I would like to transform 12V DC to 230V AC 50Hz, peak load 200W. Efficiency is not a concern. Main concern is puresness of the sine wave.

If I am searching for "pure sine wave inverters" there are 2 kinds of results:

  • consumer grade inverters which claim to generate pure sine waves but the ones I looked into have massive PWM which would need massive filtering.
  • research papers which propose fancy inverters with filtering at the end.

The arguments for PWM inverters I have read of are efficiency, but because I don't need efficiency I wonder if going with PWM and massive filtering is the right way to go or if I should build an old-fashioned inverter with some motors or linear amplifiers myself or maybe are there devices for hobbyists in this direction I haven't found?

  • 3
    \$\begingroup\$ Please tell us what kind of load you want to connect. Pure sine wave is marketing speech. At least the current is not going to be a sine wave for most loads anyway, and percussions of that will always be seen in the voltage unless there are also pure sine current loads of ten times the power connected. \$\endgroup\$
    – Janka
    Commented Jun 3, 2021 at 13:12
  • 1
    \$\begingroup\$ "have massive PWM which would need massvie filtering." "Main concern is puresness of the sine wave." You need to state your requirements as a measurable physical quantity. What's the problem with filtering? \$\endgroup\$
    – winny
    Commented Jun 3, 2021 at 13:22
  • \$\begingroup\$ To give a quick number: for my wall outlets I would need 40dB filtering at 17kHz, for the PWM inverters I tried I would need more. \$\endgroup\$ Commented Jun 3, 2021 at 13:44
  • \$\begingroup\$ the problem with non-switching semiconductors is the power loss. PWM is usually not a problem for inductive loads because the current is sinusoidal. Class D amplifiers also work with PWM. If that is not acceptable, the only thing left is the mechanical variant or a class A or AB amplifier. The efficiency for class A will be below 7% and for AB below 60%. \$\endgroup\$
    – arnisz
    Commented Jun 3, 2021 at 13:53
  • \$\begingroup\$ Note that you haven't explained what your load is as requested in the comments. Why is PWM a problem? \$\endgroup\$
    – Transistor
    Commented Jun 3, 2021 at 14:23

2 Answers 2


The old-fashioned way would be a "rotary inverter". Use a 12V motor to drive a 230V alternator. Adjust the speed of the motor, and the gearing between them, to give the desired 50Hz.

  • \$\begingroup\$ Note that the motor would have to be over 1/3 HP. \$\endgroup\$
    – AnalogKid
    Commented Jun 3, 2021 at 18:28
  • \$\begingroup\$ As this seems to be the most easy and pure approach I will accept this answer. Another word for this is motor-generator: en.wikipedia.org/wiki/Motor%E2%80%93generator#Conversions \$\endgroup\$ Commented Jun 12, 2021 at 20:53

An all-linear approach would be a simple sine wave oscillator, followed by a power amplifier, followed by a step-up transformer. Conceptually, pretty straightforward. IRL, not so much.

There are many choices for the oscillator, determined by the desired harmonic distortion level, frequency accuracy, etc. OK, that one was easy.

The power amplifier has to make 200 W after its own inefficiencies, powered by a single 12 V rail. If we assume the output cannot get closer than 1 V to the rails due to output transistor headroom, that means the output waveform into the transformer is 10 V peak-to-peak, or 5 V peak, or 3.5 Vrms. At 200 W, that's an output current of 56.6 Arms.

But wait, there's more. Because the amplifier output is centered at 6 Vdc rather than GND, it must be AC coupled into the output transformer. Moving 57 A through a capacitor without a significant voltage drop across it requires one gigantic capacitor.

One way to mitigate this is to use a BTL - bridge-tied load - configuration. This requires two linear power amplifiers running 180 degrees out of phase. This doubles the effective voltage across the transformer primary, which halves the primary current. Also, the DC offsets of the two driving signals cancel out each other, so you don't need a coupling capacitor.

Last up, the output transformer. What you can use is a conventional step-down power transformer connected "backwards", with the low voltage secondary driven as the primary. My guess is that no one makes a catalog standard item that will meet the voltage ratio and input current requirements. Also, the 120 Vac output impedance will not be nearly as low as a standard AC outlet. This will affect the output amplitude stability with load changes, as well as any downstream power line filters. Oversizing the transformer will improve this greatly, but a 300-400 W transformer is a biiiig chunk of iron.

  • \$\begingroup\$ It looks like Hammond does have a 300VA [toroidal] transformer with voltages near what you'd need. It costs about US$80 and weighs about 7 lbs. \$\endgroup\$
    – Theodore
    Commented Jun 3, 2021 at 15:25

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.