A PT100 is a platinum resistance temperature sensor with a nominal resistance of 100 ohms at 0 °C. It does not connect directly to a web application. A compatible measurement input or transmitter reads the sensor, and a collection layer carries the result to software. Measurement quality depends on that entire chain.
The sensor is not the whole system
Consider range, sensor class, sheath, placement, environment and response requirements together. Air temperature and the internal temperature of a product are different measurements. Where the sensor is installed can matter more than the number of decimal places shown on a screen.
Confirm the input module’s PT100 support, wiring arrangements and conversion behaviour from its manufacturer documentation. Two decimal places do not establish accuracy to two decimal places. Resolution, accuracy, stability and total measurement uncertainty are distinct concepts.
Wiring and field conditions
Two-, three- and four-wire arrangements address lead resistance differently. The choice depends on the input module and installation. Three-wire compensation has assumptions that need to hold; four-wire measurement does not remove every problem caused by poor mounting or installation.
| Topic | What to verify |
|---|---|
| Wiring | Supported arrangement and connection quality |
| Input module | Sensor type, range and fault detection |
| Placement | Contact, airflow, immersion and response |
| Calibration | Reference, method, date and validity |
This is not an electrical installation procedure. Field wiring should follow the device manufacturer’s instructions and be performed by competent personnel. Software should translate device fault information into understandable quality states instead of treating faults as ordinary temperatures.
Illustrative measurement chain
In a cold-room example, a compatible module reads a PT100. The collector stores temperature, unit, source time and validity. The screen shows the latest measurement and its age. A disconnected wire becomes a sensor fault, not a zero-degree reading.
When maintenance replaces the sensor, preserve the equipment identity and calibration history. A replacement at the same location may still change observed behaviour. The replacement date should be available when interpreting the trend or comparing periods.
Verify both physical and software quality
Compare the source instrument display, raw collector record and normalised interface value during commissioning. This can reveal scaling and conversion errors. Physical measurement quality requires an appropriate reference and method; two arbitrary sensors agreeing is not by itself calibration.
Test sensor faults, out-of-range values and communication loss separately. If the last valid reading stays on screen, its stale status must be visible. Alarm logic should not treat invalid readings as normal measurements, and missing values must not be quietly interpolated without an explicit policy.
Questions before selection
What physical quantity and range are required? Which response time matters to the decision? How is placement validated? Who maintains the calibration record? What quality and fault information does the input device actually provide?
The BIPM guide to industrial platinum resistance thermometers provides a primary reference for measurement principles and uncertainty. A project’s device choice and installation still require the relevant manufacturer documents, actual field requirements and competent measurement assessment. Software visibility cannot automatically correct an unsuitable sensor location or an uncalibrated measurement chain.
Separate acceptance of the measurement from acceptance of the software
Consider an illustrative installation with one sensor, one input module and a web trend. The measurement specialist establishes an appropriate reference procedure for the required range. The software team separately traces the module’s output through conversion, storage and display. Agreement between the displayed value and the raw module value confirms that part of the data path; it does not establish the sensor’s calibration or the suitability of its position.
Keep the two sets of evidence together. Record the sensor identifier, input configuration, engineering unit and the applicable calibration reference or maintenance record. If a user later disputes a trend, this information helps distinguish a conversion error from a physical measurement issue. A timestamped screenshot without equipment and configuration identity is a much weaker record.
The acquisition device may offer diagnostic information for an open circuit, out-of-range reading or internal failure. Use the actual manufacturer’s definitions when mapping these states. A fault code is not a temperature and should not enter an average as if it were one. Test what the interface and alarm logic do when the device returns such a condition, then check recovery when a valid observation becomes available again.
Evaluate response in the intended location
A sensor in a protective assembly can respond differently from a bare element. Placement, thermal contact and the surrounding medium influence what a change in the recorded value means. The application should not promise an immediate picture of product temperature merely because it redraws the chart rapidly. Define the physical quantity the sensor represents and the decision the observation is intended to support.
An illustrative maintenance comparison might show a different trend after a probe is replaced or moved. Preserve the change date and location history before interpreting the difference as a process improvement. Reusing a display label is convenient, but it must not erase the distinction between the earlier and later measurement arrangement. Where a report spans both, provide enough context for the reviewer to recognise the change.
Before choosing hardware, collect the required range, installation environment, cable route, response requirement and available input-module documentation. A competent measurement assessment can then address compatibility and uncertainty. Orvun Systems can discuss how those validated measurements enter a traceable collection and monitoring flow; the software layer does not substitute for the physical assessment.