Hi Erwin
I can fill-in for DLath and help you resolve this issue.
You will need to understand how the precision power meter computes the energies and how the ADE7933-ADE7978 chipset computes the energies. The manner in which each energy result is computed, will greatly influence the result, particularly when harmonics are present. I would recommend that you read this discussion: Re: Re: 7816 Reactive Energy measurements with nasty loads.This is with a different chip but it would convey the point that I am trying to make.
Can you take FFTs of the current and voltage inputs going into ADE7933? Based on the input signal content and the description in the datasheet about how the energies are computed, does the results from ADE7978 make sense? (Refer to the other discussion mentioned above, for a similar analysis). This chipset can also compute fundamental active and reactive energies and so you can note that down as well for comparison and verification.
Let me know if you need more help!
Regards
hmani