# How Exactly Does a Push-pull Pair Cancel Even Harmonics?

I've been studying cassette tape circuits recently and the same bias oscillator circuit keeps cropping up time and again. This one from a copy of Practical Electronics (circa 1960):

...and this one from an Aiwa cassette player circuit from the 80's:

As far as I can make out, it's favoured because it has low harmonic distortion and this, it seems, is down to it's push-pull configuration.

It's often said that a push-pull pair cancels even harmonics, but I'm interested to know exactly how this happens. I considered it's because of its symmetrical configuration and the non-linearity of one active device cancels the other, but in the simple push-pull case aren't one the devices off. So, how does the cancelling happen?

• These are simple astable oscillators which generate a series of odd harmonics because that is what a squarewave is. They don't cancel out the even harmonics because they don't produce them in the first place. Sep 9, 2016 at 17:44
• Well that explains things for these oscillators. But what of the general case of push-pull pairs? Sep 9, 2016 at 17:49
• Just as an aside Jim, when I built this circuit on the breadboard it produced a very definite sine wave. Are you certain it's a square wave generator? Sep 9, 2016 at 18:18
• Same was true when I modelled it in the simulator. Sep 9, 2016 at 18:21
• The basic circuit is an astable - the other components act as a tuned filter for the fundamental tone leaving you with a sinusoidal wave but the basic circuit configuration is an astable. Sep 9, 2016 at 18:50

If you look at a push pull amplifier with some amount of cross-over distortion it looks like this: -

In other words it creates a symmetrical shaped signal in amplitude but generates odd harmonics that can be significantly reduced by correct biasing and sometimes negative feedback.

Now take a look at an example of even harmonic distortion: -

You should be able to see that it can generate significant amplitude asymmetry and this, in even a slightly badly designed push pull amplifier, is unlikely to happen to any great extent.

It can happen (of course) if the load is more easily driven by the pushing transistor compared to the pulling transistor - asymmetry in amplitude will result but it would be a really crappy push-pull.

Compare this to a single transistor driver circuit and you will definitely witness some amplitude asymmetry: -

• Thank you Andy, that saves me having to make the same point about class A, AB and B type outputs. Sep 9, 2016 at 17:57
• @JImDearden a pleasure of course! Sep 9, 2016 at 17:58
• As a physical aside, I recalled that a musical instrument made from a tube, open on one end and closed on the other (like a bottle where air is blown across the opening) is incapable of producing even-numbered harmonics.
– jonk
Sep 9, 2016 at 18:07