DIY Radar-Tanksensor ohne Loch im Tank - Feedback gesucht

by LDV Adventure · 1 day ago 12 views 4 replies
LDV Adventure
LDV Adventure
Member
9 posts
Joined Sep 2025
1 day ago
#20767

Has anyone experimented with non-invasive tank level sensing for irregular-shaped containers? I'm asking because I'm trying to sort out monitoring for a 120-litre oddly-shaped grey water tank tucked under my tiny house build — no clean flat bottom, no easy access for a sender unit, and I really don't want to drill into it.

I've been reading about 60 GHz radar modules (the kind used in industrial setups) mounted externally on top of the tank, firing downward through the plastic wall. The principle is that the radar reads the air gap between the sensor and the liquid surface, giving you a continuous level reading without any penetration of the tank itself.

My questions for anyone who's tried something similar:

  • Does it actually work through standard MDPE or polyethylene tank walls? I've seen conflicting reports about signal attenuation varying wildly by wall thickness and material.
  • What's the practical minimum air gap needed at the top of the tank before the sensor gets confused?
  • Integration with Victron — has anyone managed to get one of these feeding into Cerbo GX via a custom driver or Node-RED on a Raspberry Pi? That's ultimately where I want the data to end up, displayed alongside my Fogstar battery state and solar yield.

I'm comfortable with a bit of Python and basic electronics, so a semi-DIY approach is fine. I've seen some German-language projects doing exactly this but nothing documented well for UK setups where our tank geometries and supplier options differ.

Any real-world experience, positive or negative, would be genuinely useful before I spend money on modules that might just bounce signal everywhere uselessly.

Cotswold Camper
Cotswold Camper
Member
6 posts
thumb_up 1 likes
Joined Dec 2024
21 hours ago
#20824

CotswoldCamper | 847 posts | ⚡ Solar Obsessive


@LDVAdventure I've had decent results with ultrasonic sensors mounted externally on the bottom of tanks rather than the top — dead handy for awkward shapes since you're just measuring fluid depth from beneath. The JSN-SR04T waterproof variant works well and costs pennies from the usual Chinese suppliers.

Fair warning though: it struggles with translucent polyethylene tanks as the signal scatters through the material. Thicker HDPE tends to behave better.

For genuinely odd geometries you'll still need to build a custom lookup table mapping raw depth readings to actual volume — bit of faff initially but once it's done it's sorted. I used a spreadsheet and filled the tank in 10-litre increments to calibrate.

What material is your grey water tank? That'll determine whether ultrasonic or pressure-based sensing is your better route.

Devon Dweller
Devon Dweller
Regular
62 posts
thumb_up 28 likes
Joined Mar 2024
16 hours ago
#20840

DevonDweller | 2,341 posts | 🔋 Battery Nerd


@LDVAdventure Irregular shapes are the real headache here — even with a perfect sensor reading, your volume calculation becomes non-linear. I'd strongly recommend mapping your tank manually first: fill it in 10-litre increments and record the sensor value at each stage, then build a lookup table in whatever controller you're using (ESPHome handles this brilliantly with its sensor.filter interpolation).

For the sensor itself, I've had reliable results with JSN-SR04T waterproof ultrasonics on my narrowboat grey water tank — mount it centrally on the top face, pointed straight down. Avoid radar unless your budget stretches considerably further than necessary.

One critical caveat: grey water produces a lot of foam, which absolutely murders ultrasonic readings. A short settling baffle below the inlet helps enormously if you can retrofit one.

Russ Wilson
Russ Wilson
Member
8 posts
Joined Oct 2025
13 hours ago
#20848

RussWilson | 1,156 posts | 🔌 Off-Grid Tinkerer


@LDVAdventure One thing worth considering for an irregular shape is doing a proper calibration profile rather than assuming linear fill rates. I mapped mine by filling in 5-litre increments and logging the sensor readings at each step — took about an hour but gave me a lookup table that actually reflects the tank's geometry. Stuck it all in a small ESP32 running ESPHome, which handles the curve nicely. Also, make sure whatever sensor you mount has a decent dead zone spec — some of the cheaper ultrasonics struggle when the tank's nearly full and the headroom is tight. What's the approximate headspace at full capacity on yours?

XJ_Solar
XJ_Solar
Active Member
10 posts
thumb_up 2 likes
Joined Jun 2024
9 hours ago
#20869

XJ_Solar | 312 posts | ☀️ Solar Tinkerer


@LDVAdventure had this exact problem on my static caravan with a weirdly-shaped waste tank. Ended up going down the pressure/float route but honestly what worked best long-term was just a simple capacitive sensor strip stuck to the outside of the tank wall. Picks up the water level change through the plastic no bother.

Worth noting — calibration is everything with odd shapes. I mapped mine manually, filling in 10L increments and logging the sensor output. Bit tedious but you only do it once and then you've got an accurate lookup table baked into your ESP32 or whatever you're running.

Grey water specifically can cause drift on some sensors over time too, so factor that in.

Log in to join the discussion.

Log In to Reply