Let's look at a CFL. (compact fluorescent or "energy-saving" lamp)
- Voltage: 220-240 V AC
- Current: 0.061 A
- Power: 7 W
PF = Real Power / Apparent Power
PF = 7 / (220 * 0.061) = 0.522
Here is the plot from the Picoscope software. Note the non-sinusoidal current waveform that is typical of this style of lamp.
Looking at the averages from the measurements we have on the Picoscope screenshot:
V = 247.6
I = 0.0475
P = 6.725
PF = 6.725 / (247.6 * 0.0475) = 0.572
The next plot is from a dehumidifier unit with only the fan running.
V = 245.7
I = 0.136
P = 33.01
From the plot we can see the power factor should be pretty good
PF = 33.01 / (245.7 * 0.136) = 0.988
A quick look inside the device with a torch reveals the presence of a 1.0uF capacitor that is providing power factor correction for the fan motor.
Now with the compressor running:
V = 244.7
I = 1.535
P = 329.8
A tell-tale sign that the power factor is no longer 1 is the presence of negative values in the power waveform (green trace).
PF = 329.8 / (244.7 * 1.535) = 0.878
That'll do for now I think. Time to go and ponder how to interface the voltage and current waveforms into a microprocessor.
Seeing as the majority of 'modern' electronic devices are non-linear I think the use of something like the dsPIC will be required to properly interpret the waveforms...