MAX31865 Overview

A PT100 sensor has a nominal resistance of 100 ohms at 0°C, while a PT1000 has a nominal resistance of 1000 ohms at 0°C. The resistance changes predictably with temperature.

The MAX31865 measures the RTD resistance relative to a precision reference resistor (Rref). Software then converts the measured resistance into a temperature using the platinum RTD resistance-temperature relationship.

Understanding the Reference Resistor (Rref)

The reference resistor is one of the most important components affecting the MAX31865 measurement. It provides the reference against which the RTD resistance is measured.

On the pictured module, the resistor near the printed Rref marking is marked 431. Using the standard three-digit resistor marking convention, this indicates a nominal resistance of 430 ohms.

Which Reference Resistor Should Be Used?

Does the Reference Resistor Need to Be Exactly 430Ω?

For a PT100 installation, 430Ω is the nominal design value. A precision resistor with low tolerance and a low temperature coefficient is preferable.

For example, a 430Ω resistor with 0.1% tolerance has an initial specified tolerance of approximately ±0.43Ω. A 1% resistor permits approximately ±4.3Ω.

A resistor marked 431 identifies the nominal value but does not, by itself, confirm the tolerance or temperature coefficient. Those specifications must be established from the actual component specification or by measurement.

Can Changing Rref Improve Accuracy?

Yes, where the fitted reference resistor has excessive tolerance, drift or instability. However, replacing an already suitable precision resistor may produce little improvement.

• Use a precision 430Ω reference resistor for PT100 or 4300Ω for PT1000.
• Prefer 0.1% tolerance or better for applications requiring improved accuracy.
• Choose a low temperature coefficient, such as 25ppm/°C or better, where practical.
• For advanced calibration, measure the actual fitted reference resistance and use that measured value in the code.
• Keep the reference resistor and electronics away from unnecessary heat.

The resistor must also be physically suitable for the PCB footprint and the circuit. The pictured board uses a surface-mount reference resistor; verify the footprint before ordering a replacement.

MAX31865 Module Connections

The pictured breakout has the following labelled connections. Refer to these labels rather than relying on the physical orientation of another manufacturer's module.

RTD Sensor Terminals

The four RTD terminal positions on the pictured module are labelled from left to right:

Connecting 2-Wire, 3-Wire and 4-Wire RTD Sensors

The number of wires affects how accurately the converter can account for the resistance of the sensor leads.

1. PT100 2-Wire Connection

A 2-wire sensor has one conductor connected to each side of the platinum sensing element.

For the pictured jumper layout, close both small triangular solder jumpers marked for 2-wire/3-wire connections, or use suitable short links across the left and right terminal pairs.

Leave the 2/3 selection bridge in its original 2-wire position; do not perform the 3-wire trace-cut modification.

In the Arduino code, select:

thermo.begin(MAX31865_2WIRE);

2. PT100 3-Wire Connection

A 3-wire RTD has two conductors connected to one side of the sensing element and one conductor connected to the opposite side.

First use a multimeter to identify the two wires that connect together through a very low resistance. Do not rely solely on wire colours.

For the jumper arrangement shown on the pictured module:

1. Close the left solder jumper marked 2/3 Wire, linking the negative force and measurement connections.
2. Locate the small selection jumper beside the 2 Wire marking and above the Rref area.
3. Cut the existing thin trace connecting the selection jumper to its 2-wire position, then check that the cut is open circuit.
4. Solder the bridge marked 3 for 3-wire operation.
5. Leave the right-hand 2-wire terminal link open so F+ and RTD+ remain separately connected to the paired sensor wires.

In software, select:

thermo.begin(MAX31865_3WIRE);

3. PT100 4-Wire Connection

A 4-wire RTD has two wires connected to each end of the sensing element. One conductor in each pair carries excitation current and the other provides a separate measurement connection.

Use a multimeter to identify the two wire pairs. The resistance between conductors on the same sensor end should be relatively low compared with the resistance measured through the RTD element.

For 4-wire operation, leave the pictured breakout's configuration jumpers in their unmodified default state. Do not close the 2/3-wire terminal links or perform the 3-wire selection modification.

