Skip to Navigation

Temperature Sensors (App Note)

This is the start of an application note about temperature sensors,

Quick Summary:

If you are measuring in the range of -50 to +150 degrees C, consider using a silicon type temperature sensor.  They are generally the cheapest solution, the easiest solution, and the most accurate solution.

Beyond that range, thermocouples are usually the best option.

Thermistors and RTDs can seem to be very accurate when you look at the raw specs, but that is just the accuracy of resistance versus temperature.  It is generally not easy or cheap to measure that resistance with enough accuracy to achieve the great stated accuracy of the sensor element itself.

 

Silicon Type Sensors:

In the range from -50 to +150 degrees C, silicon temperature sensors are generally cheaper, easier to use, and more accurate, than other types of temperature sensors. With no or minimal extra components, they provide a high-level linear voltage output that connects directly to a LabJack's analog inputs.

Signals from silicon type sensors are easy to acquire with any LabJack (U12/U3/U6/UE9).

Following are silicon type analog output temperature sensors available in through-hole style packages convenient for soldering to the end of a cable.  If you consider digital output sensors and/or surface-mount packages you can find even better sensors:

EI-1034:  A silicon based temperature probe made by Electronic Innovations and sold by LabJack.  It uses an LM34CAZ sensor element from National Semiconductor with a 10k load resistor from signal to ground.  The LM34 provides an easy-to-use 10 mV/°F.  Range with 5V/0V supply is -17 to +110 °C (0 to 230 °F).  Accuracy (max) is +/-0.56 °C (+/-1.0 °F) at room temp and +/-1.1 °C (+/-2 °F) across range.  Non-linearity is +/-0.3 °C (+/-0.6 °F) max across range, so a simple calibration can provide more accurate measurements.  Assembly has a 6ft cable, which can be extended to 25ft, or much longer if you add a 10k series resistor to prevent oscillation.  Uses a 6" x 0.25" waterproof stainless steel probe.

EI-1022:  A silicon based temperature probe made by Electronic Innovations and sold by LabJack.  It uses an LM335A sensor element from National Semiconductor with a 2k current setting resistor.  The LM335A provides an easy-to-use 10 mV/°K.  Range with 5V/0V supply is -40 to +100 °C (-40 to 212 °F).  Accuracy (max) is +/-3 °C (+/-5.4 °F) at room temp and +/-5 °C (+/-11 °F) across range.  Non-linearity is +/-1.5 °C (+/-2.7 °F) max across range, so a simple calibration can provide more accurate measurements.  Assembly has a 6ft cable, which can be extended to much longer distances with no added components.  Uses a 4" x 0.25" plastic probe.

LM34CAZ:  TO-92 package with 10 mV/°F output.  Buy from Digikey and others.  Range with 5V/0V supply is -17 to +110 °C (0 to 230 °F).  Accuracy (max) is +/-0.56 °C (+/-1.0 °F) at room temp and +/-1.1 °C (+/-2 °F) across range.  Non-linearity is +/-0.3 °C (+/-0.6 °F) max across range, so a simple calibration can provide more accurate measurements.  For lower temperatures, to -40 °C (-40 °F), you need to add a negative bias on the signal output.  Note that even if you want to measure in Celsius, the LM34 is better than the LM35 because you get more voltage per temperature (18 mV/°C versus 10 mV/°C) and you can measure lower with a single supply (-17 °C versus +1 °C).  Regardless of cable length we always recommend a 10k resistor from Vout to GND (preferably right at the sensor), and this is usually good for cables up to 25ft.  Beyond 25ft see the "Capactive Loads" section in the LM34 datasheet and consider adding a series resistor.

