FWIW: The house is pretty new construction; the development, although small, was put together by local developers and has underground wiring; the transformer serving three of the houses is literally in my back yard, about 40 yards from the house; the service entrance for power is in the garage, the portion of the house closest to that transformer; the breaker box is at the service entrance, in the garage; and, finally, I’ve got solar power, and the the inverters and such are physically next to said breaker box, disconnects and all. So, you might say I’ve got a pretty stiff supply, here!
Next: As it happens, in my career, I’ve worked on no-kidding test equipment on and off. And not just as a user of $RANDOM voltmeters and such; I’ve worked in calibration labs back in the days before getting the degrees and for two memorable summers as a) a fix-it techie on a production line and b) a design engineer on some audio analysis gear. Lots of fun.
But that background makes me very wary of any calculations based on measured electrical parameters that claim to be accurate to 1% or better. Even bandgap voltage references, used everywhere as references for ADCs, aren’t that good. Even if one spent the big bucks to get a laser-trimmed one, the ADC itself would add to the inaccuracy. And we haven’t even got to the part where sensing component values drift with time, temperature, and voltage. This is why test equipment that promises 1% or better accuracies cost $$$ and need to go out to cal periodically. And equipment that can do stuff like that often have operating temperature ranges that keep them in air conditioned labs
.
In fact, a piece of gear that can actually make 0.1% accurate power measurements over a wide range of voltage, current, and temperature is in my garage: it’s the utility-grade electric meter hooked up to the output of all those solar inverters. I had a look at the guts of the thing and its toroidal current sensors, partly because I had to install it, and partly due to interest. The thing definitely had been calibrated; all the delicate stuff was under a thick layer of hardened goo, the better to keep moisture and air pollution at bay; and you can bet that the magnetics were good from -40C to 100C. It had enough certification stamps to kill somebody. All of which makes sense: people get pretty picky about energy measurements when there’s money involved, and wanting a box like this to keep its calibration for, say, 20 years or so.
And one pays for the privilege: that box, and its single, accurate function, cost around $180 or so. Somehow, I don’t think that Tesla, what with its mass-produced cars and drive to lower costs, would spend that kind of money. I would guess that with some clever engineering they might be accurate to 2, 2.5 percent. And 5% if they were rushing.
Now, the Superchargers, where people pay Real Money for electrical energy, probably do have some seriously accurate energy meters in there. But I don’t think the cars would.