Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC 61298 — Process Measurement and Control Devices: General Methods for Evaluating Performance
Key insight: IEC 61298 is the foundational metrology standard for process instrumentation. Every performance specification on a pressure transmitter, temperature sensor, or flow meter datasheet — accuracy, hysteresis, drift, temperature effect — traces its test methodology back to this standard. Understanding it is essential for writing defensible instrument specifications and interpreting manufacturer data correctly.
1. Scope and Structure of IEC 61298
IEC 61298 is a multi-part standard that defines uniform methods for evaluating the performance of process measurement and control devices. It applies to all analog and digital instruments that measure or control process variables including pressure, temperature, flow, level, pH, conductivity, and analytical parameters. The standard is organized in four parts:
- IEC 61298-1 (General considerations): Defines terminology, reference conditions, test uncertainty requirements, and the general framework for performance evaluation. Establishes the rules for test equipment accuracy — the reference standard must be at least four times more accurate than the device under test (DUT).
- IEC 61298-2 (Tests under reference conditions): Covers basic accuracy tests including dead band, hysteresis, repeatability, and linearity determination under controlled laboratory conditions (23 °C ± 2 °C, 50-75% RH, 96-106 kPa atmospheric pressure).
- IEC 61298-3 (Effect of influence quantities): Specifies test methods for determining the effects of ambient temperature, humidity, supply voltage variation, vibration, mounting position, and electromagnetic interference on instrument performance.
- IEC 61298-4 (Drift and long-term stability): Defines accelerated aging tests, long-term drift measurement over 30-90 day periods, and procedures for establishing recalibration intervals.
The standard’s philosophy is that performance claims are meaningless without a defined test method. A manufacturer claiming “0.075% accuracy” must specify under which part of IEC 61298 that accuracy was determined — including reference conditions, the number of test points, the number of test cycles, and the data reduction method (maximum deviation vs. root-sum-square).
Critical distinction: There is a profound difference between “accuracy” and “reference condition accuracy.” IEC 61298-2 accuracy applies only under strictly controlled laboratory conditions. The actual in-service accuracy is always worse — often 3-5 times worse — due to temperature effects, vibration, and long-term drift. Smart instrument manufacturers increasingly provide “total probable error” (TPE) calculations that combine the IEC 61298-2 reference accuracy with the influence effects per IEC 61298-3, giving the engineer a realistic estimate of field performance.
2. Key Test Methods and Performance Metrics
The standard defines several critical performance metrics that every instrumentation engineer should understand:
| Performance Metric | IEC 61298 Definition | Test Method | Typical Specification | Common Pitfall |
|---|---|---|---|---|
| Accuracy (Reference) | Maximum deviation from true value including hysteresis and repeatability | 5 up/down cycles, 11 equally spaced points, calculate max deviation as % of span | ±0.075% of span | Often confused with “calibrated accuracy” which uses fewer points |
| Hysteresis | Maximum difference between upscale and downscale readings at the same input | Full-scale up then down cycle; report max difference at any point | ±0.04% of span | Often embedded in accuracy spec; must be separately requested |
| Dead band | Range through which input can be varied without producing a detectable output change | Apply small input changes around a setpoint; find minimum change for response | ±0.005% of span | Critical for integrating control loops; not always tested |
| Repeatability | Closeness of agreement between successive measurements under same conditions | 10 repeated measurements at same input; report 3σ spread | ±0.02% of span | Short-term only; does not predict day-to-day variation |
| Temperature effect | Change in output per 10 °C ambient change outside reference range | Soak at -40, -20, 0, 23, 40, 60, 85 °C; measure zero and span shift | ±0.1% per 10 °C | Dominates error budget in outdoor installations |
| Long-term drift | Change in output over specified time under constant conditions | Measure at 0, 7, 14, 30, 60, 90 days; fit linear trend | ±0.1% per year | Not always included in datasheet; critical for custody transfer |
The test sequence specified in IEC 61298-2 follows a strict protocol:
- Pre-conditioning: Apply rated power for at least 30 minutes (for electronic instruments) or 2 hours (for smart/microprocessor-based instruments) before any measurements.
- Zero adjustment: Set the zero at the minimum input value with the output adjusted to the minimum rated value.
- Span adjustment: Set the span at the maximum input value with the output adjusted to the maximum rated value.
- Measurement cycles: Apply input at 0, 25, 50, 75, 100, 75, 50, 25, 0% (5 cycles minimum). Record output at each steady-state point after stabilization time (typically 30 seconds for pressure, 5 minutes for temperature).
- Data reduction: Calculate the deviation at each point, determine maximum positive and negative errors, and report accuracy as reference condition accuracy = (max positive deviation + |max negative deviation|) as a percentage of span.
Best practice for specification engineers: When writing instrument data sheets, always reference the specific part of IEC 61298 that applies to each performance claim. For example: “Accuracy: ±0.1% of span per IEC 61298-2” and “Temperature effect: ±0.05% per 10 °C per IEC 61298-3.” This unambiguous language prevents manufacturers from using different (and often more favourable) test methods to claim better performance. Also specify the “total probable error” (TPE) at normal operating conditions — this is the specification that truly matters for process control loop performance.
