My Axpert 5kva/4kw started behaving erratically. I have had it for more than 3 years. It shuts down with fault 03 - Battery Voltage too high, but the voltage displayed is 19V while the actual voltage is 49V. This only happens when the main power is down at night with no solar input. It it sometimes recovers, but I also have to restart it by hand. During the day with normal solar input and mains it works perfectly. Any ideas? The battery is a LiFePO4  and is also more than 3 years old. The capacity is probably down to 50%, but I have not checked that.

2 hours ago, Hennie_Strydom said:

It shuts down with fault 03 - Battery Voltage too high, but the voltage displayed is 19V while the actual voltage is 49V. This only happens when the main power is down at night with no solar input.

It sounds like the inverter's battery voltage measurement resistors have gone high resistance, resulting in a low voltage reading, but that doesn't explain why the fault code is 03, battery voltage too high. The Solar Charge Controller has its own battery measurement, so when solar is available and the solar charger's processor is running, it can send the correct voltage to the DSP. At night, this is not available, hence the fault.

2 hours ago, Hennie_Strydom said:

The battery is a LiFePO4  and is also more than 3 years old.

The LFP battery at 3 years should not be badly degraded. But perhaps if it has high internal resistance, or if the BMS is disconnecting from the inverter, strange things could happen. I'd say it's much more likely that the sense resistors in the inverter are faulty.

If you are handy with electronics, these are the resistors to look for:

2 hours ago, Coulomb said:

It sounds like the inverter's battery voltage measurement resistors have gone high resistance, resulting in a low voltage reading, but that doesn't explain why the fault code is 03, battery voltage too high. The Solar Charge Controller has its own battery measurement, so when solar is available and the solar charger's processor is running, it can send the correct voltage to the DSP. At night, this is not available, hence the fault.

The LFP battery at 3 years should not be badly degraded. But perhaps if it has high internal resistance, or if the BMS is disconnecting from the inverter, strange things could happen. I'd say it's much more likely that the sense resistors in the inverter are faulty.

If you are handy with electronics, these are the resistors to look for:

Thanks so much for the information. Do you think it might also help to upgrade the firmware? The inverter also sometimes does a power glitch for about a second. During the day it often shuts down the solar input and if I do not tweak it, it remains off for the entire day. I fix this with a workaround using a raspberry Pi via the serial port of the inverter. I give a command to bump the battery CV voltage up 100 millivolts and then 1 second later bringing it down again. This has been working well for 3 years, but I believe it is a firmware bug.

6 hours ago, Hennie_Strydom said:

Do you think it might also help to upgrade the firmware?

What firmware are you running now? I'm not aware of problems like you describe.

12 hours ago, Coulomb said:

What firmware are you running now? I'm not aware of problems like you describe.

The firmware is:

Main CPU 72.70

Secondary CPU 4.12

I can see now that the problems only start when the system has been running on the battery alone with no solar or mains. The behavior is quite erratic. It sometimes messages mains power down, when it is not or even over current fault when the load is just 70W. It seems as if the measuring system is faulty in this mode.

14 minutes ago, Hennie_Strydom said:

I can see now that the problems only start when the system has been running on the battery alone with no solar or mains. The behavior is quite erratic. It sometimes messages mains power down, when it is not or even over current fault when the load is just 70W. It seems as if the measuring system is faulty in this mode.

Have you tried to disconnect the inverter from the BMS and run charging management based on battery voltage only?

21 minutes ago, Beat said:

Have you tried to disconnect the inverter from the BMS and run charging management based on battery voltage only?

There is no BMS connection between the inverter and the battery - only the 48V terminals are connected.

4 hours ago, Hennie_Strydom said:

Main CPU 72.70

That's the second most recent firmware available for that model, and I don't recall any massive problems with it.

Be aware that you can't run the latest factory firmware for your model (73.00) in unpatched form, because you'll eventually run into fault code 90. You could run the LFP flavour of patched firmware version 73.00e to see if that fixes it, but I'm not hopeful that it will.