Automatic dosing pumps overshoot in small 20- to 50-litre classroom hydroponic reservoirs because of transport delay: the time it takes for concentrated acid or nutrient solution to mix thoroughly through the tank is longer than the time it takes for the controller to decide to add more.

When an Arduino, ESP32, or commercial relay board detects a pH of 6.8 against a target of 5.8, it turns on a peristaltic pump. In a small tank with a submersible fountain pump, that acid takes between 45 and 180 seconds to disperse uniformly across the entire water volume. If your sensor code checks the probe every two seconds and keeps running the pump until the reading reaches 5.8, the pump will dump enough concentrated phosphoric or nitric acid to drop the tank down to pH 4.0 or lower once mixing finally completes. The plants then suffer root burn or nutrient lockout, and the system oscillates wildly as an automated base pump tries to correct in the opposite direction.

The Core Problem: Control Loops Faster Than Physical Mixing

In industrial horticulture, reservoirs hold thousands of litres and use high-flow venturi injectors or high-volume circulation pumps that move water rapidly past industrial sensors. In a school lab or classroom, typical reservoirs are 20-litre buckets or 50-litre storage totes fitted with small 5W to 15W submersible pumps.

Water in these containers does not mix instantly. It forms stratified layers and dead zones. If you drop concentrated nutrient A and B or pH Down into one corner of a 40-litre tote:

• Scenario A (Probe near the doser): The sensor gets hit by an unmixed plume of pure acid, instantly reads pH 4.5, shuts off the pump prematurely, and leaves the bulk water unchanged.
• Scenario B (Probe far from the doser): The acid stays in a localized plume while the probe continues to read the unmixed bulk water at pH 7.0. The controller keeps the peristaltic pump running continuously for twenty seconds. Two minutes later, once the submersible pump circulates the plume, the whole tank crashes to lethal acidity.

Fix 1: Implement Pulse-Pause Dosing Logic

The single most effective software fix is replacing simple proportional or continuous threshold logic with a strict pulse-pause cycle.

Never run a dosing pump continuously until a target threshold is reached. Instead, dose a fixed micro-volume, shut the pump off, and enforce a mandatory mixing delay before the sensor reading is trusted again.

In code, this means your control loop triggers a run time of 1 to 2 seconds on a standard 12V peristaltic pump (which typically dispenses roughly 1 ml per second), and then transitions into a waiting state where dosing is locked out for at least two to three minutes while the submersible circulation pump homogenises the solution.

Fix 2: Physical Separation and Inline Mixing Chambers

Where you place your dosing outlets relative to your probes determines whether your readings reflect reality.

• Never put dosing tubes next to sensors: Dosing lines should discharge directly into the turbulent discharge stream of your main circulation pump or return drain, never into stagnant tank corners or directly above the sensor manifold.
• Use an inline probe manifold: Instead of dangling sensitive pH and electrical conductivity (EC) glass probes directly into the open reservoir where students might knock them or where boundary layers form, place them in a bypass loop or an inline tee on the return plumbing.
• Add a mixing baffle: If your tank is a compact tote, passing the dosed return water through a small perforated container or baffled channel before it enters the main reservoir forces mechanical mixing and breaks up high-concentration plumes.

Fix 3: Dilute Your Stock Solutions

Standard commercial hydroponic concentrates are formulated for commercial farms where single doses are measured in hundreds of millilitres. Using undiluted 50% phosphoric acid or industrial nutrient concentrates in a 25-litre classroom system means that a single extra second of pump activation overshoots the target drastically.

For small school tanks, dilute your working stock bottles:

• pH Adjusters: Dilute standard concentrated pH Down (phosphoric or nitric acid) with distilled or reverse-osmosis water to a 5% or 10% working solution. A 2 ml dose of a 5% solution nudges a 30-litre tank gently rather than causing an irreversible plunge.
• Nutrient Concentrates (EC): If using two-part (A & B) feeds, dilute them 1:5 or 1:10 with pure water before connecting them to the peristaltic dosing tubes, adjusting your pump calibration steps accordingly.

Classroom Realities: Load-Shedding and Holidays

In South African schools, automated dosing faces two real environmental hurdles: power interruptions and long, unattended school holidays.

When load-shedding drops power to the system, submersible circulation pumps stop immediately. If a micro-controller reboots upon power restoration and immediately takes an EC or pH reading before the circulation pump has run long enough to clear stagnant water around the probe, it will dose based on stale, localized readings. Always program a startup delay: require the main circulation pump to run for at least five minutes after power restoration before permitting any automated dosing cycle to fire.

If you are designing automated food production systems or agricultural science teaching stations, our curriculum-aligned guides at Sheen Robotics Farming provide complete reference architectures, including tested pulse-pause firmware snippets and failsafe configurations for small-footprint builds. You can also explore our modular sensor integration hardware at Sheen Robotics Hydroponics.

By pairing diluted stock solutions with pulse-pause intervals that respect physical fluid mechanics, even a modest 20-litre desk reservoir can maintain stable pH and EC values through weekends without risking chemical shock to the crop.