Anyone else finding their MPPT drops out on cold mornings? Running a Victron 100/30 here

I've got a 400W panel array (two 200W panels in series) feeding into a Victron SmartSolar 100/30, with a 200Ah lithium battery bank. Everything's been working a treat since the summer, but now we're getting proper cold nights here in Norfolk I'm noticing the controller sometimes just... doesn't wake up properly at first light. The app shows it sitting in "Off" or occasionally throwing a brief high voltage warning, then it sorts itself out maybe 20-30 minutes after sunrise.

My panels are rated Voc of 47.2V each, so in series that's 94.4V at Standard Test Conditions — but I know open circuit voltage climbs in the cold. We've had a few -3°C nights recently and I'm now thinking I might be nudging uncomfortably close to that 100V input limit on cold clear mornings before the panels warm up.

Has anyone actually done the maths on temperature-corrected Voc for something like this? I'm trying to work out whether I need to rewire to parallel (which would halve my voltage but obviously double the current), or whether I'm worrying over nothing. Would love to know if others have hit this with similar setups, especially if you're also somewhere flat and exposed where the panels get properly cold overnight.

@NorfolkBoater — seeing exactly this on my motorhome setup at the moment. Worth checking whether your Voc is exceeding the 100V input limit when temps drop. Cold panels can push voltage significantly higher than the rated spec — there's a temperature coefficient calculation you need to run for your worst-case winter morning temp.

What are your panels' open-circuit voltage ratings individually? If they're, say, 24V Voc each, you're sitting at 48V nominal in series, but on a frosty -5°C morning that could creep up considerably.

Also — have you checked the battery temperature sensor is connected? The Victron will throttle or cut out if it thinks the lithium is too cold to accept charge safely. Are you running a Fogstar or similar with a built-in BMS low-temp cutoff? That'd do it too.