I’ve a 3-phase electricity setup with 2 single-phase solar installations, linked to phase 1 & phase 2 (phase 3 has no solar injection). I’m using 2 Shelly EM’s to measure the input of both solar installation (2 different phases, so I need to ensure each Shelly is getting power from the right phase).
The new solar installation is showing a nightly returned energy of about 14,5 W, which I find rather high. The old installation also shows a nightly returned energy, but limited to 3,5W. As a result, the energy dashboard in HA, is showing a constant return value each night.
PS: when the night is over, I see limited positive peak where the inverter is activating, showing it is (for a short moment) consuming power, which also suggests me that the inverter (Solis) is turned off at night. The old inverter (+15 year old, no datalogger) doesn’t have this behaviour.
It’s an Gen 1 EM. I don’t think it is interference, as I tried with both clamps of the EM: 1 close to the breaker, one further away. Both have the same readings. Not sure if it a callibration issue - anyway: that is not something I can influence, I guess.
Green is the wire with incoming solar energy. Goes to the breaker, which is where the Shelly gets it power (blue). Red is solar energy going to the into my homecircuit (other location below).
I’ll check what the effect is tonight when I would if I turn off the inverter (as it is in another room, it has an extra breaker over there which I’ll turn off). Probably it will all go to zero.
Measurements of this morning, including 15/20 min of positive power usage by the new inverter (yellow/red). The old one doesn’t have this. I do find this a bit weird, the solis manual doesn’t really specify this behaviour.
Confirmed, when the inverter is shutdown (via his own breaker), the power becomes 0. So it seems the inverter is passing down an constant return of 15W for no apparent reason (during the night).
I’ve asked both Solis & Shelly for a logical explanation & resolution. But both parties cannot give any other explanation then: it must be some kind of inductive/capacitive noise in the network. In short: we cannot anything about it.
As a result, I’m not going to monitor the Solis with the Shelly EM, 15W noise constant is too much of a false delta. Instead, I’ll get the value from the Solis cloud API’s. I’ll use the EM for some other circuits.
