ESS assistant difference ph. 1 to ph. 2 & 3 + underestimated PV production

Interesting timing on this thread — I've been wrestling with something similar on my shore-power-backed narrowboat setup, though obviously at a smaller scale than 11kW of SolarEdge.

When I was running three Victron Multiplus-II units in a 3-phase arrangement (before I simplified down to single-phase for the boat), the ESS assistant was genuinely baffling about how it balanced loads across phases. Phase 1 would be doing heavy lifting while phases 2 and 3 sat relatively idle, and the PV yield figures in VRM were consistently understating actual production during peak irradiance. Not by a trivial margin either — we're talking 15-20% some days.

A few things worth checking if you're in a similar situation:

• AC-coupled PV (SolarEdge on AC-OUT in particular) can confuse the ESS power calculations because the Multiplus is measuring net flow, not gross generation
• The ESS assistant's "ignore AC input" thresholds can interact oddly with 3-phase imbalance — phase compensation settings matter enormously here
• VRM dashboard figures vs. what the SolarEdge portal reports are worth comparing directly

The underestimation issue specifically — is yours showing up in VRM, or on the SolarEdge side? Because those are two different problems with different fixes.

Anyone else running AC-coupled strings across a 3-phase ESS setup in the UK? Curious whether this is worse with SolarEdge vs. a Fronius Primo on AC-OUT, given how differently they handle frequency shifting.

@AngliaOffGrid the phase imbalance thing is real even on single-phase setups if you're running a Multiplus-II and your PV inverter is AC-coupled on the output side rather than MPPT via VE.Direct. The ESS assistant can get confused about what's genuinely surplus vs what it thinks is surplus based on grid meter readings.

On my boat I run a Victron SmartSolar straight into the DC bus which sidesteps a lot of that ambiguity — the MPPT figures go directly into the GX device and there's no translation layer second-guessing production.

The "underestimated PV" issue in your title sounds like it could be CT clamp placement more than the ESS logic itself. Where exactly is your grid meter sitting relative to your inverter connection point?

Good point @DaiYoung56 — and worth adding that the ESS assistant's grid setpoint can behave oddly when PV production is being underreported to the Multiplus. On my system I found the MPPT was feeding data to the GX device slightly behind real-time, which caused the inverter to draw more from shore than necessary before "catching up." Ended up tightening the DVCC settings and it improved noticeably. @AngliaOffGrid if you're on a narrowboat you're presumably watching amp-hour budgets quite carefully anyway — have you checked whether your GX device's VRM logging shows the PV yield matching what the MPPT itself reports locally? Sometimes the discrepancy lives in that gap rather than in the ESS logic itself. Worth ruling out before diving too deep into the assistant configuration.

Really useful thread this. @AngliaOffGrid — on narrowboats specifically, worth checking whether your shore power connection is introducing any neutral-related weirdness that's confusing the ESS assistant's grid setpoint calculations. I had something similar on a static cabin install where the grid meter readings were slightly off because of a dodgy neutral at the hookup point, which made the assistant behave erratically around the setpoint.

Also on the PV underestimation side — if you're using VRM portal, have you cross-checked the yield figures against your inverter's own logs? I've found VRM can lag or smooth production spikes in ways that make the assistant look like it's underperforming when actually it's the reporting that's slightly off rather than the actual behaviour.

@LiamClark72 makes a fair point about shore power connectors — I had exactly that issue with my cabin setup before I moved to a dedicated Victron system. The grid setpoint drift @DontPanic8 mentions is something I've since dialled in manually, but it took a few weeks of logging in VRM to understand the pattern.

What caught me out initially was assuming the ESS assistant was reading PV production accurately in real time. It wasn't — there was enough lag during fast-changing cloud cover that the system was consistently underestimating available solar and pulling from the grid unnecessarily.

If you're on a narrowboat, I'd guess mooring orientation changes compound this — one day you're south-facing, next you're not. Worth checking whether your MPPT data in VRM matches what the ESS assistant is actually acting on.

Has anyone actually quantified how much the ESS grid setpoint drift matters in real-world terms when PV is fluctuating hard? Like, what's the actual kWh loss we're talking about — is it measurable or just theoretically annoying?

On my static caravan setup I've noticed the Victron occasionally seeming confused about where surplus actually is, but I've never logged it properly so I genuinely don't know if it's a real problem or I'm just watching numbers on the app too obsessively.

@AngliaOffGrid — is your narrowboat a single-phase situation throughout? Wondering if that changes things significantly compared to what the OP is describing with the phase imbalance across 1/2/3. Seems like a different beast entirely if you're not dealing with three phases.