Feed-in limit (MaxFeedInPower) not respected in "Total of all phases" mode with DC-coupled PV

Q&A and troubleshooting
, , ,
1

Hi,

i have a Problem with my system and need your help.

System:

  • 3x Multiplus-II 48/3000 in three-phase configuration
  • Cerbo GX (latest firmware)
  • 3x Victron MPPT (DC-coupled PV, 10 kWp)
  • Pylontech batteries (19.2 kWh)
  • ESS mode: Optimized without BatteryLife
  • External grid meter
  • Multiphase regulation: Total of all phases

Issue:

When the battery reaches 100% SOC and DC feed-in is enabled, the configured MaxFeedInPower limit of 300W is regularly exceeded — often reaching 1000–1600W — for extended periods of up to one hour before the system eventually returns to the configured limit.

The pattern is consistent and reproducible: the excess feed-in begins exactly when the battery is full and the Pylontech BMS reduces the CCL (Charge Current Limit) towards zero.

Important clarification on MPPT behavior:

The MPPTs do throttle down significantly — they are not running at full power. However, the throttling is insufficient to keep the feed-in within the configured 300W limit. The ESS control loop appears to be unable to regulate the DC-coupled feed-in precisely enough in this state, resulting in sustained over-export despite active MPPT throttling.

Key observation:

When switching Multiphase regulation from “Total of all phases” to “Individual phase”, the 300W limit is respected correctly and the MPPTs throttle to the appropriate level. Switching back to “Total of all phases” immediately reproduces the problem.

This strongly suggests that the feed-in limiting logic for DC-coupled PV lacks the necessary precision in “Total of all phases” mode. In “Individual phase” mode, each Multiplus regulates its own phase independently, which appears to provide tighter control over the DC feed-in power. In “Total of all phases” mode, the cross-phase power balancing seems to interfere with the feed-in cap, preventing accurate enforcement of MaxFeedInPower.

Disabling DC-coupled feed-in entirely also resolves the issue — the system holds 0W feed-in cleanly even with a full battery — which confirms the MPPT throttling mechanism works in principle, but loses precision specifically in the combination of DC feed-in enabled + full battery + “Total of all phases” regulation.

Question:

Is this a known limitation of the feed-in limiting implementation in “Total of all phases” mode? Is there a workaround that allows using “Total of all phases” (which we rely on for balanced phase compensation with unequal loads) while still reliably enforcing the MaxFeedInPower limit for DC-coupled sources?

Thanks a lot!