Thermocouple configuration in MAX

[AngryPenguin]

I'm in the process of upgrading an ancient labView system to a nice shiny new one (v4 to 2009) but am having some problems setting up my thermocouples.

The old PC runs them fine and gets sensible readings - the new one isnt so happy. I'm using the same old SCB-68 board for both and have simply upgraded the DAQ card (now a PCI-6221) and the PC.

MAX and LABView can read the voltage from the thermocouple fine and the voltage responds correctly to change in temperature of the thermocouple (ambient is 1.38v and response is 50mv/C) - however, when I try and tell it that it is a thermocouple it tells me that the room is a constant 398,000 degrees! This reading shows no response at all to change in temperature of the thermocouple.

I've got the thermocouple hooked up to AI2 on pins 65 and 31 as MAX's connection diaigram seems to suggest. Is it a case of how I have wired it up or is it my configuration that is wrong?

Thanks in advance for your help!

Pete

[jeffneuner]

not sure what you mean when you say  "when I try to tell it that it's a thermocouple" - Can you be more specific or show a snapshot of your code?

Did you create a new task in MAX for acquiring the reading? 

MAX walks you through the process of creating a new voltage input task for temperature/thermocouple.It let's you select thermocouple type (J,K,T, etc) and CJC source.  Check that your cold junction compensation (CJC) is not set to some wacky value. Try using a constant CJC set to 25.

[AngryPenguin]

Sorry yes - I wasn't clear. When I create a Max Task and say AI2 is a Voltage I get fairly normal readings but when I make a task and say AI2 is a J type thermocouple then I get the odd ones. My CJC is set to a constant 25.

[jeffneuner]

OK - sounds like MAX is obviously not scaling the signal properly -

did you try making it a K type to see how it scales?  see if MAX is scaling everything wrong, or just J-types

what version of MAX?

might be worth a call to NI

[AngryPenguin]

Mmm, all thermocouple types produce numbers of similarly wild magnitude. It's the latest version of MAX thats installed.

I'll see if I can pinch another thermocouple from somewhere tommorow and try that out.

Though I am usually very wary of customer support (and the associated waiting lines) I haven't tried NI in the past so I gave them a ring and got straight through and although they didnt help much they were very friendly and unlike most companies the engineer actually knew what he was on about which was a pleasant surprise. He said he was going to ring me back so I'll let you know how it goes.

Thanks for your help too Jeff.

[Henrik_Volkers]

50mV/°C sound like a factor 1000 to high. (coarse 51µV/K for Type J) and even at the max temp (>1200°C) for any kind of standard TC you should never read more than 80mV!!

Is there a (not existing)  1:1000 attentuation configured somewhere ??

That could also explain the CJC Temp....  

[jeffneuner]

yes - Henrik is correct - I didn't look at your units carefully

room temperature should be 1.3 mV, not volts

and change of about 50 uV per degree C

your voltage reading is off by a factor of 1000

[AngryPenguin]

Ah OK, well noticed! I'll find some alternative power sources and a voltmeter.

[AngryPenguin · Accepted solution]

Heh, I've solved it. Sadly it was me being stupid all along.

The setup of the equipment we are measuring is such that the thermocouple is about 10ft away from the PC. Its attached to a wire that goes behind some cupboards and then into the back of the PC. 

When I plugged the thermocouple straight into the board it worked fine - exactly as it should. Then when I put it back on its extension cable it went nuts again. I pulled the wire out from behind the cupboards and neatly hidden halfway along the wire is a amplifier box. It never occoured to me to check the wire - I guess even if there was an amplifier I would have expected it to be at either end! I get the impression that this kit was cobbled together by several people none of whom talked to each other so it doesnt surprise me. I guess in the old VI somewhere there must be a correction factor hidden somewhere.

Thanks for your help guys - you certainly guided me in the right direction.

Pete