Using Precision Digital Multimeters for Quick Checks on Temperature Transducers

figure
If your processes use temperature transducers, then the ability to quickly see if they're good or bad is important. This article describes how to use a digital multimeter (DMM) to perform some quick and simple tests for the most common ones—thermocouples, resistance temperature detectors (RTDs), and thermistors.

Calibration and troubleshooting are two very different requirements. Calibration maintains product quality; troubleshooting affects product quantity. Calibration happens on a schedule; troubleshooting happens in emergencies. Calibration must be precise; troubleshooting must be fast. When a production line is down, speed is of the essence. The failed component must be isolated and replaced ASAP. With a precision multimeter, you can perform quick, go/no-go checks on most temperature transducers, and while these tests tell nothing about transducer accuracy, they will tell you if a transducer has failed. And sometimes that's exactly what you need.

Does This Thermocouple Work?
Thermocouples are unpowered transducers that generate a very low voltage. When two dissimilar metals are in contact with each other, a potential is created across the junction—the Seebeck effect. This voltage across the junction of the two metals is proportional to the junction's temperature.

The "Type" of thermocouple describes the metals used to make the junction, e.g., a J Type thermocouple uses iron in one wire and a copper/nickel alloy in the other. The junction of the metals may be sheathed in various configurations or may be exposed.

The higher the temperature, the higher the voltage produced by the thermocouple. (It is somewhat misleading to use the terms "high" and "voltage" in this context, since the voltage across a common, J Type thermocouple is about 1.0 mV at a room temperature of 68°F and about 1.9 mV at body temperature, 99°F).

There are two steps to checking thermocouples. The first is to check for a short on the terminals and the second, to make sure that voltage tracks with the temperature.

The first test can be performed with any quality multimeter. Put the meter in ohms or continuity mode; on a good thermocouple, you should see a low resistance reading. If you see more than a few ohms, you probably have a faulty thermocouple. If the reading at room temperature is close to 110 Ω, then you have an RTD on your hands—read on.

The second test requires a meter that can measure down to tenths of millivolts (0.0001 V). A meter that can measure hundredths of millivolts (0.00001 V) makes it even easier to do this check, because the added resolution shows very small temperature changes.

Connect the meter to the terminals of the thermocouple. Grabbing the end of the thermocouple should cause the voltage to increase slightly, since you're warming it up. As you release the junction, the temperature (and voltage) should drop.

Multimeters with min./max. recording and the ability to graph electrical signals (similar to an oscilloscope) are also convenient for this application. Min./max. recording lets you connect the meter, walk over to the tip of the thermocouple, warm it for a few seconds, and walk back to the meter to check the results. Typical values for a good thermocouple are shown in Figure 1.

 

figure
Figure 1. Using the digital multimeter's min./max. record function allows you to monitor thermocouple voltage changes over time and make sure voltage increases with increasing temperature.

The figure shows that it took 37 s to warm the tip. Of course, if you had to walk to the end of the transducer, this time would be longer.

Does This RTD Work?
RTDs operate on the principle that the resistance of any conductor changes with temperature. As the temperature of a conductor rises, the increased molecular vibration impedes electron flow. Thus, the higher the temperature, the higher the material's resistance.

Most RTDs are of the PT-100 variety. They consist of a platinum wire coil with a nominal resistance of 100 Ω at the freezing point (or, for purists, the triple point) of water. Resistances other than 100 Ω at 32°F are less common, but do occur. It helps to know what the resistance of your RTD should be.

Sometimes copper or another metal is substituted for the platinum. For example, in some electric motors and transformers an extra set of copper windings functions as an RTD, indicating overtemperature conditions within the motor. In these special applications, and with metals other than platinum, you will probably find freezing-point resistances other than 100 Ω.

To measure an RTD or any resistance, the measurement system drives a current through the device and measures the voltage drop.

While most low-cost DMMs with a millivolt and resistance function can be used to check thermocouples or thermistors, they may not have sufficient resolution and accuracy for testing RTDs. For checking RTDs you'll need a meter that is capable of indicating changes of tenths of an ohm, and you'll want a meter that measures to hundredths—the absolute value of the resistance isn't important, but the ability to track small changes is. Look for multimeters with resolutions down to 0.01 mV or 0.01 Ω and added features such as min./max. recording or a graphical display. Since small changes in resistance reflect large changes in temperature, their additional resolution and improved accuracy give you a clearer picture of how well the RTD being tested is functioning, giving you more confidence in your results.

RTDs can have two, three, or four leads. In a two-wire configuration, simply connect the meter across the leads and measure the resistance. For a PT-100 RTD at room temperature, this should be about 110 Ω (±20%). If you grab the tip of the RTD, you should see the resistance increase. Let go, and you should see the resistance gradually settle back after you release the tip.

