“RMS Power”

Recent discussion in the rec.radio.amateur.homebrew newsgroup showed a

widespread misunderstanding of the meaning and importance of RMS and
average values of voltage, current, and power. So I’ve put together this
explanation which uses as little math as possible, and hopefully explains the
concepts in an intuitive way.

The Meaning of Average

Before I can proceed, it’s important to explain just what average means.
Fortunately, its meaning when dealing with waveforms is essentially the same as
its common meaning. If we were to sample a waveform (that is, a graph of
voltage, current, power, etc. versus time) at equally spaced times, then add up
their values and divide by the number of samples, we’d have approximately the
average value of that voltage, current, power, or whatever the waveform
represents. The smaller the time intervals between samples, the more accurate
the average will be. The mathematical operation of integration is a way to find
what the value would be if we could shrink the time interval extremely close to
zero, and it’s needed if we want to calculate the exact average value of some
waveforms. But I’m going to use square waves for this explanation, and we can
easily see their average values without any math at all.

A periodic waveform is one that repeats in an identical fashion, over and over.
The period of the waveform, or one cycle, is the time interval that repeats. If we
find the average of one complete cycle of the waveform, then the average of the
next cycle will be exactly the same as the first. If we combine the samples from
two cycles, add them together and divide by the total number of samples, we find
that not only is the average value of one cycle the same as the average of the
second, but also that the average value of two cycles is the same as the average
of one cycle. In fact, the average value of any whole number of cycles is the
same as the average value of just one cycle. So we can easily find the average
value of a very long-duration periodic waveform simply by calculating the
average value of one complete cycle.

Equivalent Power

The next important concept is that of equivalent, or heating power. Let’s suppose
we put 5 watts into a resistor for 10 seconds. The total amount of energy applied
to the resistor is 5 watts X 10 seconds = 50 watt-seconds = 50 joules. This of
course raises its temperature by some amount. The actual amount of
temperature rise, in degrees, depends on how massive the resistor is, how fast
the energy is applied, and how fast the heat can be moved away by conduction,
convection, and radiation. So we won’t try to calculate the exact temperature rise.
But the important thing here is that the total amount of heat (energy) the resistor
dissipates is the amount of energy that’s applied to it. Put another way, imagine
the resistor being immersed in a thermally insulated tub of water. If we apply 50
joules to the resistor, the water temperature will rise by an amount we can
calculate, and it won’t matter how we apply that 50 joules – suddenly or slowly –
that same temperature rise will result whenever and however we apply 50 joules.

Consider the following simple circuit, where the switch closes and opens, closes

and opens, continuously, with a ten second period:

5 seconds on, 5 seconds off

+
10 Volts 10 Ohms

The voltage applied to the resistor will look like this:

10

Volts

0
0 5 sec. 10 sec.
time ->

And the power dissipated by the resistor will look like this:

10

Power
(Watts)

0
0 5 sec. 10 sec.
time ->
This second waveform of power versus time shows the instantaneous power,
that is, the power delivered to the resistor at each instant of time. Too often, it’s
overlooked that whenever the voltage and/or current varies with time, the power
does also. But it’s obvious that when the switch is open, no power is being
delivered to the resistor. This is extremely important to realize, because it
enables us to find the actual amount of energy that’s being transferred as well as
calculate other important quantities.

As mentioned above, the amount of energy delivered to the resistor is the

product of the power and the time. This is strictly true only if the power is
constant during that time. When the power varies, the energy can be calculated
by sampling the power at frequent time intervals, calculating the power X time
product during each interval, and adding them up. The total ends up being the
area under the power waveform curve, and it gets more and more accurate as
the time interval becomes shorter and shorter. Again, mathematical integration is
necessary for all but the simplest curves if extreme accuracy is required. But for
our simple power waveform, it’s easy to see that in the first five seconds, the
area under the curve is simply 10 watts X 5 seconds = 50 watt-seconds = 50
joules. This is the total energy delivered to the resistor during the first five
seconds of the period. During the second five seconds, the power is zero, so the
energy delivered is zero. So the total energy delivered over one period of ten
seconds is 50 joules.

Now let’s find the average power during the ten second period.

I think you can see by inspection of the second (power) waveform that the
average power is 5 watts. It’s 10 watts for half the time and zero for the other
half, so if we were to sample it at regular intervals, half the samples would be 10
watts and the other half would be zero. Dividing by the total number of samples,
we’d get 5 watts for the average.

