The outlet of a shell-and-tube or plate heat exchanger is one of the worst places on a plant to put an undersized temperature instrument: the fluid is often two-phase, the line vibrates, and the reading drives the temperature control loop that protects the next unit downstream. A practical selection map starts with four independent decisions, probe type, sheath/well material, response time, and recorder class, all anchored to the actual outlet temperature range [S4].
For a steam-heated exchanger heating process water to 90 to 150 degrees C, a 4-wire Pt100 Class A element inside a stainless 316L thermowell, paired with a paperless recorder sampling at 1 second intervals, is the common baseline.
Probe Selection: RTD vs Thermocouple at the Outlet
Pt100 or Pt1000 RTDs give typical accuracy of plus/minus 0.15 degrees C at 0 degrees C for Class A elements and plus/minus 0.3 degrees C for Class B, which is roughly an order of magnitude tighter than a standard Type K thermocouple at the same point [S4]. Type K wire (chromel/alumel) covers minus 200 to plus 1260 degrees C in its published range, while Type N (nicrosil/nisil) trades a slightly narrower minus 200 to plus 1140 degrees C window for better oxidation drift in the 600 to 1100 degrees C band, which matters on fired-heater and reformer outlets.
The decision is mechanical as much as electrical. RTDs win on accuracy, long-term drift, and repeatability, and they are the default where the outlet temperature feeds a regulatory or quality control loop rather than a thermal-limit cutout [S2].
Sheath, Thermowell, and Process Interface
The sheath or thermowell is the only thing standing between the sensing element and the process, so the material grade has to match the fluid, not the temperature. Stainless 316L is the workhorse for water, glycol, mild chemicals, and steam condensate up to about 600 degrees C. For sour service, the well has to meet NACE MR0175 limits on hardness and alloy composition, otherwise stress-corrosion cracking will eat the well before the recorder notices. For acidic or chloride-bearing streams, Hastelloy C-276 or Inconel 600 thermowells are common. [S2]
Thermowell wake frequency is a separate check. ASME PTC 19.3 TW is the standard method for verifying that the well's natural frequency is safely above any flow-induced vibration, and most plants require the calculated frequency ratio to stay under about 0.8. On the recorder side, an atEX/IECEx certified head transmitter inside the connection head is now common on new builds, so the 4-20 mA or HART signal leaving the thermowell already carries a digital layer for the paperless recorder to read [S4].
Response Time, Sampling, and Recorder Class

Time constant is the spec that most projects under-spec. A bare 6 mm Pt100 in a thermowell typically lands at 7 to 15 seconds for a 63 percent step (tau), while a 3 mm mineral-insulated Type K can drop to 1.5 to 3 seconds. If the outlet loop controller is running a 5 to 10 second scan, a 15 second probe means the recorder is always showing a temperature that is 30 to 90 seconds behind reality, which is enough to mask a fouling event. [S4]
Recorder class is set by the application, not by the channel count. A paperless recorder with 8 to 16 universal inputs, 1 second base sample, on-board 4 GB storage, and Modbus TCP or Ethernet/IP export covers most exchanger-outlet duties. For FDA- or GMP-regulated plants, the recorder needs 21 CFR Part 11 compliant audit trail, electronic signature, and local archiving, which is a software feature, not a probe feature. For batch records on heat exchanger outlet temperature during pasteurization or sterilization, a 0.1 degrees C resolution recorder with redundant storage is a common floor [S2].
Hazardous Area, Enclosure, and Signal Integrity
Exchanger outlets in oil and gas, petrochemical, and many pharmaceutical plants sit inside classified areas, so the recorder and its associated head transmitter must carry the right protection level. ATEX category 2 (zone 1) or IECEx equivalent is the usual baseline for an outdoor hydrocarbon unit, with category 1 (zone 0) reserved for the inside of a process vessel or tank, which is not where an outlet temperature instrument is normally mounted [S4].
Enclosure rating is the second environmental gate. IP66 stainless steel or cast aluminum is the default for outdoor exchanger skids, and NEMA 4X is the common North American equivalent. On offshore or washdown-heavy sites, the recorder cabinet itself needs to be specified to the same rating or higher, and the cable entries should be Ex d or Ex e certified glands rather than standard cable glands. Surge protection on the 4-20 mA loop is cheap insurance on long cable runs from the exchanger outlet to the control room.
Comparison: Four Realistic Outlets and the Right Recorder