Three-wire RTDs are commonly used when a measurement system is made up of resistance bridges. The wires that connect the tip to a measuring device have a temperature-dependent resistance of their own (as do all metals). The extra wire helps the bridge balance out the effects of lead resistance. When checking a three-wire RTD with an ohmmeter, all you need to know is that two of the three wires should be shorted. Usually, the shorted wires are the same color. Between any of the shorted wires and the third wire, the transducer should act just like its two-wire counterpart. That is, at room temperature, the meter should read about 110 Ω for a PT-100 RTD, and resistance should increase slightly as the temperature at the tip increases.

Four-wire RTDs are less common than the other types. If you do come across one, it should have two shorted pairs of wire. Again, the shorted wires are generally the same color. The resistance between different colored wires should have a reasonable value at room temperature and should increase if you heat the tip.

Does This Thermistor Work?
Thermistors are made of semiconductor material and they work in a way opposite to RTDs. While RTDs experience an increase in resistance with increasing temperature, thermistors tend to exhibit a lower resistance with higher temperatures. This is because semiconductor materials tend to conduct more electrons as the temperature goes up.

Although many types of thermistors are available, two-wire thermistors are the most common for general-purpose temperature measurement. Checking a thermistor involves performing a resistance measurement. Using the DMM's resistance function, you should be able to watch the resistance of the transducer stabilize at room temperature and drop as the tip of the transducer is heated.

 

Thermistors generally have a large change in resistance per degree of temperature, so just about any meter can be used to quickly test a thermistor's response. Graphical multimeters can take advantage of this property by graphically displaying the changing resistance. Figure 2 shows a plot of resistance over time for a thermistor that was heated briefly.

 

figure
Figure 2. Using a multimeter with a graphing function lets you see how the thermistor behaves when the temperature changes—this thermistor was heated briefly, causing its resistance to drop.

Words to the Wise
Temperature transducers usually fail in a big way. Rather than drifting, they usually just stop working. While there can be no substitute for regular calibration and certification, in a pinch a precision DMM can work for you as a solid troubleshooting tool.


Reader Comments

I received my November 2003 issue of Sensors today and was shocked to read the misinformation about how thermocouples work in the article, page 33, "Using Precision Digital Multimeters for Quick Checks on Temperature Transducers". It is incorrect to state that a voltage is produced across the junction of the two dissimilar metals of a thermocouple wire. Rather a thermocouple circuit contains two junctions, a measuring junction and a reference junction. The voltage is not produced at the measuring junction or across the measuring junction. Rather the voltage is produced along the portions of the thermocouple wires that experience temperature differences. And the voltage produced is related to the difference in temperature between the measuring junction and the reference junction.

The author then gives the example that a type J thermocouple would produce about 1 mV at a room temperature of 68F and 1.9 mV at a room temperature of 99F. This would be true only when the reference junction is at 32F. If we connect a thermocouple to a multimeter to see if it is working, we would need to chill the multimeter voltage input terminals (the reference junction in this instance) to 32 F for the voltages used in this example to be valid.

On page 34 the author suggests that a good thermocouple will measure only a few ohms and if it reads more it is probably faulty. And he states that if it reads 110 ohms then it is an RTD. The resistance of a thermocouple is simply the series resistance of the two wires of dissimilar metals in the circuit. This wire resistance simply depends on the wire materials, the wire cross sections, and the total length of the wires. Small thermocouple wires can be perfectly accurate and functional and yet have a resistance of more that a few ohms. I assure you that I have a perfectly functional 36 gage Type T thermocouple that is approximately 100 inches long and its resistance measures 110 ohms. And I assure you that it is not an RTD.

It is likely that many of your readers look to Sensors magazine for correct information on various technical subjects. Those who read this incorrect information about thermocouples are subject to become mislead and confused. I would agree that the author was correct with his concepts in a very general way but his specifics are way out of whack.

Jerry Gaffney
Gaffney Engineering
Gainesville, FL

 


Author David Pereles responds:

Mr. Gaffney, I'm OK with you addressing the issues you have with the article. I agree that I should have specified a length when determining if a transducer is a resistive device. I was working with transducers of a few meters, connected to patch panels. Of course any long conductor, especially with a small cross-section, will have a significant resistance.

I don't agree with your other two issues in the context of this article. The article does not claim that the voltages given in the article would actually be observed. I used voltages from the tables, rather than measured voltages, since I could not predict precise values. They are given simply to illustrate the point that we are talking about thousandths of volts.

Your comments do uncover an assumption in the article -- since the meter itself is acting like a reference junction, I am assuming that the the thermocouple measuring junction is being warmed relative to the meter. If ambient temperature is high, this technique falls flat. So while this works ok in the lab, on a hot factory floor or out in the field, you might need to use a more aggressive heat source to see a significant voltage increase. I will add something to this effect to the article in case we use the information again.

Referring to the voltage "across the junction" seems like a reasonable simplification, especially in the context of the article and the voltage measurement being performed.

While I do believe the article could be improved, I think we have different ideas regarding the specific issues. I think it would be best if you addressed your issues in your own words. I appreciate your careful reading and technical integrity. I will support you in any way I can.

Regards,
Dave Pereles