Bob Minchin coughed up some electrons that declared:

Tim S wrote:
Andrew Gabriel coughed up some electrons that declared:


It would be needed at the sensor end.

Not necessarily. If the base unit could synchronise time (delta time, not
absolute) with the sensor over the 433 link, at least to the millisecond
level, the sensor could send a batch of current readings with timestamps
to the base.

The base being driven from an unregulated AC wallwart would take a series
of hi res voltage readings with timestamps and, having found the
intersection of current readings and voltage readings, could in a
leisurely manner do the maths.

It assumes:

1) You can achieve the timing accuracy with a weedy microcontroller - I
think you could manage mS without any problem.

2) You can time lock over a 433MHz link to the same resolution - not
sure.

3) You assume that your wallwart transformer is maintaining a stable
ratio bewtween mains voltage and output voltage and the waveform comes
through largely uncorrupted.

Note - I didn't say it would work, just that it would be an interesting
exercise

Cheers

Tim
Much easier to do the measurements at one point and there is mains power
for the processor.

I agree - *if* you have a socket near the meter...

I do here (Pembury) but in Robertsbridge the meter is outside, which is
common with many properties. I suspect the designers conceeded that a large
portion of th users wouldn't be able to source a direct connection, short
of a pin in the cable.

Although having said that, I wonder how well a capacitative pickup would
work - with a sticky electrode on the sheath of each meter tail?...

many samples of the both waveforms are necessary as
the voltage 'flat tops' regularly - or at least it does here. Just have
a look with a 'scope (and suitable probe!!) sometime.
So many domestic devices with SMPSUs these days to blame.

Bob