ooh statistics knowledge alert:

RMS is, im presuming the root mean square as said above. The forumal for

sinusoidal (or other periodic) functions is:

√((Integral[f(t)², Lims: T1, T2])/(T1-T2)) where T1≤t≤T2 (the time T1-T2 should be a whole number of cycles of the function), f(t) is the function being investigated.

EDIT after some investigation: it appears in electronics that the function is generally for ac

power, in which case:

RMS= √((Integral[(Imax sinωt)², Lims: T1, T2])/(T1-T2)) where

Imax is the amplitude of the fluctuations, ω is the angular frequency (radians per second)

EDIT after some more investigation and some subsequent algebra (gotta love calculus):

RMS=(

Imax)/√2

The formula for non-sinusoidal functions is:

√((Σxi²)/n) where xi is a value of x, n is the number of items (obviously Σ is the summation sign).

RMS and other such measures, are only likely to be used if they are BLUEs (Best Limit Unbiased Estimators), which to the best of my knowledge they arent, the average (officially, mean (of the sample)) is the BLUE of the mean (of the population). However, your statement "provide a reference... for non-sinusoidal signals" suggests to me that the

RMS is used to estimate the average

voltage coming from an ac

power supply, which if i understand correctly tends to be more towards one of the two limits rather than being completely in the middle of the limits.[EDIT: the stuff in edits above was obviously added after this comment, i feel a bit silly now]

Presumably, other methods of investigating the

power,

voltage and

current are to investigate the maximum, which is useful so you dont blow fuses

etc, or the minimum, so you get at least a certain

voltage all the time. IEDIT: apparently the peak

voltage is given by Vpeak= Vrms x √2, which it seems could be as high as 339V in the UK!

I cant think of any other stuff to put here at the moment, so i will go off for an investigate...