How to Diagnose a Platen Zone That Has a Slowly Drifting Temperature Calibration?

One crucial heating zone on a platen has been running the same product for a year. The parts are beginning to show a slow, minor change in quality, a softening of the material that indicates they are being somewhat overheated. But the display on the controller says the same precise setpoint it has always said. What it says about the temperature is a lie. The reality is gradually slipping away. The thermocouple, that little welded bead of two metals, is the controller's only window on the heat of the platen, and it is ageing silently. Its metallurgical structure is changing and its millivolt output for a particular temperature is steadily going out of calibration, silently.

Thermocouple Drift: An Insidious, Slow, Cumulative Error
Heated platen systems most often employ thermocouples as the temperature sensor. They are simple, robust and cheap. But they are not flawless. The thermoelectric characteristics of the two dissimilar metal wires vary with time due to:

Grain development in the metal crystal structure at high temperatures

Oxidation of the wire surfaces (particularly if the protective covering is damaged

Contamination from the surrounding environment (e.g. chemical vapours or moisture leaking into the thermocouple well)

Repeated temperature cycling causes a mechanical strain at the welded connection

Such ageing processes change the millivolt output of the thermocouple at a given actual temperature. This change is called the drift. The drift is usually slow-on the order of 0.1-0.5°C each month for Type K thermocouples at 200°C-but cumulative. After a year, reading can be down 2-5C or more.

The temperature calibration drift platen zone diagnosis problem is insidious in that the controller continues to report a stable set point. Operator notices no alarm. You can only tell that something is amiss by the quality of the product or by a separate verification measurement.

Why is Drift Often Temperature-Dependent
A thermocouple may read accurately at room temperature (where it is seldom used) yet drift dramatically at the normal operating temperature. This is because heat activates the mechanisms of ageing. Grain growth and oxidation increase dramatically with temperature. A thermocouple that drifts 1°C at 150°C might drift 5°C at 300°C. Thus calibration tests must be done at the real process temperature, not merely at ambient.

A Step-by-Step Diagnostic Procedure for Detecting Calibration Drift
The process below is executed during a scheduled maintenance outage. It needs a calibrated, traceable reference standard.

Necessary Equipment.
Traceable temperature reference: Precision resistance temperature detector (RTD) with calibrated readout or dry-block calibrator with built-in reference probe. The reference must be accurate to ±0.1°C or better and have a current calibration certificate traceable to national standards (e.g. NIST, UKAS).

Access tooling: A way to position the reference probe in the same location as the suspect thermocouple. This may require a separate test well or a temporary surface-mounted probe near the sensor for a platen containing embedded thermocouples.

Documentation forms: Log to record date, zone identity, setpoint, controller reading, reference reading and calculated offset.

Step 1: Calibrating the Platen
The platen is heated to its typical operating temperature (or equivalent setpoint). The process is halted, and the platen is allowed to settle. The controller must be in a stable state (no power oscillation, temperature stable to ±0.5°C for at least 10 minutes).

Step 2: Insert the Reference Probe:
The reference probe is positioned as close as feasible to the suspect thermocouple tip. If the platens have drilled thermocouple wells, the reference probe may be placed in a neighbouring well or in a well dug especially for calibration. If a well does not exist, a surface-mount RTD with thermally conductive paste is mounted at the surface of the platen above the embedded thermocouple.

Step 3: Take the Readings
After the reference probe has equilibrated (usually 2–5 min), the following data are recorded:

Suspect zone temperature as displayed on the controller (T_control)

reference probe temperature (Tref)

Setpoint platen ( T_set )

The difference (offset) is calculated as follows:Offset = T_control – T_ref

If the offset is positive, the controller reads hotter than the genuine temperature (the platen is actually cooler than displayed). A negative offset means the controller reads lower than the real temperature (the platen is actually hotter than what you see).

Step 4: Compare against Baseline/Specification
The measured offset is compared to: 1.

The baseline offset measured at the latest calibration (e.g. 6 months ago) A drift is substantial if the offset changes by greater than 0.5 °C.

The tolerance of the manufacturer (normally ±0.5°C for precision platens). If the offset is more than this tolerance the thermocouple is out of specification.

Step 5: Watch the Trend
The offset is recorded in the calibration history chart for that zone. A steady and continual increase of the offset over time is a sure indication of thermocouple ageing. For instance:

Date T_set (°C) T_control (°C) T_ref (°C) Offset (°C)
Jan 2025 180 180.2 179.8 +0.4
Jun 2025 180 180.5 179.2 +1.3 Dec 2025 180 181.0 178.0 +3.0
The ageing thermocouple is a slowly disappearing memory of actual temperature, its metallic voice slowly shifting pitch.

