The 30-Second Answer
If you need accuracy better than ±1 °C, your process temperature is below 600 °C, and response time is not critical: use an RTD — almost always a PT100 in a 3- or 4-wire configuration.
If you need to measure above 850 °C, you need response under one second, lead-wire length is long, or the process is hostile to fragile elements (vibration, mechanical shock, transient extremes): use a thermocouple — typically Type K or Type N for general work, Type R/S/B for high-temperature precision.
The Engineering Trade-Off in One Line
RTDs are accurate, stable, and slow. Thermocouples are fast, robust, and forgiving — at the cost of accuracy and long-term drift.
Choose an RTD When…
• Range is < 600 °C
• Accuracy ≤ ±0.5 °C is needed
• The measurement is repeated over months/years
• You are validating a pharmaceutical, food, or aerospace process
• You can afford a slower (1–10 s) response
Choose a Thermocouple When…
• Range exceeds 850 °C
• You need response < 1 s
• Lead-wire runs are long (> 30 m)
• Environment is mechanically harsh (vibration, shock, thermal cycling)
• Budget constraints rule out 4-wire RTDs
Property-by-Property Breakdown
| Property | RTD (PT100 / PT1000) | Thermocouple (Type K typical) |
|---|---|---|
| Temperature Range | -200 to +850 °C | -270 to +1820 °C (type dependent) |
| Accuracy (typical) | ±0.1 to ±0.5 °C | ±1 to ±5 °C |
| Long-Term Stability | Excellent — < 0.04 °C/year at 200 °C | Good to fair — 1–4 °C drift per year at high temperature |
| Repeatability | ±0.05 °C typical | ±0.5 to ±2 °C typical |
| Response Time (63%) | Slower — 1 to 10 s typical (sheath dependent) | Faster — < 1 s ungrounded, < 0.1 s exposed-junction |
| Linearity | Highly linear (low-order polynomial) | Non-linear — table or polynomial lookup required |
| Output Signal | Resistance (Ω) — typically 100–400 Ω | Voltage (mV) — typically 0–70 mV |
| Sensitivity | ~0.385 Ω/°C (PT100), ~3.85 Ω/°C (PT1000) | ~41 µV/°C (Type K), ~76 µV/°C (Type E) |
| Excitation Required | Yes — constant current (typ. 0.1–1 mA) | No — passive (Seebeck effect) |
| Self-Heating | Yes — small, ~0.05 °C in still air at 1 mA | None — passive device |
| Lead Wire Effect | Significant — 4-wire eliminates, 3-wire compensates | Minor — but extension wire must match thermocouple type |
| Cold-Junction Compensation | Not required | Required — major error source if neglected |
| Element Cost (bare) | $15–80 (thin-film), $50–400 (wire-wound) | $2–10 (base metal), $80–400+ (noble metal) |
| Assembly Cost (typical) | $80–500 | $40–250 (base metal) |
| Best For | Precision, validated, long-life measurements | High-temp, fast-response, hostile environments |
Detailed RTD Advantages
1 · Accuracy
A Class A PT100 conforms to ±0.15 °C at 0 °C — three times tighter than the best (Class 1) Type K thermocouple. A 1/3 DIN element narrows that further to ±0.10 °C. With NIST-traceable calibration, individual RTDs are routinely accepted in metrology laboratories with uncertainties below ±0.05 °C.
2 · Stability Over Time
Platinum is a noble metal with a well-characterized resistance-temperature relationship that does not drift with use. A well-installed RTD operating below 200 °C will typically drift less than 0.04 °C in a year. Thermocouples, by contrast, experience grain growth, oxidation, and (in Type K) magnetic ordering effects that produce 1–4 °C drift per year at elevated temperatures.
3 · Linearity
The PT100 R(t) curve deviates from a straight line by less than 1 °C across most of its working range. Linearization is trivial — usually a single second-order polynomial. Thermocouples require multi-segment polynomial fits or look-up tables to convert millivolts to temperature with reasonable accuracy.
4 · Interchangeability
One Class A PT100 is interchangeable with another to within the tolerance band — no per-sensor characterization required. Two random Type K thermocouples may disagree by 1–2 °C at the same temperature.
5 · No Cold Junction
RTD electronics only need to measure resistance, not generate a temperature reference. This simplifies signal conditioning and removes the largest source of error in thermocouple work.
Pharmaceutical & FDA Validation
Pharmaceutical manufacturing under cGMP and 21 CFR Part 11 essentially demands RTDs for product temperature monitoring — the validation paperwork for thermocouples is prohibitive given their drift. Autoclave qualification, freeze-drying, and sterile fill-finish are RTD applications by default.
Aerospace Engine Inlet
Engine inlet total air temperature (TAT) and outside air temperature (OAT) measurements for flight control are exclusively RTD — the long-term repeatability requirement (less than ±0.2 °C over a 5-year overhaul interval) eliminates thermocouples.
Calibration Laboratories
Every calibration bath, every chamber, every transfer standard above the LN2 boiling point uses a PT100 or PT25.5 SPRT. Type S thermocouples appear only above 600 °C and as transfer standards in industrial heat treating.
Detailed Thermocouple Advantages
1 · Temperature Range
RTDs are useful to about 850 °C. A Type B thermocouple is useful to 1820 °C — more than double the range. Glass manufacturing, sapphire growth, semiconductor wafer processing, and ceramic firing are not RTD applications.
