incorrect thermocouple data from 6034e cb-68lp labview 5.1

[MarvinTsoi]

I've setup the 6034e, cb-68lp, and labview 5.1 with a lm34 temp sensor for the cjc.  I thought it was working fine but i noticed when my thermocouples are exposed to room temperature that it wasn't at the same temp as the cjc at room temp.   It is about 5 fahrenheit higher. My thermocouples are GRSE with no bias resistors.  They are uninsulated and approximately a few inches long for testing purpose.  

I also noticed that my data fluctuates +/- 1 fahrenheit every data point.  Is this normal?  Is noise causing this fluctuation?

I there a way to fix/improve my situation without resorting in purchasing better equipment?

[Jason_D]

Hi Marvin,

According to this document, the lm34 has 1 degree fahrenheit accuracy there for fluctuating by 1 degree is not unreasonable.  How long have you had this system?  It is very possible that the 5 degree difference you see is due to calibration of the sensor and/or the 6034.

[MarvinTsoi]

This system was a hand-me-down and probably has not been used for about a year before me.

I would understand the inaccuracy of 1degree but is it reasonable that it fluctuates +/- 1 degree (as in, if it's -1 from the average, it can jump to +1 above the average, total of 2 degrees)  every data point i collect. 

I attached a small sample of what i am seeing.

[Margaret F.]

Hi Marvin,

After looking at your data, it seems that it mostly stays within a +/- 0.5V range. There are a couple of outliers and these could be coming from noise in the line or a board or sensors that are out of calibration. My first recommendation is to run a self-calibration from Measurement and Automation Explorer by right-clicking on the 6034e under Devices and Interfaces and selecting Self-Calibrate. You should do this with nothing hooked up to the board. You can see when your 6034e was last calibrated by clicking on the board under Devices and Interfaces in Measurement and Automation Explorer and going to the Calibration tab at the bottom of the screen. Our boards are guaranteed for one year. If you would like more information on getting the board calibrated at NI, please visit ni.com/calibrate. I'm not sure of the manufacturer of your sensors, but I did find this document which guarantees accuracy at or above a certain temperature. Check the specs of your sensor to see if that applies here as well.

Regards,

Message Edited by Margaret F. on 12-22-2008 04:55 PM

Message Edited by Margaret F. on 12-22-2008 04:55 PM

[MarvinTsoi]

I do not see the self-calibrate option.  I have MAX 2.0, all i see when i right click on the 6034e, is the test panel or properties.

[MarvinTsoi]

I've done a self calibration on my setup and it seems to correct the 5 degree offset, but it still doesn't help with the fluctuations.

I've used the Test feature for the channel with the thermocouple with a constant CJC temperature value and i'm still getting a +/- 1 degree fluctuation. 

I believe i am using a shielded cable, it is 2 meters long.  I've moved it around the room and found that it does not affect readings, i've even changed the thermocouples both type and length, and still no change in the fluctuations.

Do i need a signal conditioning hardware?  Is it possible to get steady readings using my setup without signal conditioning?  What am i doing wrong?

[RT4CY]

Hi MarvinTsoi,

What physical pins are you wiring your thermocouple to?  Since you mentioned that your sensor is a grounded signal source, you should be wiring your positive lead to AI and your negative lead to AI Sense.  This should correspond to pins 68 and 62 (AI 0, AI Sense).  You should not wire to AI Gnd since this may introduce a ground loop.  Check the following tutorial for more information. 

Do you have more than one sensor wired to the DAQ device?  If so, let's try with just one sensor to see if we get these fluctuations.  I may have to read up a bit on your sensor to see if this is expected.  Have you worked with this hardware before and have seen better results?

A simple test for the DAQ device is to put a known voltage across the AI channels.  You can bring in a known voltage from any battery, and wire a differential input.  This is a common troubleshooting procedure to verify that our DAQ device can read a voltage properly.  If you have a function generator, that would be even better (since we can also check an AC signal).  You might also try different channels on the DAQ device, or different hardware (DAQ device, terminal block) altogether (if that is an option).

Also, I believe you mentioned that this fluctuation happens when reading room temperature.  Are you saying that temperatures above/below ambient are not showing fluctuation?  Say you are measuring a hot water bath at 100F, you do not see a temperature fluctuation of 1F? 

Thanks, look forward to your next post!

[MarvinTsoi]

My original Thermocouple was a grounded signal source.  I have moved locations in an attempt to see if this was noise from my environment.  Now i am using a thermocouple sensor that is not grounded, there is no other sensors or wires connected to the board.  So it is connected as a referenced single ended sensor at pins 68 and 67.  Originally i have tried boiling water, ice water, room temp water and still room air.  All of them have identical fluctuations.  Now i attempted to see if updating the software and computer will help, it resulted in worst fluctuations, as seen in the attached picture, thermocouple is placed in room temperature water.  Maybe my VIs wrong?

I have another temperature sensor, LM34, which requires a 5V input voltage.  This is connected to channel 15, pin 23, and this is giving me correct voltage reading.  I mean the output voltage converted to temperature is correct.

[RT4CY]

Hi MarvinTsoi,

Thanks for the reply, and the image!  I did not realize you weren't using a Temperature task.  Before going further, I strongly recommend you take a look at the following link.

Getting Started with NI-DAQmx: Main Page

This is a great resource for users who are new to NI-DAQmx.  I believe you may have an issue with the Input Range.  This is probably causing your resolution to be insufficient for the small voltage values that your thermocouple is giving you.

Since you are now using a floating signal source, you need to use bias resistors and a differential input configuration.  We have several on-line resources available that will help you understand why this is necessary.

Why Do I Need to Ground My Thermocouples When I Am Already Measuring Them in Differential Mode?
Reducing Thermocouple Noise to Improve Measurements
Taking Thermocouple Temperature Measurements

I am not sure what you are doing in the formula node.  My suggestion is that you use a temperature task.  For troubleshooting, let's do this in MAX first.  Simply right-click on your device and choose Create Task.  You will choose Acquire Signals»Analog Input»Temperature»Thermocouple.  But please make sure that you use the bias resistors, as you not get reliable measurements without them.  There is also quite a bit information about this in the tutorial I linked in my previous post (Field Wiring and Noise Considerations for Analog Signals)

[MarvinTsoi]

I have done what you have suggested, placed the thermocouple in differential, pins 68,34.  With resistors of 100k ohms from each lead to the ground at pin 67.  Using a task in MAX i have aquired the following graph at 1000 samples at 1000 Hz.  Maybe my resistors are connected incorrectly.