Interim Corrective Action: Controller Offset Correction
If the drift is below acceptable limits (e.g. < 2°C) and a scheduled replacement cannot be carried out immediately then a controller offset (also termed input trim or calibration offset) can be used. Most temperature controllers provide the ability to add a defined offset to the thermocouple measurement, or to subtract from it. The offset value is the inverse of the measured error. Such as:

Measured error: Controller reads 3°C too high (offset = +3°C)

Set controller offset: -3.0°C

When the offset has been applied, the controller reading should agree with the reference probe (within instrument limitations). But the offset is merely a temporary Band-Aid. It corrects the reading at one temperature, but it may not be accurate at other temperatures, since the drift can be non-linear. Furthermore, as the ageing process proceeds, the offset will have to be re-established from time to time.

Note: The offset adjustment should be documented and the access to this setting should be limited to authorised staff of the controller. The offset value and date of adjustment are recorded together with the calibration log.

Permanent Corrective Action: Replace Thermocouple
If a continuous drift is proven, the thermocouple should be scheduled for replacement. The sensor is a degrading asset . It will drift further till it goes out of bounds . The only cure is replacement.

Replacement of embedded platen thermocouples may include:

Taking the platen off the press

Removing the old thermocouple from the well

Insert a new calibrated thermocouple of identical kind

use of high temperature thermal paste for good thermal contact

Reconstruction and revalidation of the zone

It may be necessary to ship the platen to a qualified service center to replace the platens with thermocouples potted or welded in place. In such circumstances, this downtime should be included in a maintenance schedule.

Technical Accuracy: Which Thermocouple Types Are Most Susceptible to Drift?
Type MaterialsTypical Use Drift Susceptibility
Type K (Chromel-Alumel) High High – particularly susceptible to oxidation and "green rot" at high temperatures in reducing atmospheresGeneral use industrial platens up to 500 C
Type J (Iron-Constantan) Moderate Moderate – iron oxidises rapidly Lower-temperature applications (< 400°C)
Type T (Copper-Constantan) Low Low – copper is stable but limited to < 300°C Cryogenic and moderate temperature
N (Nicrosil-Nisil)Low Low– intended as a drift-resistant alternative for type KHigh temperature and critical processes
RTD (Pt100)Very low RTDs are much more stable with time but require a more complicated apparatusApplications that require long-term stability in precision
For important heated platen applications (e.g. semiconductor or medical device manufacturing) a Type N thermocouple or precision RTD is suggested over Type K to minimise drift. The type K is still widely used in existing platens because of affordability and availability. In such instances, regular calibration (e.g. every 3–6 months) is critical.

Why It's Important to Calibrate at Operating Temperature
A thermocouple may have a zero offset at 20°C (using an ice bath or ambient reference), yet drift considerably at 180°C. So the calibration check has to be done at the temperature of the process. For this purpose a portable dry-block calibrator is appropriate. This instrument consists of a heated metal block which has a well for the reference probe and another well for the suspicious thermocouple. The block is set to the target temperature and both sensors are placed side by side. You can see the difference straight from the data .

A dry-block may not be accessible and the platen itself can be utilised as the heat source with the reference probe placed adjacent to the thermocouple as specified in Step 2.

Calibration Schedule
The number of calibration tests varies on:

Type of thermocouple: Type K thermocouples require more regular calibration (every 3-6 months) than Type N or RTD (annual).

Operating temperature: Drift is faster at higher temperatures. If the platen runs above 200 °C, checkups every 3 months are advisable.

Process criticality: For processes producing safety-critical parts (such as medical devices, aerospace components), calibration should be checked before each production campaign.

A basic, documented calibration log for each zone will allow calculation of the drift rate. For example, if a zone drifts at a predictable 0.2°C per month, the operator knows that the thermocouple will go out of tolerance within a known time frame and replacement can be proactively scheduled.

Conclusion: Listening to the Silent Cry of the Ageing Sensor
Ageing temperature sensors generally have a sluggish, chronic process drift-their silent scream. Routine calibration detects it, and planned replacement is the treatment. While the controller display is steady, the platen zone may be making parts of inconsistent quality. The likely cause is not the heater or the process, it is the thermocouple slowly silently drifting out of calibration. The drift is due to metallurgical ageing (grain growth, oxidation, contamination) which changes the output of the sensor. The only viable defence is a rigorous calibration check against a traceable reference standard, carried out at the normal operating temperature. The drift is shown by a confirmed offset trend. The permanent solution is to replace the thermocouples but a temporary solution can be a controller offset adjustment.

The precision of a process is only as good as the last time the sensors were calibrated. In the heated platen, only the tiny bead of the thermocouple is the eye to see the temperature that makes every product. Only a traceable reference and a planned replacement can restore the truth when its testimony begins to float away.