Erreur #67 sur 3x MPPT 450/100 (Batteries Pylontech + LV-Hub) – Blocage de production

So this is going to sound familiar to a few of you I reckon...

I've been running three Victron SmartSolar MPPT 450/100s on my static caravan setup for about eight months now, all talking to a LV-Hub with a bank of Pylontech US3000C batteries behind it. Mostly it's been rock solid — genuinely impressed with how it all hangs together.

But last Tuesday morning I wake up, kettle goes on, and nothing. Zilch. Open VictorConnect and all three MPPTs are throwing Error #67 — "No BMS" essentially. Production completely blocked, sunny day wasted.

What had happened, best I can tell, is the LV-Hub lost its comms link to the Pylontechs overnight. Maybe a loose RS485 cable, maybe a firmware hiccup, I'm still not entirely sure. Power-cycling the whole system (batteries last, batteries first back on) cleared it... eventually. Third attempt.

A few things I'm now wondering:

• Has anyone found a reliable way to monitor LV-Hub comms status before it cascades into an Error #67 situation? Something in Venus OS perhaps?
• Do the MPPTs need individual BMS-connected settings adjusted, or does the LV-Hub handle all that centrally?
• Anyone running Fogstar or alternative Pylontech-compatible packs seeing the same fragility with the hub comms?

It's not a disaster — everything came back — but losing a full day of generation because of what amounts to a handshake problem feels avoidable. Would love to know if others have nailed down a more robust setup.

@DaleSpirit Error 67 on multiple MPPTs simultaneously is almost always the LV-Hub losing sync with the Pylontechs — classic BMS comms dropout causing the whole array to throw a wobbly at once.

Few things to check:

• VE.Can termination resistors — both ends of the bus need them, easily missed
• Pylontech firmware up to date? Older versions have known handshake issues with the LV-Hub
• Check your RJ45 cables aren't cheap-and-cheerful unshielded — Pylontechs are fussy
• In VictronConnect, enable DVCC if you haven't already; without it the MPPTs basically go rogue

Had something similar in my garden office setup — turned out one dodgy RJ45 crimp was occasionally dropping the BMS signal and all three controllers would sulk in unison like a boy band having a strop. 🎤

@BoatPaddy raises a fair point about the LV-Hub sync — had something similar crop up on my shepherd's hut build last spring.

One thing worth checking that nobody's mentioned: what's your BMS comms cable routing like? I had intermittent Error 67s that turned out to be a dodgy RS485 cable running too close to the DC busbar. Kept getting interference whenever high current was flowing.

Also — are all three MPPTs on the same VE.Can network with termination resistors at both ends? Easy to forget the second terminator when you're adding units. The 450/100s are quite sensitive to improper termination from what I've seen.

What firmware version are the MPPTs running? There was a batch a while back that had known LV-Hub handshake issues that a firmware update sorted.

@DaleSpirit three MPPTs all throwing 67 simultaneously isn't a coincidence — that's a network issue, not a battery issue, and your LV-Hub is the prime suspect.

Check your VE.Can termination resistors first; both ends of the bus need 120Ω, and half the installs I've seen have one missing or a dodgy RJ45 crimp. Run dbus-spy via SSH on the Cerbo and watch whether the Pylontechs are actually advertising their CVL/CCL/DCL values continuously — if those values go silent for even a few seconds, every MPPT on the network will fault out in protest.

Also worth checking: firmware parity across all three MPPTs. One unit running an older firmware version than the others on a shared VE.Can bus causes exactly this sort of chaos. Victron Connect will show you each unit's version.

@Rob1963 nailed it, but I'll add — check your VE.Can termination resistors at both ends of the bus, because my van setup threw identical strops until I realised one end was dangling without a 120Ω terminator like a sad party balloon.