thermistor excitation and NI6224 DAQ board / anything quick & (relatively) cheap?

[groomit]

Hi everybody,

I need to measure temperature in the range 15 °C to 40 °C and to accurately resolve 0.1 °C changes. My thermistor has nominal resistance of 22 kOhm at 25 °C and resistance of 14 KOhm at 37 °C. Dissipation constant in air at 25 °C is 0.3 mW/°C - but the thermistor will be used in water-based solution. I have a NI6224 DAQ board with a SCB-68 connector and Labview 7.

How can make my NI6224 work with my thermistor with minimal cost? 

In particular:

the NI6224 has no analog output. Can I use a digital output as a current source? I would say no because digital output current is 24 mA, which would make 24 mA times 14 KOhm >> 10 V, which is the input range of the board. Is that correct?

Could I buy a NI 9265 module + chassis + cables etc to provide current excitation to the thermistor, and connect its two leads to a voltage differential input of the NI6224 board? Is there any cheaper option with other NI or third-party equipment? 

I thank you in advance for your time and your help

(please, note that this message was originally posted as a reply to another thread named "thermistor". I am following a suggestion to move it here to a new thread to increase visibility and try and get some help...)

[johnsold]

Apparently you have little electronics training. The digital outputs would not serve as a current source. They are more like a voltage source with a current limit. The 14 kohm thermistor connected to a 5 V digital output would draw ~360 uA.

What you can do is to make a voltage divider with a fixed resistor and the thermistor. Measure both the supply voltage and the voltage at the junction of the resistor and thermistor. From Ohm's Law you can determine the current through the fixed resistor. That current and the voltage across the thermistor allow you to calculate the resistance of the thermistor.

Use the +5 V power output of the device as the excitation source.  Depending on the actual voltage, you might need to use a 10 V scale on the analog input to measure the excitation. The power supply could be a few tenths of a volt above or below 5.00 V. The value of the fixed resistor should probably be similar to the values of the thermistor at the temperatures you will be measuring.  Values of 15 kohm or 22 kohm might be good starting points. At 37 degrees the 15 kohm resistor would result in a dissipation in the thermistor of 416 uW. In water that will likely produce too little self heating to notice.

If you have noise from the power supply and it becomes a problem, adding and R-C filter may help.

Lynn

[groomit]

Hallo Lynn and everybody,

thanks for your reply and useful info. I am a MD PhD doing basic research in a life science lab and thus have NO electronic training at all... but I have some high-school physics knowledge, and I like to learn...

Do I understand correctly that I would need to take 2 single-ended voltage measurements, one connecting the 5V output of the DAQ board to an analog input, and the other connecting the junction between the fixed resistor and the thermistor to a second analog input?

Would it be better in terms of noise reduction to take a differential measurement of the voltage difference across the thermistor? Please, consider that I will in any case construct a calibration curve of each thermistor by repeated measurements in a water bath at different temperatures.

Should I connect one lead of the thermistor to the DAQ board ground?

Thank you in advance!

[johnsold · Accepted solution]

Yes.

You are correct about two single ended measurements. The other end of the thermistor will be connected to ground.

There is no advantage to a differential measurement in this situation because there is no true differential signal.  Both voltages are referenced to ground.

Calibration is always a good idea. Here it will compensate for differences in thermistors and any deviation from the nominal value of the fixed resistor.

Even though your temperatures probably change very slowly, it can be useful to sample at a higher rate and average groups of samples to produce a single reading. This can reduce noise significantly. One common noise source is the AC power line. By sampling at a multiple of the power line frequency, such as 1200 Hz, and then averaging 1200 samples you get one reading per second with substantial reduction in 50 or 60 Hz noise.

Lynn