RTD Self Heating: 5 Proven Ways to Get Better Accuracy

Share:
Temperature
RTD Self Heating: 5 Proven Ways to Get Better Accuracy

Every resistance thermometer needs a current to read its resistance, and that current warms the element a little above the process it is trying to measure. Knowing how large this error is, and how to shrink it, separates a good temperature loop from a merely working one.

I Squared R mW per °C Excitation Current Pulsed Measurement

RTD self heating is the small temperature rise caused by the measuring current flowing through the sensing element. This guide shows how to calculate it, how installation changes it and how modern transmitters keep it negligible.

Hello everyone, today we are going to learn what RTD self heating is, how to calculate the error from excitation current and the self heating coefficient, and how to reduce it in real plant loops.
RTD self heating

What Is RTD Self Heating?

RTD self heating is the rise in temperature of a resistance thermometer element caused by the electrical power dissipated when the measuring current flows through it. The element then reads slightly higher than the true process temperature, because it is measuring its own warmth as well as the process.

An RTD is a passive sensor, so it cannot give a signal without an external current or voltage, as explained in what is an RTD and how it works. The instrument pushes a small current through the platinum element and reads the voltage across it, and that power, P = I²R, turns into heat.

Current excitation circuit for ratiometric RTD measurement from Analog Devices
Image credit: Analog Devices. Diagram courtesy of Analog Devices, shown here for educational reference.

The size of the error depends on three things: the current, the element resistance and how easily heat escapes to the surroundings. Resistance follows Ohm law, so doubling the current makes four times the heat.

Do You Know?

Analog Devices shows that a Pt100 driven at 10 mA can read about 0.5 °C high because of self heating. Dropping the current to 1 mA cuts that error to about 0.005 °C, a hundred times smaller.

Advertisement

The Self Heating Coefficient in mW per °C

Manufacturers describe how well an element sheds heat with a self heating coefficient, written in mW per °C. RDC Control defines it as the power in milliwatts that raises the internal temperature of the sensor by 1 °C, so a higher number means a smaller error.

Some data sheets give the inverse figure in °C per mW. Always check which one is printed, because mixing them up gives an answer that is wrong by a huge factor.

2.5 mW/°CSmall thin film element, TI
65 mW/°CLarge wire wound element, TI
50 mW/°CPt100 in water at 1 m/s, RDC
100 timesWorse in still air than water, RDC

Texas Instruments states that typical coefficients range from 2.5 mW per °C for small thin film elements to 65 mW per °C for larger wire wound elements. The same element also behaves very differently in flowing water, still water, oil or still air.

RDC Control notes that static air can produce errors about 100 times higher than flowing water for the same sensor. That is why a coefficient is meaningless unless the test medium and flow velocity are stated with it.

RTD Self Heating Error Formula

Power: P (mW) = I² (mA²) × R (Ω) ÷ 1000
Error: ΔT (°C) = P (mW) ÷ Self heating coefficient (mW/°C)

With pulsed excitation, multiply P by the duty cycle

Example, Pt100 at 0 °C:
I = 1 mA, R = 100 Ω, coefficient = 2.5 mW/°C, duty 100 %
P = 1² × 100 ÷ 1000 = 0.100 mW
ΔT = 0.100 ÷ 2.5 = 0.040 °C

The resistance R is the value at the process temperature, not at 0 °C, so use the table or the RTD resistance to temperature calculator for hot service. A Pt100 at 100 °C is about 138.5 Ω, which raises the power by the same ratio.

RTD Self Heating Calculator

Self Heating Error of a Resistance Thermometer
Result
Average power 0.100 mW, self heating error 0.040 °C
Quick Tip

When a data sheet gives no coefficient at all, assume the worst case for a small thin film element, about 2.5 mW per °C. If the result is still tiny, you can safely ignore self heating for that loop.

Excitation Current Versus Error at a Glance

CurrentPower in 100 ΩError at 2.5 mW/°CError at 25 mW/°C
0.1 mA0.001 mW0.0004 °C0.00004 °C
0.5 mA0.025 mW0.010 °C0.001 °C
1 mA0.1 mW0.04 °C0.004 °C
2 mA0.4 mW0.16 °C0.016 °C
5 mA2.5 mW1.0 °C0.1 °C
10 mA10 mW4.0 °C0.4 °C

The table shows why current matters far more than anything else for RTD self heating. Going from 1 mA to 5 mA multiplies the error by 25, while changing the element type moves it by a factor of ten at most.