The cleanest way to shortlist is to line up four typical outlet conditions against the four decision criteria above. A district-heating plate exchanger with 110 degrees C water on the secondary side takes a 4-wire Pt100 Class A in a 316L thermowell, a 10 second time constant is fine, and a 4-channel temperature recorder with 1 second logging is more than enough. An ammonia refrigerant evaporator outlet around minus 30 degrees C takes a Pt100 or Pt1000 element (Type T or K thermocouple also works), still 316L well, but the recorder must accept negative ranges and the well must be sealed against moisture ingress [S4].
A refinery crude preheat train outlet at 280 to 350 degrees C is the boundary case: Pt100 with a high-temperature mineral-insulated insert is still acceptable, but a Type J or Type K thermocouple in a 316L well is more common for cost and vibration reasons.
Who Should NOT Pick the Mainstream RTD Option
There are three situations where the default Pt100 in a 316L well is the wrong call. The first is any exchanger outlet that sees two-phase flow at high pressure drop, where steam or vapor pockets drive the local temperature reading up and down: a faster-response thermocouple, or a thermowell with a thermoconductive paste filling, will track better than a thick-wall RTD well. The second is a hydrogen-rich atmosphere, where hydrogen can diffuse into the sheath and embrittle a standard 316L well; an Inconel or Hastelloy sheath, or a vented sheath design, is the fix. [S2]
The third is a hygienic or sanitary exchanger in a dairy, beverage, or pharma skid, where the well has to be tri-clamp or aseptic-flanged, surface finish Ra below 0.8 micrometers, and CIP/SIP compatible up to 140 degrees C with caustic and acid cleaning. In that case the recorder and probe have to be qualified as a system, including the temperature and humidity data recorder if ambient compensation is needed, and the calibration certificate has to trace back to a national lab. Picking a standard industrial Pt100 here is a documentation problem waiting to happen.
Sourcing, Standards, and Validation Trail

Traceability is what turns a recorder from a box on a desk into a defensible measurement. Every probe should ship with a calibration certificate stating the actual resistance or millivolt table at three or more points (0, 50, and 100 degrees C for a Pt100, or 0, 500, and 1000 degrees C for a high-temperature thermocouple), and the recorder's analog input should be re-checked annually against a dry-block calibrator with a reference PRT. A useful cross-reference is how dry block calibrator selection for steam sterilization lines lays out the same probe-vs-loop thinking for a different process duty. [S2]
On the standards side, IEC 60751 governs the platinum resistance relationship for Pt100/Pt1000 elements, ASTM E230 covers thermocouple tables and limits of error, and ASME PTC 19.3 TW is the reference for thermowell mechanical design. For hazardous areas, ATEX 2014/34/EU and the IEC 60079 series set the equipment-level rules. Recorder software for regulated industries has to meet 21 CFR Part 11 for electronic records and signatures. None of those standards is optional on a heat exchanger outlet that feeds a safety or quality loop [S4].
Shortlist logic in one line: pick the probe type from the outlet temperature, pick the sheath from the fluid chemistry, pick the time constant from the control loop scan time, and pick the recorder from the data integrity and hazardous-area rules. Two trackable signals on the next review: any move by a major recorder OEM to publish higher-density thermocouple input modules (32 to 64 channels per unit), and any plant-side incident report that ties a fouling event to recorder response time rather than to the heat exchanger itself.