LM34AH:  TO-46 package with 10 mV/°F output.  Buy from Digikey and others.  Range with 5V/0V supply is -17 to +150 °C (0 to 300 °F).  Accuracy (max) is +/-0.56 °C (+/-1.0 °F) at room temp and +/-1.1 °C (+/-2 °F) across range.  Non-linearity is +/-0.4 °C (+/-0.7 °F) max across range, so a simple calibration can provide more accurate measurements.  For lower temperatures, to -40 °C (-40 °F), you need to add a negative bias on the signal output.  Note that even if you want to measure in Celsius, the LM34 is better than the LM35 because you get more voltage per temperature (18 mV/°C versus 10 mV/°C) and you can measure lower with a single supply (-17 °C versus +1 °C).  Regardless of cable length we always recommend a 10k resistor from Vout to GND (preferably right at the sensor), and this is usually good for cables up to 25ft.  Beyond 25ft see the "Capactive Loads" section in the LM34 datasheet and consider adding a series resistor.

LM135A: TO-46 package with 10 mV/°K output.  Buy from Digikey and others.  Range with 5V/0V supply is -55 to +150 °C (-67 to 300 °F).  Accuracy (max) is +/-1.0 °C (+/-1.8 °F) at room temp and +/-2.7 °C (+/-4.9 °F) across range.  Non-linearity is +/-0.5 °C (+/-0.9 °F) max across range, so a simple calibration can provide more accurate measurements.  Very long cables can be used with no added components.  The other sensors are 3-terminal series sensors, but the LM135A is a 2-terminal shunt sensor that requires a resistor to control current (self-heating might need to be considered).

LM60BIZ:  TO-92 package with 6.25 mV/°C output (plus 0.424V offset).  Buy from Digikey and others.  Range with 5V/0V supply is -25 to +125 °C (0 to 230 °F).  Accuracy (max) is +/-3 °C (+/-5.4 °F) across range.  Non-linearity is +/-0.6 °C (+/-1.1 °F) max across range, so a simple calibration can provide more accurate measurements.  Specified for use with long cables without added components.

 

Thermocouples:

If you can't use a silicon type sensor, a thermouple is usually the next best option.  They are not particularly accurate, but often fine when you are measuring hundreds of degrees.  See the Thermcouples App Note for more detail.

 

Thermistors:

Thermistors often have great-looking accuracy specs for low cost, but that is the accuracy of resistance versus temperature.  Trying to measure the resistance with an accuracy that matches the accuracy of the sensor spec can be difficult and/or costly, but likely not as bad as an RTD.

The U6/T7 are good at handling thermistors, because of the capabilities of the analog inputs and the available current sources.  For other devices you generally use a voltage divider circuit, which requires 1 resistor and 1 excitation voltage, and then you get a voltage signal that you can measure.  The LJTick-Divider can be configured for such resistance measurement.

 

RTDs:

RTDs often have great-looking accuracy specs at reasonable prices, but the problem is that is the accuracy of a small resistance change versus temperature.  Trying to measure resistance with an accuracy that matches the accuracy of the sensor spec can be difficult and/or costly.

The U6/T7 is good at handling RTDs, because of the capabilities of the analog inputs and the available current sources.  See Section 2.5 of the U6 User's Guide, and the information is similar for the T7.  For other devices, and often even with the U6/T7, you need signal conditioning for RTDs.

Here is a good forum topic related to RTDs.

 

tags:

Comments

#1

I want to measure temperatures across a length at several intervals. Which kind of temperature sensors allows probes to be lined up in series so that I only have to use input channel on the T7? Also, I don't want the length of wire leading back to the T7 to interfere with the reading.

#2

What is the needed temperature range?

About how many sensors?

Can you give an idea of the distance to the test specimen and the distance between sensors?

#3

So if I am using your labjack USB DAQ and I need to monitor temperature what options do I have? I have a temperature controller that I believe handles RTD type sensors (it was the cheapest temp controller I could find). Is that possible given your DAQ? Or do I have to do some signal condition like this instruction suggests?

Thanks!

#4

If a silicon type sensor works for your temperature range and other requirements, they are the easiest and best.

Are you going to use a LabJack instead of the temperature controller, or are you wanting to use a LabJack in addition to the temperature controller?