RDC Control notes that traditional instruments used 5 mA, while modern practice is 100 µA or lower. With a 50 mW per °C Pt100 at 100 °C in water flowing at 1 m/s, they calculate about 0.07 °C error at 5 mA and a negligible 0.000027 °C at 100 µA.

Advertisement

Second Worked Example: Old Panel Meter on a Thin Film Sensor

Suppose an old indicator drives 5 mA through a thin film Pt100 with a coefficient of 2.5 mW per °C, measuring water at 100 °C where R is 138.5 Ω. Power is 5² × 138.5 ÷ 1000 = 3.46 mW, and the error is 3.46 ÷ 2.5 = 1.39 °C.

That is far larger than the Class A tolerance of plus or minus 0.35 °C at 100 °C under IEC 60751. Replacing the indicator with a modern transmitter at 0.5 mA reduces the error to about 0.014 °C, well below the tolerance, as discussed in temperature measurement errors.

5 Proven Ways to Reduce RTD Self Heating

1
Lower the Current
Use 1 mA or less for Pt100 and 0.1 to 0.5 mA for Pt1000.
2
Pulse the Excitation
Switch the current on only while the ADC converts.
3
Improve Heat Transfer
Use a snug thermowell, thermal paste or a spring loaded insert.
4
Pick a Larger Element
Wire wound and bigger elements shed heat better than tiny chips.
5
Measure Ratiometrically
Lets a small current still give full ADC resolution.

Lower current means a smaller voltage signal, so the measurement electronics need more gain or a better converter. A modern delta sigma ADC easily resolves a few hundred millivolts from 1 mA through a Pt100.

Pt1000 elements have ten times the resistance, so at the same current they create ten times the heat. The benefit of Pt1000 is that a much smaller current still gives a useful voltage, as compared in Pt100 vs Pt1000 differences.

Heat transfer from the element to the process is the second lever against RTD self heating. A loose insert with an air gap inside a thermowell behaves like the still air case, so both self heating and response time get worse.

Do You Know?

Texas Instruments picked 1 mA for its three wire Pt100 reference design to maximise signal while keeping self heating low. At the maximum element resistance of 400 Ω, the dissipation stays below 0.4 mW.

How Transmitters Pulse the Excitation

IdleNo current flows, element cools to process
ExciteCurrent source switches on for a short window
ConvertADC reads RTD and reference voltages
Switch OffCurrent stops, average power stays tiny
RepeatCycle repeats at the update rate

Many smart temperature transmitters and PLC RTD cards switch the excitation on only during conversion. If the current flows 10 percent of the time, the average power and the RTD self heating error both fall to one tenth, which you can see by entering a duty cycle in the calculator above.

Some inputs rotate the current through several channels, so each sensor is excited only for its own time slot. The settings are normally fixed in firmware and are not user adjustable, but check them during transmitter configuration and ranging.

Quick Tip

When replacing a transmitter brand in a critical loop, compare its excitation current in the manual. A switch from 0.2 mA to 1 mA can shift a small thin film sensor by a measurable amount.

Ratiometric Measurement and Lead Wires

Analog Devices explains that ratiometric measurement reads the RTD as a ratio of a reference resistor, using the same current for both. Any drift in the excitation source cancels out, so the current can be small without losing accuracy.

Lead wire resistance adds its own error and also a little heat along the cable, though that heat is outside the sensing element. Use three or four wire connections, explained in 2 wire, 3 wire and 4 wire RTD, to cancel it.

Measuring Self Heating With the Two Current Method

Calibration labs measure the real coefficient by reading the sensor at two currents in a stable bath. If the power doubles, for example going from 1 mA to 1.414 mA, the reading rises by the self heating amount once more.

The zero power temperature is then the first reading minus the difference between the two readings. This correction is part of good Pt100 RTD calibration and is also included in the uncertainty budget described in measurement uncertainty in calibration.

