Load output stuck OFF after LVD event, even with battery well above 'on' threshold

Been catching up on some threads elsewhere about this and it's triggered a memory of something I had with my garden office setup last winter that I never fully resolved.

Running a Victron SmartSolar MPPT with a 100Ah Fogstar Drift LiFePO4. I had a genuine low voltage disconnect event one evening — battery had drained down overnight with no solar coming in. Next morning, even with the battery sitting at 13.1V (well above my load reconnect threshold of 12.6V), the load output just stayed stubbornly off.

Had to manually disconnect and reconnect the battery to kick it back into life. Felt like the controller had latched into a fault state and wasn't polling voltage properly to trigger reconnection.

A few things I've been wondering about since:

• Is there a hysteresis timing issue at play? Some controllers seem to require the voltage to stay above threshold for a sustained period, not just momentarily
• Could the BMS on the Fogstar cutting out briefly have confused the MPPT into thinking the battery was still absent?
• Is this more common with LiFePO4 than AGM because of the flatter discharge curve?

I've since tweaked my reconnect threshold up slightly and added a small delay via a VE.Direct rule, which seems to have helped — but I'm not 100% sure that was the actual fix or just coincidence.

Anyone else seen this with Victron kit specifically, or is it more of a universal MPPT behaviour? Curious whether it's a firmware thing or something more fundamental in how the load output logic works.

@Smithy98 — classic "LVD latch" behaviour, and it catches people out more than you'd think.

The SmartSolar's load output doesn't simply re-enable the moment voltage recovers — it requires the battery to hold above the re-connect threshold for a defined settling period. If your battery is genuinely depleted (rather than

@Smithy98 the SmartSolar won't auto-recover the load output until it sees a proper charge cycle complete — specifically, it needs to reach the absorption voltage threshold first, not just creep above the LVD "reconnect" voltage.

Easiest fix: in VictronConnect, check your Load Output settings and switch the algorithm from "BatteryLife" to a fixed voltage mode, so you have predictable, manual control over exactly when it reconnects.

Worth noting — if your battery is a Fogstar or similar LiFePO4 with its own BMS, the BMS may have tripped independently and is holding the load low regardless of what the MPPT thinks it's doing. Two separate latches stacking on each other is a right faff to diagnose without looking at both simultaneously via the app.

Worth adding a practical gotcha here — if you're running LiFePO4 (which I suspect you might be given the context), the BMS can disconnect under LVD and the MPPT sees the voltage snap back almost instantly once load is removed. So the controller thinks the battery recovered, but it hasn't completed a proper absorption cycle, so the load output stays latched off.

The fix that worked for me on my static caravan setup: either use the "Load output always on" mode temporarily to break the latch (obviously monitor it closely), or just force a manual charge cycle via the VictronConnect app — go into the charger settings and bump it through bulk/absorption manually.

Also check your re-connect voltage setting in VictronConnect. If it's set too close to your LVD threshold, you'll trigger this repeatedly.

Good point from @BayJason about LiFePO4 — this is where it gets interesting with lithium. The SmartSolar's load output algorithm was essentially designed around lead-acid charge curves. With LiFePO4 sitting at, say, 13.2V resting, the MPPT can genuinely think the battery hasn't completed a proper absorption cycle and keeps the latch engaged.

I had exactly this on my motorhome last February. Battery was at 95% SOC per my Victron BMV, load output stubbornly dead.

Fix was twofold:

1. Manually disconnect/reconnect via the VictronConnect app to break the latch
2. Reconfigure the charge profile so absorption voltage/time actually gets reached and recognised

If you're on a default lead-acid profile with LiFePO4 cells, the algorithm never sees the "all-clear" signal it expects. It's not a bug exactly — just a square-peg-round-hole situation.