How to Commission a Thermocouple Loop with CJC Verification
A thermocouple loop is commissioned, not just calibrated: the sensor, the extension wiring, the transmitter configuration, and the cold-junction compensation all have to agree before the number on the operator screen means anything. This walkthrough is for the technician putting a new or replaced thermocouple into service, and it treats the CJC check as part of the job rather than an afterthought, because a reference-junction error offsets every reading the loop will ever produce.
Commission a Thermocouple Loop in one line: To commission a thermocouple loop, confirm the sensor and extension wire are the same thermocouple type with correct polarity at every junction point, configure the transmitter for that type with the intended burnout direction, verify the cold-junction compensation against a known reference temperature, and prove the value lands correctly in the control system. The CJC step matters because the instrument adds its own terminal temperature to the measured millivolts, so a CJC error shifts every reading by the same amount.
What You Need
Have the loop documentation showing the sensor type and range, the sensor and transmitter datasheets, a reference thermometer you trust, a multimeter, and the transmitter configuration tool. Confirm from the datasheet whether the junction is grounded, ungrounded, or exposed - the differences are covered in thermocouple junction types - because a grounded junction changes how you handle shield grounding later.
Confirm Type and Polarity End to End
A thermocouple circuit only reads correctly if the same pair of alloys runs unbroken from the measuring junction to the transmitter terminals. Check the sensor type stamped on the head against the loop sheet, then walk every junction box between the head and the transmitter confirming the extension or compensating cable matches the sensor type and that polarity is preserved at each landing. The full identification procedure, including the magnet check and the heat test, is in verifying thermocouple polarity and extension wire.
Any point where the alloy pair changes - a copper jumper, a reversed terminal, a type J leg spliced into a type K run - creates an unintended junction that adds or subtracts millivolts according to the temperature at that spot. The cruel part is that the loop still produces a plausible number, and the error changes as the junction-box temperature changes through the day, which is why wiring faults get mistaken for process instability.
Configure the Transmitter and Burnout Direction
In the transmitter's configuration menu, select the exact sensor type so the right voltage-to-temperature curve is applied, set the range and units the control system expects, and set a modest damping value only if the point genuinely needs it. Configure the failure behavior deliberately: thermocouple burnout detection should drive the output in whichever direction is safe for this specific loop, which is a decision about the process, not a default to accept blindly.
Leave the cold-junction compensation source set to the transmitter's internal terminal sensor unless the site has a documented reason to use a fixed or remote reference. A fixed CJC value is only correct while the terminals actually sit at that temperature, which is rarely true in a field enclosure that bakes in the afternoon and cools overnight.
Verify the Cold-Junction Compensation in Place
A thermocouple only measures the difference between its hot junction and the instrument terminals; the transmitter measures the terminal temperature and adds it back, which is the whole story of cold-junction compensation. To verify it, hold the sensor tip at a temperature you know - alongside your reference thermometer in ambient air before insertion works well - and confirm the loop reads that temperature closely. Because a type K thermocouple produces roughly 40 microvolts per degree Celsius, a CJC error of five degrees shifts the reading about five degrees at any process temperature.
A second, quick confidence check: gently warm the transmitter housing or terminal area by a few degrees and watch the reading. A healthy loop holds essentially steady because the compensation tracks the terminal change; a loop with a failed or fixed-value CJC follows the terminal temperature instead of the process.
Verifying the Result and Common Mistakes
Finish by proving the whole chain: insert the sensor, confirm the local reading is plausible against process conditions, then confirm the same value, units, and range appear in the control system and historian. If you want a bench-grade check of the transmitter electronics themselves, the millivolt-injection procedure in calibrating a thermocouple and checking CJC complements this in-place commissioning.
The mistakes that come back as work orders: copper wire run between the head and the transmitter, one reversed pair in a junction box that makes the reading wander with the weather, the wrong sensor type selected in the transmitter so the curve is subtly wrong across the range, and trimming the transmitter to hide a wiring fault instead of finding it. If you find yourself adding an offset larger than a degree or two to make commissioning numbers agree, stop and find the physical cause.
Frequently Asked Questions
How can I verify CJC without an ice bath?
Hold the sensor tip at ambient next to a trusted reference thermometer and confirm the loop reads the same temperature; the transmitter's CJC is doing all the work in that comparison because the hot junction and terminals are near the same temperature. Then warm the terminal area slightly and confirm the reading stays put. An ice bath adds precision but is not required to prove the compensation is alive and roughly right.
Why does my new thermocouple read fine at ambient but wrong at process temperature?
That pattern points at the curve or the wiring rather than the CJC. If the transmitter is configured for the wrong thermocouple type, the error grows as the millivolt signal grows, so it hides at ambient and appears at temperature. A mismatched extension-wire segment does the same because its junctions sit in a temperature gradient once the process heats the run. Recheck the configured type first, then walk the wiring.
Should burnout drive the reading upscale or downscale?
Whichever direction moves the process toward safety for that specific loop. On a heater control loop, upscale burnout makes the failed sensor look hot so the controller cuts heat input; on a loop where a low reading trips protection, downscale may be right. It is a process-safety decision to make with the loop owner, then document and test during commissioning.
Automation services
Need help turning this into a working system?
Merobix integrates SCADA, programs Allen-Bradley and Siemens PLCs, and designs and fabricates industrial control panels.
Meeting requests are reviewed before confirmation.