Bad Data Costs More Than No Data
A controller dosing to a wrong reading is worse than a farm with no controller at all, because the errors are systematic and invisible. Probes drift, junctions clog, buffers expire, and a reading that looks plausible gets trusted for weeks. Then the crop tells you, and by then the damage is in the tray.
Calibration is not a chore to be delegated and forgotten. On a commercial site it is a documented, scheduled, auditable process with tolerances, spares and a paper trail you can show an auditor or a buyer. This guide sets out what to calibrate, how often, how to do it properly, and how to build a routine that survives staff turnover.
What to Calibrate, and How Often
| Sensor | Commercial calibration interval | Verification between calibrations | Replacement life |
|---|---|---|---|
| pH probe | Every 2–4 weeks in recirculating systems | Weekly buffer check | 12–24 months, storage-dependent |
| EC probe | Every 4–8 weeks | Monthly standard solution check | 3–5 years for body; cell constant drifts with fouling |
| Dissolved oxygen (optical) | Every 3–6 months | Air-saturation check monthly | Cap life 1–2 years |
| Temperature sensor | Every 6–12 months | Ice-point or reference check quarterly | 5+ years |
| Flow meters | Annually | Bucket-and-stopwatch cross-check quarterly | 5–10 years |
| Pressure transmitters | Annually | Gauge comparison semi-annually | 5–10 years |
| Dosing pumps | Quarterly output verification | Weekly visual and draw-down check | Peristaltic tube life is the real interval |
Intervals tighten with conditions: high temperature, aggressive sanitisers, high EC and heavy biofouling all shorten the stable period. If a probe needs correction beyond tolerance at two consecutive calibrations, shorten the interval rather than accepting the correction.
pH: Where Most of the Drift Happens
- Two-point calibration minimum at pH 4.01 and 7.00 (or 6.86); add 10.01 if you run above 7
- Temperature compensation matters. Buffers and samples shift with temperature; calibrate at the process temperature or use an automatic temperature compensation probe correctly
- Never reuse buffer. Pour a fresh sample every time; a contaminated buffer invalidates the calibration and the contamination is invisible
- Check slope and offset, not just the reading. A probe reading 7.00 in buffer but with 70% slope is failing; most controllers will show this if you know where to look
- Store in storage solution, never in water or dry. A dry-stored glass bulb is a dead probe
- Clean by type: protein or biofilm with a cleaning solution, inorganic scale with dilute acid, oil with a mild detergent — never wipe the glass bulb, and never use an abrasive
EC: The Sensor People Forget Is a Physical Device
| Problem | Sign | Fix |
| Cell fouling | Reading drifts low, slow response | Clean cell, re-verify against standard |
| Bubble trapped in the cell | Erratic reading, sudden jumps | Remount at an angle, ensure flow past the cell |
| Wrong cell constant | Linear offset across the range | Re-enter constant from the certificate, verify with two standards |
| Temperature compensation error | Reading moves with temperature despite stable solution | Check the temperature sensor and the compensation setting |
| Cable or connector corrosion | Noisy, intermittent signal | Inspect connectors; replace cables as a consumable |
Verify with at least two standard solutions at different conductivities — typically one near your working setpoint and one well above it. A single-point check will not reveal a slope error.
DO: Optical vs Electrochemical
- Optical (luminescent) probes are the right default for continuous duty: no electrolyte, no membrane changes, minimal drift, low flow dependence
- Calibrate in water-saturated air at known barometric pressure, or use a zero-oxygen solution for a two-point check
- Caps age. Plan on a replacement cap every 12–24 months and keep spares; a cap failure silently flattens the reading
- Watch for coating. Biofilm or precipitate on the cap surface reads as low oxygen; clean on the same schedule as the probe body
- Electrochemical probes remain viable for spot checks but need membrane and electrolyte service, and drift within days of continuous immersion
Build the Routine So It Survives Staff Turnover
| Element | What good looks like |
|---|---|
| Written SOP | Step-by-step with photos, buffer lot numbers, tolerances and pass/fail criteria |
| Calibration log | Date, sensor ID, before/after reading, slope, technician, action taken |
| Spare inventory | At least one spare pH and EC probe per 10 installed, plus buffer stock with expiry tracking |
| Reference check | One handheld meter kept as a known-good reference, calibrated on a separate schedule |
| Alarm rules | Alarm on out-of-range, on rate-of-change, and on sensor fault — not just on the process value |
| Audit trail | Records retained for the period your certification scheme requires |
Cross-check sensor data against a physical reality at least weekly: measure EC and pH by handheld in the same sample the loop probe sees, measure flow with a bucket, and measure temperature with a calibrated thermometer. Digital data and physical truth diverge, and the gap is the useful signal. For the controller side, see our controller guide, and for the data layer, see the IoT monitoring guide.
FAQ
How often should pH probes be calibrated on a commercial farm?
Every two to four weeks in recirculating systems, with a weekly buffer verification between calibrations. Shorten the interval if the probe needs more than a small correction at two consecutive calibrations, or if it runs hot, high-EC or sanitised water.
Can I calibrate a pH probe with a single buffer?
It confirms the offset but hides slope error, which is the usual failure mode as a probe ages. Use at least two buffers bracketing your working range, and record slope as well as offset.
Why does my EC reading keep creeping up with no change in the solution?
Usually a fouled or partially blocked cell, a trapped bubble, or a wrong cell constant. Clean the cell, confirm the constant against two standards, and check that the cell has flowing solution past it rather than sitting in a dead leg.
Should I use optical or electrochemical DO probes?
Optical for anything continuously immersed: no electrolyte or membrane service, far less drift, and cap replacement every one to two years. Keep an electrochemical or handheld unit for spot verification.
What spares should a farm keep for sensors?
One spare pH probe per ten installed, a spare EC probe, at least one spare DO cap, buffer and standard solutions with expiry tracking, and one calibrated handheld meter kept purely as a reference instrument.
How do I know if the problem is the sensor or the solution?
Take a sample and measure it with the reference handheld meter. If the handheld agrees with the loop sensor, the solution changed; if it disagrees, calibrate or replace the sensor before touching the recipe.
Get a Calibration Routine That Holds Up in an Audit
Send your sensor list, system count and certification requirements through the quote form and we will return a calibration schedule with intervals, tolerances, a log template and a spares list matched to your installation.