MPPT temperature value doesn't update since 3.80

Been scratching my head over something similar on the boat this past week or so.

Since updating my Victron SmartSolar MPPT 100/30 to firmware 3.80, the temperature reading via MQTT has been completely frozen. Shows a rock-solid 36°C regardless of whether I'm moored up in the shade or the controller's been baking in afternoon sun. Meanwhile, the SmartSolar Control Display shows a perfectly sensible reading that actually changes throughout the day — so the sensor itself is clearly working fine.

It's the MQTT side that's gone peculiar. Values just aren't refreshing. Stuck like a broken clock.

This matters more than it might seem for anyone running any kind of automation off their monitoring. I've got a basic Node-RED setup on a Raspberry Pi that uses temperature alongside solar yield to make decisions about the battery management on my narrowboat — Fogstar Drift lithium if anyone's curious. When the temperature feed freezes, the whole logic goes a bit sideways.

My suspicion is it's a publish interval or null-value bug introduced in 3.80, rather than anything hardware-side. The display rendering fine suggests the raw data exists — it's just not making it through to the MQTT broker properly.

Anyone else seen this? Particularly interested if you're running VenusOS on a Cerbo GX or a Pi-based setup. Wondering whether it's isolated to certain configurations or whether it's widespread enough that Victron's already aware and there's a fix incoming.

Would love to know if rolling back to 3.79 sorted it for anyone, though I'm always nervous about firmware downgrades on working kit.

Seeing this exact issue on my garden office setup with a SmartSolar 100/20. Updated to 3.80 a couple of weeks back and the temperature value in MQTT just freezes — it'll show one reading and then stop updating entirely until I restart the controller.

Worth checking whether it's specifically the Battery Temperature topic or the Ambient Temperature that's sticking? On mine it seems to be both, but curious if @DorsetCamper is seeing the same.

Also — is anyone pushing data through VRM Portal seeing the same freeze, or is this purely an MQTT/local issue? Trying to narrow down whether it's the firmware's reporting logic or something specific to the MQTT broker side.

Raised it with Victron's community forum but no official response yet.

Same here on my shepherd's hut SmartSolar 75/15 — temperature's been stuck on whatever value it grabbed at startup. Not updating at all over MQTT.

Worth checking if it's just MQTT or whether the VictronConnect app also shows it frozen. Mine does, so it's not purely a publishing issue — something's gone wrong at the firmware level with the sensor polling itself.

Rolled back to 3.79 yesterday and it's reading properly again. Bit of a faff but straightforward enough if you've still got the old firmware file. Victron Community forum has a thread on it — looks like others are hitting this too so hopefully a fix lands soon.

Yep, seen this one — it's a known regression in 3.80. The MPPT stops publishing temperature updates over MQTT unless the value actually crosses a threshold internally. Basically the delta-filter is too aggressive now.

Few things worth trying:

• Force a re-read by disconnecting/reconnecting the VRM MQTT session
• Check if your dbus-mqtt service is actually running properly (svstat /service/dbus-mqtt)
• Roll back to 3.79 temporarily if the live data matters to you

@DorsetCamper on a boat especially, I'd want accurate temp data for charge compensation — don't let it sit.

Victron community forum has a thread on this already, someone flagged it to the devs so a fix should come in the next point release. Worth subscribing to that ticket if you want a heads-up when 3.81 drops.

Worth adding from my motorhome setup — I've confirmed this affects the SmartSolar 100/50 on 3.80 as well, so it's not isolated to the smaller units.

One thing worth checking: if you're polling via MQTT and the value appears frozen, cross-reference it against the VictronConnect Bluetooth reading directly. In my case the local BT value was updating correctly — it's specifically the MQTT publish path that's broken post-3.80.

@MarineGaz is right that this is a regression. I'd log it formally via Victron Community if you haven't already — they do track MQTT-specific bugs separately from general firmware issues. The more units confirmed affected, the faster it tends to get prioritised.

In the meantime I've bodged around it by triggering a service restart on my Cerbo GX every few hours via a cron job. Not elegant, but keeps the temperature data flowing for my battery thermal monitoring.