Start by selecting a platinum-based resistive sensor with a reference resistance of 100 Ω at 0°C. This choice ensures linear response up to 850°C, reducing calibration drift across industrial ranges. Connect the sensor in a four-wire configuration–power leads and sense leads must be separared at the connector to eliminate lead resistance errors, which can exceed 0.5°C in 2 km cable runs. For signal conditioning, use a 1 mA constant current source; anything above risks self-heating (>0.2°C), while below 0.5 mA degrades signal-to-noise ratio to less than 60 dB.
Apply a precision instrumentation amplifier with a gain of 100 to amplify the millivolt-level sensor output. Select an op-amp with input offset voltage below 10 µV and input bias current under 1 nA–OPA188 or equivalent– to maintain accuracy within ±0.1°C over 0–200°C. Include a low-pass filter with a cutoff frequency of 10 Hz; standard 60 Hz mains interference can introduce 0.3°C peak-to-peak ripple without filtering.
Convert the amplified signal using a 24-bit ADC running at 10 samples per second. Ensure the ADC operates in ratiometric mode, referencing the same excitation current to eliminate gain errors. For noise immunity, shield all conductive paths with grounded copper foil, reducing capacitance-coupled interference by 80%. Ground the shield only at the amplifier terminal; dual end-grounding creates ground loops. Use twisted pair for all signal wires–24 AWG minimum– to reduce magnetic pickup by 95% over single conductors.
Implement cold junction compensation only if thermocouples coexist; for pure resistive sensing, ambient temperature changes below 0.05°C per minute can be ignored. Store calibration coefficients in EEPROM, recalibrating the sensor every 500 hours of operation–drift typically reaches 0.3°C after 1,000 hours at 650°C. Deploy firmware linearization algorithms but evaluate each segment–third-order polynomials introduce nonlinear residuals up to 0.02°C compared to lookup tables.
Practical Guide to Resistance Temperature Sensor Circuit Layouts
Start with a 4-wire configuration for precision measurements above 0.1 °C; parasitic lead resistances cancel out, leaving only the sensing element’s true resistance.
Place the excitation current source as close as possible to the sensor terminals–ideally within 30 mm–to minimize voltage drops across copper traces at 2 mA excitation.
Use a differential amplifier with input impedance above 10 MΩ to prevent loading errors; a common-mode rejection ratio of 90 dB filters mains-frequency interference typical in industrial environments.
Ground reference should tie to the analog common near the amplifier, not the sensor, to avoid ground loops; star grounding reduces noise pickup on the signal path.
Add a 100 nF ceramic capacitor directly across amplifier inputs to suppress high-frequency spikes; position it within 5 mm of the op-amp pins to maintain stability during transient events.
Calibrate the circuit at two fixed points–0 °C (ice bath) and 100 °C (boiling water)–recording offset and gain adjustments digitally; verify linearity between these points with a secondary standard (±0.05 °C) to detect drift.
Choosing the Right Resistance Thermometer Configuration for Precise Temperature Measurement
For industrial applications requiring ±0.1°C accuracy, a 3-wire platinum element with a 100Ω Class A sensor is the minimal viable configuration. Copper or nickel variants introduce nonlinearity above 150°C, while platinum maintains linear response to 850°C with a temperature coefficient of 0.00385 Ω/Ω/°C. The third wire compensates for lead resistance up to 10Ω, critical in environments where cable lengths exceed 30 meters.
For cryogenic measurements (−200°C to 0°C), a 4-wire setup eliminates parasitic resistance errors entirely, offering resolution better than 0.01°C. Silver-soldered connections degrade below −70°C, so welded or gold-plated terminations are mandatory. A 1000Ω platinum element increases sensitivity tenfold over 100Ω models, reducing the impact of noise in high-impedance circuits–ideal for superconducting experiments or liquefied gas monitoring.
Signal Conditioning Requirements
Bridge circuits with precision resistors (0.01% tolerance) are necessary when measuring microvolt-level signals. A 24-bit ADC with programmable gain (PGA) extends dynamic range, but watch for offset drift: 0.1 μV/°C in bare-die amplifiers. For DIN IEC 60751 compliance, excitation current must not exceed 1 mA to prevent self-heating errors–0.5 mA is optimal for 100Ω elements. Differential wiring shields reduce common-mode noise by 80% compared to single-ended setups, especially near variable-frequency drives.
In high-vibration environments, thin-film sensors outlast wire-wound types by 3:1, but their 1 mA excitation limit demands current sources with flat-pack design with grounded foil reduces thermal lag by 60% versus glass-encapsulated versions. Avoid adhesive bonding if thermal gradients exceed 5°C/mm–mechanical clamping with thermal grease ensures uniform contact.
When selecting between 2-, 3-, or 4-lead arrangements, model lead resistance first: 0.4Ω/m for 20 AWG copper. A 5-meter cable adds 2Ω–enough to shift a 100Ω Class A reading by 5°C without compensation. For explosive atmospheres, intrinsic safety barriers impose 1 W maximum power dissipation; a 2-wire 10 mA loop with a 120Ω resistor suffices, but resolution drops to 0.5°C. Always cross-verify with a handheld calibrator at −25°C, 0°C, and 100°C–the three-point check catches 90% of linearity errors.
