connecting thermocouple to DAQ board

[trishtrishtrish]

Hi,

I have a NI USB-9162 with NI cRIO -9215 (DAQ board) and I want to aquire temperature data with a J type thermocouple. I have read (http://zone.ni.com/devzone/cda/tut/p/id/4237#toc2) that you can't just connect the thermocouple to the DAQ board since you need to compensate for the cold-junction. Could someone explain to me how to do this using labview software? Thank you

[Bassett_Hound]

Hi trishtrishtrish,

I assume you are asking how best to connect to the AI module and how this affects your system.  There are three noise sources: the noise of the measurement board, the thermal drift of the measurement board, and the thermal noise of the cold junction measurement. The first two you should be able to get from the specs of the board, leaving the hardest number to estimate being thermal noise of the cold junction measurement.

When you make a thermocouple measurement, your "resolution" is limited by system noise (quantization and random) just as with any other measurement.  However, because of the slow nature of temperature measurements, gain and offset drifts with temperature might also be considered noise and hence limit resolution.  More specifically to thermocouple measurements, fluctuations in the cold junction measurement caused by changes in the temperature of either the CJ sensor or the actual CJ terminal temperature can also be considered drifts or noise, and further limit the effective resolution when considered over a long time period.  NIs TC products, (which you might consider looking into) like the 9211 are designed with very good offset and gain drifts to limit that effect, but the magnitude of fluctuations in the CJ measurement error depend on the thermal characteristics of the installation.

The dominant source of noise (or resolution) in this application will be in the cold junction.  Changes in the temperature of the wiring terminals that make up the cold junction will look like measurement noise, and so they will have the effect of reducing resolution.  How much noise (or resolution) the you will end up with will depend on (1) how stable the temperature of the cold junction terminals can be kept and (2) over what length of time the signal is looked at.  Over a period of hours, it will probably be very difficult to maintain the noise (or resolution).  Over a period of one second, it should be rather easier. If we attempted cold junction compensation, the noise (or resolution) of the measurement will be dominated by the difference between the CJ Sensor measurement and the actual cold junction temperature. This should reduce the noise over long terms, but increase the noise over short periods.

If you are looking to make good measurements you will probably just need to build a old-fashioned cold junction terminal block sitting in an ice bath, and use 0 C as the cold junction value without actually measuring anything, then the noise will just be the stability of the terminal block temperature.   If the terminal block temperature fluctuates by 0.1 deg C rms, then you would have an additional 0.1 deg C rms of noise.  Check here for an image of how the ice bath would look: http://zone.ni.com/devzone/cda/tut/p/id/4218  You could also look into an insulated oil bath with an RTd as a CJ sensor.  I've seen both of these techniques working well in the past but feel they can be a hassle to deal and you might be better off by looking into a device designed for thermocouple measurements.

Regards,

Steve

[davey31415]

I thought part of the idea behind signal conditioning for thermocouples involved measuring the temperature at the cold junction and compensating for it.  Am I wrong here or do only certain thermocouple inputs do this?  I'm going to take a look at the 9211 manual and see what I can learn from that.

I started browsing posts looking for a way to connect the thermocouple I have to a DAQ device.  I believe ths connector it uses is standard, it has two narrow, flat tips which resemble a 110VAC socket if both plugs were rotated 90 degrees.  Do I need to use a certain type of converter depending on the materials the thermocouple is made up of?

[gwd]

What you are describing is a standard thermocouple connector- actually the smaller of two standard connectors. NI refers to this as a mini-thermocouple connector.

They do indeed come with different metals to eliminate thermal voltages when connecting particular types of thermocouple wire. You need to purchase the right connectors (and you need to be sure to connect the wires correctly!) My favorite source for all this stuff is omega.com.

Now how you make the connection depends on how accurate you need to be. For some measurements in a room with stable temperature it may be good enough just to measure the room temperature and correct the measured voltage. NI does offer hardware solutions for doing better compensation (http://sine.ni.com/nips/cds/view/p/lang/en/nid/10688, for example).