Anyone else struggling to get accurate SOC readings with a PWM controller on AGM batteries?

I've been running a fairly basic setup in my static caravan up in the Dales for about two years now — two 200W panels going into a 40A PWM controller (a Renogy Wanderer), charging a pair of 110Ah AGM batteries wired in parallel. It's done the job well enough, but I've never once trusted the state of charge readings on the controller's little display. It seems to jump around all over the place, especially first thing in the morning or on cloudy days.

I know the general advice is to move to an MPPT and add a proper battery monitor like a Victron BMV-712, but I'm trying to decide whether the dodgy SOC is a controller problem, a PWM-vs-MPPT thing, or just the nature of AGMs being a bit unpredictable. I've checked resting voltage against known SOC tables and the readings are plausible, but the controller's percentage display just doesn't match up. At one point it was showing 80% while resting voltage said I was closer to 60%.

Has anyone found a decent workaround short of replacing the whole controller? Or is the BMV-712 the only sensible answer here? Curious whether anyone's had similar grief with the Renogy specifically or if this is just PWM controllers in general.

@JaneReid68 yeah PWM + AGM is a rough combo for accurate SOC. The controller's basically just guessing at voltage, and AGMs have a notoriously flat discharge curve so voltage tells you almost nothing useful mid-range.

Honestly the single best upgrade you can make is a decent battery monitor — something like a Victron BMV-712. Tracks amp-hours in and out, gives you actual SOC rather than voltage-based guesswork. Made a massive difference on my van setup.

Longer term, swapping to an MPPT controller (even a basic Renogy one) gets you better charging profiles which at least means your AGMs are properly topped up, making voltage slightly more meaningful. But the shunt monitor is the real fix.

@JaneReid68 your Renogy Wanderer is essentially doing SOC by horoscope — ditch the guesswork and bolt a Victron BMV-712 onto those AGMs, it'll actually measure coulombs in and out rather than squinting at voltage like a bloke reading a menu without his glasses.

@JaneReid68 Worth adding to what @KentCruiser and @LutonNomad have said — AGMs are particularly awkward because their resting voltage curve is quite flat in the middle range, so voltage-based SOC estimation becomes genuinely meaningless between roughly 50-80% charge. A coulomb-counting monitor like the BMV-712 will transform your confidence in the readings overnight.

One thing to check first though: make sure your battery bank is properly equalised and your AGMs aren't already suffering from sulphation after two years of potentially inaccurate charging. A static caravan setup that's been undercharging could mean the batteries themselves are no longer performing to spec, which would skew your readings regardless of what monitoring you add.

Had this exact headache in my shepherd's hut. The thing that tripped me up for ages was resting voltage — you can't trust any SOC reading if the batteries have been charging or under load within the last few hours. Even with a proper shunt monitor, I was getting daft readings until I started waiting a solid 4–6 hours after last charge before treating voltage as meaningful.

Also worth checking: are your AGMs properly equalised? Mine were showing false "full" readings at around 80% actual capacity because they'd sulphated slightly from repeated partial charging cycles — very common with PWM setups.

@TimKnight curious what you were about to say re AGM resting voltage profiles — I've noticed mine sit noticeably higher than flooded lead-acid at the same actual SOC, which caught me out initially.

Great points already covered above. One thing I'd add — with PWM specifically, your panels are being dragged down to battery voltage rather than operating at their optimal Vmp, so you're often getting a partial charge that looks complete on the controller display but really isn't. The batteries sit at a surface charge voltage that reads falsely high. Try disconnecting the panels and any loads for a good few hours before trusting any voltage reading — that proper rest period @BatteryJackie mentioned makes a real difference. Also worth checking your controller's AGM voltage profile is actually set correctly; the Wanderer defaults to flooded sometimes depending on firmware. A 14.7V absorption for AGMs is what you're after up in the Dales, especially in colder months when charging efficiency drops off noticeably.

Something that caught me out for a long time — and nobody mentioned it yet — is temperature compensation. My garden office setup uses AGMs and during last winter I was scratching my head at wildly optimistic SOC readings. Turned out the colder the battery, the higher the resting voltage relative to actual charge state. 12.5V means something very different at 5°C versus 20°C.

If your Renogy Wanderer has a temperature sensor port, it's genuinely worth using one. Made a noticeable difference to how my readings tracked reality once I plugged it in.

Also worth pairing any of this with a proper battery monitor like a Victron BMV-712 — shunt-based coulomb counting is miles more reliable than voltage-based SOC guessing, especially with AGMs which have that notoriously flat discharge curve.