Ek het my vlotskakelaars vir ultrasoniese vlaksensors verruil — twee jaar, Een Byna-Vloed, Nul Spyt
The tank that nearly bit me sits next to my shed — 2,500 gallons of stored well water that feeds both the house and the garden irrigation. For years it was guarded by the same technology people have trusted since the 1970s: a twelve-dollar float switch dangling off a wire.
Last spring it stuck. Not stuck off — stuck on, which is somehow worse. The tank filled up, the overflow pipe did its job, and then the overflow pipe’s job became flooding the yard for six straight hours while I was at a trade show an hour away. I came home to a swamp, a waterlogged shed floor, and a pump controller that had been cycling hot until it finally gave up. That replacement pump cost me $380 and a weekend I’ll never get back.
And here’s the part that really made me upgrade: the replacement float did the exact same thing two months later, only in reverse. It hung up on a bit of algae, never told the pump to stop, and the pump ran itself dry and burned out. Weer. That’s when I swore off mechanical floats for good and started reading about ultrasonic level sensors like a man possessed.
Why I finally dumped my float switches
Look, float switches aren’t evil. Hulle is goedkoop, they’re dead simple, and honestly? For a sump pit in the basement, they’re fine. But for a tank you actually depend on, they have three problems that I learned the hard way:
Eerste, they’re mechanical, and mechanical things jam. Skaal, Alge, Puin, a wire that snags the arm — give it a year and it will find a way to lie to you. Tweede, a float is binary: it tells you “full” or “empty” at exactly one point, and nothing in between. There’s no trend, Geen waarskuwing nie, no “hey, you’re at 60% and your pump is acting weird.” Third, you can’t see it from anywhere. I spent six hours driving home from that trade show with zero idea whether my yard was flooding. That’s not monitoring — that’s hoping.
What the thing actually does (no engineering degree required)
The ultrasonic sensor I eventually installed is embarrassingly simple in concept. Inside the housing there’s a piezoelectric transducer — basically a tiny speaker that can also listen. It fires a high-frequency sound pulse down at the water surface, the pulse bounces back, and the microprocessor times the round trip like a very precise stopwatch. Since the speed of sound in air is a known quantity, the distance to the water — and therefore the level in the tank — is just arithmetic.
The clever part is what happens when conditions aren’t textbook. Sound travels faster in warm air and slower in cold air, and the air above a tank is never a constant temperature. So these units carry a built-in thermistor and compensate automatically. A heatwave in July or a freezing night in January doesn’t throw the reading off. I checked mine against a tape measure a few times during the first week — it was within a quarter inch every single time.
Two things I wish I’d known before buying:
The dead band. Every ultrasonic sensor is blind for the first few inches right below its face — it can’t fire and listen at the same time. If the tank fills up into that zone, the readings go erratic or just wrong. Mount it high enough that the maximum fill level stays clear of this.
Nothing touches the liquid. That’s the whole point. It’s why the same tech is used on chemical tanks — a buddy of mine stores sodium hypochlorite solution and he’d never put a float or a probe anywhere near that stuff. No contact means no corrosion, no contamination, no moving parts to fail. Maintenance is basically “wipe it off once a year.”
Float vs. ultrasonic vs. radar — the honest version
Every engineer I talked to kept bringing up radar, so let’s settle it. Radar uses microwaves instead of sound, which makes it immune to the stuff that bothers ultrasonic: Swaar dampe, stof, Skuim, wild temperature swings. My friend who runs a municipal water plant uses radar on the digester tanks, where there’s a permanent layer of foam — ultrasonic just can’t see through that.
But radar costs real money and it’s fussier to set up. We’re talking $2,000-plus per loop for a decent unit, and you need someone who knows what they’re doing to configure it. Vir water, Olies, Ligte chemikalieë, irrigation reservoirs, rainwater tanks — the boring-but-important 90% of the world — ultrasonic gives you the same continuous, accurate level data at a fraction of the price. Foam is genuinely its weak spot; if your tank foams, skip straight to radar or fit a stilling well. Otherwise, don’t overthink it.
What it costs, and what I’d actually buy again
Prices move around, so treat these as ballpark US street numbers as of when I was shopping:
The float switch I started with: $12–25 on Amazon. You get what you pay for.
DIY route: a waterproof JSN-SR04T ultrasonic module runs about $12 if you enjoy Arduino tinkering. I enjoyed it in theory. In practice I wanted something that worked without a soldering iron at 11 n.m..
Where I landed: a generic IP68-rated 4-20mA ultrasonic transmitter, Ongeveer $220 on Amazon — adjustable beam angle, built-in false-echo suppression, Die werke. Paired it with a ~$130 cellular gateway that feeds an app on my phone. No Wi-Fi at the shed, so cellular was the only honest option.
The name-brand industrial tier: if you need certifications and a support line, the Siemens Sitrans LU150, the Endress+Hauser Prosonic FMU30, or a VEGA VEGASON 61 run anywhere from $700 na $1,000+, and that’s before you buy a single cable.
The lazy-but-smart option: battery-powered LoRa tank monitors (the YoLink-style ones) go for about $50–70 plus a $30 hub. No wiring at all, and the app is decent. You lose the 4-20mA output that a PLC or controller wants, but if your goal is just “check the tank from my phone,” it’s honestly the best value on the list.
My raad: don’t buy the $12 no-name, and don’t buy the $1,000 name-brand unless you genuinely need it. Read the datasheet for three numbers — dead band, beam angle, and IP rating — and you’ll be fine.
Installasie: where I screwed up so you don’t have to
I’ll be straight with you: my first install was wrong. I mounted the sensor too low, basically at the level I thought was the max fill — which meant the moment the tank got genuinely full, the water entered the dead band and the app started showing garbage. I spent a whole evening convinced I’d bought a defective unit before I re-read the manual, felt like an idiot, and moved it up six inches.
Once mounted properly, it’s a two-hour job: drill a hole, fit a cable gland, mount the sensor high enough to clear the dead band, and run the two-wire 4-20mA loop back to the controller. An electrician quoted me $350 for that cable run; I did it myself with a buddy, a drill, and YouTube, and it was fine. Calibration is just menu work: set the zero point (distance to the bottom of the empty tank), set the span (distance to max fill), and then run the “teach” routine that maps the tank’s permanent echoes — ladders, Pype, the wall where the beam bounces — so the sensor learns to ignore them. Took me about ten minutes.
Een eerlike voorbehoud: after a heavy rain I got a false high reading for about an hour when the surface got foamy. The app history showed it clearly, nothing shut down, and it hasn’t repeated. Annoying, not dangerous — but now you know.
The surprise discount from my insurance agent
I almost didn’t include this because it sounds like a scam, but it’s real. After a neighbor’s basement flood (washing machine hose, $14,000 in skade), I asked my agent whether any of my gear mattered for my policy. Turns out a lot of carriers offer a 3–5% discount on homeowners coverage if you can