26 Dec
2012
26 Dec
'12
5:19 p.m.
? "1 volt peak to peak rms"?? Well, which is it? ;)
On Wed, 26 Dec 2012, Chris Tofu wrote:
C: I'm trying to recollect my "thinking"
or lack thereof on this. I
think I thought I read something to the effect that the "peaks" were RMS
values. In other words the "1 volt" part wasn't really a volt, but 1.414
or whatever (and I know I said something different in a different e-mail
- is what happens when you think too hard I guess!).
Think of it as a type of AVERAGE. You get a certain height peak,
but are more interested in the average voltage.
Sometimes the range is most useful, sometimes the average.
BUT, these are AC signals, and some are positive, and some are negative.
If you were to take an "ARITHMETIC MEAN", and just add up the voltages
and divide by the number of samples, you would get ZERO. (assuming no
DC component offset)
You COULD take the absolute values and average those.
BUT, spikes, or odd shaped curves can give you a useless average.
Likewise averaging "BillG" and "Billy in the box under the bridge",
which is why most people talk about MEDIAN home price, NOT MEAN.
OR, you could SQUARE each number, average [MEAN} those squares,
and then take the square root. Hence, ROOT of the MEAN of the SQUARES,
or RMS.
(Square root of square is one way of getting absolute value).
THAT gives a type of average that is more useful in terms of how much work
you can get out of it, etc.
(cf. "Variance", "standard deviation")