It is unfortunate that your system does not work in the way you expect.
Mine works following the charge state reported by the shunt and the inverter charger settings (charging curve and times). I chose this way because the shunt is a more accurate measure of voltage and current as verified by other measuring instruments.
Please let us know if/when you find with a solution
Do not use Jkbms data, just the BMVā¦ itās enough and simplify everything. Just use the proper parameters for each Jk you got to protect cells.
This is the very same problem as in JK V17 BMS. Itās not really flipping between charging and loading, itās just a display result of bad software in the JK BMS. This happens not only at absorption, but every time when PV-power and load+selfconsumption is almost the same.
When you measure the current to or from your battery with another tool, you will probably notice that charge currents are measured quite accurate, but discharge currents are measured 20% to high. When my BMS displays a discharge current of 18 A my independent amperemeter says itās only 14 A.
This behaviour has been introduced with one of the last V17 firmwares. Since then the BMS reports the SOC quite correctly.
Thanks for replying. In the end, I gave up looking for the problem (I use BMV-712 and the bms can is disconnected from cerbo).
The only concern is that when starts the behavior with charging/discharging, the delta on the cells jump from 0.010-0.012V to 0.080V.
Iām pretty sure that the change in the voltage difference between the cells has nothing to do with the jumping. Youāre currently in the absorption phase, and thatās where the balancing of the cells takes place.
My cells were already very well balanced from the factory, and to even begin balancing, I had to reduce the trigger difference from 10 mV to 5 mV. I set my system to start balancing as soon as the first cell reaches 3.45 V and the cell difference is at least 5 mV.
But to be clear: the problem of the display jumping between charge and discharge has nothing to do with Victron. Itās coming from the JK BMS.
On my side, the unbalance is starting when the cells reach absorption and current is under 20A. Until there everything is fine.
The problem is somehow related on victron ecosystem. For example if I use mppt (standalone) to charge the battery, everything works great. Bulk, absorption, float, all perfect. Means that jk is working correctly until is added to multiplus, cerbo, Ess.
At least this is the behavior from my testing.
In this case you probably donāt have an inverter running and therefore you donāt have a load on the battery. I have already told you that the jumping in my system can not only bee seen at a full battery, but also always when AC load + self consumption is almost the same as PV power. In both cases the system is close to zero power from the battery, and should only jump between a few watts in and out. But because of the wrong current measurement of the JK BMS the jump distance is much larger.
Please: just do an independend measurement of the current which is going in or out the battery. In my video you can see the charge/discharge power is jumping from +500W to -500W in not time. In my grid the Multiplus-II is not allowed to follow the load faster than at 400W/s, so this alone would be a good sign that there is no real power jump. But just see at the values from the MPII/cerbo in my video: neither the AC load nor the power going to the grid nor the solar power are changing. And this is impossible when the battery power is jumping with 1000 W difference.
No, the values of the battery in my case as well as in yours are reported by the JK BMS. Victron has nothing to do with it, and the victron numbers clearly state that there is no real power jump!
Did you het this all figured out?
Im looking at buying an
EEL 12V MB56 Lifepo4 Battery DIY Box Kits with Bluetooth 300A JK B2A8S30P BMS 2A Active Balance
Wondering if Iām going to have issues with the jk?
The JK BMS has problems measuring small currentsā¦ mostly during balancing.
However, since you want to set up a 12V system, there is nothing standing in the way of the JK BMS. With this low voltage, you have larger currents flowing through the BMS, which can then be measured more reliably.
With larger 48V systems, you have small currents in the base loadā¦ which can cause problems with the SOC over a longer period of time.
There is now also a current FW that fixes minor problems.
But if the shunt was poorly constructed, even that wonāt help 
Hi
No solution from my side. Iāve reported to JK support and I hope they can improve the firmware in the future. Until then is working like this. The bms can is disconnected from cerbo, and I use BMV-712 for controlling the charging.
Maybe Victron developers can do something to improve the compatibility with Jk. Are many users with same problem. Iāve built some batteries for Deye and Easun inverters and everything works great there. Somehow only with victron are the problemsā¦
I have the same problem. And i have a Deye.
I have tested a v15 and a v19.
2 bms. Same error.
I thought Deye was the problem.
I have read current with a clampā¦and and Deye is same as clamp.
Currentās BMSs go crazy when current is low..or in floatā¦..
Hi
Here is from bms
And here is from BMV-712. I donāt use CAN connection between bms and cerbo. Better without and smart shunt make a good job.
