Dyness powerbrick (float voltage 56,5V) + recently updatedd firmware
DVCC enabled, dyness powerbrick as BMS
I noticed some “high DC ripple warnings” when my battery SOC reaches 100% (independent of actual AC load)
I then checked the usual fixes (retorque bolts, check sizing,…) but still getting some alarms and then took a dive in the graphs/settings.
Graph 1:
ESS + excess grid feed in enabled + DVCC enabled
MPPT vs battery voltage similar below 100% SOC, max 0.2V ripple and independent of load or MPPT power = OK
As soon as SOC reaches 100% → 2V difference between MPPT and battery voltage, ripple spikes to 1V (and this is usually the point I get alarms). Ripple decreases as MPPT power decreases.
What is the advice to fix this issue? I would like to feed in surplus if possible.
For me it looks like the issue is due to big difference of Dyness BMS cutting of feed (typically around 54V) & float voltage (set in multiplus + MPPT as 56,5V per Dyness manual)
When excess feed in is enabled, the MPPT voltage is increased by 0.4V over the set value. The inverter runs to try to export this power, However, this can also lead to a rise in DC Voltage. Start by reducing the Voltage limit by 0.4V, and then slowly increase it until the problem re-appears, then reduce it slightly again.
We have seen moans about these batteries before. How many do you have?
iirc one powerbrick does not recommend discharge above 50 or 60A.
A 5kVA inverter feeding back unrestricted is way more than that so you are likely asking too much of the battery, hence the ripple, other cabling/installation issues not withstanding.
One powerbrick, according to technical specs recommended charge discharge rate is 140A. With peaks up to 200A. Therefore I assumed it would be acceptable to charge at 100A.
Interesting from the graph is that the voltage of MPPT0 was increased with 3V (!!!) from 54V to 57V, not 0.4V, in comparison with the battery monitor voltage, which jumped only 1V, when the feed-in was enabled… So the deltaV is about 2V…
Sure the cables and contacts are OK?
Not really. The charge is restricted to 100A which is pretty much the near 5000W the MPPT is producing. The voltage ramps up in the tail end of charge as the battery reached 100%, which is normal for lithium. It then adds the 0.4V offset to enable exporting.
Looks pretty normal from here.
I am talking about the second graph, from 13:30 on, when the SOC is 100% now for more than an hour…
Are you saying is normal to have more than 2V difference between the battery monitor and MPPT output voltage?
Red and orange graphs, I believe…
ah ok. That is weird that there is such a discrepancy between the MPPT and BMS readings. Something is off there.
Would he helpful to overlay the CCL/DCL limits and charger, inverter and BMS current readings
Same for CVL and all the other voltages.
Indeed…
But even on the first graph, up until a point they are going together, but then, after 100% SOC, they diverge a lot, having more than 2V difference, not only 0.4V…
Perhaps the BMS has disconnected.. Hence wanting to see the BMS limits. The battery details tab at the time also helpful.
These new Dyness batteries seem to have some emotional issue.
One easy (Neanderthal) test you can do, set your battery setting (in the inverter and GX) to whatever but not lithium and no DVCC and disconnect all Comms from the battery BMS to the GX and or etc, leave the battery without coms basically.
Set in the inverter, e.g. absorption 52V, and float to 51V, now use and see.
Then, connect first it coms, and see again, etc, step by step, and evaluate the behaviour.
To my untrained eye it looks like bms want to cut off charging (ccl becomes zero). Doesn’t work due to dvcc, then cuts internally? (But didn’t get any alarm/warning/… from battery itself untill now) and then Victron following the voltage set (which is quite a lot higher than battery voltage).
Reducing maximum charge voltage to 55V seems to improve ripple a lot. But unsure if this is good as a permanent solution (as float in dyness manual is 56,5V)
Reducing maximum charge voltage in dvcc from 56,5V to 55,8V has reduced ripples from >1V to around 0,5V (and alarms have decreases) and voltage gap between BMS - mppt voltage has reduced.
Dont forget their bms can override the internal programming of components. You have overriden their max voltage in dvcc which is the only way to do it.
