AC Loads overview: separate Shelly Pro EM50 measurements and unexplained 25 W difference

Been down a similar rabbit hole with my setup lately and it's doing my head in a bit.

Running a MultiPlus-II 48/5000 with a Cerbo GX and I added a Shelly Pro EM50 a few months back to get better visibility on my AC loads — specifically because the Cerbo's built-in figures never quite matched what I was expecting. Thought the Shelly would clear things up. It... sort of has, but now I've got a consistent ~20-30W discrepancy between what the Cerbo reports as AC loads and what the Shelly's actually measuring.

Not a massive deal in the grand scheme of things but when you're trying to nail down your daily consumption for battery sizing (running 280Ah of Fogstar Drift 48V cells), even 25W unaccounted over 24hrs is a chunk of watt-hours you're just... losing track of.

Few things I've been wondering:

• Is Venus OS doing something weird with the way it calculates AC loads from the MultiPlus-II's internal measurements vs what the Shelly reports?
• Could it be the CT clamp positioning on the Shelly throwing things off slightly?
• Anyone else noticed this get worse or better between Venus OS versions?

I've seen some chat about whether the MultiPlus-II's own current sensing is just inherently a bit rough at low loads — wouldn't surprise me tbh.

Anyone else running a separate energy monitor alongside their Victron kit and seeing similar gaps? Curious whether it's a calibration thing or something more structural in how ESS reports consumption.

@WezFrost that 25 W ghost load is almost certainly your Cerbo GX itself plus the MultiPlus-II's internal consumption — the inverter draws around 15–20 W just sitting in standby, and the Cerbo adds another 4–5 W. Neither shows up cleanly in VRM's built-in calculations because they're measured after the shunt but before your Shelly's CT clamps, depending on exactly where you've clamped them.

Worth double-checking:

• CT clamp position relative to your busbar and any direct DC loads
• Whether your Shelly is measuring the full AC output or just downstream distribution
• PowerAssist mode can introduce brief measurement skew too

On my setup I've got the Shelly Pro EM50 clamped at the consumer unit rather than at the MultiPlus output terminals, which eliminated most of my unexplained delta. Took a bit of rewiring but worth it for clean figures in VRM.

@WezFrost welcome to the forum — great first post and a genuinely interesting one to dig into!

One thing worth checking beyond what @DodgyRoamer mentioned: the Shelly Pro EM50 measures apparent vs real power differently depending on your power factor. If you've got any switching supplies or inductive loads on that circuit, you can easily see a 20-30W discrepancy that's essentially power factor related rather than a "missing" load.

Also worth logging your MultiPlus-II's own VRM data overnight with zero AC loads switched on — that baseline figure tells you exactly what your system is consuming just to exist. Mine sits around 18-22W at idle on a 48V setup.

The Cerbo itself pulls roughly 4-5W, so it all adds up fast.

@DodgyRoamer beat me to the Cerbo being a sneaky little power vampire, but nobody's mentioned that the Shelly Pro EM50 itself sits between your meter and reality — CT clamp positioning relative to your neutral can introduce exactly that kind of offset if it's not dead-on.

Had the same "where's my 20-odd watts going" drama in my van until I realised my clamp was reading at a slight angle like it was auditioning for abstract art. Repositioned it properly and suddenly the universe made sense again.

Worth also checking if your MultiPlus-II firmware is current — older versions reported internal consumption slightly differently to the Cerbo, which can skew your VRM dashboard figures versus what the Shelly sees.

@WezFrost one thing nobody's touched on yet — the power factor of your AC loads can skew this significantly. The Shelly Pro EM50 measures true power (watts), but if your MultiPlus-II is reporting apparent power (VA) internally and there's any translation happening through the Cerbo's DVCC layer, you'll get a systematic offset that looks like a ghost load but is actually a measurement methodology mismatch.

Worth checking in VRM whether your AC loads figure is derived from a direct CT measurement or calculated as [inverter output] - [known loads]. On my motorhome setup with a Multiplus-II 12/3000, I found a persistent ~18W discrepancy that traced back to exactly this — the Cerbo was doing subtraction arithmetic rather than reading a dedicated sensor, so any rounding across the chain accumulated.

Set your Shelly to report active power only and compare against the VRM AC Loads widget directly.

@Liam1990 raises a good point on PF, but don't overlook CT clamp placement — if your Shelly's clamp is sitting on the wrong side of a junction or not fully closed around the conductor, you'll get systematic errors that look suspiciously like a fixed 25 W offset every single time.

Also worth noting: the MultiPlus-II itself draws ~15–20 W just ticking over at no load, and depending on exactly where your Shelly is measuring versus what the Cerbo's pulling from the VE.Bus data, you could easily account for that mystery 25 W without any actual measurement error at all — just two instruments measuring genuinely different things.

Pull up VRM Portal and compare your MultiPlus AC-in vs AC-out figures; that'll tell you whether the inverter's own consumption is the culprit.