Excellent amount of detail thanks!
Indeed. It’s hard to believe that could comply with Euro standby power regulations. Is there any chance there’s still a fan or something running during those periods or is the unit completely quiet? It would be interesting to know how that 110W is spread across the three phases. Any chance you could get a 3x Wattage/PF display of it in that state please?
Out of interest, which of those three phases is the one connected to the IotaWatt’s “official” voltage input jack? The schematic shows a 10uF series cap on that voltage input that presumably doesn’t exist on the other two that come in via the adaptor cable and CT inputs. If I’ve done the maths correctly, that would introduce a 1.4° phase shift on that voltage signal that’s not on the other two (although that may be compensated for somewhere else). Don’t be tempted to add a 10uF cap to the other two as the series resistance on those inputs is much lower (312R Vs 13K) so doing so would introduce a massive phase error.
A general rule of thumb for meter design (I think I first heard it from the energy IC manufacturers) is to put the exact same filters on all your inputs: I and V, across the board. That way any phase error introduced by the filters will be nominally the same on every channel (Vs and Is).
One really simple test would be to choose one of the phases that comes in via the adapter cable and move both its VT and its heatpump CT to the phase that comes in via the “official” VT input jack. Don’t just re-assign the CT to the other VT, but rather leave the config unchanged and physically move both the VT and the CT. That way we know both VTs are measuring the exact same voltage signal and both CTs are measuring the exact same current signal. Since it’s a class 1 meter, they should each be
±1.5% of reality, so worse case should differ by 3%, so about 23W if it’s the phase with the electronics or 18W ±1% of reality, so worse case should differ by 2%, so about 15W if it’s the phase with the electronics or 12W if it’s one of the other two (do it while the heatpump is running rather than standby). That would rule out any basic calibration/configuration issues. Unfortunately, it would completely screw with the power readings for any other circuits on that phase, so you might want to turn logging off while you do the brief experiment.
[EDIT] Since you have to disturb your logging to do this experiment, it might even make sense to do it for all 3 phases at the same time. Put all three VTs on the one phase, and all three CTs on the heatpump feed that corresponds to that phase.
Assuming that all checks out ok, there’s then a 60W difference between your two phases that don’t power the electronics and accessories. That might be down to a voltage imbalance, so knowing the three phase voltages at the time might also be useful. If it turns out there is a voltage imbalance, then the 60W imbalance may just be the way the 3-phase compressor reacts to that, or it may be that the VTs are sliding around on their PhaseError Vs Voltage graph. You can see pictures of that in various reports here: