"Richard Harrison" wrote in message
...
[...]
Steve also wrote:
"Also ok, but not sure how it plays into the RMS discussion."

My speculation is that the effective value of a nonsinusoidal waveform
could be found by summation of its sinusoidal constituents.

That's what I thought you were going for. My gut feel is that this must
be valid, otherwise "Fourier ain't an exact solution after all". But it
sounds like a lot of work and you have to be able to extract all the
significant Fourier components (harmonics) and there is an element of
approximation here, just like my "mathematicall function" method, no?

....Of course doing the integration on the waveform takes some time to crank
through the math and get all the quantities collected correctly.


But, it`s not difficult to find an effective value for ...
nonsinusoidal periodic waveforms as well. One can graphically take

For what I wanted to do, the integration was easier. I may be wrong,
but Fourier also needs a function and you still need to integrate, no?

Now, to answer Richard Clark's comments about accuracy in another
post... I did assume that the wave shapes I was interested in were defined
(I think the word is) explicitly. I assumed a mathematical function. When
I compared the three waveforms which were appropriate (sine, triangle,
trapezoid) the results were very much the same (for ratio of RMS to
Average). (gotta remember to keep my sentences shorter).
From this I came to the conclusion that for wave shapes which differed
slightly from the ideal (assumed shapes) the values would be well within
acceptable bounds. Easily 5%. Yes, not exact, but much better than Bob
Shrader had assumed.


large number of equally spaced ordinates of the form, using at least one
complete alternation, Richard Clark.

I don't think they have to be equally spaced since the actual time
enters into the calculation (as you described later). This would make long
sections with a constant value easier.


Both alternations are not needed
but could be used as a minus times a minus is a plus and each of the

This assumes that the waveform is symmetrical around zero. I think in
general, one whole period is required.



or[...] using a planimeter.

Richard H. clearly prefers graphical solutions and that's ok.

Unless I`ve opened a new can of worms with this posting, I don`t know of
any difference of opinion I have with Steve.
Best regards, Richard Harrison, KB5WZI

Right now it appears that this half vs. full period is the only
difference I can see.

--
Steve N, K,9;d, c. i My email has no u's.