Power mismatch

[lachesis]

I'm using esphome with my CircuitSetup 6 channel meter to monitor my home power usage. I live in the USA, with split phase 120V/240V. I have one phase of the mains connected to a CT but I cannot fit a second CT in my rat's nest of a panel right now, so I am concentrating on tracking particular loads instead.

For example, my home AC compressor uses 240V. I have a CT on only one of the legs of this appliance. Right now it is reporting a current of 19.10A and a power usage of 1814.44W. My single voltage sensor reports 121.3V. I happen to know (from another monitoring system) that the other phase of my power is considerably lower, around 116V. Still, the split phase voltage should be ~237V. Given this, I would expect 19.1A * 237V to give an apparent power of 4526W. Given the reported (I presume) real power of 1817W, that's a power factor of 0.4. Now I know the power factor of inductive loads can be pretty poor, but I find myself surprised that it would be that bad.

For another example, I have my electric resistive stove (with a presumed power factor of 1) connected through a different CT. That one is reading 35.72A when the stove is on, but only 4285W. I would expect it to read closer to 237V * 36A => 8532W.

So, how does power calculation work in this system? Why am I seeing such strange values? And how can I get some more confidence in these readings?

Here is my esphome yaml config.
esp_powermeter.yaml.txt

Here are the readings I'm receiving.

[CircuitSetup]

Sorry I wasn't able to reply to you sooner!

In your config you have a multiply: -2 for power on CT1, CT2. If you're getting a negative value, the fix should be to turn the CT around on the wire. This is probably why you are seeing low power factors. ESPHome is able to output power factor from the meter, so you can also add that in your config to see what is happening.