3. Engineering Design Insights
Uncertainty propagation in control loops: The accuracy of a process measurement per IEC 61298 is only the starting point. In a typical PID control loop, the measurement uncertainty propagates through the controller, the final control element, and the process itself. Using root-sum-square (RSS) combination of the transmitter accuracy, the DCS analog input card error, and the control valve positioner error yields a loop accuracy that is 2-3 times worse than the transmitter alone. For critical loops (e.g., reactor pressure control, custody transfer flow measurement), the total loop uncertainty must be verified against the process requirements, not just the transmitter datasheet.
Temperature effect budgeting: For outdoor installations, the temperature effect per IEC 61298-3 is often the dominant error source. Consider a pressure transmitter with ±0.075% reference accuracy and ±0.1% per 10 °C temperature effect, installed in a northern climate with an annual ambient temperature range of -30 °C to +40 °C (63 °C swing from the 23 °C reference). The temperature-induced error is ±0.63% — eight times the reference accuracy. Options for mitigation include: (a) selecting a transmitter with active temperature compensation (reduces effect to ±0.04% per 10 °C), (b) installing the transmitter in a temperature-controlled enclosure, or (c) using remote diaphragm seals to locate the electronics indoors.
Common engineering error: Assuming that “digital accuracy” or “smart sensor accuracy” eliminates the need for periodic recalibration. All process instruments drift over time regardless of technology. IEC 61298-4 provides the framework for determining recalibration intervals based on drift testing. A common mistake is to set annual recalibration intervals based on manufacturer recommendations without verifying that the instrument’s actual drift over one year is less than the process tolerance. Always demand drift test data from the manufacturer or perform your own 90-day drift test per IEC 61298-4 for critical applications.
Environmental influence testing for harsh applications: For instruments used in high-vibration environments (near compressors, pumps, or agitators), the vibration test in IEC 61298-3 specifies 10-150 Hz sweep at 1g (or 2g for severe service), with the instrument mounted in its normal orientation. Output variation during vibration must not exceed the specified tolerance — typically ±0.1% for general-purpose instruments. For steam sterilisation applications (biotech, pharmaceutical), the humidity test at 95% RH and 60 °C for 48 hours reveals condensation vulnerabilities that cause intermittent failures in the field.
Drift testing and recalibration optimization: IEC 61298-4 describes two approaches to drift assessment. Method A (accelerated) uses 168 hours at elevated temperature (85 °C for electronics, 125 °C for sensors) to predict 1-year drift. Method B (real-time) measures drift at reference conditions over 30, 60, and 90 days. For custody-transfer and safety-critical applications, Method B is strongly preferred despite the longer test duration, because accelerated aging does not always reproduce the same degradation mechanisms as real-time aging — particularly for thin-film strain gauge sensors where passivation layer migration is time-dependent rather than temperature-dependent.
4. Frequently Asked Questions
Q1: What is the difference between “accuracy” as defined in IEC 61298 and “calibrated accuracy”?
Per IEC 61298-2, reference condition accuracy includes hysteresis, dead band, and repeatability over 5 full-range cycles at 11 test points. “Calibrated accuracy” is a narrower term — it usually means the maximum deviation after a 5-point calibration (0-50-100-50-0%) at room temperature only. Calibrated accuracy is typically 2-3 times better than full IEC 61298-2 accuracy because it excludes hysteresis and uses fewer test points that may miss worst-case deviations. Always verify which definition a datasheet uses.
Q2: How should I determine the recalibration interval for a process transmitter?
Use IEC 61298-4 Method B: measure drift at 30, 60, and 90 days under reference conditions. Plot drift versus time and extrapolate to the point where drift equals the allowable error budget. Divide this time by a safety factor of 2-4 to establish the initial recalibration interval. After two years of field data, adjust the interval based on actual performance. For most industrial pressure and temperature transmitters with analog output, a 12-24 month interval is typical; for smart transmitters with digital compensation, 24-36 months may be achievable.
Q3: Does IEC 61298 apply to wireless and IoT process measurement devices?
Yes — the standard applies regardless of the communication protocol. However, wireless devices introduce additional performance considerations not covered by IEC 61298: (a) update rate affects the effective measurement bandwidth and response time; (b) battery voltage variation over the discharge cycle can affect measurement accuracy (addressed by testing per IEC 61298-3 with supply voltage variation); and (c) radio-frequency interference from the wireless transceiver itself can couple into the measurement circuitry. For wireless devices, supplement IEC 61298 testing with application-specific communication and power-cycle tests.
Q4: How do I compare performance specifications from different manufacturers?
First, verify that both manufacturers reference the same edition of IEC 61298 and the same test conditions. Differences to watch for: (a) test point count (5-point vs. 11-point), (b) number of cycles (1 vs. 5), (c) whether hysteresis is included in the accuracy spec or listed separately, (d) the temperature range over which the temperature effect is specified (per 10 °C vs. per 20 °C vs. full range), (e) data reduction method (maximum deviation vs. RSS vs. 2-sigma). Many manufacturers publish “best-case” specifications under favourable conditions. Request complete IEC 61298 test certificates — not just datasheet summaries — for critical applications.
