@CrazyRabbit So, yes, I have taken electronics, and I have been responsible for power quality and provisioning professionally, though I am not an electrician. I do know one end of a screwdriver and DVM from the other, and I have had front row seats for more than a few power oddities over the years, some of which I diagnosed and corrected. This is not one that I understand.
The power factor of the load that was added that caused the voltage rise was 1.0, and it was small, between 0.5kW and 0.6kW, and the rise was about 3V. I don't know where the nearest large capacitor could be that might be connected on the distribution circuit, but it isn't here, or within at least 1/3-1/2 mile, and several distribution transformers. All of my neighbors are individually fed from a transformer, all 15kW, and I can't think of anyone who could have a bad capacitor that would be connected. As noted above, PG&E did have a capacitor bank on the circuit, and removing it dropped the grid voltage by about 3V, but didn't change the behavior.
The meter to main service panel is 18", 3/0 copper. House load only hits 5kW when the dryer runs, and a few times a year when the AC runs. Typical draw with lighting, and vampire drains is 0.2-0.3kW, but with the refrigerator, it can be 0.6-0.8kW. So, not a high draw home, and not much is left connected routinely. Nothing weird electrically, at least routinely, and certainly nothing new in the last nine months when the problem cropped up.
Yes, exporting 4-7kW does cause a 3V or so rise in the meter voltage due to the impedance here, but the grid voltage wanders up and down 242-254V during the day, with similar export or loads here, and the pattern doesn't seem to correlate with time of day particularly, e.g. there isn't a reproducible 6:00am drop when some load kicks on the line. There are only two non-residential accounts on this distribution line, and one uses next to no power. The other is ten miles away, with a number of transformers and loads (computers, small HVAC, nothing too big that I am aware of beyond a well pump). The three big well pumps that I know of on this distribution circuit are three phase, and two programmed to kick in overnight for cost reasons, and the third is programmed for 4-7am use when needed. Voltage drop on the latter is not really detectable, as it is only a 2.7kW load.
Solar, and powerwalls export power by advancing the phase. As
@wwhitney points out the export may cause a voltage rise due to the impedance of the transformer and HV line, but the voltage rise is not intrinsic to the export.
All the best,
BG
