Every time DESS triggers higher charge/discharge powers, Essential Loads measurement reported is predictably increased or lowered.
When the Multiplus is exporting/discharging, the Essential loads reported power increases to ~120-125% and when it charges the batteries it decreases, to ~80% of actual load.
Have anyone else observed similar behavior?
And the meter is assigned correctly in the GX itself as well as the loads on the i out and output assignment?
Yes, everything else works correctly and all measurements are within expected tolerances until there’s a large charge/discharge cunnret.
As LX already mentioned, it looks like a setting issue.
Make sure that:
ESS Grid metering is set for “External meter”
Where are the AC loads positioned ? Set it in “Position of AC loads”
Grid meter role is set to “Grid”
Looks good in settings.
Is the meter correctly installed on the grid side ?
What grid meter is it ?
DESS just started to charge at full power, and the “Essential Loads” dropped from 200-220W to about 150-160, I disconnected the external meter and changed the ESS settings to Inverter/Charger and the reported Essential Load power stayed at the same low reading when charging, so I think this is something else in the Multiplus II and me installing the external meter approximately at the same time was just a coincidence.
Edit:
It’s a Victron ET112 connected to my Venus OS over RS485
Is the increase/decrease also proportional to AC loads ?
What does it show if loads are 1000-1500W and trigger a charge/discharge ?
Do you have a MK3 interface to directly monitor the MP II ?
I have not observed any fluctuation due to high AC loads (between Grid meter and Multiplus), but I can test that later.
Yes, I have an MK3 Interface between the Raspberry Pi and the Multiplus currently.
The wiring between the Multiplus and grid meter is not optimal, it’s on a 10A (230V) breaker with aprox 15 meters of 1.5mm2 cable, so there’s some voltage drop over it when the Multiplus is on full 1700W charge/discharge.
I would consider the 1.5mm2 wiring below minimum.
It is definitely the main culpritt in your issue.
Is there any chance to test, or better change, with at least 2.5 or ideally 4 mm2 ?
The 4 mm2 might seem overkill but there is also a distance.
Edit after your post. At 3000+ W of power and mentioned distance, no less than 4 mm2.
All my installations of 3000 or 5000 MP II have 6 mm2 wiring and 32A MCB.
I need to clarify:
Incoming through the meter and to the distribution board is 6mm2. From the board to the 2kW load there’s a separate 2.5mm2 and to the Multiplus (only shared with Essential Loads behind it) is the 1.5mm run.
It’s not below minimum. Sure I would have used 2.5mm2 to the Multiplus if I re-did it, because I always rather overkill, but there’s only a 4V drop (1.7%) now on full load. (234.7V at meter, 230.8 at Multiplus).
At max load with full charge (the multiplus is a 3000 48V limited to 32A or ~1750W charge power) and my 220W Essential Loads the current in the 1.5mm2 run is under 10A. On discharge the Essential Loads is redacted from the current over the 1.5mm2 run so then it’s even less.
I would still try a thicker cable to the MultiPlus and watch the numbers.
Even a 2.5mm extender cable, for testing, will show a difference.
Besides the wiring story.
The essential loads may be including the inverters own self consumption figures and when it switches to grid charging, the self consumption load is now being fed by the grid.
Theory only thought right now. I have no way to check this.
This sounds more feasible! I only doubted the wiring as I thought the issue started with the external meter, but since it’s now tested on internal metering only the wiring should have no play unless the voltage goes out of spec low. Some fluctuations right when large loads start/stop is expected, but that’s not what’s we are discussing.