If we had applied 5 watts continuously for the whole 10 seconds of the period,
we’d have delivered 50 joules of energy to the resistor. The resistor would
dissipate a total of 50 joules, exactly the same amount as it does with the 10
watts of power being delivered for half the time. The effective or heating power
of our circuit is 5 watts, which is the average power.

It’s easy to see that if we delivered four times the power for ¼ of the time, we’d
still have the same average power, and the same total delivered energy – and
therefore the same amount of resistor dissipation.

Now let’s look at the voltage and current more closely. Recall the voltage
waveform:

10

Volts

0
0 5 sec. 10 sec.
time ->
Ohm’s law tells us that the current will be:

Amps

0
0 5 sec. 10 sec.
time ->

By inspection, the average voltage is 5 volts, and the average current is 0.5
amps. But the average voltage times the average current (= 2.5 watts) is not the
average power. In fact, that product is meaningless. We need some other
measure of voltage and current if we’re going to calculate average power without
first finding the instantaneous power.

The Meaning of RMS

RMS is a mathematical function, like average, that reduces a complex function to

a single value. And, like average, it has a precise definition. The definition is
revealed by the name – it’s the square Root of the Mean of the Square of the
function. Mean is the same as average, so we calculate the RMS value of a
function or waveform by first squaring it, then taking the average of the squared
function or waveform, then taking the square root of the average. But it has to be
done in the proper order! Let’s calculate the RMS value of the voltage waveform
from the example. Recalling that the voltage was

10

Volts

0
0 5 sec. 10 sec.
time ->

The first step is to square the waveform:

 100

2
Volts

0
0 5 sec. 10 sec.
time ->

Then we take the average of this waveform, which we can see by inspection is
50 volts2 . The average is no longer a waveform – that is, it doesn’t vary with time
– it’s just a single value. Finally, we take the square root of the average, and get
7.071. . . volts. That’s the RMS value of the original voltage waveform. Doing the
same for the current waveform, we calculate an RMS value of 0.7071. . . amps. It
turns out that when we multiply these together, we get 5 watts – the average
power. This is always true, for any voltage and current waveforms (assuming a
resistive load for simplicity). The reason comes from the basic mathematical
definitions of RMS and average. It’s easiest to see by looking just at the voltage
v2
or current. The instantaneous power is where v2 is the instantaneous voltage.
R
2
v Avg(v 2 )
Therefore, the average power is Avg( ) = . (R can be moved out of the
R R
average since it doesn’t change with time.) The RMS voltage V RMS is Avg(v 2 ) ,
VRMS2
so (V RMS)2 = Avg(v2), and the average power is . A similar analysis of the
R
current shows that the average power is also I 2 RMS R . The importance of RMS
voltage and current are that they can be directly used to calculate the
average power.

RMS Power

The RMS value of a power waveform can be calculated just like the RMS value
of any other waveform, although it doesn’t represent heating power or any other
useful quantity. I’ll do it just to show how it would be done.

Recall that the power was:

10

Power
(Watts)

0
0 5 sec. 10 sec.
time ->
Squaring it, we get

100

2
Watts

0
0 5 sec. 10 sec.
time ->
The average of the square of the power is clearly 50 watts, and the square root
of that is 7.0711. . . watts. We found earlier that the equivalent heating power of
our circuit – the average power -- was 5 watts, not 7. The RMS value of power
is not the equivalent heating power and, in fact, it doesn’t represent any
useful physical quantity. The RMS and average values of nearly all waveforms
are different. A notable exception is a steady DC waveform (of constant value),
for which the average, RMS, and peak values are all the same.

It should be noted that the term “RMS power” is (mis)used in the consumer audio
industry. In that context, it means the average power when reproducing a single
tone, but it’s not actually the RMS value of the power.

Summary

I’ve shown that:

• The equivalent heating power of a waveform is the average power.
• The RMS power is different than the average power, and therefore isn’t the
equivalent heating power. In fact, the RMS value of the power doesn’t
represent anything useful.
• The RMS values of voltage and current are useful because they can be used
to calculate the average power.

For those who are interested, mathematical derivations are readily available and
can be found on various web sites and in textbooks. What I’ve tried to do here is
to present the basic concepts with as little mathematics as possible. Questions
are welcome, either directly to me at [email protected] or posted on the
rec.radio.amateur.homebrew newsgroup. And I’d appreciate very much if anyone
finding an error would notify me so it can be corrected.

Roy Lewallen, W7EL

Modified 2004-11-18.
Thanks to Steve Nosko, K9DCI, for suggestions.