In software, select:

thermo.begin(MAX31865_4WIRE);

Arduino UNO Wiring

The example below uses software SPI so that each signal pin is explicitly defined in the sketch. This makes wiring easy to check and does not require changing hardware SPI configuration.

*Use 5V only when the particular breakout's 5V VIN and logic-level compatibility have been confirmed. The MAX31865 IC itself must not be connected directly to 5V logic or supply pins.

ESP32 Wiring

The ESP32 can communicate with the MAX31865 through SPI. This example uses the familiar GPIO numbers of a standard ESP32 development board and software SPI for compatibility with the example sketch.

*The example assumes the breakout supports operation with 3.3V on VIN. Confirm its supply requirement before wiring. A regulated breakout requiring a higher VIN must be powered according to its own specifications while ensuring every ESP32-facing signal remains 3.3V compatible.

Arduino IDE Setup and Example Code

Install the Required Library

1. Open the Arduino IDE.
2. Select Tools > Manage Libraries, or open Library Manager from the sidebar.
3. Search for Adafruit MAX31865.
4. Install Adafruit MAX31865 library by Adafruit.
5. Install any dependencies requested by Library Manager, including Adafruit BusIO where required.
6. Select the correct controller board and serial port.

For ESP32, also install the current Espressif Systems ESP32 board package through Arduino IDE Boards Manager if it is not already installed.

Complete MAX31865 Temperature Test

The following example works with either Arduino UNO or an ESP32 using the wiring tables above. Select the correct controller and update the sensor settings before uploading.

#include <Arduino.h>
#include <Adafruit_MAX31865.h>

// Select the RTD connected to your board:
// 100.0 for PT100, 1000.0 for PT1000
const float RTD_NOMINAL = 100.0;

// Actual fitted reference resistor:
// Normally 430.0 for PT100
// Normally 4300.0 for PT1000
const float RREF = 430.0;

// Select the sensor wire configuration:
const max31865_numwires_t RTD_WIRES = MAX31865_3WIRE;
// Other options:
// MAX31865_2WIRE
// MAX31865_4WIRE

// Software SPI pin definitions
#if defined(ARDUINO_ARCH_ESP32)
const int PIN_CS   = 5;
const int PIN_MOSI = 23;
const int PIN_MISO = 19;
const int PIN_CLK  = 18;
#else
// Arduino UNO example
const int PIN_CS   = 10;
const int PIN_MOSI = 11;
const int PIN_MISO = 12;
const int PIN_CLK  = 13;
#endif

Adafruit_MAX31865 thermo(
  PIN_CS, PIN_MOSI, PIN_MISO, PIN_CLK
);

void setup() {
  Serial.begin(115200);
  delay(1000);

  Serial.println("MAX31865 RTD Test");

  if (!thermo.begin(RTD_WIRES)) {
    Serial.println("MAX31865 initialisation failed");
    while (true) {
      delay(1000);
    }
  }

  // Use 50 Hz noise rejection for Australia
  thermo.enable50Hz(true);

  Serial.println("RTD initialised");
}

void loop() {
  // Read the raw RTD ADC value
  uint16_t raw = thermo.readRTD();

  // Convert reading to resistance
  float resistance = (raw / 32768.0) * RREF;

  // Calculate temperature in degrees Celsius
  float temperature = thermo.calculateTemperature(
    raw, RTD_NOMINAL, RREF
  );

  // Read the fault register
  uint8_t fault = thermo.readFault();

  if (fault) {
    Serial.print("Fault detected: 0x");
    Serial.println(fault, HEX);

    if (fault & MAX31865_FAULT_HIGHTHRESH)
      Serial.println("RTD high threshold");

    if (fault & MAX31865_FAULT_LOWTHRESH)
      Serial.println("RTD low threshold");

    if (fault & MAX31865_FAULT_REFINLOW)
      Serial.println("Reference input fault: REFIN- high");

    if (fault & MAX31865_FAULT_REFINHIGH)
      Serial.println("Reference/force connection fault");

    if (fault & MAX31865_FAULT_RTDINLOW)
      Serial.println("RTD input/force connection fault");

    if (fault & MAX31865_FAULT_OVUV)
      Serial.println("Overvoltage/undervoltage fault");

    thermo.clearFault();
  } else {
    Serial.print("Raw RTD: ");
    Serial.println(raw);

    Serial.print("Resistance: ");
    Serial.print(resistance, 4);
    Serial.println(" ohms");

    Serial.print("Temperature: ");
    Serial.print(temperature, 2);
    Serial.println(" C");
  }

  Serial.println("--------------------");
  delay(1000);
}

Expected Result

Open Arduino IDE Serial Monitor and select 115200 baud.

