ESS suddenly started oscillating after 2 years of stable operation
My ESS installation has worked fine for about two years. The problem appeared suddenly, without any changes to the configuration.
When I switch ESS on, it starts oscillating almost immediately. With ESS off, everything is stable.
I attached some VRM plots showing the MultiPlus output, ET340 power per phase, battery power and VE.Bus voltage/current.
I also tried reconnecting the USB connection between the ET340 and MultiPlus and reverted the MultiPlus to firmware 3.73. The behaviour did not change.
Any ideas what could cause this?
I am wondering whether the ET340 could be the problem. Would replacing it with an EM24 Ethernet be a sensible test?
thank you for your suggestion
I checked the jumper and it was solid
no external current sensor was connected. An ET340 is used to measure power to and from the grid.
To be honost I don’t see self-oscillation on your charts, only normal activity. An ESS waits for deviation from the grid setpoint to adds or subtracts power. It can only react if a deviation is present and the response will overshoot linear to the change of load to balance the totalizers. I believe you worry too much.
Thanks, but just to clarify: we weren’t at home, so the normal load was around 400 W. The battery discharge limit is 800 W and the charge limit is 1800 W. The oscillations occurred mainly at night, when there was hardly any change in the household load. That is why I’m puzzled by the oscillations.
Here a picture of the console. The ESS is on, the battery is almost full. Normally the battery would discharge to get the grid to zero.
So no oscillations here now, but ESS not doing what it is supposed to do.
(I will make a video of the console as soon as the ESS shows oscillations. Thank you for the suggestion).
I switched ESS off before stopping the recording. The purpose was to show what happens when ESS is running and then show the stable situation after switching it off.
Without ESS, any conditions or conflicts that would affect it, cease to be an issue.
It is also not trying to balance the phases.
There have been a couple of topics on oscillations the last while, though no apparent indication why.
My System is as follows
• Multiplus II 48/3000/35-32
◦ VEbus 2639510
• Cerbo GX
◦ Firmware v3.75 (version 3.14 when installed)
• ET112 meter
• A touchscreen
• 3 Pylontech US2000C batteries
• 2 x 400W solar panels with an Enphase microinverter on ACOUT
• SmartSolar MPPT 250/85 with 4 x 400W solar panels
◦ Firmware 1.61
• Part of my consumer unit is connected to ACOUT, while the high-power appliances remain on the EDF grid.
• ESS system operational since 2024. The Cerbo is connected to my home Ethernet network, so I don’t know how the updates are carried out.
A few days ago, I accepted the update and saw the new interface. Then I went away for a few days. When I returned, the Multiplus was showing severe instability.
A 2,000W consumption was detected on ACIN, then on the AC loads, and then on the essential loads, even though nothing unusual was running (just two fridges).
When I switch off the fuse supplying the Multiplus’s ACIN input, everything works normally.
In the Cerbo’s system menu, I am offered the option to revert to version 3.70; the problem is the same. I don’t know what the previous version was, but I don’t remember accepting an update.
Interesting — our systems show quite a few similarities.
I also have a MultiPlus II, Cerbo GX, Pylontech batteries and an ET340 (same type of connection to the MultiPlus as your ET112).
My ESS had been stable for about two years. The problem appeared suddenly after I had been away for several days. The MultiPlus then started oscillating between charging and discharging, sometimes far beyond the configured 800 W limit.
I downgraded Venus OS from 3.75 to 3.73, but the problem remained. So in my case it does not seem to be simply a 3.75 issue.
I am currently waiting for my installer to replace the ET340 with an EM24 Ethernet. I hope this will help determine whether the problem is related to the ESS/grid-meter control loop.
It would be interesting to compare our observations and see whether others are seeing the same behaviour with 3.75.
