Thermocouple wiring with a Voltage Divider

[JLocke]

Hello,

I am trying to measure the temperature through a K-type thermocouple by making a voltage divider connection. As you can see in the picture attached, I am wiring a resistor ( 10 kOhm) with the pins of the thermocouple to the DAQ (NI-6009).

The problem that I am facing is when running this with LabVIEW. I am attaching a file I created in order to control a heating element depending upon the temperature measurement. The result I get is -360.425 C which totally makes no sense, and makes me wonder what is wrong with either the .vi file or the wiring (I have checked the polarity of both the resistor and the thermocouple and that part should be solved already).

I am attaching a picture of the result since to obtain something everything needs to be wired first.

Thanks in advance

JL

[johnsold]

It is hard to see how you have the voltage divider connected in the picture.  Can you post an image of the schematic?

I never heard of anyone uisng a voltage divider on the output of a thermocouple. The voltage is quite small and dividing it makes little sense.

Next, the USB-6009 is marginally useful for measuring the output of a thermocouple.  The resolution is about 1 degree.

Your formula for converting voltage to temperature is suspect.  I do not have any hardware so I just generated an array of values from zero to 0.05. This results in temperatures on the chart of -97 to -87 degrees.  For a Type K thermocouple 0 to 50 mV should represent temperatures from 0 to 1232 C.

Lynn

[JLocke]

Here is a very similar schematic that can help you understand

http://zeus.phys.uconn.edu/wiki/images/LabViewThermistor.pdf

They also mention a different formula to convert the voltage to temperature (F), which might be a better solution than mine.

I am using a voltage divider since I found it could work well with a thermocouple instead of having to buy new hardware.

Regards

JL

[johnsold]

The schematic you posted is for use with a thermistor, NOT a thermocouple.  The two are very different devices.  The thermocouple generates tens of microvolts per degree (with a source impedance of a few ohms or less) while the thermistor changes resistance by tens or hundreds of ohms per degree from an initial resistance of thousands of ohms.

The thermistor circuit will not work with a thermocouple.

Lynn

[JLocke]

Okay I didn't know, so do you know the equivalent circuit for a thermocouple or just a way to implement that voltage divider idea to a thermocouple?

[johnsold]

Just remove the resistors. Connect the thermocouple to a pair of differential inputs.  The internal bias networks of the USB-6009 will take care of any bias currents required.  Configure the device for differential inputs on the most sensitive input range. Look at omega.com for thermocouple voltage tables to get started on the voltage to temperature conversion process.

Lynn

[JLocke]

So precisely, from the schematic I firstly posted, you would simply remove the resistor and connect directly the thermocouple to AI0 and GND directly? Then from the .vi file you mentioned the temperature ranges, which would be then the proper formula to convert the voltage to temperature? From my opinion the one mentioned in the link before should be fine but at this point I am not sure.

Thanks

JL

[johnsold]

Connecting from AI0 to GND only works for single ended configurations.  Look at Table 4 in the manual. You want to connect the thermocouple to AI 0+ (terminal 2) and AI 0- (terminal 3). You will not have a ground connection to the thermocouple.

Look at Omega for information about thermocouples.  For starters just use a linear approximation such as 39 microvolts per degree. It is not accurate over a wide range but is close enough to see it things are working.

Lynn

[JLocke]

Okay, so for my perspective I need a Signal Conditioner in order to convert that voltage signal into temperature, since I need quite high accuracy, do you recommend me any from Omega which can be precise enough?

[johnsold]

First, accuracy and precision are quite different things.  You used both in the same sentence as though they were interchangeable.

The standard limits of error for Type K thermocouples is 2.2 degrees C or 0.75% for temperatures above 0 C.  Special order thermocouples can get that to 1.1 or 0.4 degrees. Then you need to include any error due to the signal conditioning circuit and any error in the voltage measurement system.

Temperature measurement is not as easy or as simple as most other types of measurements such as voltage.

Please tell us what you are measuring and what your requirements for accuracy and precision are.

Lynn