How to Commission an RTD in a Thermowell
An RTD that bench-tests perfectly can still read wrong in service if the installation is careless: an air gap at the well tip, a half-immersed sensing element, or a wiring scheme that defeats lead compensation. Commissioning is the set of checks that turns a good sensor and a good thermowell into a good measurement. This page covers the mechanical fit, the insertion-depth reasoning, the wiring, the transmitter configuration, and the final verification that proves the point reads true.
Commission an RTD in a Thermowell in one line: To commission an RTD in a thermowell, verify the element reaches the bottom of the well with spring loading holding the tip in contact, confirm the well's immersion puts the sensing element fully into the flowing process, wire the sensor 3-wire or 4-wire so lead resistance is compensated, configure the transmitter for the correct sensor type, curve, and connection scheme, and verify the installed reading against a reference. Most installed-RTD errors trace to tip contact, immersion, or wiring, not to the element itself.
What You Need
Have the sensor and thermowell datasheets on hand so you can compare element length against well bore depth, plus the loop documentation showing sensor type and connection scheme, a multimeter, and the transmitter configuration tool. If the element has not already been proven, a quick resistance check against the IEC 60751 table - the procedure in how to bench-test an RTD - is cheap insurance before you bolt everything together.
If the well is already installed in live process, the well itself is the pressure boundary and swapping an element does not open the process; that is the whole point of a thermowell. But any work that disturbs the well or its flange on a live line is a site-procedure and permit matter, and the well's own specification - material, insertion length, wake-frequency review - is an engineering decision made before commissioning, not during it.
Fit, Tip Contact, and Insertion Depth
The element must reach the bottom of the well bore and stay in contact with it. Spring-loaded assemblies exist precisely for this: the spring compresses as the fitting seats, pressing the tip against the well bottom so heat conducts into the element instead of across an air gap. Check that the assembly actually compresses - the element should push back slightly before you tighten the fitting. An element that is short for the well, or a spring that never loads, leaves an insulating air pocket at the tip, and the sensor then reads a lazy, low version of the process temperature.
Immersion depth is the other half of the story. The sensing element occupies a length at the tip, and that whole length needs to sit in the flowing process, not in the nozzle or the pipe wall's thermal boundary layer. A well that is too short for the line leaves part of the element measuring metal and ambient instead of process, an effect called stem conduction error, which shows up as a reading biased toward ambient that worsens in cold weather. Follow the manufacturer's immersion recommendation for the well and service; if the reading tracks weather more than process, immersion is the first suspect.
Wire It So Lead Resistance Cancels
RTD accuracy depends on the input electronics knowing which resistance is element and which is copper wire. A 3-wire connection lets the transmitter measure and subtract the lead resistance, on the assumption that all three leads are equal; a 4-wire connection measures the element directly and removes lead effects entirely, which is why it is the choice where the input supports it. A 2-wire hookup adds the full loop resistance of both leads straight onto the element and should only ever appear on very short runs where the error is genuinely negligible.
During commissioning, land the conductors per the sensor's wiring diagram - the commoned pair on one side of the element matters - and verify with the meter that the two commoned leads read near-equal resistance to each other. Terminate cleanly: a corroded or loose terminal adds resistance that a 3-wire scheme cannot fully cancel and that wanders with temperature and vibration, producing the kind of slow flicker that gets misdiagnosed as a noisy element.
Configure the Transmitter to Match
The transmitter must be told exactly what is connected: the sensor type and curve, per IEC 60751 for standard platinum elements, and the connection scheme, since 2-, 3-, and 4-wire inputs are handled differently. A mismatch here produces a plausible-looking reading with a built-in bias, which is far more dangerous than an obviously broken one. Set the range, units, damping, and the burnout direction - whether the output drives upscale or downscale on sensor failure - to match the loop sheet and the alarm philosophy, so a failed element trips something rather than parking at a believable value.
If the transmitter supports sensor matching with Callendar-Van Dusen coefficients from the element's calibration certificate, commissioning is the moment to load them, because doing it later invalidates whatever loop checks were done in between. Record the final configuration in the loop documentation; the next technician should be able to reconstruct the setup without opening the head.
Verifying the Result and Common Mistakes
Prove the installed measurement, not just the parts. Compare the commissioned point against a reference: a calibrated portable probe in an adjacent well, a process temperature known from operating conditions, or a redundant installed sensor. Then watch the trend for the first days in service; a healthy point tracks process moves promptly and does not follow the weather. A monitoring platform such as Merobix makes that check easy by trending the new point alongside its neighbors, where a stem-conduction bias or a loose-terminal flicker stands out immediately.
The recurring mistakes: an element that never touches the well bottom, spring loading defeated by a wrong-length sensor, a well too short for the line, 2-wire connection on a long run, mismatched or corroded lead terminations, the wrong curve or wire-count configured in the transmitter, and burnout direction set opposite to what the alarm scheme assumes. None of these require exotic tools to catch - they require the commissioning checks in this list actually being done.
Frequently Asked Questions
Why does my RTD read low and sluggish after installation in a thermowell?
The usual cause is poor thermal contact at the well tip: the element is short for the well or the spring loading never compressed, leaving an air gap that insulates the sensing element. The other frequent cause is insufficient immersion, where part of the sensing element sits in the nozzle or wall boundary layer and conducts ambient temperature into the reading. Check tip contact and immersion before suspecting the element, which likely bench-tests fine.
Should I wire an installed RTD 3-wire or 4-wire?
Use 4-wire wherever the transmitter or input card supports it, because it measures the element directly and eliminates lead-resistance error including mismatch between leads. Use 3-wire where the input only supports that; it subtracts lead resistance on the assumption all leads are equal, which is usually close enough with matched conductors in one cable. Reserve 2-wire for very short runs, since it adds both leads' full resistance to the element.
Do I need to recalibrate an RTD after installing it in a thermowell?
Installation does not change the element's resistance curve, so a bench-proven element does not need recalibrating just because it was inserted. What installation does change is the thermal path, so the commissioning verification compares the installed reading against a reference to catch immersion and contact problems. If the transmitter is configured with sensor-matching coefficients or a new range, verify the loop end to end after those changes.
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.