Anyone else struggling to get accurate SOC readings with a PWM controller and AGM batteries in winter?

Been having a nightmare lately trying to trust my battery state of charge readings. Running a pair of 100Ah AGM batteries with a 20A PWM controller (Renogy Wanderer), two 175W panels on the roof of my Transit conversion. Works a treat in summer but since October the SOC display has been all over the shop — showing 80% then dropping to 50% within an hour of mild load.

I know PWM isn't the most efficient in low light and the panels are flat-mounted so I'm losing angle in winter anyway, but I'm not sure if the dodgy readings are a controller issue, a temperature compensation issue, or whether my batteries are just starting to go. They're about 3 years old and have lived through two Scottish winters, so they've had a hard life.

I've been thinking about adding a proper battery monitor like a Victron BMV-712 to get shunt-based readings rather than relying on the controller's voltage estimates. Has anyone gone down that route and actually found it made a meaningful difference to how you manage your system day to day?

Also curious whether anyone's found a practical workaround short of replacing the whole setup — I'd rather not jump straight to MPPT and new batteries if there's a cheaper fix that'll get me through this winter at least.

WildMechanic | 847 posts

@BorderNomad Yeah, PWM controllers are notoriously poor at accurate SOC estimation - they're basically just doing simple voltage-based calculations, and AGM voltage curves are really quite flat in the middle range, making it genuinely difficult to pinpoint where you actually are.

Winter compounds everything too. Cold batteries show artificially higher resting voltages, so your controller thinks you're fuller than you are. Then you've got reduced panel output meaning your batteries might rarely hit proper absorption, so the controller never gets a reliable "full" reference point to reset against.

My honest recommendation - grab a decent dedicated battery monitor like a Victron BMV-712 or even a cheaper Renogy one. Coulomb counting is far more reliable than voltage guessing. You'll wonder how you managed without it. The Victron's also got Bluetooth which is handy for checking from inside when it's freezing out! 🙂

DougDixon56 | 312 posts

@BorderNomad Worth remembering that AGM voltage curves flatten out considerably in cold weather, which throws SOC readings off even further than usual. A battery sitting at 12.5V in January isn't telling you the same story as it would in July.

What I'd suggest is getting a proper battery monitor with a shunt - something like a Victron BMV-712 will give you Coulomb counting rather than relying on voltage alone. Makes an absolute world of difference. Yes it's an extra outlay but you'll actually be able to trust what you're looking at.

Also check your controller's temperature compensation is set correctly for AGMs - incorrect settings will mess with charging endpoints and knock your readings sideways from the start. The Wanderer does have temperature compensation built in but it needs configuring properly.

AshDweller | 203 posts

@BorderNomad Had almost identical issues last winter in my shepherd's hut setup — two 110Ah AGMs with a basic PWM controller. The voltage-based SOC estimation is genuinely useless in cold conditions because AGM resting voltage shifts with temperature, so the controller's thresholds are calibrated for maybe 20°C and you're nowhere near that.

The single best thing I did was add a Victron BMV-712 shunt-based monitor. It measures actual coulombs in and out rather than guessing from voltage, so temperature-related drift becomes largely irrelevant to the SOC reading. Not cheap (~£90-100) but transformative for peace of mind.

Also worth checking — is your PWM controller applying any temperature compensation to its charge voltage? The Wanderer has a port for a temperature sensor. Without it, your batteries are probably being undercharged significantly right now.

Amy2000 | 156 posts

@BorderNomad One thing nobody's mentioned yet — check whether your controller is actually using the correct battery type setting. PWM controllers often default to flooded/wet settings, and if yours is set wrong, the charge thresholds will be off, which throws SOC estimates out from the start.

Also worth investing in a proper dedicated battery monitor like a Victron BMV-712 or even a cheaper Bayite unit. They track actual amp-hours in and out rather than relying on voltage guesswork, which is genuinely transformative for AGMs. Your PWM controller's SOC display is essentially just a rough voltage-based estimate — fine in summer, pretty useless once temperatures drop and the voltage curves shift.

Longer term, an MPPT upgrade would help too, but the battery monitor is the quicker win and much cheaper!

Andy1979 | 847 posts

@BorderNomad Something worth adding to what's already been said — PWM controllers are genuinely quite poor at SOC estimation compared to MPPT units, as they don't have the same voltage regulation accuracy. Honestly, the most reliable thing I ever did was fit a dedicated battery monitor like a Victron BMV-700. It measures actual amp-hours in and out rather than relying on voltage curves, which as @DougDixon56 touched on, become pretty meaningless in cold weather anyway.

Also make sure your battery temperature compensation is set correctly if your controller supports it — AGMs typically want around -20mV/°C/cell adjustment. In a van or mobile setup over winter that can make a surprisingly big difference to how the controller behaves. What temperatures are you seeing inside your battery compartment overnight?

Stacey28 | 412 posts

This thread is basically the story of my first winter running a garden office setup. I spent weeks second-guessing my readings before I finally invested in a Victron BMV-712 battery monitor — genuinely life-changing for understanding what's actually happening rather than relying on the controller's voltage-based guesses.

The thing that really surprised me was how dramatically cold temperatures affect AGM capacity. My 100Ah batteries were effectively behaving like 75Ah batteries on bitter mornings, so even an "accurate" SOC reading was misleading because the reference point had shifted.

@Andy1979 is right about PWM limitations — once I upgraded to an MPPT I got noticeably better winter performance, but honestly the dedicated battery monitor made the biggest immediate difference. Knowing actual amp-hours in and out tells a proper story rather than a voltage snapshot.

CrafterWanderer | 1,203 posts

Something nobody's touched on yet — AGM self-discharge accelerates significantly below 10°C, so if your batteries are in an uninsulated compartment, your resting voltage readings are being skewed by temperature before PWM even enters the equation. Most PWM units lack proper temperature compensation unless you've wired in an external sensor.

On my static caravan setup I added a Victron BMV-712 purely as a dedicated SOC monitor, completely independent of the controller. Coulomb counting gives you far more reliable data than voltage-based estimates ever will — especially in winter when voltage curves flatten and become almost meaningless for AGMs under partial load.