USB 6009 thermocouple noise

[johndg]

I'm using NI-DAQmx 7.5, Labview 7.0, and a USB-6009.

I'm trying to get a more accurate reading from my j-type thermocouple.  It is now distorted with noise and jumps rapidly +/- 20 degrees around room temp without any heat being applied.  When heat is applied, it climbs as it should but with the same variation.  I haven't moved to the labview side of it yet, I'm still trying to get a usable reading in MAX.  I have it wired as a differential input with a floating signal source.  I've installed 10k ohm resistors for bias current return as instructed.  Installing the resistors helped but still leaves me with the readings shown in the attached pic.  I've also tried grounding my USB 6009 with no results.  Please help. john

[Kenny_K]

I have tried to do the same with a USB6008, but I got the same results initially.  We figured out that what you had to do was amplify the output of the thermocouple slightly (along with some other resistors, etc) in order to get more accurate results.  The signal out of the thermocouple is usually very weak and is suceptible to any sort of noise in the system.

Check out this post about thermocouple and the usb6008/9

If you want to be able to use the USB6008/9 to measure temperature, try using the AD590 chip by Analog Devices.  It is a really nice chip and can be read directly by the USB6008 and some simple math.  I have a program written for the AD590 and the USB 6008/9 if you would like to look at it. 

Kenny

[2c worth]

You may want to take a look at this US$79 low cost USB DAQ module EMANT300.
You need the 24 bit ADC and programmable gain for measuring the uV readings
from the thermocouple. It comes with LabVIEW 7.0 examples

http://www.emant.com/318004.page


"johndg" <x@no.email> wrote in message
news:1137100256211-310101@exchange.ni.com...
> I'm using NI-DAQmx 7.5, Labview 7.0, and a USB-6009.
> I'm trying to get a more accurate reading from my j-type
> thermocouple.&nbsp; It is now distorted with noise and jumps
> rapidly&nbsp;+/- 20 degrees around room temp without any heat being
> applied.&nbsp; When heat is applied, it climbs as it should but with the
> same variation.&nbsp; I haven't moved to the labview side of it yet, I'm
> still trying to get a usable reading in MAX.&nbsp; I have it wired as a
> differential input with a floating signal source.&nbsp; I've installed 10k
> ohm resistors for bias current return as instructed.&nbsp; Installing the
> resistors helped but still leaves me with the readings shown in the
> attached pic.&nbsp; I've also tried grounding my USB 6009 with no
> results.&nbsp; Please help. john
> &nbsp;
> &nbsp;
> &nbsp;
> &nbsp;
> &nbsp;
> &nbsp;
>
>
> thermocouple reading.jpg:
> http://forums.ni.com/attachments/ni/170/160861/1/thermocouple reading.jpg

[Otis]

Hi John,

What you are running into is a problem with the actual voltage resolution of your USB-6009.  This device has 14-bits of resolution over a minimum range of +/- 1V.  That means that the voltage resolution is (1 - (-1))V / (2^14) = 2V / (16384) = 122uV.  While this is an excellent resolution for such a low cost USB device, you will still notice problems when trying to make a thermocouple measurement.  The reason for that is that Thermocouples make a very small output voltage (see image below).



Common thermocouple types have the following response:
K type: 41 microVolts per degree Celsius
J type: 56 microVolts per degree Celsius
T type: 52 microVolts per degree Celsius

There are a few options for ways to get around this limitation.

• Good: You could read the signal in as a voltage and then manually convert that voltage into a temperature.  This would be the least accurate method, but it should allow you to average out your signal and then use the Convert Thermocouple Reading.vi to show the voltage as a temperature.
  
• Better: If you want to keep using the USB-6009 to measure thermocouple, what you really need to do is first amplify your signal before you connect it with your device.  This should increase your resolution of the thermocouple allowing you to get more accurate temperature specs.  This will require some more complex programming to properly scale your measurements, but it will give you more accurate readings for relatively little cost.
• Best: Opt for getting a USB device that is designed for taking temperature measurements such as the USB-9211A.  This device would give you a voltage resolution of 160mV/2^24 = 9.54 nV!  That's over 12500x the resolution of the USB-6009 that you currently have.

Hopefully the above information helps you better understand what is going on and why.  If you have any other questions feel free to post them here.

Message Edited by Otis [DE] on 01-13-2006 11:37 AM

Message Edited by Otis [DE] on 01-13-2006 11:38 AM

[johndg]

Since that level of programming is beyond me, looks like I'll be purchasing a different DAQ card.  Thanks for the advice. 

John

[David_49]

Hello Kenny,

I'm intersted in having your program for the AD590 and USB6008.

Can you send it to me at this email : dlemarechal@bmp.fr ?

Thanks in advance,

David.

[Adzel]

In the above posts could someone please explain to me what the following means:

"I have it wired as a differential input with a floating signal source.  I've installed 10k ohm resistors for bias current return as instructed. "

I realize this is over a year old. I am new to LabVIEW. I have my thermocouples installed differentially. When I put the 10 k resistor across the positve thermocouple wire input to a GND the readings did improve. Obviously this cuts down on the noise I was seeing. The readings look better. But bias current return seems arcane to me.

Please explain if possible (or direct me to an appropriate link)

Adz

[Jarrod_B.]

Hello Adzel,

Please refer to the following KB.  It explains the concept of bias current.

Why Do I Need to Ground My Thermocouples When I Am Already Measuring Them in Differential Mode?

Let us know if you have any further questions.

Best regards,

[josh2.0]

Hello Kenny K, I actually hav purchased the AD590 and still get random values when tryiing to read the temperature.

 could I take a look at that program you said you have for the 6008/9??

Thanks,

Josh

[josh2.0]

 I would be interested in having this program.

Here is my email : joshua.goddard86@gmail.com

Thanks,

Josh