So here it goes, I will keep this issue as simple as possible. I have this Class AB amplifier over here, that brings me a 60 Hz sinusoidal signal up to aprox 15Vp. Must be said that the sinusoidal wave is being generated by PWM and a MFB 2nd order low-pass filter (with cutoff frequency at 350Hz).

Class AB Amplifier Driver Stage

I've been attempting to match this stage with a voltage multiplier through a transfomer. The thing is, I'm using this transformer as an elevator. For this, I've selected a common 120V to 24V transformer. Let me know if I'm wrong but, as long as the voltage connected to the secondary winding is lower than that one that's obtained when using it in common application (24Vrms), I will be able to get an amplified signal in the primary winding without reaching saturation, similarly, as long as I keep under the 120Vrms.

Having said that, I'm trying to use this transformer as a 1:5 elevator to avoid using several stages in the voltage multiplier. The current voltage multiplier is:

enter image description here

And the summary of what i'm trying to do looks like this:

enter image description here

What's the problem I'm facing then? All this arrangement must be getting me a 750Vdc output (mulitV reference connector), at least when the load impedance is over 10kOhms, but it doesn't, I'm not getting more than 7V.

enter image description here

Some notes:

  • C_ladder is the capacitance of every capacitor in the voltage multiplier. It equals 1000uF, 250Vrated.
  • Even I'm using a feedback stage to control the output voltage through frequency. The frequency can reach up to 150Hz (for now) to keep the voltage level.
  • The load that will be connected to this power source isn't even a simple resistive load, it's a Amplifier Class D with a LC filter before being connected to another transformer, so I don't even know if this approach will be enough to acomplish the task of amplify another sinusoidal wave and feed it to the transformer with the Amp class D.
  • I know if I keep the capacitors this large, the time to reach the desired voltage level will be huge, but I haven't found a way to keep the input impedance low enough so the frecuency control works with frequencies lower than 350Hz. If I make them of 1uF, settling time will be smaller but I would need a sinusoidal input over the 10kHz to reach the hipotetic 750V.
  • I am basing myself on this design that I found on the internet: https://demolab.phys.virginia.edu/demos/documents/PASCO_SF-9586.pdf
  • \$\begingroup\$ Have you measured the AC voltage present across the transformer's output winding? Also, the transformer winding ratio is 5:1. Shouldn't that be 1:5? \$\endgroup\$ Commented Jul 7, 2023 at 5:30
  • \$\begingroup\$ The green signal is actually the voltage at the output of the transformer. I forgot to say that the component that I'm using in the simulator is indeed one with a 5:1 ratio, but I have connected it backwards, this only to make it more in line with how it would be in reality, sorry about that. \$\endgroup\$ Commented Jul 7, 2023 at 18:02

1 Answer 1


Your problem is the amount of capacitance required for this type of approach. The voltage multiplier circuit capacitance has to be large to operate at 60 Hz, as you have correctly pointed out. Since you are using a 5:1 transformer, this impedance is reflected back through the transformer as and appears as five times the input capacitance, and you are coupling to the transformer through a much higher impedance (the 10 uF capacitor).

There is no free lunch: 750 Volts at 10K is 56 watts, so if your class AB amplifier could supply 28 volts, it must supply about 2 amps, even ignoring losses in your multiplier. But at 60 Hz, a 10 uf capacitor has an impedance of 265 ohms. In any case, the load capacitance reflected through the transformer is so much higher than this coupling capacitor that the circuit cannot supply sufficient voltage to the transformer.

If you try simulating the transformer and multiplier circuit by itself with a sine wave source directly driving the transformer (no coupling capacitor) and measure the current, you will see the problem clearly. This is the reason that voltage multipliers are seldom used at low frequencies or high currents. The component sizes become too large. A 750-volt, 75-milliampere switching power supply would be a much less challenging design approach.

Good luck!

  • \$\begingroup\$ Yeah, just wondering if this design could still work in a way. Thank you very much. I've been stuck on this for a while and the truth is that everything I now understand about design has been more by trial and error. I am not an expert in power electronics at all, so if you have any reference that could help me design switching power supplies, I would greatly appreciate it. The ones I find are very general explanations and with examples dedicated to voltages no greater than 100 volts. \$\endgroup\$ Commented Jul 7, 2023 at 18:28
  • \$\begingroup\$ Here's an example of a boost converter monolithicpower.com/en/… . It's much lower power than what you need, but the principle can be the same. You are not starting out with an easy one - your output capacitance will make this a dangerous circuit so I doubt you will find any reference circuits at this power and voltage level. Please be careful - you might start out with a lower power while you learn the ropes. \$\endgroup\$ Commented Jul 8, 2023 at 1:33
  • \$\begingroup\$ Thanks for helping. This design is supposed to be just a reference to test it with all security measures in the future, so, as long as I keep simulating it, I'll be safe ;) \$\endgroup\$ Commented Jul 8, 2023 at 19:53

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.