Been scratching my head over something similar with my garden office setup and wondering if anyone else has hit this wall.
Running a dual-Cerbo configuration — one master in the main house distribution board, one slave out in the office — and I've got a couple of temperature/energy monitoring sensors tied to the slave unit. The data feeds through Node-RED into a dashboard I've cobbled together, which mostly works fine.
The issue I keep running into is around Virtual Devices on the slave Cerbo. When I pull the sensor readings through Node-RED, the units are coming through inconsistently — sometimes watts, sometimes kilowatts, no obvious pattern I can pin down. It's making my consumption graphs look absolutely daft.
A few specific questions for anyone who's done this:
- Does the slave Cerbo handle unit scaling the same way as the master, or does something get lost in the MQTT relay between the two?
- Has anyone successfully mapped custom sensor data from a slave device into a coherent Virtual Device that the master then presents cleanly on VRM?
- Is there a Node-RED flow pattern that normalises units before they hit the dashboard, rather than relying on the Cerbo to do it?
My sensors are fairly basic clamp-type units, nothing exotic. The Victron ecosystem generally plays nicely with Node-RED in my experience, but this master/slave arrangement seems to introduce some quirks I wasn't expecting.
Would be especially interested if anyone running a van conversion alongside a fixed installation has tried bridging two Cerbos — seems like a niche but relevant use case here.