Various aspects of home brew inverters

Author	Message
Warpspeed Guru Joined: 09/08/2007 Location: Australia Posts: 4406	Posted: 05:21am 15 Apr 2018       The big problem with PWM is that the high frequency switching frequency (and its odd harmonics) need to be removed, to recover the desired 50Hz pure sine wave power buried underneath. This obviously requires some low pass filtering. Now here we must introduce some reactive components, both inductive and capacitive to achieve the low pass filtering effect. Some of these components are deliberately introduced and obvious, and others are hidden and less obvious in that they represent stray inductances and stray capacitances due to construction and layout. Designing an efficient low pass filter is fairly straightforward, as long as you can define the EXACT source and load impedances. The problem with an inverter is that there is NO WAY to define the load impedance. There is no way to anticipate what may be connected as a load. It can vary between a complete open circuit, to a very low pure resistance, to crazy non linear loads that switch on and off at various points in the cycle, and highly inductive or capacitive loads of very low power factor. So we have a big toroidal transformer that may typically have a self resonance of a few Khz, due to the very high inductance of the secondary winding, and capacitance back to the metal in the core and between turns. Then we add a series choke in the primary that will add another self resonance. And maybe even more LC filtering after that. Then we fire up the inverter, and wonder why the output waveform has some wobbles and discontinuities on it. Kind of like a jelly in an earthquake, it will wobble around according to its own self resonant nature, pretty much unpredictably. If you load the bastard right down with a heavy purely resistive load, it will probably work fine and produce a nice clean sine wave. But the real world loads are not often perfectly linear or resistive, and you are very likely going to see some kinks, wobbles, and strange weirdo output waveforms. Trying to make an efficient low pass filter that will work into absolutely any kind of extreme load is an exercise in futility. But in practice the grid waveform is pretty crappy and distorted anyway, so why worry ? All your "stuff" is very likely to work perfectly well on a wobbly 230v waveform. Cheers,  Tony.