Measurement & Control

Calibration

Definition

The process of confirming and, if necessary, adjusting the accuracy of a measurement device by comparing it with a known reference standard. For humidity sensors, salt solution reference environments or chilled mirror dew point meters are used as calibration references. Calibration is recommended every 6 months in the pharmaceutical sector and annually in general industrial applications.

Detailed Explanation

Calibration is the process of comparing the values reported by a measurement device against a known reference standard, and adjusting the device if necessary. Etymologically, the word comes from the Latin "calibre" (gun barrel diameter, precise measure) and is one of the fundamental concepts in metrology.

Calibration pyramid (traceability): 1. National/international standards (BIPM, NIST, NPL, TÜBİTAK UME) 2. Primary standards laboratory (national metrology institute) 3. Accredited calibration laboratory (TÜRKAK-accredited) 4. Facility-level reference instrument (e.g., chilled mirror dew point meter) 5. Working standard (field calibration setup) 6. Measurement device (field sensor)

Each level must be traceable to the previous one; this chain is called metrological traceability. A NIST-traceable calibration guarantees that the measured value is tied to the international SI unit system.

Types of calibration: • Verification — checks if the device is within tolerance; no changes made • Adjustment — the device value is adjusted to match the reference (offset, gain) • Certified calibration — results are documented and an accredited certificate is issued • Cross-check — comparison of two measurement devices under the same conditions

Calibration methods for humidity sensors: • Salt solution: saturated salt solutions produce constant RH (LiCl 11%, MgCl₂ 33%, NaCl 75%, K₂SO₄ 97% at 25°C) • Chilled mirror dew point reference: most accurate at ±0.1°C • Two-temperature method: controlled temperature + known vapor source • Climate chamber: special HVAC for reference environment generation

Calibration Calculation

Calibration error and correction:

Error = Device reading − Reference value

Linear adjustment (offset + slope correction): Corrected = a × Raw + b

a: slope (gain coefficient, ideal 1.000) b: offset (zero shift, ideal 0)

Calculating a and b from two-point calibration: a = (R₂ − R₁) / (M₂ − M₁) b = R₁ − a × M₁

R₁, R₂: reference values M₁, M₂: values measured by the device

Saturated salt solution RH values (at 25°C): • LiCl: 11.3 ± 0.3% RH • MgCl₂·6H₂O: 32.8 ± 0.2% • Mg(NO₃)₂·6H₂O: 52.9 ± 0.2% • NaCl: 75.3 ± 0.1% • KCl: 84.3 ± 0.3% • K₂SO₄: 97.3 ± 0.5%

Calibration uncertainty (uncertainty budget): U = k × √(u_ref² + u_method² + u_drift² + u_temp²)

u_ref: reference instrument uncertainty u_method: method uncertainty (application, readings) u_drift: drift between calibrations u_temp: temperature effect k: coverage factor (typically 2 = 95% confidence interval)

Practical Example

A 6-month periodic calibration of a pharmaceutical warehouse RH sensor:

Device: industrial-class RH/T sensor, installed 18 months ago Reference: salt solution kit (NIST-traceable, certified by an accredited lab) Conditions: 25°C, stable

Measurement: • Point 1 — LiCl solution (reference 11.3%): device reads 12.7% Error: +1.4% RH • Point 2 — NaCl solution (reference 75.3%): device reads 76.1% Error: +0.8% RH • Point 3 — K₂SO₄ solution (reference 97.3%): device reads 95.4% Error: −1.9% RH

Evaluation: • Sensor specification: ±1.5% RH (manufacturer) • Out-of-tolerance point: point 3 (97.3%) • Cause: polymer saturation in the high RH band, mild creep

Decision: adjustment. • Slope and offset calculated by linear regression • Offset and gain corrections entered through the device's configuration interface • Re-measurement: all points within ±0.5%

Documentation (GMP requirement): • Calibration certificate (date, measured values, adjusted values, uncertainty analysis) • Device serial number, calibration equipment serial number • Calibrator's name and signature, visible authority evidence • Next planned calibration date (6 months later) • All records archived for 5 years (FDA 21 CFR Part 11-compliant)

Cost structure: field calibration service is a mid-segment service fee per sensor. For multi-sensor facilities, an annual periodic maintenance contract is the most economical approach — a single GMP audit failure can result in a compliance penalty 10–20× the calibration service fee.

Engineering Note

Considerations in calibration management:

• Calibration interval decision — sector/application-based: – Pharmaceutical (GMP): 6 months – Lithium battery (IEC): 12 months – General industrial: 12–24 months – Comfort HVAC: 24–36 months – If drift between calibrations exceeds 50% of specification, shorten the interval • "Adjustment" mistake — some operators tend to adjust the device at every calibration; however, if the device is within tolerance, only verification should be performed. Continuous adjustment can corrupt device behavior. • Uncertainty reporting — a certified calibration must include uncertainty information; "measurement = 12.7% RH" is insufficient; "12.7% ± 0.3% (k=2)" is the correct format. • In-situ vs ex-situ calibration — field calibration (sensor at its installed location) reflects real-world conditions but has lower accuracy; lab calibration (sensor removed) provides higher accuracy but misses field effects. Ideal: a combination. • Calibration records — manual records are prone to human error; an automated record system (NKT BMS integration) shows calibration dates, drift trends, and upcoming planned actions for all sensors. • ISO 17025 accreditation — for critical processes, services should only be procured from TÜRKAK or equivalent accredited laboratories; certificates from non-accredited labs are invalid in audits. • Maintenance contract — including calibration service in equipment purchases (OPEX vs CAPEX) reduces total lifetime cost and provides compliance assurance.

Related Academy Articles

← Back to full glossary