This is to document something observed twice (by me) now, which by the tip-of-the-iceberg principle means there might be more to it. Not necessarily to do something about it immediately.
When the ESS mode is set to Keep Batteries Charged, the Multi relies on creating an overvoltage of 0.4V, and feeding this to the loads. That way the battery remains full, while the excess is used to power the loads.
If the battery sends CCL=0 before the CVL (charge voltage) is reached, this makes it impossible for the solar chargers to create an overvoltage.
Additionally, the DC-consumption of the Multi is always added to the CCL sent by the battery, in an attempt to roughly approximate the requested CCL. When CCL=0, this means the solar chargers are limited to the DC consumption of the Multi. But this is not an accurate measurement, which generally causes the MPPTs to be limited more than is required, causing solar production to death-spiral.
The scenario usually culminates with the solar charger(s) limited to 1 Ampere, with most of the energy fed from the grid as almost no overvoltage can be created.
At the moment the only solution is to adjust the DVCC charge voltage down.
Currently seems to affect Dyness batteries more than others. One solution might be a quirk for Dyness batteries.
Another solution might be to handle this as part of a better way to handle CCL, when feeding in the excess DC-coupled PV is requested. At the moment, CCL is ignored at any site where excess DC-coupled PV is fed in.
This is to document something observed twice (by me) now, which by the tip-of-the-iceberg principle means there might be more to it. Not necessarily to do something about it immediately.
The scenario usually culminates with the solar charger(s) limited to 1 Ampere, with most of the energy fed from the grid as almost no overvoltage can be created.
At the moment the only solution is to adjust the DVCC charge voltage down.
Currently seems to affect Dyness batteries more than others. One solution might be a quirk for Dyness batteries.
Another solution might be to handle this as part of a better way to handle CCL, when feeding in the excess DC-coupled PV is requested. At the moment, CCL is ignored at any site where excess DC-coupled PV is fed in.