With a correctly connected PT100 at approximately 25°C, a typical output might look like:

MAX31865 RTD Test
RTD initialised
Raw RTD: 8339
Resistance: 109.4412 ohms
Temperature: 24.55 C
--------------------

The displayed values are illustrative, not a calibration reference. An actual room-temperature PT100 will typically measure slightly above 100Ω. Warm the sensor gently with your fingers and the reported temperature should gradually increase. Allow it to return to ambient temperature and the reading should decrease.

If the Serial Monitor reports a fault rather than a temperature, check the RTD wiring, solder jumpers, supply and SPI connections before relying on the measurement.

Using Two MAX31865 Modules on One Controller

Two modules can be connected to the same controller. With hardware SPI they can share CLK, SDI and SDO, but each module needs a different CS pin. For software SPI, the same shared wiring arrangement can also be used with separate CS pins.

On an ESP32, for example, connect module A to GPIO5 for CS and module B to another suitable output such as GPIO17. Keep the SPI signal connections and ground common.

Why Two MAX31865 Modules Give Different Readings

Two MAX31865 boards connected to similar RTDs may give slightly different temperatures, even when both systems are operating correctly.

The difference can come from the reference resistor, RTD tolerance, sensor wiring, temperature gradients, calibration or the converter itself.

1. Reference Resistor Tolerance

The MAX31865 measures a ratio between the RTD resistance and Rref. If one module's actual Rref differs from the nominal value entered in the software, the reported resistance and temperature will be shifted.

For example, consider two PT100 modules where the reference resistors measure:

If the rest of the system is ideal, using 430.0 in software for Module B despite its actual resistance being 431.0Ω can cause an error of approximately 0.66°C near 0°C.

Using the correct measured Rref in software can reduce this error, provided the measurement is accurate and the resistor is stable.

2. Sensor Tolerance

PT100 and PT1000 probes are available in different tolerance classes. Two sensors placed at the same temperature are not guaranteed to report exactly the same resistance.

Check whether the RTD is specified as IEC 60751 Class AA, Class A, Class B or another tolerance. Sensor class, temperature range and calibration all affect how closely the readings can agree.

3. Lead Resistance

Two-wire PT100 sensors can show a temperature offset because the connecting-wire resistance becomes part of the measured resistance.

For improved results, use a correctly wired 3-wire RTD with matched lead resistance or a 4-wire RTD where practical.

4. Different Actual Sensor Temperatures

Two probes next to each other may not be at exactly the same temperature, especially when touching different surfaces or exposed to moving air.

For comparison tests, place the probes in a stable, well-mixed environment and allow enough time for both to settle.

5. Electrical Noise and Heating

Electrical noise, poor contacts and unstable power can affect measurements. RTD excitation current can also produce some self-heating in the sensor.

The Adafruit MAX31865 library uses one-shot measurements and switches off the measurement bias between conversions, helping reduce self-heating compared with unnecessarily continuous excitation.

Improving Accuracy and Calibrating the MAX31865

Calibration is most effective after confirming that the resistor selection, wiring and software settings are correct. A software offset should not be used to conceal an incorrectly configured sensor or a faulty module.

Step 1 – Confirm the Correct Rref

1. Disconnect power from the MAX31865.
2. Identify the reference resistor fitted to the board.
3. Check whether it is nominally 430Ω or 4300Ω.
4. Where higher accuracy is required, measure the resistor with suitable precision equipment or verify its value from a reliable component specification.
5. Update the RREF setting in software to the measured value.

An ordinary handheld multimeter may not have sufficient accuracy to characterise a 0.1% reference resistor. An in-circuit resistance measurement may also be influenced by other circuit paths. For precision calibration, isolate the component appropriately or use a suitable calibrated measurement method.

Step 2 – Use a Known Resistance to Test Each Module

A useful diagnostic technique is to temporarily replace the RTD probe with a known precision resistor, such as a calibrated resistor near 100Ω for a PT100 configuration.

