How can the total consumption be splitted in REAL consumption & power consummed by the EV charger?
Both are counted in one number but when splitting these by substraction off the EVC power the value you get is not correct often even negative.
Reason is different speed & timing off publish.
Tried to calculate grid+sun-ev-battery should be consumption but also this doesn’t give correct values.
To have that maths work out, you would probably need to look at the EVCS builtin consumption calibration table and adjust that (per current setting) so its reported values match your system.
If you end up with a negative value when doing “consumption - ev”, your evcs is overreporting it’s consumption.
It may also be possible that the consumption figure is wrong due to meter settings, wrong phase settings etc.
Ac inverters allocated to the wrong AC-Port may result in wrong values, as well. (AC Inverters not known by venus obfuscate real values as well)
So, first step would be to verify the consumption without evcs running is right.
@dognose Yes it is rather correct without EVCS & Sun production. The reason why getting abnormal values is the fact that the values are not taken at the same moment (different measuring speed). At night (no sun) then the values are more accurate or realistic.
This is the Math I do
Consumption = N/+/grid/40/Ac/Power + N/+/pvinverter/31/Ac/Power - N/+/evcharger/40/Ac/Power - N/+/vebus/276/Dc/0/Power
Also tried to Calculate the EV power (Current x Actual Voltage) as this is faster (update voltage is same speed as the AC Power) but that isn’t the solution.
But the cause is that the Solar Meters are slower updating then the grid meter.
But Consumption is also a calculated value in the System. Any idea how this is calculated & how they get all values at the same moment because only then the math can be correct?
While checking deeper I see I am not the first one noticing the EVcharger power is not that accurate.
What I found is that the charger just measures current & then makes calculation to power. As long he is taking a correct voltage (measured by the grid meter for example) this should give relative correct values. If it is a fix voltage then it is logic the result doesn’t correpond with the reality.
Found this Mqtt path N/+/evcharger/40/Current if this gives me the real current value then I can try to recalculate the power and compare to see if its more accurate. Yet no car to test …
Its’s not calculated event-based as mqtt message handling would be.
It calculates at a certain interval with a correlating data set “as of now”.
Very slow devices may still cause some inaccuracy, if they dont change between multiple calculations.
So if I understand correct : calculation is done with the last known value off each topic. As every value updates in the system at a different moment you can never have a correct consumption value.