of this circuit as both MOSFETs remain on state hence not producing PWM waveforms. I could not understand why. Does anyone have tried to simulate this ciruit before?

Dumbdummy77

 

Well I simulate this Zero Voltage switch circuit in LTspice and in SIMETRIX.
See the attached zip file

 

I have tried to simulate the circuit and I could not get any useful waveform. I could not understand why.

Simulators are not perfect. Sometimes, they aren't even very good.

 

I use LTSPice. With self oscillating push pull circuits, you may have to introduce an imbalance in the currents (e.g. a low value series resistor) in one of the sides of the push pull circuit in order to start the oscillation. Otherwise, since the parts are modeled as ideal parts, the oscillations can't start.

 

The MOSFETs are n-channel, that is they need positive pulse to be turned on. This means that we need positive charges at the gate, the diode shows the conventional direction of current so they let positive charges go to the gate only or in reality they let only electrons to leave the gate of the MOSFETS.

 

I use LTSPice. With self oscillating push pull circuits, you may have to introduce an imbalance in the currents (e.g. a low value series resistor) in one of the sides of the push pull circuit in order to start the oscillation. Otherwise, since the parts are modeled as ideal parts, the oscillations can't start.

You can create an initial imbalance by setting an initial condition, e.g.,

.ic i(L1)=1u

You may have to play with the value of the ic to get a decent start-up waveform.