Hydroponic sensors range from two dollars to several hundred, and the expensive one isn’t always more accurate. In a 2024 study (Nandi and Shrestha, published in Sensors), a $2 capacitive moisture sensor paired with a $5 external analog-to-digital converter was off by no more than 1.87% volumetric water content after calibration. A $350 Decagon 5TM was off by no more than 0.95%. The authors found no statistically significant difference between them.
That result comes with two caveats, covered below. After that we go through substrate moisture (VWC), EC, pH, and dissolved oxygen (DO) sensors: the types, how long they last, what fails first, and where you can save money. How to read the numbers together and which set you need to start with is covered in the wiki entry sensors as one system. This article is about choosing the hardware.
Hydroponic sensors: what to measure in your system
Every hydroponic system needs solution EC, pH, and temperature. The rest depends on the system:
- DWC and rafts: roots sit in the solution, and aeration is their only source of oxygen. You need a DO sensor.
- Coco and rockwool on drip: you need a substrate moisture sensor and drain EC readings. The tank can be fine while salts build up in the slab.
- NFT: the solution runs as a thin film and mixes constantly, so EC, pH, and a thermometer come first.
How the systems compare on cost and risk is covered in NFT vs DWC vs Kratky vs drip.
Substrate moisture sensor: $2 or $350
Nandi and Shrestha compared four sensors: a resistive one at about $1.50, a capacitive one at about $2, a Vegetronix VH400 at about $42, and a Decagon 5TM at about $350. Each was calibrated to its own soil, with moisture set from 0 to 50%.
| Sensor | Price, $ | Maximum error (RMSE), % VWC |
|---|---|---|
| Resistive, Arduino built-in ADC | 1.50 | 4.79 |
| Capacitive, Arduino built-in ADC | 2 | 3.78 |
| Resistive + ADS1115 | 1.50 + 5 | 2.64 |
| Capacitive + ADS1115 | 2 + 5 | 1.87 |
| Vegetronix VH400 | 42 | 1.8 |
| Decagon 5TM | 350 | 0.95 |
Most of the gain came from the 16-bit ADS1115 converter. With the Arduino’s built-in ADC, the capacitive sensor’s signal wandered by about 27 mV at a constant moisture level; with the ADS1115, by less than 1 mV. Now for the caveats.
First, a cheap sensor needs calibration, and each unit separately. In a 2025 Italian study (Abdelmoneim and co-authors, Sensors), researchers took 12 identical DFRobot SEN0193 capacitive sensors. Above 30% soil moisture, their readings differed from one another by 10-16% (coefficient of variation). A single calibration curve for the whole batch gave an error of 4.5-4.9%.
Second, both studies used soil, not coco or rockwool. You’ll have to calibrate to your own substrate by weighing the slab or block. Nobody knows yet how long such a sensor lasts in wet, fertilized substrate: Nandi and Shrestha say plainly that the long-term durability of low-cost sensors still needs field testing. Target VWC ranges for coco and rockwool, and how to place the sensor in a block, are in the wiki entry VWC: volumetric substrate moisture.
EC sensor: the one that lasts longest
For its lab-grade EC probe with a K 1.0 cell constant, Atlas Scientific lists ±2% accuracy and a life expectancy of about 10 years. According to the datasheet, recalibration is needed roughly once every 10 years too. The manufacturer says a conductivity probe has no parts that wear out. You will need to recalibrate if you change the cable length or move the probe to a different controller, since either changes the electrical properties of the circuit.
The main source of error in an EC sensor is temperature. The Atlas datasheet gives an example: a calibration solution rated at 12,880 µS/cm at 77°F (25°C) reads 9,330 at 50°F (10°C), 28% lower. A meter with temperature compensation converts the reading to 77°F (25°C), but the compensation coefficient is built in and may not match your solution. So buy a sensor with a built-in thermometer and measure the solution at a consistent temperature. How to check a meter against a standard is covered in the wiki entry EC meter calibration.
With substrate, one sensor in the tank isn’t enough. The EC of the solution draining from the slab tells you whether salts are building up in the root zone. How to collect it and how to read the difference from the feed is in the wiki entry drain EC.
pH sensor: a consumable
A glass pH electrode ages even on the shelf. For the same maker, Atlas Scientific, the working life of a pH probe is about 2.5 years, and it keeps in storage for about 5 years. An EC probe outlasts four pH probes, so budget for a new pH electrode every two to three years.
Never let the electrode dry out: between readings it should sit in storage solution. For hydroponics, the manufacturer recommends recalibrating once a year for the first two years and about every six months after that. Those figures are for one lab-grade probe. A weekly check against a buffer solution tells you whether your own meter is holding its calibration. The procedure and common mistakes are in the wiki entry pH meter calibration.
Dissolved oxygen sensor: galvanic or optical
DO sensors come in two types. A galvanic (electrochemical) sensor has a membrane with electrolyte inside. It consumes a little oxygen itself, so water has to move past the membrane: Atlas Scientific specifies about 60 mL/min (2 fl oz/min). In still water the oxygen next to the sensor gets depleted and it reads low. When the electrolyte runs out, the sensor reads very low numbers even though the solution is fine.
Manufacturers disagree on maintenance. Atlas Scientific recommends replacing the membrane and electrolyte in its galvanic probe every 1-2 years. YSI says membranes on electrochemical sensors get changed every 2-6 weeks. An optical sensor has no membrane; instead it has a cap containing a luminescent dye. According to YSI, the cap lasts 15-18 months (with a one-year warranty) or 24-28 months (with a two-year warranty). An optical sensor needs no stirring and no warm-up.
If you check DO once a week with a handheld meter, a galvanic one will do. For a permanent sensor in a DWC pond, an optical one costs more up front but needs less upkeep. DO targets, the link to temperature, and where exactly to measure are covered in the wiki entry DO: dissolved oxygen.
Sensors: service life and upkeep
| Sensor | Service life | Upkeep | Source |
|---|---|---|---|
| EC, K 1.0 | ~10 years | recalibrate ~once in 10 years | Atlas Scientific |
| pH, glass | ~2.5 years in use | keep wet; recalibrate every 6-12 months | Atlas Scientific |
| DO, galvanic | 5+ years | membrane and electrolyte every 1-2 years (Atlas) or every 2-6 weeks (YSI) | Atlas Scientific, YSI |
| DO, optical | cap 15-28 months | replace the cap | YSI |
| Moisture, $2 capacitive | not tested | calibrate each unit to the substrate | Nandi and Shrestha, 2024 |
Once sensors stay in place and control dosing, you also need fail-safe logic. What should shut the pumps down is covered in the wiki entry pH and EC automation. How to build that controller on an Arduino or ESP32, and why an EC sensor throws off pH in the same tank, is covered in Arduino pH and EC controller for hydroponics.
Sensors for a small farm: key points
- A $2 capacitive moisture sensor with a $5 ADS1115 converter was off by 1.87% VWC after calibration; a $350 sensor by 0.95%. The difference wasn’t statistically significant, but the test was in soil.
- Calibrate cheap moisture sensors one by one: units of the same model differed by 10-16%.
- An EC probe lasts about 10 years, a pH probe about 2.5. Plan on replacing pH probes regularly.
- Without temperature compensation, EC reads 28% lower at 50°F (10°C) than at 77°F (25°C).
- A galvanic DO sensor needs water flow, an optical one doesn’t, and the optical type needs less maintenance.
