How to Calibrate a Temperature Sensor: Practical Guide for Singapore Labs and Industry

Temperature sensor calibration verifies that your sensor is accurately reporting the true temperature of the process or environment it monitors — and documents any deviation from the true value so that corrections can be applied or out-of-tolerance sensors can be adjusted or replaced. In Singapore's pharmaceutical, food, calibration laboratory, and industrial sectors, regular temperature sensor calibration with traceable documentation is a regulatory and quality requirement. This guide covers the practical procedure using a portable Fluke dry block calibrator.

Understanding Temperature Traceability

When you calibrate a temperature sensor, you are comparing it against a reference that has itself been calibrated against a higher-order standard, which was in turn calibrated against the national primary standard. In Singapore, the primary temperature standard is maintained at A*STAR's National Metrology Centre (NMC), traceable to the International Temperature Scale ITS-90. The traceability chain: your sensor → Unitest reference probe (SINGLAS certificate) → NMC → BIPM. Every link in this chain must have documented uncertainty — this is what makes the calibration "traceable."

Equipment Required

• Fluke 9142 or 9144 dry block calibrator (temperature source) — with current calibration certificate showing traceability
• Reference platinum resistance thermometer (PRT) — calibrated with SINGLAS certificate
• Readout instrument for the reference PRT (e.g., Fluke 1523 or 1524 thermometer readout)
• Sensor under calibration (UUC — unit under calibration): thermocouple, RTD, or thermometer
• Readout instrument for the UUC (the instrument it normally connects to, or a reference readout)
• Calibration worksheet (for recording all readings)
• Appropriate insert for the dry block (matching the sensor probe diameter)

Step-by-Step Calibration Procedure

Step 1: Set Up and Stabilise

Power on the dry block calibrator and set to the first calibration temperature. Allow minimum 15–20 minutes to stabilise — the calibrator display will indicate when stability is achieved (some Fluke models flash a "READY" indicator). Connect the reference PRT and the UUC to their respective readout instruments. Record ambient temperature and humidity in the calibration worksheet.

Step 2: Prepare the Inserts

Select the correct insert block for the calibrator, matching the probe outer diameter of both the reference PRT and the UUC. Both probes should be inserted to the same immersion depth — at least 15× the probe outer diameter (e.g., 6 mm diameter probe = 90 mm minimum immersion) to ensure the sensor tip is at the calibration block temperature, not influenced by the cooler ambient air above the insert.

Step 3: Record Calibration Points

For pharmaceutical cold chain sensors (2–8°C range), typical calibration points are: 0°C, 5°C (midrange), 10°C. For room temperature storage sensors (15–25°C), use: 15°C, 20°C, 25°C. At each calibration point: wait for the reference PRT reading to stabilise (< 0.05°C variation over 2 minutes), then record five readings of both the reference PRT and the UUC, taken at 30-second intervals. Average each set of five readings.

Step 4: Calculate Error

Error = UUC average reading – Reference PRT average reading at each calibration point.

Example: Reference PRT reads 5.02°C; UUC reads 5.18°C → Error = +0.16°C (UUC reads high by 0.16°C).

Step 5: Determine Compliance

Compare the |Error| at each calibration point against the UUC's accuracy specification (from the manufacturer datasheet). If |Error| ≤ specified tolerance, the sensor is IN TOLERANCE (pass). If |Error| > tolerance, the sensor is OUT OF TOLERANCE (fail) and must be adjusted (if adjustable) or replaced.

For Singapore pharmaceutical applications (typical tolerance ±0.5°C), a Rotronic logger reading ±0.16°C error would PASS. A Rotronic logger reading ±0.62°C error would FAIL and require adjustment or replacement.

Step 6: Document

Complete the calibration certificate with: UUC identification (model, serial number, asset tag), calibration instrument identification and certificate number, date and operator, ambient conditions, calibration points, reference readings, UUC readings, errors, tolerance, and pass/fail determination. The certificate is signed by the calibration technician and (for SINGLAS-accredited calibrations) reviewed and approved by the designated signatory.

Calculating Measurement Uncertainty

For formal calibration certificates, measurement uncertainty must be calculated and reported. The main uncertainty sources for dry block temperature calibration include:

• Reference PRT calibration uncertainty (from its SINGLAS certificate)
• Dry block uniformity (temperature variation across the insert holes — from calibrator specification or characterisation)
• Dry block stability (temperature variation at one point during measurement)
• Readout instrument resolution and uncertainty
• Repeatability of UUC readings (from the standard deviation of the five repeated readings)

These are combined as root-sum-of-squares (RSS) of standard uncertainties and multiplied by k=2 to give the expanded uncertainty at 95% confidence. The expanded uncertainty must be smaller than the UUC tolerance / 4 for a 4:1 test uncertainty ratio.