Before adding another dose to a recirculating nutrient film technique (NFT) system, check three things. Did the pump deliver the intended volume? Did the solution have enough time and flow to mix? Does a second measurement agree with the first? If any check fails, hold the next dose and investigate. A slow or poorly placed probe can otherwise turn one correction into several.
This guide is for growers, greenhouse maintenance leads and controls integrators. It starts with a monitored reservoir and an operator-approved correction, then shows how to decide whether a dosing controller is ready for tighter supervision or automation.
Verify what the probes measure
Electrical conductivity (EC) describes how readily dissolved ions carry current. It is useful for tracking changes in the overall strength of a known nutrient solution. Different mixtures can have the same EC, so the reading cannot identify a nitrogen shortage, show whether the A and B concentrates were delivered in the right ratio or distinguish fertilizer salts from salts already in the source water. Record the incoming water’s EC, review its laboratory analysis where appropriate and track the approved formulation separately. Oklahoma State University Extension’s hydroponic EC and pH guide describes how source water and its alkalinity affect interpretation.
A glass pH electrode measures a small voltage relative to a reference electrode. It needs a compatible high-impedance pH input. Connecting it directly to a microcontroller ADC can load the signal and produce unreliable readings. Use fresh calibration buffers that bracket the operating range, and check the manufacturer’s limits for response time and slope. Keep the probe hydrated and clean the glass and junction using its approved procedure. Check the conductivity cell against a suitable conductivity standard and verify its cell constant, temperature reading and compensation setting.
Record solution temperature with both measurements. For EC, a suitable compensation setting makes readings more comparable at a stated reference temperature. Its coefficient depends on the solution. For pH, automatic temperature compensation adjusts for the electrode’s changing response. The solution’s actual pH can still change with temperature, so compare the online probe and an independent meter at similar temperatures. Hamilton’s process measurement guidance and its pH temperature guidance explain the distinction.
Sample recirculating solution
Place the probes where fresh, representative solution reaches them at both normal and low operating levels. A well-mixed sample chamber can make cleaning and checks easier, provided its supply flow is proven. Keep the measuring point away from concentrate injection, incoming make-up water, trapped bubbles and stagnant corners. Identify an accessible point for taking an independent sample. During commissioning, compare points in the reservoir and the return so the check can detect short-circuit flow or a tank that mixes unevenly.
Track reservoir level and actual recirculation. A pump status bit can stay on with a blocked line, air lock or failed impeller. A verified flow switch or meter can confirm circulation where the measurement matters. A reservoir measurement may also remain normal when one NFT channel loses flow. If blocked channels are a concern, check representative branch returns separately. Record low level, lost flow, stale samples and calibration failures as distinct conditions. Each should be visible locally and should block further dosing when it prevents reliable measurement or mixing.
Keep probe signals separate from motor wiring and validate the installed system with pumps switching on and off. Test whether an EC circuit or motor shifts the nearby pH reading. A bench prototype can use dedicated pH and EC circuits with manufacturer-supported isolated interfaces. For a wet production installation, evaluate supported industrial transmitters, connectors and enclosures with the controls installer.
Measure pump delivery and mixing time
Each peristaltic pump advances a small amount of liquid by compressing a tube. Delivery changes with tubing wear, suction lift, discharge head, air in the line and the liquid. Measure each pump’s actual output under approved conditions. Repeat the test at the smallest run duration you expect to use. A motor that turns can still move too little or nothing, and a stopped motor alone does not rule out unintended flow. Have the installer evaluate compatible tubing, backflow and siphon prevention. Treat a tubing change, prime loss or plumbing change as a reason to recheck delivery.
In an illustrative bench test with plain water, three 20-second runs deliver 9.8, 10.0 and 10.2 mL. That averages 10.0 mL or 0.50 mL per second for that setup. The result checks repeatability. It gives no chemical dose and does not establish delivery with a different liquid or installed head. The site needs its own approved verification.
Consider a 120 L operating volume with 30 L/min measured recirculation. Dividing volume by flow gives a nominal turnover of four minutes. A separate approved mixing test finds that samples from representative points first agree after seven minutes, so the pilot uses an eight-minute minimum hold and also requires stable readings. Check mixing at different levels and allowed flow rates. The time can change after filter, plumbing or return changes. Bluelab’s controller guidance explains why a dosing interval must allow the reservoir to mix.
Hold when measurements disagree
Start with one operator-approved correction inside limits set for the crop and equipment. Log the requested delivery, observed pump evidence and actual circulation. During the mixing hold, display the readings and suppress new requests. After the hold, compare fresh stable readings with the independent check required by the operating procedure. Escalate an unexpected response, a missing sample or a discrepancy outside the site’s agreed tolerance. Do not automatically retry a dose after a restart or a loss of delivery evidence.
In the illustrative pilot, an online pH reading of 6.4 at two minutes is too early to use. At the eight-minute check, it reads 6.1 while an independent sample at comparable temperature reads 6.4. The 0.3-unit difference exceeds the pilot’s example agreement limit of 0.15. Dosing stays inhibited while the operator checks sample flow, probe fouling, calibration and the second sample. These values illustrate a hold decision. They are neither crop targets nor universal tolerances.
Have a qualified controls and safety team approve per-action and cumulative limits, mutual exclusion between incompatible actions, a local stop and an independent means to inhibit chemical delivery. Assign recovery to a trained operator. Review chemical storage, handling, containment and response using product safety data sheets and the site’s applicable procedures. For U.S. workplaces, OSHA’s Hazard Communication resources cover labels and safety data sheets.
Pilot locally before adding control
A supported fertigation controller is a good starting point when it can enforce the site’s mixing hold, limits and alarms. An ESP32-S3 can add read-only telemetry and local event logging. For example, it can collect timestamped pH, EC, temperature, level, flow and fault states from compatible equipment over Modbus RTU through an isolated RS-485 interface. Confirm the register map, scaling, validity flags and supported topology with the equipment vendor. Espressif documents Modbus support. A network dashboard can report what happened while the approved local controller and independent inhibit retain authority over dosing.
Start on one reservoir with alerts and no control writes. Compare the trend to independent readings before and after a managed correction and across normal top-up and recirculation cycles. With the site’s approval, test stopped flow, low level, failed and stale sensors, empty pump supply, interrupted power and lost telemetry before enabling any control. Track agreement after mixing, repeat requests held for review, misses found during manual checks and the time it takes staff to resolve an alert. A useful pilot shows that the team can explain why a correction was accepted, why another was held and who decided it was safe to resume.