Oscillating discharge in Multiplus 5000

Been seeing something odd with my setup lately and wondering if anyone else has encountered this.

Running a Multiplus II 5000 with a 15kWh Fogstar Drift LiFePO4 bank managed by a JK BMS, AC-coupled Fronius Primo on the output side, and a Cerbo GX pulling everything together via VE.Can. I've had ESS running in optimised (with BatteryLife) mode for about eight months without drama.

Over the past few weeks I've noticed what I can only describe as a rhythmic oscillation in the discharge current — roughly every 45–90 seconds the inverter seems to pull a sharp surge from the battery before settling, then repeats. It's visible on the VRM portal as a sawtooth pattern on the battery current graph, usually between midnight and 06:00 when PV is obviously nil.

My suspicions so far:

• ESS assistant feedback loop between the Fronius and the Multiplus creating instability at low loads
• BMS communication latency causing the DVCC current limiter to hunt
• A grid setpoint drift issue (I'm grid-tied but with island capability)

I've already tightened the grid setpoint to ±30W and updated to the latest Venus OS, which reduced the amplitude slightly but didn't eliminate it.

Has anyone seen similar behaviour — particularly with AC-coupled systems where the inverter is also managing a Seplos or JK BMS over CAN? Curious whether this is fundamentally an ESS tuning problem or something deeper in how the Multiplus handles reactive load at partial discharge.

Would be useful to compare VRM logs if anyone's up for it.

@ZoeGrant what firmware version is your JK BMS on? I had something similar with my Multiplus II (3000 not 5000) and it turned out the BMS communication interval was slightly out of sync with what Victron expected via the GX device.

Few questions that might help narrow it down:

• Are you seeing this on the VRM portal as actual oscillation in the discharge graph, or just on the CCGX/Cerbo locally?
• What's your DC bus voltage set to in VenusOS?
• Is the Fogstar bank all one string or paralleled?

The Multiplus can sometimes hunt if the charge/discharge floor voltages in the BMS don't quite match what you've programmed into ESS. Had to tighten mine up to within about 0.2V before it settled down properly.

@ZoeGrant classic symptom of the Multiplus and JK BMS not playing nicely over the charge current limits. The Multiplus is probably seeing the BMS request a current reduction, overshooting, then overcorrecting — hence the oscillation.

A few things worth checking:

• DVCC enabled in VenusOS? Should be, if you're running a GX device
• Reduce SVS and SCS charge voltage/current limits — sometimes tightening the allowed range settles things down
• JK BMS UART comms to GX can be flaky; worth checking your cable and baud rate settings

I had similar grief with my Fogstar bank early on. Turned out my BMS firmware was sending erratic SoC values which confused the Multiplus. @TrevorRoberts75 is probably onto something with the firmware question.

What GX device are you running, and is the BMS connected directly or via a Cerbo?

Worth checking your ESS assistant configuration specifically — if you're AC-coupled and running ESS, the frequency shifting behaviour can cause the Multiplus to oscillate when the BMS is simultaneously trying to throttle charge current via DVCC.

On my narrowboat I had something loosely similar when I first commissioned my Victron setup. The fix was ensuring DVCC was fully enabled in VEConfig and that the BMS had sole control of charge current limits rather than fighting the Multiplus's own charge profile.

Check in VictronConnect that DVCC > SVS and SCS are both enabled. If the JK is communicating over CANbus rather than VE.Direct, also worth verifying the BMS profile selected matches your actual battery chemistry — wrong profile can cause erratic current target behaviour.

What's your comms setup between the JK and the Multiplus?

@ZoeGrant had almost the exact same oscillation on my van build before I moved it into the cabin — turned out my Victron charge current was set so close to the JK BMS limit that they were basically arm-wrestling each other 24/7 like a couple of caffeinated toddlers.

Drop your max charge current in VictronConnect to about 80% of whatever your JK is set to — gives enough headroom that neither device feels the need to throw a wobbly. Also worth checking your DVCC settings in Venus OS if you're running a GX device; letting DVCC handle the current control rather than the BMS asserting it directly smoothed mine right out.