  • Note the excitation current of the transmitter or card.
  • Find the element resistance at the working temperature.
  • Read the self heating coefficient and its test medium.
  • Check the insert fits tightly in the thermowell.
  • Compare the calculated error with the tolerance class.
  • Consider pulsed or lower current if the error is large.
  • Record the correction in the calibration certificate.
Myth: Self heating only matters in laboratories.
Fact: Old 5 mA meters on thin film sensors can read over 1 °C high in plant service.
Myth: A higher current always gives better accuracy.
Fact: It improves signal size but raises heat with the square of current.
Myth: Pt1000 always has less self heating than Pt100.
Fact: At the same current, Pt1000 dissipates ten times more power.
Myth: Four wire connections remove self heating.
Fact: Four wire cancels lead resistance, not the heat in the element itself.
Benefits of Controlling Self Heating
  • Readings match the true process temperature.
  • Calibration results transfer to the field.
  • Small sensors can be used without bias.
  • Pulsed methods keep power very low.
Trade Offs and Limitations
  • Low current gives a smaller signal to measure.
  • Needs a better ADC or more amplifier gain.
  • Coefficients vary with medium and flow.
  • Pulsing adds design complexity.
Advertisement

Where RTD Self Heating Errors Show Up

Still Air and Gas
Room, duct and HVAC sensors with poor heat transfer.
Surface Sensors
Pipe clamp and bearing sensors with small contact area.
Thin Film Chips
Tiny elements with low mW per °C values.
Laboratory Baths
Precision work where 0.01 °C matters.
Legacy Indicators
Old meters and bridges using several milliamps.

Old analogue meters often used a Wheatstone bridge with a fairly high bridge current. When these are replaced, the apparent process temperature may drop slightly, and operators should know the reason.

Thin film elements are small, cheap and fast, but their low mass also means low heat dissipation, as described in thin film RTD advantages. For thermistors the effect is even stronger, as covered in NTC and PTC thermistor working.

TI Three Wire Pt100 Design Note

PDF
Three Wire PT100 RTD Measurement Circuit With Low Side Reference and Two IDAC Current Sources
Texas Instruments circuit note, excitation current and self heating

RTD Self Heating Error Video

To confirm a suspected self heating bias in the field, compare the installed sensor with a reference probe in a dry block calibrator. Test at the transmitter current, not at a meter current.

RTD Self Heating FAQ

What is RTD self heating?

It is the rise in element temperature caused by the power dissipated by the measuring current. The sensor then reads slightly above the true process temperature it is meant to measure.

The power equals current squared times resistance. How much RTD self heating raises the reading depends on how easily the heat escapes into the process around the element.

What is the self heating coefficient?

It is the power in milliwatts needed to raise the sensor temperature by one degree Celsius. A higher value means the sensor sheds heat well and the error is smaller.

Texas Instruments quotes a typical range from 2.5 for small thin film elements to 65 for large wire wound ones. The medium and flow used for the test must always be stated.

How do I calculate the error?

First find the power as current squared times resistance, and use milliamps and ohms divided by 1000 to get milliwatts. Then divide that power by the self heating coefficient.

For 1 mA in a 100 Ω Pt100 with 2.5 mW per °C, the power is 0.1 milliwatt. The RTD self heating error is therefore about 0.04 °C in this case.

What excitation current is normal?

RDC Control says traditional instruments used about 5 mA, while modern practice is 100 µA or lower. Many transmitters use well under 1 mA for a Pt100 to keep RTD self heating negligible.

Pt1000 sensors usually run at lower currents because of their higher resistance. Always read the actual value in the transmitter or input card manual before comparing sensors.

Does a four wire RTD remove self heating?

No, a four wire connection only cancels the resistance of the lead wires. The heat is created inside the element itself, so the connection method does not change it.

To reduce self heating, lower the current, pulse the excitation or improve heat transfer to the process. Four wire wiring and low current together give the best accuracy.

Why is still air worse than water?

Air carries heat away from the element far more slowly than flowing liquid. The same power therefore raises the element temperature much more in air.

RDC Control notes errors can be about 100 times higher in static air than in flowing water. Duct, room and surface sensors need the lowest practical current for that reason.

How do labs measure self heating?

They read the sensor in a stable bath at two currents whose powers differ by a factor of two. The change in reading equals the self heating at the lower current.

Subtracting that change from the first reading gives the zero power temperature. The correction and its uncertainty are then included in the calibration certificate.

Advertisement

Related Articles

External References

What We Learn Today

  • RTD self heating comes from the I squared R power of the excitation current, and the error equals that power divided by the mW per °C coefficient.
  • Texas Instruments quotes coefficients from 2.5 mW per °C for thin film to 65 mW per °C for wire wound elements, so element style matters a lot.
  • Lower current, pulsed excitation, ratiometric input design measurement and good thermal contact keep the error far below IEC 60751 tolerance classes.
I hope you like above blog. There is no cost associated in sharing the article in your social media. Thanks for reading!! Happy Learning!!

Leave a Reply

Your email address will not be published. Required fields are marked *