Battery not discharging although shortage of power

New user (1 à 2 months) of Victron Equipment.
Situation:
Belgian home, 3 fase N400V.
3 SMA inverters distributed over all phases
4kW on L2 (3360Wp SouthEast, 3360Wp SouthWest)
1.5kW on L3 (1800Wp on SouthWest)
2.5kW on L1 (3520Wp on NorthWest)
Alfen Pro Line EVSE, connected with one step phase rotation “L2L3L1”. Meaning that L1 of the Alfen is connected to the L2 of the grid/house. This is done to most efficiently make use of the PV power as the 4kW is the biggest PV generator and I can instantly use that power when generating power without going extra through the meter (“Nul Op Meter”)

Victron MP2 5000VA connected to L1
All SMA PV inverters are connected to GRID, not to the Victron AC OUT
Cerbo GX
EM24 as Grid meter
Smartshunt
Battery 15kWh with JKBMS
DVCC enabled
Configured, ESS Assistant enabled

During the day (lately we had some nice sunny days) I noticed that the Victron is not discharging to what it can on other times.

The following behavior occurs for example and we can’t explain why this is happening:

PV generating on all phases.
Usage (heavy) on phase L2 because the car is charging.
The car is pulling about 5000W but the Victron system is not compensating the usage of the grid at that time.

Does anybody know of any setting that might prevent discharging while there is already quite some PV power being generated? No amp limits are being hit.
Afaik all relevant settings (but I would like to hear from you what I might have missed) are correct:

As you can see above the system is more than capable to delivering hundreds of watts. I have seen discharges of 4200W (and even more I think).

@independer

Are you pulling 809W from net constantly? Elle thus might be temporary due to change PV earnings.
The MP2 is changing 400W/s with done grid codes.
The “total of all phase” will select the most efficient inverter set point to generate the needed amount of energy summed up on all phases. If your grid meter does not measure sun of phases, you should change this.

Yes, this was continuously (I mean: pulling from the grid).
It might have been some more or less but always a portion being pulled from the grid and most certainly not being compensated for every second.
The weird thing is also it keeps at around -70W, as if something is blocking or restricting more power.

PV Earnings/PV kW generated at that point were pretty constant, no fluctuations on that level.
I am/was aware of the per 400W steps on grid.
Grid meter is an EM24. Afaik this meter is fully supported according to https://www.victronenergy.com/upload/documents/Datasheet-Energy-Meters-Selection-Guide-EN.pdf.

During the night there is absolutely no issue, the MP2 is perfectly able to keep up with all the fluctuations.

Screenshot shows: Grit Setpoint = -50W
You instructed the system to push 50W all times.
20W can be be steering latency depending on your energy meter and location of loads. I have the 10Hz Victron em-3p75ct. EM24 has 1Hz.

Screenshot shows: Grit Setpoint = -50W

Correct, that is indeed what I want

You instructed the system to push 50W all times.

Correct, that is what I expect it to do, I agree.

20W can be be steering latency depending on your energy meter and location of loads.

What is the 20W? I understand your previous point but where does this 20W thing come from? I am having some trouble understanding this. Can you further clarify?

I have the 10Hz Victron em-3p75ct. EM24 has 1Hz.

OK, so you are implying the EM24 is not fast enough to have the Victron steer more power towards grid although I am using said power for multiple seconds?
What surprises me is that it is capable of dealing with power usage and keeping a zero on the meter, just not when I am also generating quite a lot of PV power while also needed battery power/discharging.

Just making sure I am understanding you, because I selected this meter as it is on the list of Victron, it is a passthrough meter and it is accessible over ethernet. So it checked all boxes in my analysis(?)
If this would be the case, the solution would be to install a 3P75CT and re-verify?

Can we verify this first via logs via Victron? If so, how?

@independer

The 20W is the diff between the -50 und -31 in the screenshot. Sorry mixed it up.

I saw the 809W grit screenshot shows different voltages from BMS/MP2 vs SmartShunt.

My MP2 was not using full potential, when MP2 and BMS hat 0,5V diff.

I installed a dedicated voltage sense line from MP2 L1 master to battery poles. This compensated the voltage drop in my 50qmm battery cables at 70A.

Maybe select JK-BMS as BMS not the SmartShunt.

Ok, so the 20W was a mistake and not relevant.

Yeah, the Smartshunt is “just there for comparison” but is not used as battery monitor (see screenshot) as I like to gather some data between JK BMS and Shunt (for fun now).

Voltage sense line: OK, but looks like a long shot as it works quite ok for 4000W at other times.