3-phase Quattro ESS: L2 stays in passthrough on-grid, works off-grid + abnormal rattling noise

Bit of an odd one and wondering if anyone's seen similar with their Quattro setups.

Running a 3-phase system across my off-grid cabin site — three Quattro 48/10000s managed through a Cerbo GX, with a Fogstar Drift lithium bank behind it. When I'm grid-tied (I have a DNO connection for backup), L2 stubbornly sits in passthrough and refuses to play ball with the ESS logic. L1 and L3 are behaving perfectly, shifting between grid and battery as expected. The moment I island the whole thing and go pure off-grid, all three phases work flawlessly. Bizarre.

I've been through the ESS assistant settings twice, checked phase compensation is set correctly in VEConfigure, and the firmware is current on everything.

The second issue — and this one's harder to ignore — is an intermittent rattling from the L2 inverter itself. Not a buzz, more of a mechanical rattle, almost like something's come loose on the transformer or a capacitor is vibrating against the casing. It's subtle but definitely wasn't there 18 months ago.

My gut says the two faults might not even be related, but I can't rule it out.

Has anyone had a specific L2 passthrough issue in a 3-phase ESS config that turned out to be a grid code or phase rotation problem rather than a hardware fault? That rattling is making me nervous about running it hard until I know what's causing it.

Would love to hear from anyone who's diagnosed something similar — especially if you've dealt with Victron support directly on a 3-phase ESS fault like this.

@VoltWendy Classic L2 passthrough-on-grid mystery — seen it before, usually a dodgy AC input current sensor on that unit or phase rotation mismatch confusing the Cerbo.

Check your VEConfigure settings — if L2's AC input current limit is set to 0A it'll just sit there like a useless brick in passthrough forever.

The rattling though... that's a separate headache. On my static caravan setup the Quattro fan bearing started doing exactly that before it died spectacularly on the coldest weekend of the year, naturally.

Pull the front panel and see if it's the cooling fan or something vibrating loose internally. Victron support are actually decent on warranty fan replacements if it's still in period.

Firmware version? Some older builds had phase-sync issues that caused exactly this passthrough behaviour on multi-unit 3-phase configs.

@Rob1963 interesting about the current sensor — but can I ask a daft question first? Has anyone ruled out the phase rotation being off on the incoming supply?

I'm only running a single Quattro in my tiny house so I've never dealt with 3-phase myself, but I've read that Victron's ESS is quite particular about detecting correct L1-L2-L3 rotation before it'll allow all phases to actively switch rather than passthrough.

Worth checking in VEConfigure what the AC input assistant is actually seeing on L2 specifically.

On the rattling — is it coming from inside the unit or could it be a loose busbar or cable terminal vibrating? Curious whether the rattle only happens under load or constantly, as that might help narrow it down.

@ZoeRoss not a daft question at all — phase rotation is absolutely worth ruling out first, it's a 5-minute fix vs hours of head-scratching.

@VoltWendy on the rattling — don't ignore that. Had a similar issue on my boat install, turned out to be a loose DC busbar connection vibrating against the casing. Not dangerous until it was. Worth pulling the front panel and checking torque on all connections while you're in there.

Also worth checking your VEConfigure settings for that L2 unit specifically — sometimes after a firmware update the AC input limit resets per-phase and causes weird passthrough behaviour. Check all three units independently rather than assuming they're synced.

What firmware version are you on? Some older builds had quirks with 3-phase ESS coordination that Victron patched a while back.

@BatteryRay @ZoeRoss solid points on phase rotation — worth confirming with a clamp meter or the Cerbo's grid meter readings before pulling anything apart.

One thing nobody's mentioned yet: on the Quattro 10000s specifically, there's a known interaction between AC input voltage setpoints and phase asymmetry. If L2's input voltage is sitting even slightly outside the acceptance window configured in VEConfigure, it'll hold passthrough rather than switch to inverting — but critically it won't flag a visible alarm, it just silently degrades.

Check your AC Input voltage range settings in VEConfigure3 across all three units and compare them. They should be identical. I found a subtle discrepancy on my shepherd's hut system where one unit had drifted to a narrower window after a firmware update.

The rattling is a separate concern — I'd be isolating that mechanically first. Loose DIN rail mounting or transformer lamination resonance are the usual culprits.