1 grid-forming Quattro cluster + multiple Mode 3 follower clusters on islanded bus with AC source?

Been mulling this one over for a while, so figured I'd throw it open here.

My shepherd's hut setup is relatively modest — single-phase Quattro 5kVA — but I've been helping a friend design something considerably more ambitious: multiple independent Quattro battery clusters sharing a common islanded AC bus, with one cluster acting as grid-former and the others operating as followers.

The core question is whether this actually works reliably in practice, or whether it's one of those architectures that looks clean on paper but causes grief in the real world.

Victron's documentation suggests you can have follower inverters synchronise to a grid-forming Quattro on an islanded bus, essentially treating that formed grid as their AC input source. What's less clear is:

• How tightly do the followers actually track frequency/voltage transients from the grid-former during load steps?
• Is there a practical limit on how many follower clusters you can hang off one forming cluster before phase stability becomes an issue?
• Does anyone have real-world experience with this using VE.Bus or ESS configurations, rather than just bench testing?

The application here isn't three-phase — it's multiple single-phase clusters on the same bus, each with independent battery banks, acting as a sort of modular capacity expansion. The grid-former would be tied to a small diesel genset for backup sync purposes.

Keen to hear from anyone who's actually deployed something similar in a UK off-grid context, particularly with Fogstar or CALB cells on the follower banks where BMS communication adds another layer of complexity.

Has anyone hit unexpected issues with protection relay coordination in this kind of arrangement?

@BatteryEmma cracking question and one I've been noodling on myself. The short answer is yes, this can work — Victron's ESS documentation covers exactly this topology — but the devil is absolutely in the detail of your VE.Bus configuration.

Your grid-forming Quattro needs to be the undisputed "king" of that islanded AC bus. The follower clusters must be set to External Control / AC Slave mode, otherwise you'll get competing frequency references and the whole thing turns into an expensive argument.

Key gotchas:

• Exact firmware parity across all units is non-negotiable
• Synchronisation cabling between clusters needs to be screened and kept short
• Make sure your Mode 3 EVSE is seeing a stable 50Hz ±0.5% before it'll play ball

What's the intended load profile? That'll determine whether your grid-former needs paralleling or a beefier single unit.

@BatteryEmma @Titch the bit nobody mentions is what happens when your AC source (gen, whatever) has a wobbly moment. The follower clusters are sat there merrily syncing to the grid-forming Quattro, but if the source gets hiccups, you need your grid-forming cluster's AC input settings extremely well tuned — otherwise the whole islanded bus has a meltdown and everything falls over simultaneously. Done this on the narrowboat with a cheap Chinese genny and it was... educational.

Key thing in VEConfigure: set your grid-forming unit's LOM detection appropriately for a genuinely islanded bus, because default settings assume utility grid characteristics. The followers will inherit whatever the bus looks like, so rubbish in, rubbish out. Victron's ESS documentation is misleading here — you're not running ESS, you're running something considerably more bespoke. Test under load before you rely on it.

@BatteryEmma worth flagging that your follower clusters need to be properly configured for Phase Compensation in VEConfigure, otherwise each cluster tries to optimise for its own battery bank and you get real imbalance issues on the bus.

Done something similar (smaller scale) on my boat — two Multiplus units, one grid-forming, one following a shared AC bus. The key thing I found was setting tight AC input current limits on the followers so they're not all fighting to absorb from the source simultaneously.

Also: comms between clusters matter more than people think. VE.Bus or at minimum a shared CCGX/Cerbo giving a unified view. Without that you're flying blind when something goes sideways.

What AC source is your friend planning — gen or grid-tied inverter? Makes a difference to how aggressive you can be with the follower response settings.

@BatteryRay raises a solid point on Phase Compensation, but I'd add that AC coupling limits deserve equal attention here. Each follower cluster's grid-tie inverters are still subject to the frequency-shift curtailment mechanism — so if your battery banks on the follower clusters are already full when the AC source kicks in, you can hit a situation where Quattro's frequency ramping conflicts with what the follower clusters expect.

In my van build I run a single Quattro but I've tested multi-unit behaviour on a friend's three-phase cabin setup. The ESS Assistant vs Micro-Grid Assistant choice matters enormously on the follower units — wrong assistant and the handover sequencing under source instability (which @Boycie25 touched on) gets unpredictable fast.

What firmware versions are you targeting? VE.Bus firmware mismatches between grid-forming and follower units have caught people out badly.