Question
Your datasheet for the DAE 7753 recommends that voltage and current register
readsings are syncronised with zero crossings on the voltage channel to
eliminate noise. Is this still the case when a reactive load is used and the
voltage and current waveforms are not in phase.
Answer
The RMS calculation is not done using a fixed time period - it is implemented
using a low pass filter (see figure 53 on page 24 for example). That means that
there is some residual ripple at twice the line frequency contained in the root
squared input signal.
If you were to sample asynchronously, rather than synchronizing to the zero
crossings, then successive readings will exhibit a random amount of the
residual ripple component and will jump around more.
Synchronizing the reading to the voltage zero crossing means that you always
see around the same amount of ripple on each reading. Averaging over several
readings will further reduce the value of the ripple. There may be some error
due to the ripple but successive readings will give a similar answer.
When a reactive load is being measured, the fact that the current signal is out
of phase with the zero crossing on the voltage signal will cause the ripple
waveform to be out of phase as well. However, it is still a good idea to
synchronize readings with the voltage zero crossing, simply to reduce the
variation between one reading and the next.