Interesting one this — I've seen similar discrepancies crop up a few times with Victron MPPT units and it's worth digging into properly before assuming the controller itself is faulty.
The gap between what the Victron app reports and what your multimeter is reading is massive here — 4+ volts isn't a calibration drift, that's something more serious. A few things worth checking:
Multimeter placement matters enormously. Are you measuring directly at the controller's battery terminals, or further down the cable run? If there's a shunt, fuse block, or any other inline component between the MPPT and where you're probing, you may be picking up a floating voltage rather than actual system voltage. Disconnect the battery entirely and probe the controller terminals in isolation.
Check the battery voltage sense wiring. The SmartSolar 100/30 has a separate VSense input — if that's misconfigured or picking up noise, the controller can start chasing a phantom voltage target and genuinely push output higher than intended.
Firmware — worth checking you're on the latest via VictronConnect. There were some known MPPT regulation quirks on older builds.
From my own setup running a 24V system with a couple of Fogstar Drift LiFePO4 batteries, I always run the VSense lines properly terminated and the app readings sit within 0.05V of a decent multimeter every time. When they don't, it's almost always a connection or sense wiring issue rather than the unit itself going rogue.
That said — 16.5V on a 12V system is potentially damaging your battery if it's sustained. Worth isolating the array until you've got this sorted.
Has anyone else seen genuine over-voltage output from a SmartSolar