I whakakapia e au taku whakakā mānu papu - anei ngā mea i kitea e au i muri 6 Ngā Marama
The moment water started rising and my old float switch just... didn't
It was 2 AM on a Tuesday in March when I woke up to a strange sound coming from the basement. Not a drip — a puna wai. My sump pump had failed, and by the time I found it, the water had already crept under the door to the utility room and was making its way toward the water heater. Total damage: mō $1,800 in ruined drywall and a ruined weekend.
Te tangata hara? He $4 mechanical float switch that had simply jammed shut with mineral buildup. After that night, I went down a research rabbit hole that led me to optical water level sensors — and I haven't looked back since.
If you're dealing with a sump pump, an espresso machine reservoir, a hydroponics setup, or any kind of fluid management system, this one's for you. I've spent the last six months testing optical sensors in real-world conditions. Let me walk you through what actually works — and what the reviews on Amazon won't tell you.
How Optical Sensors Actually Work (No Jargon, I Promise)
Before we get into the comparison, I want to make sure you understand he aha these things work the way they do, because it genuinely is kind of cool.
Te Nuinga pūoko taumata wai ōmata use what's called a prism-tip probe — a small glass or clear plastic tip at the end of the sensor. Inside the housing, there's an infrared LED that fires a beam of light straight into the base of that prism, and a receiver sitting nearby waiting to catch it.
Here's the clever part: when the prism is sitting in air, the light bounces around whakaroto the prism due to total internal reflection and makes it back to the receiver. But when liquid covers the prism, the refractive index changes — the light escapes out into the liquid instead of bouncing back. The receiver notices the sudden drop in returning light and flips its output state in about 50 hēkona.
Koirā. Kāore he wāhanga nekehanga, no floats, no shafts. Just light and physics.
I tested this with a cheap Arduino setup — a KEYTHON optical sensor I grabbed for about $8 i Amazon — and watched the serial monitor flip from HIGH to LOW the exact second I dipped the tip into a glass of water. It was genuinely satisfying.
Why I Ditched My Mechanical Float Switch (And You Probably Should Too)
Kia pono tātou: mechanical float switches are fine — they've been doing the job for 50 tau. Engari "fine" me "pono" are two different things, and after my basement flood, I wasn't interested in fine anymore.
Here's the core problem: moving parts wear out. A mechanical float switch has a physical float that rides up and down a shaft as the water level changes. That float triggers a magnetic reed switch at a certain height. Sounds simple, Matau? It is — until debris, putunga konupūmā, or just years of use cause the float to get stuck, the shaft to corrode, or the reed switch to seize.
I pulled my old float switch out after the flood. The shaft was coated in a thick white calcium crust about 2mm thick. The float itself was partially sunken — it had been slowly losing buoyancy for months without me knowing.
Now compare that to optical sensors:
No moving parts = no jamming. The sensor just sits there. It can't get stuck, can't corrode on a shaft, can't lose buoyancy.
They fit anywhere. My old mechanical float needed at least 4 inches of vertical clearance to operate. The KEYTHON sensor I installed is about the size of a AA battery. I mounted it directly to the side wall of my sump pit with a zip tie and a bit of silicone. Kua oti.
They last longer in real conditions. I spoke with a HVAC technician friend of mine who's been installing optical sensors in commercial applications for about three years. His words: "Once you go optical on a critical application, you don't go back."
That said — mechanical float switches are still significantly cheaper, te tikanga $3–$10 ia wae, and for non-critical applications like a garden irrigation reservoir, they may be perfectly adequate. You don't need a Ferrari to drive to the grocery store.
Real Sensors, Real Prices — What I Actually Bought and Tested
Rather than just reading spec sheets, I went out and bought a handful of the most commonly discussed sensors to test myself. Here's what I found:
KEYTHON Optical Liquid Level Sensor — ~$8–12 The budget hobbyist pick. I used this for my Arduino sump pump project. It's compact, works at 5V DC, and has three wires (VCC, GND, PUTA). Dead simple to wire up. The plastic prism tip isn't rated for high temperatures or harsh chemicals, so I'd stick with this for indoor, room-temp water applications only. Great value for the price.
Flosun IP68 Water Level Sensor — ~$15–25 This one is what I ended up installing in my sump pit long-term. It's fully potted in epoxy (true IP68), has a glass prism, and handles continuous submersion without issue. I paid around $22 with shipping. Installation was straightforward — just connect the three wires to a relay module, ā, kua mutu koe. The manufacturer claims operating range is -10°C to 85°C, which covers most residential use cases.
Omron FL7M Series (ahumahi) — ~$55–$120 depending on configuration I didn't buy this one (I don't have an industrial application), but I researched it extensively because it's widely cited in engineering forums. The Omron FL7M uses a stainless steel housing with a sapphire prism tip, rated for pressures up to 1 MPa and temperatures up to 120°C. If you're working with chemicals, kohu, or high-pressure systems, this is the tier you're looking at. The price is obviously in a different league, but for a commercial chemical tank or a boiler system, the cost is trivial relative to potential damage.
For reference — Ultrasonic sensors (for comparison): If you need continuous level measurement — not just "is the water at this point, yes or no" — then an ultrasonic sensor like the JSN-SR04T ($10–15) or a Siemens ultrasonic level transmitter ($150–400 industrial grade) is the right tool. Ultrasonic bounces sound waves off the liquid surface and tells you how full the tank is from 0–100%. Te whakawhitiwhiti: they have a minimum detection distance (he "pito matapo"), can be affected by foam and turbulence, and cost more. For overflow protection and run-dry prevention, heoi, point-level optical sensors are faster, cheaper, me te pono ake.
