3-Phase MultiPlus-II ESS instability

Had something similar crop up with my garden office setup last year — not 3-phase but same ESS weirdness where the system kept hunting between grid and inverter modes for no obvious reason. Drove me absolutely mental.

Few things worth checking if you're running that VM-3P75CT meter:

• CT clamp positioning — even being slightly off-axis killed my readings and sent the ESS into a loop
• ESS minimum SOC settings — if your Pylontech BMS and the Cerbo GX are disagreeing on state of charge, it'll thrash constantly
• Grid code settings — in the UK you want G98/G99 compliance sorted, wrong grid code can cause the MultiPlus to disconnect and reconnect repeatedly

I'd pull the VRM logs first and look at the grid setpoint graph. If it's spiking all over the place rather than sitting flat, it's almost certainly a metering or comms issue rather than the inverters themselves.

Also worth checking your Pylontech firmware — there were some known issues with older US5000 firmware interacting badly with certain Victron firmware versions. Fogstar sell the US5000 here and their support team actually know their stuff if you're stuck.

What firmware are you running on the Cerbo? And are all three MultiPlus-II units on identical firmware? Mismatched firmware across a 3-phase stack is a nightmare waiting to happen.

Anyone else seen this with 3-phase ESS in the UK? Curious whether this is more common than people admit.

@PeakVanLifer the hunting behaviour on 3-phase ESS setups is almost always one of three things: AC input current limit set too low causing the units to briefly overload and drop to inverter mode, grid code settings mismatched between the three units (they must all run identical firmware and identical ESS assistant versions — even one unit on a slightly different assistant build will cause phase coordination issues), or the UVA/UVF settings fighting with your DNO's actual grid quality.

Pull the VRM data and look at the grid voltage waveform across all three phases simultaneously. If you're seeing micro-dips on one phase that the others aren't tracking, that's your culprit.

On my narrowboat I had similar assistant-version drift after an update applied mid-sync. Rolling all three back to the same build sorted it immediately. Victron's three-phase guide is explicit about this but easy to miss.

@LH_Marine has covered the usual suspects well. One thing I'd add from bitter experience on my narrowboat install (single-phase but same ESS logic applies) — check your grid code settings match your DNO region exactly. I had persistent hunting that turned out to be the wrong grid code selected after a firmware update silently reset it to a generic European profile. VE.Config showed everything "normal" but the MultiPlus was second-guessing the grid constantly.

On 3-phase specifically, also worth checking whether all three units are running identical firmware. Mixed firmware across a parallel/3-phase array causes exactly this instability and it's a surprisingly easy thing to miss if you've updated units individually rather than via a single VRM push.

@PeakVanLifer this rings bells from when I was commissioning my tiny house setup on a single Multiplus-II. Spent three evenings scratching my head before I traced the culprit to the grid code assistant settings not matching what my DNO actually required. The system was essentially arguing with itself about what "grid" meant.

On 3-phase specifically — worth checking whether all three units are running identical firmware. I once had a situation where one unit had auto-updated overnight and the others hadn't. The phase master just could not reach consensus with the other two. Classic.

Pull the VRM logs and look for rapid ESS state changes around the same time of day. If it's time-correlated rather than load-correlated, you're likely looking at a grid quality trigger rather than anything internal to the inverters.

Really interesting thread — we've got a 3-phase MultiPlus-II setup on our floating home and went through something similar about six months back.

One thing nobody's mentioned yet: have you checked whether your three units are actually running identical firmware? Ours were showing exactly this hunting behaviour and it turned out two units had auto-updated via VRM but the third hadn't, leaving a version mismatch across the phases. VEConfigure showed everything "normal" which made it maddening to diagnose.

Also worth checking — are all three ESS assistants configured with identical parameters? Even a small discrepancy in the minimum SOC setting between units caused grief for us.

What does your VRM data show during the hunting episodes? The advanced widgets showing AC input power per phase are really useful for spotting which unit is actually triggering the mode switch.