1

I was always dissatisfied with inverters because it seems like an LC tank and transistors would be all you need. However, after much time spent simulating and a few false positives experimenting, I am beginning to think it is flat out impossible. I have no way to prove it, but it seems that you need at least 4 passive components to create AC from DC. It seems as though it is often that there are a pair of passive components for each phase of the duty cycle.

Is this indeed the case?

Alex Eftimiades
  • 831
  • 1
  • 11
  • 25

4 Answers4

4

A LC tank circuit is just a inductor and capacitor connected to resonate at a particular frequency. These are not active components. They can not make a frequency that doesn't already exist. They can filter, sometimes with quite high Q, a mix of frequencies to pick out or reject a narrow range.

In other words, A L-C tank can be used as the governing elements that set the frequency, and also to filter out spurious frequencies to leave you largely with a sine, but the original signal with the desired frequency content can only be made with something that consumes power. This requires at least one active device somewhere.

If it were a desing criterion for some reason, it should be possible to create a circuit that oscillates and drives a L-C tank with a single active element. It will be easier with multiple active elements or multiple windings around the same core. If you are looking to have it produce significant power, I wouldn't worry about the number of active elements and think about drive level and efficiency instead.

Olin Lathrop
  • 310,974
  • 36
  • 428
  • 915
  • The goal here was to limit the number of passive elements to a single LC tank--not the number of active elements. I know I need active elements, and I am fine using as many as I need. I was however interested to know whether one could design an inverter using an LC tank as the only passive element. – Alex Eftimiades Apr 04 '12 at 22:34
4

A blocking oscillator is able to generate a 'dirty' sine wave with only one transistor, a capacitor and an inductor. However this isn't an inverter - it produces a free running signal, but that signal is tuned by the LC time constant.

trayres
  • 598
  • 2
  • 12
2

It seems that one of the most simple configurations for what you're aiming at is a Royer Converter. However, with a Royer, it is not only the LC resonance that determines the circuit's behavior. The saturation characteristics of the transformer's core play also a role. In this application note by Zetex/Diodes, Fig. 3 and the text around it have a lot of information about a Royer's secrets, in a nutshell.

Related: What is the use of transformers with 3 pairs of wires?

zebonaut
  • 18,170
  • 4
  • 60
  • 104
  • 1
    Thank you for introducing me to the Royer converter. I have actually continued to research this question since I posted it. After a lot of time, I think the Royer converter is probably best suited for my needs. I found the wikipedia article most helpful because it contained the simplest implementation. If I understand correctly, the extra winding acts as a test probe for when the tank kicks back (due to the frequency being off resonance.) The winding is connected to transistors which switch the direction of current on the primary, thus staying in phase with the load on the secondary. – Alex Eftimiades Apr 20 '12 at 13:42
0

Astable multiviberators used for induction heating convert directly from DC to AC at resonant frequency. The capacitance at the required frequency can be calculated.

evildemonic
  • 8,698
  • 2
  • 25
  • 44