Variable Tank Level

Not strictly marine, but this cropped up in a slightly different context with my static caravan setup and I reckon the underlying issue is the same.

My fresh water tank is a 600-litre unit buried under the decking and I was seeing swings of 80-100 litres on the level sensor readout depending on time of day — worse in summer. Took me ages to work out it wasn't a leak or a calibration drift, it was thermal expansion and contraction of the plastic tank body itself distorting the ultrasonic sensor readings.

A few things that helped:

• Shading the tank — even basic insulation wrap made a noticeable difference to the amplitude of the swings
• Averaging the readings — I'm running a Victron Cerbo GX and used a simple Node-RED flow to average readings over 30 minutes rather than displaying instantaneous values
• Recalibrating at a consistent time of day — early morning when temps are stable gives a more reliable baseline

For narrowboats I'd imagine the problem is amplified because the hull geometry means tanks are often awkwardly shaped and subject to more temperature variation than a static install. Also worth checking whether the sender is resistive float-type or ultrasonic — they behave very differently when the tank flexes.

Anyone else running tank monitoring integrated into their Victron setup? Curious whether others have tackled this through the GX device or gone down a separate route like a Raspberry Pi or dedicated tank panel.

Would be good to hear from the boat folk on whether this is a known issue with certain narrowboat builders or tank suppliers specifically.

Hey @MuddyTrekker! Interesting one. With a buried tank you're likely getting ground temperature fluctuations affecting your pressure readings if you're using a pressure-based level sensor — the density of water shifts just enough to throw things off, especially with seasonal changes.

Also worth checking whether you've got any slight groundwater ingress causing the tank to partially float — even a well-installed buried tank can shift subtly, which would affect a float-based sender's calibration.

What type of level sensor are you running? If it's ultrasonic, debris or condensation on the sensor face can cause erratic readings too.

A simple sanity check — manually dip the tank with a marked rod and compare against your gauge readings over a few weeks. That'll at least tell you whether it's a genuine level issue or purely a sensor/display problem. Good luck with it!

@MuddyTrekker this is right in my wheelhouse — I've got a 500L static caravan water tank and wrestled with exactly this nonsense for two seasons.

The buried aspect actually works against you here. Your resistive sender is almost certainly reading different impedances as soil moisture changes around the tank walls, which alters the earth reference your gauge is using. It's not measuring water level — it's measuring the difference between two increasingly unreliable references.

Worth checking:

• Dedicated earth return wire directly from sender to gauge (not chassis/frame earth)
• Ultrasonic sender mounted top-down completely bypasses the resistance issue
• Veglia and Wema both do sealed senders that cope better with buried installations

I eventually ditched the resistive sender entirely and fitted a cheap ultrasonic unit wired through a Victron BMV shunt panel — overkill obviously, but I had the hardware spare and it's been rock solid since.

@MuddyTrekker had something similar on my narrowboat with a 200L tank — the float sender would give wildly different readings depending on whether the bilge pump had recently run or I'd been cruising (vibration shifts the float arm slightly).

Worth asking: is your sensor a float type or pressure/ultrasonic? On the boat I switched to a pressure transducer at the base of the tank and it's been far more consistent, as it's measuring actual water column weight rather than a mechanical float position.

Also — are your readings drifting gradually or jumping suddenly? That distinction might help narrow it down. Gradual drift on a buried tank does make me think thermal expansion like @ValleySolar mentioned, whereas sudden jumps usually point to a sensor or wiring issue.

What display/monitor are you using to read the level?

Hey @MuddyTrekker, great topic. One thing worth checking that nobody's mentioned yet — are you using a resistive float sender or a capacitive sensor? Resistive types with a float arm can give drifting readings if the pivot point gets any mineral build-up or corrosion, especially in buried tanks where you can't easily inspect them. The resistance value creeps over time and your gauge starts lying to you.

If you're feeding the signal into any kind of controller or display with an adjustable calibration curve, try logging the raw sensor output against a physical dipstick measurement over a week or two. That'll tell you pretty quickly whether you're dealing with a sensor drift issue versus something genuinely affecting the water volume, like @ValleySolar's ground temperature point. Worth ruling out methodically before throwing money at a new sensor. 👍