Blog · · 6 min read
DIY automatic plant watering system: design it to fail dry, not wet
Most DIY plant watering builds work until one part fails. How to choose pump, sensors and safety limits so that any single fault costs one small dose, not a flooded shelf.
A DIY automatic plant watering system looks like a weekend project: a moisture sensor, a pump, a tank and a small controller. Most builds work on the first evening. The real question comes three weeks later, when you are away and one part stops behaving. A plant survives a few dry days. A wooden shelf does not survive eight litres of water.
So the useful design rule is not "keep the soil moist". It is fail dry: no single fault may release more than one small dose of water. This article walks through what that rule means for the pump, the sensors and the controller. We are developing exactly this device as a public project, and you can follow it at DIY Automatic Plant Watering System.
How much water are we talking about?
Less than most people expect. Measured values for common houseplants are mostly between 35 and about 100 grams of water per day. If you plan with 60 mL per plant and day, six plants need about 360 mL daily. A tank of 8 to 10 litres then lasts two to three weeks.
Two things follow from these numbers. First, the pump can be tiny. It runs for seconds, not minutes. Second, the tank holds twenty or more days of water in one place. Everything that could empty it at once is the real hazard.
The faults that flood a room
Write the list before you choose parts. For a watering system it is short:
- Siphon. If the tube outlet is lower than the water level in the tank, water keeps flowing after the pump stops. Gravity does not need power.
- Stuck valve. A solenoid valve with a grain of dirt in its seat stays slightly open.
- Sensor that reads "dry" forever. A pulled plug, a broken wire or a corroded probe can all look like dry soil. The controller answers with water, again and again.
- Stuck switch. The transistor or relay that drives the pump fails in the "on" state.
- Hung controller. The software stops while the pump output is on.
- Loose tube. The pump works perfectly, and the water goes next to the pot.
- Power cut or lost Wi-Fi. Not dangerous alone, but they must not cause one of the faults above.
A design that only works when none of these happen is a demo. A design that survives each of them is a product.
Pump choice does half of the safety work
Many builds use one submersible pump and a row of valves. It is cheap for many plants, but it has two weak points from the list: valves can stick, and a tank above the plants can siphon through a leaking valve.
The alternative is one small peristaltic pump per plant. In this pump type, rollers squeeze a soft tube and push the water forward. When the motor stops, the rollers still pinch the tube shut. There is no valve that could stick, and a stopped pump blocks a siphon by its construction. The tank can stand where you like, and the plants can stay in their own pots and saucers.
The price is one pump per plant, and the quality of low-cost peristaltic pumps varies. The tube inside wears and should be easy to replace. For six plants, the parts for this layout come to roughly 110 to 150 in dollars or euros with low-cost pumps, and clearly more with branded ones.
Two other layouts are worth knowing. With one pump and valves, put the tank below the plants and add a small vent hole near the highest point of the line. Then water only moves while the pump runs. With a catch tray, all pots stand on one tray that drains into a basin larger than the clean tank. That protects furniture best, but the plants must stand together. One detail matters here: do not pump the run-off back into the clean tank. Recycled irrigation water is a known way for root diseases to travel from one plant to all of them.
Sensors: make the broken state look wet
Cheap moisture probes with two bare metal pins measure resistance. The pins corrode in wet soil within weeks. Capacitive probes have no exposed metal and last much longer. Use those.
More important than the probe type is what the controller sees when the probe is gone. Give each sensor input a resistor to ground. An unplugged sensor then reads zero volts. Define that in software as "fault", never as "dry". Add two plausibility checks:
- After a dose, the reading must move within some minutes. If it does not, the water went somewhere else, or the sensor is dead. Stop and lock this plant.
- A reading that is outside the range you saw during calibration is a fault, not a measurement.
Treat the sensor as a hint, not as a command. The dose is a fixed small amount, for example 20 to 30 mL, followed by a waiting time. The sensor only decides whether a dose is allowed.
The safety layer that does not trust the software
Software limits are good. They do not help when the software has stopped. So add a few limits that work without it:
- A hardware timer for the pump. A simple timer circuit between controller and pump driver cuts the power after a few seconds, whatever the controller says. A hung program or a stuck output then gives one dose, not a flood.
- A leak sensor in the saucer or tray. Two contacts at the lowest point. Water there stops all pumps.
- The tank on a scale, or a level sensor. If the tank loses more water than the doses add up to, something leaks. If it is empty, the pumps should not run dry.
- A daily budget per plant. After a fixed number of doses per day, the plant is locked until a person looks.
- A watchdog that restarts a hung controller, with all pump outputs off by default at start.
With these in place, each fault from the list above costs at most one small dose. Then the system locks the affected plant and sends an alert.
Power: a small trap with USB-C
Small pumps often want 12 volts. It is tempting to use a USB-C charger. But 12 V is an optional voltage in USB Power Delivery, and many chargers do not offer it. Either use a plain 12 V adapter, or choose pumps for 5 to 6 V and stay with a normal USB supply.
Test it by breaking it
Before you leave the system alone, try every fault on purpose, with the plants replaced by measuring cups:
- Pull each sensor plug while the system runs.
- Cut the power in the middle of a dose and watch the tube.
- Bridge the pump driver so that it is permanently on.
- Stop the controller while a pump runs.
- Pull a tube out of its pot.
- Pour water into the saucer.
Write down how much water came out each time. If the answer is ever more than one dose, the design is not finished.
What makes it a product
The watering logic fills ten lines. The work is in the faults: knowing them, choosing parts that remove some of them by physics, and limiting the rest in hardware. That is true for most small machines that run without supervision. Our public project applies this list step by step, from concept to printed parts and electronics: DIY Automatic Plant Watering System.
Watch hardware get built by AI. Follow public projects on makemachine.ai, read every result in full and get an email when there is something new.
More from the blog
- Will my electronics enclosure overheat? A five-minute heat estimate for a sealed box
- Joining 3D printed parts: screws, heat-set inserts, snap fits or glue?
- Free or paid AI 3D modeling? What you really pay for, and a 30-minute test to compare tools
Open-hardware projects built with AI
- DIY Automatic Plant Watering System: An open-hardware watering system for indoor plants: a small pump, a refillable tank and a soil moisture sensor in each pot keep up to six…
- DIY BBQ Temperature Controller: A temperature controller for charcoal grills and smokers that makers can build themselves. A pit probe, two meat probes and a small fan on…
- ZählerSchau: ZählerSchau is a small, battery-powered unit that clips onto an electricity meter in Germany (water meters via a clamp adapter), takes a…
- Roadstoy: Roadstoy is a remote-controlled toy version of the new Tesla Roadster that can briefly jump and flip using compressed air. It drives on an…
- The forever Washing machine: The forever washing machine is a washing machine offered as a service instead of a product. Customers do not buy the machine but pay for…
- Bike Light: I want to build a bike light that needs no battery. Magnets sit in the spokes close to the hub, and a coil on the fork picks up their…
- Rooficle: I want to develop a light, efficient and affordable e-bike that can be used legally on public roads in Singapore. Its core is an…
- BBQ Modular Grill: I want to design a modular grill. Instead of one sealed block, it is made of separate parts that can each be bought on their own online…
- Go FANUC Go: A mobile base that moves an existing FANUC M-710iC industrial robot (about 560 kg, with a milling spindle at the wrist) out of a room and…
- Jump & Run Robot: I want to build a small, low-cost toy humanoid robot that people can put together themselves. It should be about 30 cm tall, or as small…