PRE-Posting-EDIT: It's not that I am committed - it's more a case of I should be committed. Or need to be.
But I will split my insanity, er... ramble, over 2 posts. (I think the R-word was originally Reply. I know later, Rhetoric springs up.....) hic!
Yeah... I forget the trivials - or details.... We have 230VAC, but USA is 120V (if I recall....).
So 120 VAC is 120V RMS which means things like 108V average (is it pi over root-8 or vice versa - I did this the other day...!) and 170V peak (120V / root-2?).
And 170V peak means 340V p-p (peak to peak), or is it 170p-p with +/- 85V peak...
Holly MadCow - this is so basic!
I do know we needed battery banks of about 430 to 460VDC for 240 VAC...
But forget the details...
An AC supply is a sinewave which has an average value, a slightly different RMS value, and a reasonably higher peak value. From memory, the ratios are 63% & 71% with 100% being peak; ie - 2/pi, 1/root2 (= root-2 / 2) etc.
For a squarewave, the average and RMS and peak are all the same.
Hence for primitive square-wave inverters, what do you design for?
Probably the RMS value - so for a 120VAC RMS, you have a 120V peak = 120V RMS square wave.
But what if you are powering a switched-mode supply (SMPS), or something that requires a peak of 170V - maybe a starter for a fluorescent tube? It probably won't work on a mere peak of 120V - that's only ~70% of its required peak voltage.
Fine - so we output a square wave of 170V instead. But now it's 170V RMS and that may fry the load.
Enter the modified square wave - usually a two-level output that has a particular step timing to provide an RMS and peak voltage similar to a sinewave.
It's average will not be the same as a sinewave, but average values have little relevance for loads (except... was it moving coil meters etc?).
Note that the 2 voltage levels form 3 steps - like a winners podium - highest in the middle and flanked by the two lower levels. (Lets ignore the negative cycle - it's the same thing, just inverted....)
So what are the two levels?
Lets say the highest is 170V to give the 170V AC peak.
The other level(s) will probably be half of that = 170/2 = 65V simply because 2:1 is easy to achieve.
The timing of each level is calculated so it gives the same RMS & peak as the sinewave.
The designs might vary - maybe the peak is dropped because whilst a sinewave is at its peak for an infinitesimal small period (of time), the modSquare has to be there longer. So instead of maybe 170V peak, we use 160V peak - say with duration the same as the sinewave is between 150V & 170V (ie, they both roughly average or RMS to 160V during that period).
Then that 160V peak means 160/2 = 80V flanks, and they stay on long enough to provide the overall RMS value of 120V.
Of course there is a 3rd level - namely off or 0V... Over a 90 degree 1/4 cycle, it might be 0V for 0-10 degrees, then 80V for 10-70 degrees, then 160V from 70 to 90 degrees; then the reverse from 90-180, and then the inverse (negative) from 180 to 360 degrees.
Crikey mate - this is NOT easy in words. It is much easier pictorially!
Anyhow, hopefully you can picture (oooh - bad pun?)..... you can picture the artificial "sinewave".
(Designs could vary. If it's easier to have fixed 15 or 30-degree step intervals, you may vary the voltage levels.)
So that's the cheap solution. A squarewave is easy, but unacceptable.
[ I didn't mention the off time before - it can be +V, then 0V, then -V, then 0V etc. But intuitively, a full-on, zero, full(negative)-on is pretty doggone rough compared to a smooth sinusoid (like ripples/waves across water). ]
The step or modified squarewave (aka pseudo sinewave) is better and generally acceptable, and most things work on it.
Now to get to the catch!
But I'll do that in the next post....