I would have thought that the filter was
fairly generic so many would work instead, but whether they are pin compatible
is a different issue. Maybe the PCB could be redrilled to fit others. Or fit a common
mode choke from an old SMPS and add the capacitors. All with care due to the mains
voltage. All the values are in the photograph. The current consumed is tiny.
I was a bit surprised they did not take the
DC supply from transformer T1 using a linear regulator to save some money. I
have a commercial meter that does this but it does considerably oversize the transformer
so it’s lightly loaded, thus the current into the regulator does not over distort
the secondary voltage waveform. Revenue grade meters can’t use this approach
but the project is not aiming for that accuracy. Maybe they could have used one
6V secondary for the DC and the other for voltage monitoring to limit the distortion.
Saves all the SMPS hash so maybe the filter would then be unnecessary.
I have a problem with the words on middle
of page 17 of the article which say “They (V
and A samples) …. finally, if they are of the same polarity, also
accumulated into a fourth location for true power”. I designed from scratch a similar
meter and for true power multiplied together all sample pairs, even the opposing
polarity ones, summed the lot and divided by the number of samples per mains cycle
to get watts. Then cleared the register after each cycle.
If you did what the article says you’d get
large positive watts in the case of graph C on page 20, where the phase lag is
90 degrees so watts is in fact zero, as they say there.
Am I right?
