On Feb 17, 7:04 pm, "Jim Wilkins" wrote:
ulator would work far better.

Would it be better to regulate the peak, average or RMS voltage if the
load draws non-sinusoidal current?

I became concerned with the peak after a poorly regulated generator
burned out the transient voltage suppressors in an outlet strip, in a
cloud of purple smoke.
jsw

If the output is non-sinusoidal alternating current, it would
depend on the use. If equipment is being used that uses
the entire waveform, such as a heater, the RMS would be
the choice. For a motor, I'm not sure, but I'd go with RMS
unless better info is found. If the device is electronic, which
uses a peak detector to form a DC for a power supply, the
peak would be reasonable....
However, it would be necessary to look at the output
waveform to make a really informed choice. For instance,
if the output is a square wave the choice would be different
from a stepped sinusoid.
Note that rotating machinery will put out a sinusoid but
many generators today use a speed lower than 3600 rpm and
use an inverter to make the frequency controlled output
voltage, often a "stepped sinusoid", which has the RMS
of the sinusoid AND the peak level of the sinusoid, but
synthesized in steps.

Also, if the output is DC or pulsed DC, a capacitor
stabilizes the voltage by reducing the ripple, the same
function as the battery in an automobile. For this
appication a big ass capacitor in parallel with a resistor
can, in many cases, take the place of the battery. If a
load is always present, the resistor won't be needed. \
Such stabilizing loads are often used in stabilizing the
DC system in large RVs if a battery isn't used.


The requirements are different for an AC system.
That accounts for some of the different solutions posted
here.