I've read the Vout of a resonant inverter is typically lowered by increasing the oscillator frequency, pushing it above the resonant frequency of the LC tank.

Instead, why not lower Vout by reducing the amplitude of the square input oscillator, (and keep frequency the same)?

 

The power stage is not a linear amplifier.
The transfer function between the input and output looks more like a logic gate than an amplifier, massively non-linear.

If you tried, you would also find that the losses would become very high and the output transistor would fail.

 

As the input to the resonant tank consists of switching the power supply with MOSFETs, the input amplitude will be fixed by the voltage of your power supply.

 

The power stage is not a linear amplifier. The transfer function between the input and output looks more like a logic gate. losses would become very high and the output transistor would fail.

I should have said, vary the power supply voltage at the mosfet source (while keeping the gate voltage fully on / off).

As ian0 said -

input amplitude will be fixed by the voltage of your power supply.

 

I should have said, vary the power supply voltage at the mosfet source (while keeping the gate voltage fully on / off).

As ian0 said -

The voltage at the MOSFET source will be the voltage at the MOSFET drain i.e. your power supply voltage.
Yes - it would work if you could vary your power supply voltage, but it's rare that you get that option - power-factor correction front-ends is the the only exception.

 

it would work if you could vary your power supply voltage, but it's rare that you get that option

Unclear. Why "rare"? What's to prevent anyone from inserting a converter to adjust Vsupply?

 

If the resonant mode system is not able to vary the pulse width of the drive to the filter so as to control the output voltage, then it is a poor design indeed. Adjusting the pulse width is simpler than changing the supply voltage and is usually far more efficient. Duty cycle adjustment is much more controllable. Adjusting the frequency to vary the effectiveness of the filter is not a good idea at all.

 

Duty cycle adjustment is much more controllable. Adjusting the frequency to vary the effectiveness of the filter is not a good idea at all.

I've read the Vout of a resonant inverter is typically lowered by increasing the oscillator frequency. Have you built or seen a resonant circuit that uses PWM to vary output voltage?

 

I believe that I saw such in a Jim Williams article published quite a few years ago. And please consider that Williams certainly had much more insight into the subject than must others. (That fact is never challenged.)
Of course, adding control to a resonant mode system certainly makes it more complex, and thus more expensive to produce, and thus much less common.
And the short answer as to why not adjust the amplitude of a square wave signal is that adjusting amplitude is a linear control method and thus not as efficient. Just consider how would one adjust the amplitude? That is usually done in an analog arrangement, which implies resistive losses. How else would you do it??
The linear adjustment of the length of a pulse is quite effective, and creating the two pulse strings out of phase is fairly simple. And the response of a resonant circuit to a reduction of pulse width is fairly well understood.

 

Insightful, @MisterBill2

response of a resonant circuit to a reduction of pulse width is fairly well understood.

What is the response?

Thx!

 

Driving a resonant circuit with pulses at the resonant frequency, the amplitude drops as the pulse width narrows. This is because the power delivered is reduced. Just like dimming an LED with PWM. Te apparent brilliance drops as the pulse width narrows. Except that with an LED the exact frequency does not matter .