Strong (power) oscillating 3 phase MP2 installation (up to 2kW swings) @0,5 Hz

DIY
, , , , ,
1

I do have a strong oscillating installation and I cannot seems to get to the problem.

Installation information
System: 3 phase MP2 48/5000/70-50 230V (initially fw 552, later 555)
Batteries: 4x Pylontech US5000 (~19kWh)
DC: each battery with own pair of wires to Victron Lynx, 3 pairs 70mm2 to MP2’s
Controls: Cerbo GX (initially v3.31, later 3,54), wired ethernet
AC: only connected to AC in on 3x16A (system limit set to 13A, inverters limit @ 9000W)
Solar: to AC in, ~7kWp with 3 different systems. Measurements over ShellyEM’s to MQTT to Cerbo.
AC measurement: initially over Smart meter P1 and MQTT to Cerbo (1 sec update), later VM-3P75CT over modbus-UDP (100ms update)
ESS mode (with DESS)

Symptoms
strong change in ‘hum’ of inverters (with 0,5 Hz frequency), lot of changing/blinking of inverter led (because swing can be + and - 1000W). I cannot remember this behaviour after the installation (~1 year ago), but I am noticing it for about 6 months now.

Step 1 taken in trying to solve issue
Changed AC measurements to VM-3P75CT over modbus UDP
Upgraded Cerbo to 3.54 with factory reset
No change in situation. With the VM-3 data I was able to create the following graph:

Schermafbeelding 2025-03-11 152818
Schermafbeelding 2025-03-11 1528181920×592 120 KB

These measurements are from the VM3 (upper graph) with the 3 phases and total power (blue). Settings Cerbo: Grid measurement, setpoint 0W. Solar still connected.

Step 2 taken in trying to solve issue
Upgraded MP2 to FW 555
Tried setup with 1 battery at the time
Disable DESS
Results: No change

Step 3 taken in trying to solve issue
Disconnected PV
Results: No change

Step 4 taken in trying to solve issue
Switched from external AC measurement to internal AC measurments with different setpoints + and -
Results: No change, system keeps oscillating

I am slowly running out of options. Sometimes the system seems to be stable but I cannot find a relation with anything (yet), I will add more monitoring to search for relations with other events.

If I am correct, the Cerbo sends setpoints to the MP2’s, so I am suspecting the Cerbo somehow. It might receive wrong readings or processes the readings wrong. Can anybode guide me how to troubleshoot the Cerbo’s behaviour? Do I overlook something or are there other things to research?

Any help appreciated.

Frank

2

The 1 second update seems to be the problem.

Meter says “something is to high”, Victron reacts , Meter says “too low”

Could you programm a filter somewhere or set a deadband ?

3

Hi Ludo,

Thanks for your reply. I did not expect that to be a problem, since the oscillation also occurs when I select Internal meter (=MP2) in stead of grid meter (in the ESS settings of the Cerbo). To rule the 1 second readout being the cause of the oscillation completely, I installed the VM-3 grid meter from Victron as the first step (which is a recommende setup). This gives a 100 ms readout, but nothing changed in the oscillation…

Frank

4

The siemens sentron pac 3200 ( and probably others) can integrate measurements over programmable intervals.

And the janitza meters are probably even more powerful…and cheaper…

5

I do know the Janitza meters, but I specifically choose a Victron grid meter to prevent a discussion on meters which might or might not be compatible.

I still do not think it is a metering issue, since the same oscillation appears also when using the internal MP2 meters in stead of the grid meter. Furthermore, I bought a platform with advanced software specifically for (D)ESS, the Cerbo GX, and would expect that any filtering of input signals (if required) would be performed by the Cerbo.

6

Sorry, you’ll probably have to live with the oscillations until more people complain.

7

I notice your voltage also seems to be swinging with the same frequency, by 3 or 4 volts which is a lot for 500W (or 2a). Is that a cause, a symptom, or just coincidence?
Also, in other oscillation discussions (e.g. this) it was suggested to look at the dbus round trip times in VRM; in my installation, that is in the single digit ms range.

8

Hi Christian,

Thanks for your message. I started monitoring the dbus round trip times and they are fine (single digit) over a period of 3 days (also when the oscillation is occuring). I did not find out wheter the voltage swings are a sympton, cause or coincedence.

Since I did not have time to do detailed research the last few days I tried the following steps together:

  • completely flashed new firmware in the Cerbo
  • installed the supplied ferrite clamps around the cable for the GX touch 50 (installed a couple of weeks ago without clamps)
  • Set grid setpoint to 100W (in stead of 0W)
  • forced charging the batteries to 100% (first time after winter)

Oscillations seems to have dissapeared for know, Now I only see regular min/max peaks of 200W when the system adapts to the ever changing load (or higher peaks when a higher load is switched) but no oscillation anymore.

Finally I think there also might be a possible connection to construction works in the surrounding area and/or grid works. It seems the grid in my area is on the limits of its capacity and there are preperations for new transformer stations.

9

The oscillation is still there. My feeling is that I have grid issues, but are not able to monitor the grid voltage quick enough at the moment (I need better than 100ms sample time and the VM-3P75CT only reports this for the power, not for the current and voltage).

I am logging for a couple of days now and notice that the oscillations start just around sunrise (even with my own PV disconnected), so I suspect something is happening on the grid resulting in oscillation). Furthermore I also notice oscillation at about 2,5 Hz with lower amplitude.

Coming days I will do the following to figure out more:

  • switch off all loads (groups) at home, running logging and PC from seperate UPS and see how ‘clean’ my supply is and than slowly switching on different groups in the house again
  • logging grid voltage with scope to check for voltage peaks, oscillations and hormonics
  • get in contact with the DSO

Hope to get this solved