Step-by-Step Wiring of 2-Wire, 3-Wire, and 4-Wire Resistance Temperature Sensor Circuits
Start with proper cable selection for 2-wire setups. Use shielded twisted pair wires with a gauge between 22 AWG and 18 AWG, depending on distance. Keep lead lengths under 10 meters to minimize resistance errors. Connect the sensor directly to the measurement instrument without intermediate splices, as each junction adds 0.1°C–0.5°C of potential error. Terminate the shield at the instrument end only–never at both ends–to avoid ground loops.
For 3-wire configurations: Identify the matched resistance wires–typically color-coded red, red, and white. Wire one red lead to the excitation source, the second red lead to the high-input terminal, and the white to the low-input terminal. Ensure the instrument supports 3-wire compensation; if not, measurement errors will exceed ±0.5°C. Use a precision multimeter to verify lead resistances differ by less than 0.1Ω before finalizing connections.
Four-wire setups demand strict adherence to polarity. Connect the outer pair to the current source, ensuring 1 mA ±0.1% excitation. Route the inner pair to the voltage measurement terminals–these leads must carry no current. Avoid mixing gauge sizes; even a 0.01Ω difference introduces detectable error. For distances over 20 meters, switch to 20 AWG wires with individual shields per pair to reduce crosstalk.
Strip and crimp properly: Expose no more than 2 mm of conductor; excess bare wire skews readings. Use gold-plated ferrules for copper wires below 20 AWG to prevent oxidation gaps. Torque terminal screws to 0.5 Nm–overtightening deforms copper, while undertightening causes intermittent contact. Re-check all connections after 24 hours; thermal cycling can reveal hidden weaknesses.
Validate the circuit with a substitution method. Replace the sensor with a precision resistor–values of 100Ω, 500Ω, or 1kΩ (±0.01% tolerance)–and compare instrument readings. For 2-wire setups, the error should be
Log baseline readings at room temperature before deployment. Note the offset–typically 0.2°C–0.8°C–between the instrument display and a certified reference thermometer. Store this value for future diagnostics; drifts exceeding 0.5°C over 90 days indicate lead degradation or sensor drift.
Critical Components and Their Roles in a Resistance Temperature Sensor Conditioning Path
Begin with a precision current source delivering 100 µA–1 mA to the sensing element; values above 1 mA risk self-heating errors exceeding 0.1 °C. Use a low-drift (±5 ppm/°C) active current sink such as the REF200 from Texas Instruments–avoid passive resistors whose tolerance widens with ambient swings.
A three-wire configuration eliminates lead resistance errors up to 10 Ω per conductor if excitation currents are balanced. Failing this, parasitic voltages–typically 2.5 µV/Ω for 24 AWG copper–distort readings by 0.06 °C per Ω when uncompensated. For platinum elements (α = 0.00385), a 0.1 °C error arises from just 1.7 Ω discrepancies.
Select an instrumentation amplifier with less than 0.1 µV/°C offset drift and a CMRR exceeding 120 dB at 60 Hz; the INA333 achieves 130 dB. Unity gain bandwidth must exceed 1 MHz for 16-bit resolution at 10 samples/s–oversampling by 3× reduces noise floor by √3 accordingly.
| Component | Minimum Specification | Error Budget (1 °C FS) |
|---|---|---|
| Current Source Drift | ≤ ±15 ppm/°C | 50 m°C |
| In-Amp CMRR | > 120 dB | 30 m°C |
| ADC ENOB | ≥ 15 bits | 20 m°C |
| PCB Leakage | < 1 nA stray | 100 m°C |
Decouple the analog ground plane from digital returns using a ferrite bead (47 Ω @ 100 MHz) and 10 µF ceramic capacitors placed within 3 mm of the amplifier’s supply pins. Any shared trace longer than 5 mm introduces 1.2 °C error per 1 mA digital switching current–route high-frequency lines orthogonal to sensitive paths.
Bridge excitation voltages should settle to 0.1 % within 500 µs; faster settling demands slew rates above 0.5 V/µs. The AD8221 meets this with 1.5 V/µs, while common rail-to-rail op-amps (TLV9062) stall at 0.2 V/µs, risking 0.4 °C hysteresis at 20 °C/min ramps.
Implement a sigma-delta ADC sampling at ≥ 20× the excitation frequency; 128 kSPS suffices for 1 kHz excitation. Post-filtering with a 20-tap FIR reduces quantization noise by 16 dB, sufficient for 0.02 °C resolution–avoid IIR filters due to phase distortion near Nyquist frequencies.
Store calibration coefficients in EEPROM with CRC16; a 0.05 % error in gain coefficient propagates linearly to span errors–±0.12 °C for a 250 °C span. Use a separate 3.3 V LDO for the analog section, bypass with 1 nF X7R ceramics to suppress 50 Hz mains artifacts below 2 µV.