For example:

• Use a precision resistor with a documented actual value and suitably low temperature coefficient.
• Connect it in the correctly configured 2-wire RTD circuit with short, low-resistance test connections.
• Set both modules to MAX31865_2WIRE and use the correct RREF for each board.
• Compare the measured resistance displayed by each module with the known resistor value.

This test helps separate differences in the MAX31865 module from differences caused by actual temperature probes. It does not establish full-system calibration over the complete temperature range.

Step 3 – Correct a 2-Wire Lead Resistance Offset

For a 2-wire PT100, the resistance of both sensor leads adds to the measured RTD resistance.

If the total lead resistance is known, it can be subtracted from the measured resistance before calculating temperature.

For example, if the total round-trip lead resistance is 0.50Ω:

// Read raw data and calculate resistance
uint16_t raw = thermo.readRTD();

float measuredResistance =
    (raw / 32768.0) * RREF;

// Total resistance of both leads
const float LEAD_RESISTANCE = 0.50;

// Corrected sensor resistance
float correctedResistance =
    measuredResistance - LEAD_RESISTANCE;

// Convert corrected resistance back to RTD code
float correctedRaw =
    (correctedResistance / RREF) * 32768.0;

// Calculate corrected temperature
float correctedTemperature =
    thermo.calculateTemperature(
      (uint16_t)(correctedRaw + 0.5),
      RTD_NOMINAL,
      RREF
    );

This code fragment is intended to replace the resistance-to-temperature calculation in the complete sketch above. Use it only after a valid reading without an active fault and after establishing a reliable lead-resistance value.

A lead-resistance correction assumes the lead resistance stays sufficiently stable. Large temperature changes along the cable can change its resistance and reduce the effectiveness of a fixed correction.

Step 4 – Check Temperature at a Known Reference Point

A carefully prepared ice-water bath provides an approximate 0°C reference for basic checks.

1. Use crushed ice and a small amount of clean water to create an ice-water slurry.
2. Immerse the sensing portion of the RTD sufficiently, without wetting any components or joints that are not waterproof.
3. Allow the sensor to stabilise.
4. Observe the reported temperature.
5. Compare the result with the expected reference and record the difference.

For example, a sensor reporting +0.4°C at a true 0°C reference has a +0.4°C error at that point.

A fixed software correction of -0.4°C may improve agreement near that calibration point. It does not guarantee accuracy across a wide temperature range because sensor, wiring and reference errors may vary with temperature.

Step 5 – Apply a Temperature Offset Where Appropriate

For a basic single-point correction, add an offset after the temperature calculation:

// Replace the uncorrected temperature output
// with a calibrated temperature value.

const float TEMP_OFFSET = -0.40;

float correctedTemperature =
    temperature + TEMP_OFFSET;

Serial.print("Corrected temperature: ");
Serial.print(correctedTemperature, 2);
Serial.println(" C");

This example uses a -0.40°C offset for illustration. Determine the actual offset from a suitable reference measurement rather than copying the example value.

Step 6 – Check at Multiple Temperatures

If accurate readings are required over a wide range, calibrate at several relevant temperatures using a suitable reference thermometer or calibrated temperature source.

For example, an installation operating between 0°C and 100°C may benefit from verification near the lower, middle and upper operating temperatures.

A multi-point calibration can reveal offset, scale and nonlinearity errors that cannot be reliably corrected with one fixed temperature adjustment.

Tips for Stable Measurements

• Use the 50Hz filter: In Australia, enable 50Hz noise rejection using thermo.enable50Hz(true);.
• Keep connections secure: Poor terminal contacts can introduce resistance errors and intermittent faults.
• Separate sensor leads from power cables: Avoid routing long RTD leads alongside motor cables, relays, mains wiring or other electrical noise sources.
• Choose suitable cable: For long runs, use cable appropriate for RTD measurements, with shielding where required by the installation.
• Average stable readings: Averaging can reduce random fluctuations but will not correct reference resistor or lead-resistance bias.
• Allow the RTD to stabilise: The sensing element and protective probe housing may take time to reach the temperature being measured.
• Protect the electronics: Keep the breakout away from moisture, excessive heat and vibration.

Troubleshooting