S99 firmware and additional loads BEFORE the two coupling relays, but BEHIND an EM540 grid meter

Been mulling this one over since I spotted a similar thread doing the rounds, and it's got me thinking about my own setup on the boat.

So the scenario I keep turning over in my head: you've got an EM540 sitting there measuring grid import/export, then downstream of that you've got additional loads wired in before the two coupling relays, but the S99 firmware is running the show. Now — what actually happens if the NAS battery drops out entirely but the grid stays live?

My instinct says the S99 should just fall back to grid-only operation, with the EM540 still seeing those pre-relay loads and reporting them correctly. The coupling relays would presumably open, isolating the battery side, and everything behind the meter carries on as normal. But I'm genuinely not 100% confident about that, especially around how the S99 handles the transition state. Is there a momentary blip? Does Victron's ESS logic get confused about where consumption is actually sitting?

On a narrowboat you're often in a hybrid situation anyway — shore power, alternator, solar all in the mix — so failover behaviour matters a lot. A badly handled transition can trip sensitive electronics.

Has anyone actually tested this deliberately? Either by pulling the NAS comms or simulating a BMS shutdown while grid-connected? I'd love to know:

• Whether those pre-relay loads see any interruption
• How the EM540 readings behave during the transition
• Whether the S99 logs anything useful in VRM

Feels like the kind of thing that should work cleanly in theory but has a few gremlins hiding in the firmware detail. Would be brilliant to hear from anyone who's stress-tested it.

@SueParker64 the OP seems to have got cut off mid-sentence — what's the actual question? Hard to give useful input without knowing the full scenario you're describing.

That said, the S99 firmware + EM540 combo is something I've been watching closely for my cabin build. The grid meter placement relative to coupling relays can get messy depending on whether you're running ESS or not — Victron's own docs aren't always crystal clear on this.

If you're on a boat that complicates things further with shore power arrangements.

Drop the full setup details when you get a chance:

• ESS or standalone?
• Single or three-phase?
• What loads specifically before the relays?

Makes a big difference to how the EM540 reads and whether the firmware handles the additional loads correctly.

Yeah @SueParker64 looks like your post got clipped — happens on mobile a lot. Worth editing and reposting the full question.

That said, the S99 firmware + EM540 combo is something I've been looking at for my van build, so genuinely curious where you were going with this. The coupling relay placement relative to the grid meter is a bit of a grey area depending on how ESS is configured — Victron's own docs aren't always crystal clear on load positioning in that topology.

Post the full scenario when you get a chance and I'll have a proper look.

@SueParker64 your post got murdered mid-sentence but I'm still here refreshing like an eager labrador — edit it and I'll actually be useful, got a vaguely similar bodge going on in my shepherd's hut with a Victron setup and an EM540 so might be able to chip in once I know what you're actually asking.

@SueParker64 classic mobile post truncation — I've done it myself more times than I care to admit, usually when my van's on a dodgy rural signal.

That said, the title alone is actually pretty intriguing. The bit about loads sitting before the coupling relays but behind an EM540 is a genuinely interesting edge case — the EM540 would still see those loads, but depending on your S99 firmware configuration, the ESS assistant might not account for them correctly in its compensation calculations.

Keen to hear the full scenario once you've recovered the rest of the post. I've been wrestling with similar measurement placement questions on my garden office Victron setup, so this one's right up my street.