Inverter daily reboot?

[erikgeurts]

I noticed that every day, a few minutes before midnight, there are a few entries in the HA system log from this integration.

For example, there is an error like:

Error fetching SolarEdge Coordinator data: Modbus/TCP connect to 192.168.86.98:1502 failed.

and a couple of warnings:

Failed to connect
Failed to connect [Errno 111] Connect call failed ('192.168.86.98', 1502)

and

Cancel send, because not connected!
No response received after 3 retries, continue with next request

I also see that the binary sensor 'grid status' is briefly unavailable around that same time.

Does anyone know if perhaps the inverter does some sort of daily reboot just before midnight, and if that's expected behavior or something to be concerned about?

[RaAdNe]

@erikgeurts , I am having the same issue, with both devices inverter and smart meter going "unavailable" around 22:00 every day. About 17 seconds later after beeing back it seams all the inverter sensores are available and provide valid data. However only the smartmeter sensores show device type with its attributes and SW version show valid data. Also sensor xxx_b1_energy_import show a gap then until next charing from solar start (charging from grid disabled due to local regulations).

Due to an recent battery upgrade, SolarEdge SW updates were initiated by an electican with installer app access, now:
Inverter 0004.0023.0036,
Smartmeter 79,
Battery 48V DCDC 3.5.8 (all individuall blocks loaded the same version, correctly reported as one device)

It takes between minutes and several hours till the smart meter data sensors get back from "unknown" and provide valid data. This affects all self generated sensors based on Modbus Multi sensor xxx_m1_ac_power.
When back with valid data Modbus Multi sensor xxx_m1_ac_energy_imported data do not make sense either, i.e. it's missing energy consumption when substracting the new and last valid values (sensor data add to about 10Wh per hour, while my base power load is about 200W giving 200Wh per hour - a few network devices adding up ;-).

Since I migrated my system recently from HA in a docker container to HA OS on new HW and the same time changed the DB, I can't say if this issue was there before the current SolarEdge FW versions.

As workaround I tried an automated re-load of the ModbusMulti integration (20s after sensor xxx_m1_meter_events is back with valid data) to enforce an refresh of all sensors, but it didn't help. The issue seem to be with the SolarEdge Firmwere.

Hope somebody knows how to solve the issue.

[WillCodeForCats]

You can check if this is the reason by downloading your diagnostic file before it goes unavailable and then download it again when it's showing unavailable then comparing the values.

If energy sensors come back with a value lower than before they will show unavailable state since these sensors should not be allowed to go backwards (Home Assistant expects total increasing value). Once they start increasing again it will come out of unavailable state.

[RaAdNe]

@WillCodeForCats thanks for your prompt reply, I'll try to perform these befor and during unavailable asap.
However it not just the energy sensors affected which "copy" the integrated values of the source and check for such backward counting issue. In with my system smart meter sensores for power, frequency, voltage and current sensors are also affected.

I'll post an update as soon as I have the diagnosic files of the next event..

[RaAdNe]

Today it got even more strange, missed the pre-event time by minutes (21:56:05), so data extract only generated after the inverter and smart meter came back to from unavailable:

• back from unavailable (21:56:20) : meter type, SW Version and events - correct, all other sensors in state "unknown"

• reload integration (22:04:05 - 22:04:25): false data for meter type and SW version - meter event still 0, all other sensors remain in state "unknown"
  After integration re-load - config_entry-solaredge_modbus_multi - limited to meter data.json

• restart HA (22:15:55 - 22:16:14): false data for meter type and SW version - meter event still 0, all other sensors remain in state "unknown"
  After Inverter and meter return - config_entry-solaredge_modbus_multi - limited to meter data.json

• 22:35 sensor data for sensor.solaredge_1_i2_m1_ac_energy_exported and sensor.solaredge_1_i2_m1_ac_energy_imported back with resonable data, all other sensors still unchanged, i.e. false data or "unknown"

• 22:38:59 the meter frequency sensor came back and reported values, however HA UI got very slow nearly not reacting, UI not showing all parts of smart meter in the Lovelace card (SolarEdge PV Anlage), I restarted the whole HA host system i.e. Raspi (on CM5). Result, while nearly all meter sensors have usefull data, the meter type and SW verison still are wrong.

So is this an smart meter SW (or even HW) issue or is this an issue with the Solar Edge Modbus Multi integration?

- 09:25 Update:

• until 09:01 Smart meter device type was still false, assuming it's an SolarEdge SW or HW issue I switched off restarted the whole PV (takes about 10 minutes for the whole system):
• 09:03:29 inverter capacitor discharged, smart meter state "unavailable"
• 09:07:40 smart meter back while inverter still in start-up procedure. Smart meter device type remain false!
• 09:10:13 inverter finisched start-up and in normal operation (I_STATUS_MPPT), meter sensor data still missing (except device type and SW version - both still false data)
• around 09:23 re-downloaded SolarEdge Modbus Multi integration via HACS and restarted HA, 09:23:54 smart meter sensor device type and SW Version - now back stating correct values,

I am not sure if correct, but it seams that something with the SolarEdge Modbus Multi integration went wrong, may be with latest HA updates 2025.8 to 2025.8.3.

Hope this helps others having similar issues and to find the cause if its with HA or the SolarEdge Modbus Multi integration.