So right, last three days, IoTaWatt total power has been ~80% of meter power.
To recap where we are:
Given the spaghetti wiring, it has been difficult to determine the phase/load relationship and we’ve been concentrating on trying to get the mains correct. We determined with resonable certainty that the VT was connected to Main1 here. So Main1 is phase A.
We determined that the hot water heater was connected to Main3 here.
We determined that Mains3 is phase B using unity power factor of hot water here. So by elimination Main2 is phase C. This is somewhat supported by the order in which the Mains are presented to the meter here. Note that the sequence right to left on the back of the meter is B, A, C. I don’t know if 3ph meters work like 3ph motors, but if so wiring it A, B, C would have made it turn backwards, so I’m guessing that without color codes, they just connect them up arbitrarily and if it goes backwards, they swap A and B.
So the last piece was to insure that all of the CTs are facing the same way with respect to line/load, and that was corrected. So now I believe that the mains measurements are setup correctly. Yet the readings have been ~20% low for several days.
Going Forward:
Expectation for the mains with derived reference is at least within 5% hopefully better. So I’d like to continue to pursue that 20% difference. My suspicions, in order of interest, are:
- One or more CT is reporting low.
- There may yet be an error in the configuration/setup.
- There are gross voltage deviations between the phases.
- The meter is wrong.
So to further explore the obvious #1 there are a few things that can be done.
Take all of the CTs, configure then for phase A. Mark each with a unique number using a felt tip marker, and clamp them on the Main1 line at the same time. Observe the power recorded on each and note any outliers. Do this at both lower and higher power. Also note variations in power factor when above a few hundred watts. Reinstall and configure with the most consistent CTs on the three mains. Wait a couple of 24hr cycles and see if there is better agreement.
If you have a decent true-rms clamp on ammeter like a Fluke, you can configure outputs for each of the mains with Amps as the units, then compare the meter to the IoTaWatt Amps. They should be within 1% on all phases.