Thermocouple k type fluctuation reading

[marlin12]

I am new to Labview please anyone can help would appreciated.

I am try to read temperature using thermocouple K type with a module NI-9213 in Labview.

I used two individual channel.  One is an ambient air temperature.

The second is one is mount inside of PVC tubing to read the temperature of air flow that also has moisture.

Ambient air temperature reading is working perfectly fine.

The reading in the inside PVC tubing is fluctuating.  The temperature fluctuate between 4-7 degree Celsius randomly.

My question is what cause this fluctuation reading.  Is it air flow inside the tube.  Vibration?  Moisture?  CJC?

Please help anyone?

Thank you so much in advance

Danh

[Finnabon]

Hi Danh,

A quick way to eliminate the thermocouple as the source of problem would be to swap the two thermocouples and see if you're seeing the same behavior. You might also try securing the wires of the thermocouple to reduce noise.  

Is the airflow in the PVC tube a constant temperature? The movement of the air could be making the varying temperatures in your environment more apparent.

[tbob]

What kind of wire are you using to connect the thermocouples?  You should be using twisted pair with a shield.  A bad solder connection could cause fluctiations. 

If you still have a problem, try blocking the PVC pipe at both ends to stop air movement to see if the fluctuation stops.  If the fluctuations are caused by air movement, you could take an average of the last few readings, or you could acquire several readings, convert the data to a waveform, and run it through a low pass filter.  This would smooth out the fluctuations.

[marlin12]

Thank you for the quick response.

I am truly agreed that it has to do with the air flow.  I will try your idea and get back with a respond.

The wiring that I am using is  Type K Extension Grade Wire, 24 Gage Stranded, Polyvinyl Insulation.

Thanks again.

[NIquist]

It sounds like you have the right extension wire.  This is very important because the dis-similar metals are what generate the TC signal.  If you connect TC wire with copper wire at any point you have a new thermal junction that will add or subtract from the correct value (unless the temperature at that point is iso-thermal). 

If your problem is turbulent thermal flow you can just glue the TC onto a small block of metal and put that in your pipe (but don't smoke it!).  The mass of metal will act as thermal capacitor and naturally low-pass filter your output measurements.  Of course you can fix it in the software as Tbob mentioned.

[aeastet]

If you add mass to the thermocouple it will cause the response time for the to slow down. The more mass the slower the reaction time. It is like averaging the data on using a mechanical filter. The size of the mass will have a big effect on the reading.

[NIquist]

---

@aeastet wrote:

If you add mass to the thermocouple it will cause the response time for the to slow down. The more mass the slower the reaction time. It is like averaging the data on using a mechanical filter. The size of the mass will have a big effect on the reading.

---

Well, yes.  That was my point.  In this case it's a desired behavior, not a limitation. 

[aeastet]

---

@NIquist wrote:

---

@aeastet wrote:

If you add mass to the thermocouple it will cause the response time for the to slow down. The more mass the slower the reaction time. It is like averaging the data on using a mechanical filter. The size of the mass will have a big effect on the reading.

---

Well, yes.  That was my point.  In this case it's a desired behavior, not a limitation. 

---

Well I thought you would know that but I wanted to make sure that the orginal poster understood that.

[marlin12]

Hello,

Just want to say thank you everyone for their helpful inputs.  I have found a solution for the erratic TC reading.  What I did is grounded the negative TC wire and that seemed to help tremendously.

Thank you again for yours help.

Danh

[tbob]

By grounding the negative wire, you've effectively changed a differential output into a single ended output.  But that's OK if it works.  It depends on the input circuit and what it expects.