Wrong AC loads calculation?

Bit of a rabbit hole this one, but bear with me — had something similar crop up with my Multiplus II last winter that had me questioning my own sanity (and my 20-odd years of marine electrics experience, which is frankly worse).

The short version: my AC loads were reading noticeably higher in VRM than what I could account for by physically measuring each circuit with a clamp meter. Talking a solid 15-20% discrepancy, not just rounding noise.

Turns out the MP II's internal consumption — transformer losses, idle draw, the whole lot — was being lumped into the "AC Loads" figure rather than shown separately. Victron's own documentation is, shall we say, interpretively flexible on this point.

A few things worth checking if you're seeing similar:

• ESS assistant vs. no assistant — changes how loads are calculated and displayed in VRM entirely
• Grid metering vs. inverter metering — are you using an external EM24 or relying on the internal figures?
• Multiplus input current limit — if it's throttling AC in, the maths gets creative

Running a Fogstar Drift 48V LiFePO4 bank here, so I'm particularly twitchy about accurate load readings feeding into my SOC calculations — garbage in, garbage out as they say.

Curious whether anyone else has cross-referenced their VRM AC loads against a proper sub-meter on the distribution board? I'd love to know if this is purely a display/calculation quirk or something worth raising with Victron support directly.

@OldSailor curious timing as I've just been through exactly this with my shepherd's hut setup — Multiplus II 48/3000 running a mix of resistive and motor loads.

The bit that trips people up consistently is that the MP-II's AC load reading in VRM is measuring apparent power (VA) not real power (W) when you've got inductive loads on the circuit. If your power factor is sitting at 0.7-0.8 (fairly typical with motors, fridges, etc.) you can easily see a 25-30% discrepancy versus what your energy meter reports.

Worth checking in VRM whether you're comparing the right parameters. The AC Consumption L1 figure versus what a calibrated meter like an Owl or Shelly EM shows inline should tell you definitively which instrument is lying.

What loads specifically are running when you see the biggest variance?

@OldSailor worth checking whether the Multiplus is including its own idle consumption in the AC loads figure — mine on the narrowboat was showing ~80W "phantom" load that turned out to be the inverter's self-draw being double-counted through the way I'd wired the current transformer.

Also, are you running a Cerbo GX? There's a known quirk where the DVCC settings can skew what gets reported as AC load vs DC load in VRM. Took me a good weekend with the Victron manual and a decent clamp meter to unpick it.

@SomersetVanLifer the 48/3000 seems particularly prone to this — something about the power assist thresholds on that model. Check your ESS assistant configuration if you're grid-tied at all.

@OldSailor this caught me out on my boat build too. One thing neither @PennineNomad nor @SomersetVanLifer has mentioned — check your VE.Configure transformer settings. If the transformer ratio isn't set correctly for your installation, the Multiplus can misreport AC loads quite significantly because it's calculating from the internal current sense rather than true measured output.

Also worth pulling the data from VRM Portal if you're connected — the raw logs will show whether it's a display/reporting issue or an actual measurement discrepancy. Big difference between the two in terms of how you fix it.

On my 12/3000 I had a persistent ~150W phantom load that turned out to be the unit running in a weird hybrid state during grid fluctuations. Victron support were actually helpful when I sent them the VRM export.

Right, so jumping in here because this is basically my life right now with my static caravan setup —

Has anyone checked whether VE.Bus System is reporting the AC loads before or after the transfer switch kicks in? I spent an embarrassing amount of time convinced my Fogstar cells were somehow conjuring phantom loads, when actually I was looking at figures from two different measurement points in VRM depending on how I'd navigated there. 🤦

The dashboard widget and the advanced tab can show subtly different things during grid-tied operation — anyone else notice that, or have I just completely lost the plot?

@RayWatson81 what firmware version are you running? I wonder if this is one of those things that got quietly "fixed" at some point without anyone mentioning it in the changelog.