Tāutanga: He ngāwari kē atu i taku i tūmanako ai
I want to address the elephant in the room — are these actually hard to install?
Kāore. Genuinely, Kāore.
For a basic Arduino/Raspberry Pi/ESP32 project, here's literally all you do:
- Connect VCC to 5V or 3.3V on your microcontroller (check your sensor's rating first).
- Connect GND to ground.
- Connect the Signal (PUTA) wire to any digital input pin.
- Upload a 10-line sketch to read the pin state.
I had mine running in under 20 meneti, including time spent rewiring because I initially connected the relay module backwards. (My fault, not the sensor's.)
For a sump pump installation like mine, the wiring is similar but uses a relay to switch the 120V pump circuit. Make sure you're using a DPST relay (double-pole, single-throw) so both the hot and neutral legs are disconnected when the sensor triggers. That's a safety must — never rely on switching just one leg.
Total installation time for my sump pump setup: mō 45 meneti, including drilling a mounting hole and running the wire through a cable gland. Cost of installation materials beyond the sensor itself: roughly $8–15 for a relay module, cable gland, me te waea.
One thing worth noting: if you're installing this in a potable water system or any application subject to building codes, check your local regulations. Most residential sump pump and water heater tray installations don't require permits, but commercial and industrial applications often have specific requirements.
Ngā Whakahekenga Inihua: A Bonus You Might Not Know About
Here's something I didn't expect to find during my research: several home insurance providers offer premium discounts for policyholders who install water leak detection systems — and some specifically list water level sensors in their qualifying equipment.
After installing my optical sensor system, I contacted my insurer and asked. They required:
- A signed letter from a licensed plumber (I had mine out for an unrelated issue and got him to sign a one-pager)
- Photos of the installed sensor
- The sensor's product specification sheet
Hua: 8% whakahekenga utu on my annual premium, which comes to about $34 ia tau. The sensor and installation cost me around $40 total. Payback period: mō 14 marama, then pure savings after that.
It's not a fortune, but it's something. Call your insurer — many don't advertise this, but they have a list of approved devices. If your sensor is IP68-rated and specifically marketed as a leak/flood detector, there's a good chance it qualifies.
The Nasty Bits: What Optical Sensors Can't Do (And Common Problems)
I want to be straight with you because this isn't a sponsored post — I actually had some headaches along the way.
Foam and bubbles will fool them. This is the biggest limitation. Because optical sensors rely on light refraction, a thick layer of foam or a cluster of bubbles can look exactly like liquid to the sensor. If you're dealing with a tank that generates a lot of foam (some chemical processes, certain cleaning solutions, even vigorous fermentation), this is a real problem.
Te whakatika: look for sensors with software-adjustable delay settings or specialized prism geometries designed to shed bubbles. The Flosun I installed has a 0.5–5 second adjustable delay via a small trimmer potentiometer on the board — I set it to 2 seconds and it eliminated false triggers from sump pit turbulence. The KEYTHON budget sensor does not have this feature, so I had to handle the delay in code instead.
Hard water scaling. After about 4 months in my sump pit (which has moderately hard water), I noticed the sensor occasionally stuttering — triggering and untriggering rapidly for a second before stabilizing. I pulled it out and there was a faint calcium film on the prism. A wipe with a cloth dipped in white vinegar, 30 hekona, done. Total issue. But it is something to be aware of if you're installing this in a high-mineral-content application.
They're point-level, not continuous-level. You need to understand this distinction clearly. An optical sensor tells you "liquid has reached this exact point." It does not tell you "ko te kura 62% Kua kī." If you need continuous monitoring, you need ultrasonic, kaha, or pressure-based sensors instead. This isn't a flaw — it's just a different tool for a different job.
Which Sensor Should You Actually Buy?
Here's my honest take, based on what I tested and what I've read from others in the same situation:
For a home sump pump, water heater pan, or basement leak detector: → Flosun IP68 (~$15–25) — the sweet spot of price, mauroa, me te whaikiko. Install it, set it, Mahi atu.
For a DIY Arduino/ESP32 hobby project: → KEYTHON (~$8–12) — dead cheap, works great, just don't expect industrial-grade durability.
For any commercial, ahumahi, or safety-critical application: → Omron FL7M or equivalent (~$55–120+) — the additional cost is absolutely justified by the reliability ratings, awhe pāmahana, and pressure tolerance.
For continuous level monitoring (not just point detection): → Add an ultrasonic sensor to your setup on top of the optical — or go full ultrasonic if point-level detection isn't sufficient.
Te Raina Raro
After six months with an optical sensor running my sump pump alarm system, I can say with confidence: this technology is genuinely mature and reliable. The physics are solid, the components are inexpensive, and the failure modes are predictable and easy to troubleshoot.
If you've been putting off upgrading a sketchy float switch because you assumed it would be complicated or expensive — it isn't. You can be up and running for under $40 and about an hour of your time.
And maybe more importantly: call your insurance company. That discount alone might be the nudge you needed to finally do it.
*I purchased all sensors mentioned in this article with my own money. No sponsorships, no affiliate links. The opinions are entirely my own.