FoxESS Community - Owners & Installers Forum

Has anyone noticed that the PV Power sensor seems to underreport with recent firmware (KH7, 1.37 to 1.58)?

I have a Riemann integral which provided a more precise value for solar generation today, but ever since the firmware updated, it has been underreporting by about 10% compared to the 'solar generation' (total and today) sensor, and I think this is because sensor.pv_power is low.

The definition is here:
Here it shows before & after the update, note that the Riemann tracks the solar generation today sensor quite well prior to the update (although I do purge it into long-term-statistics after a day or two)
So in this example, note that pv_power is reporting zero, while the solar generation today sensor is already starting to increase.
Another 'before'

It looks like calculating power from sum(voltage*current) gives about 80w higher than the pv_power sensor. As far as I can see, calculating it from voltage*current results in the same energy figure as from the Solar Generation Today sensor, while the Riemann sum directly on the inverter power sensors gives a significantly lower reading.

I wonder if they have moved the parasitic load overhead of the system from the load sensor to the pv sensors. I had noticed something similar with the load power vs the load energy sensors on the old firmware, where the Riemann sum on load power gave a different result to the load energy sensor.
For Load Energy, the Riemann is closer after the update than before.