2 · Response Time
An exposed-junction Type K bead can respond in under 100 ms. Even a mineral-insulated grounded thermocouple typically reaches 63% of a step change in under 1 second. A typical RTD assembly, with its larger thermal mass, takes 2–10 seconds. For combustion monitoring, transient testing, and fast control loops, thermocouples win without competition.
3 · Cost
Base-metal thermocouple wire is cheap. Type K spool of 100 m of 24 AWG runs around $80 — equivalent PT100 sensors are an order of magnitude more. For multi-point monitoring on a power plant boiler with 200 measurement points, the TC cost advantage is decisive.
4 · Durability
A thermocouple is just two wires welded together. There is no brittle ceramic substrate, no fine platinum element, no anchoring solder joint to crack under thermal cycling. Mineral-insulated thermocouples survive immersion in molten metal, missile-engine exhaust, and re-entry vehicle heat shields — environments that destroy any RTD.
5 · Point Measurement
The sensing point of a thermocouple is the weld itself — small, well-defined, and easy to locate. RTDs sense over the active length of the element (5–30 mm), which is occasionally a problem in tight-tolerance assemblies or in steep temperature gradients.
6 · Lead Wire Length
Thermocouple signals are voltages, not resistances. The wire itself contributes only a small series resistance that the instrument ignores. RTDs lose the lead-resistance correction advantage as wire length grows, especially in 2-wire configurations.
Gas Turbine EGT
Exhaust gas temperature on a modern turbofan runs up to 900 °C steady-state with 1500 °C+ transients. Type K is the standard, with redundant probes feeding the FADEC. RTDs would not survive the first start cycle.
Furnace & Kiln Work
Glass and ceramic kilns at 1400–1600 °C use Type R or S noble-metal thermocouples in alumina protection tubes. Above 1700 °C Type B is the only standardized option.
Surface Temperature
Welded-foil "ribbon" thermocouples can be glued to a metal surface and respond in milliseconds — ideal for transient thermal testing of brakes, heat exchangers, and IC packages. No RTD geometry approaches this.
Recommendations by Industry
A condensed view of what we typically specify for common industries. Custom requirements always trump general rules — talk to applications engineering for borderline cases.
| Industry / Application | Recommended Sensor | Reasoning |
|---|---|---|
| Pharmaceutical autoclave | PT100 Class A, 4-wire | FDA validation, ±0.5 °C requirement, range ≤ 135 °C |
| Biotech freeze-dryer | PT100 Class A, 4-wire | Cryogenic stability, repeatable cycle validation |
| HVAC supply/return air | PT1000, 2-wire | Low cost, good accuracy, long lead wires |
| Gas turbine EGT | Type K, mineral-insulated | Range > 1000 °C, fast response, vibration tolerance |
| Reciprocating engine exhaust | Type K | Up to 1100 °C transient, low cost per channel |
| Glass-melt furnace | Type R or S | Range 1300–1600 °C, oxidizing atmosphere |
| Steel reheat furnace | Type N or Type S | Long service life at 1100–1300 °C |
| Distillation column tray | PT100, 3-wire | Below 400 °C, ±0.5 °C process control accuracy |
| Cryogenic LN₂ storage | PT100, 4-wire, ungrounded | -200 to -100 °C, repeatable, no CJC complication |
| Heat treat salt bath | Type K with Inconel sheath | 400–950 °C, fast response on parts |
| Aerospace engine inlet (TAT) | PT100 Class A | Long-term stability across maintenance intervals |
| Reactor jacket / batch chemistry | PT100, 3- or 4-wire | Sub-degree control, slow process dynamics |
| Combustion / flame zone | Type B or Type R | Above 1600 °C, noble metal stability |
| Surface temperature, transient | Type K or T ribbon | Milli-second response, no thermal mass |
| Bearing & motor winding | PT100 Class B | ±1 °C accuracy, vibration rated, embedded |
A Step-by-Step Selection Procedure
- 1
What is the maximum measurement temperature? If above 850 °C → thermocouple. Type K up to 1100 °C, Type N to 1300 °C, Type R/S to 1600 °C, Type B to 1800 °C.
- 2
What accuracy do you need? Better than ±1 °C absolute → RTD (PT100 Class A or 1/3 DIN). Coarser → thermocouple is fine.
- 3
How fast must the sensor respond? Need response under 1 second → thermocouple, ideally exposed or ungrounded mineral-insulated. Slower process dynamics → RTD fine.
- 4
How long are the lead wires? Longer than 30 m and budget-limited → thermocouple. Lab/short run → RTD 4-wire eliminates the issue entirely.
- 5
Will the measurement be validated or audited? Pharmaceutical, aerospace, nuclear, FDA → RTD with NIST-traceable calibration. Production process control with periodic check → either is acceptable.
- 6
What is the chemical environment? Sulfur, halogens, or hydrogen → avoid Type K, use Type N or RTD with Inconel/Hastelloy sheath. Vacuum or reducing → Type J or RTD. Neutral oxidizing → Type K or RTD.
- 7
What is the budget per measurement point? Under $50 per channel → thermocouple. Under $200 → either is feasible. Above $200 and accuracy critical → RTD with appropriate transmitter or precision DMM.
Ready to Spec
Your Sensor?
Whether you need a standard catalog configuration or a fully custom engineered assembly, our technical team responds to all quote requests within one business day.