Dbus-fronius: DC-Werte je MPPT (Pv/x/V, Pv/x/P) um Faktor 10 falsch

Been scratching my head over something similar on my narrowboat setup and wondering if anyone else has hit this wall.

Running a Cerbo GX with Venus OS (latest stable) alongside a Fronius inverter, and I keep noticing the per-MPPT DC values — voltage and power — showing figures that are clearly off by a factor of ten. So where I'd expect, say, 340V I'm seeing 3400V displayed in VRM. The total AC output from the Fronius looks perfectly sensible, so it's specifically the granular MPPT channel data that's gone sideways.

My suspicion is it's something to do with the scaling factor interpretation in dbus-fronius — the way it handles SunSpec registers when the inverter is set to report using integer + scale factor mode. Feels like somewhere in the chain, the scale factor is being applied twice, or the sign is being misread.

A few things I'm curious about:

• Has anyone seen this with non-Fronius inverters, or is it Fronius-specific?
• Does changing the SunSpec model register offset in the Fronius Datamanager settings make any difference?
• Is this a known bug in recent dbus-fronius versions, or am I chasing a config gremlin?

I mostly care because I'm trying to do proper yield monitoring per string — two roof arrays running into separate MPPTs — and the duff data makes that completely pointless.

Would love to know if there's a workaround short of waiting for a fix upstream. Anyone had a dig around in the dbus-fronius source who might point me in the right direction?

@Midge — ooh this rings a bell! When I was fiddling with the Cerbo on my shepherd's hut build, I noticed some of the DC values from the MPPT readings looked wildly off — turned out to be a decimal place issue in how the dbus driver was interpreting the raw data. Factor of 10 out is classic for that.

Is it just the Pv/x/V and Pv/x/P paths that are wrong, or are you seeing it on the aggregate totals too? That might help narrow down whether it's the Fronius-specific dbus service or something upstream.

Also — which version of dbus-fronius are you running? There were some oddities around this in earlier Venus OS builds that I thought had been patched. Worth checking the Victron community changelog if you haven't already dug through that rabbit hole!

Interesting one. Is this specifically the per-MPPT DC values being out by 10x, or is the total PV power reading also affected?

On my shepherd's hut I've got a Victron MPPT rather than Fronius, so can't replicate exactly — but I've seen scaling factor bugs in dbus drivers before where the unit multiplier gets applied twice.

Worth checking:

• Which Venus OS version introduced this for you?
• Are the values wrong in VRM portal as well, or just locally on the Cerbo display?
• Is it all MPPT channels or just Pv/1/V and Pv/1/P specifically?

If it's a dbus-fronius driver issue, the Venus OS GitHub would be worth raising it — they're pretty responsive to issues with reproducible data attached. A dbus-spy dump showing the raw values versus what's displayed would give them something concrete to work with.

@RussScott — good question, and worth separating those two things out. From what I've seen on mine, the total PV power on the overview screen looks fine, but the moment you drill into the individual MPPT tracker values (the Pv/x/V and Pv/x/P paths specifically) that's where the 10x discrepancy shows up.

Worth checking in VRM whether the raw logged values match what you're seeing locally on the Cerbo screen — sometimes that helps isolate whether it's a display/scaling issue in dbus-fronius or something deeper.

@Midge — are you using the Fronius integration over Modbus TCP or the solar API? That might narrow it down. My boat setup uses Modbus and I've not hit this yet, but I've got a suspicion it could be unit scaling — millivolts vs volts perhaps?

@FenlandOffGrid that's useful context. Quick question though — does this 10x error only show up in the dbus path values (i.e. Pv/x/V and Pv/x/P specifically), or does it also bleed through into what VRM Dashboard actually displays?

Reason I ask is I've got a shepherd's hut build in progress and I'm planning to integrate a Fronius with a Cerbo GX. If the VRM graphs are also affected that'd be a real headache for monitoring. I'd rather know now before I'm committed to the hardware.

Also — is this a known Victron bug with a tracker on GitHub, or more of a "it's always been like this" situation? Wondering whether Venus OS 3.x has addressed it at all or if there's a manual workaround needed on the dbus side.