B&R's Temperature Input Module

[hisaurabh]

Hello,

Our machines are using B&R's Temperature Input Module X20AT6402which takes the inputs from a theromcouple connected at a point at our machine. The measured values is being displayed on the HMI.

There have always been a difference of 2~3 degreeC between the value I directly measure at the point where the thermocouple is connected and what is being displayed on the HMI- which is now worrying me. For example, if I measure the temp value by a multimeter on the point where the thermocouple is connected I get 37.9degreeC but the value being displayed on the HMI is 39.9degreeC.

Our earlier process would allow this much variation but the newer one doesn't. Can somebody help me in understanding why there is this difference and what is needed to be done?

I have done following troubleshooting steps yet :

(1) Changed thermocouple many times
(2) Changed the male-female connectors between thermocuple and Input module.
(3) Changed all the cables.
(4) Even replaced the Input module with new one couple of times.

But all in vain and I still get the variation.
Please help.

Thanks,
Saurabh

 

[gclshortt]

I see from the literature that it has an internal terminal temperature compensation factor. This is sometimes also referred to as a cold junction compensation.
Are your thermocouples being directly wired to the card? If extensions to the wires are made this needs to be in the appropriate wire type.

Regards,
Garry
http://www.accautomation.ca

 

[shooter]

a thermocouple is measuring the temperature difference between the two ends of the wires. On the card is also a temperature sensor (Pt100 or LM35 or a diode or anything but a tC.
Then the sensor and your TC are added to find the temperature on the end of the wire.
So calibration is needed.
A TC is never better as 2 C, but after calibration around 0.3 is achievable.

 

[danw]

A) You should be using RTD's for temperature measurements at these values.

1) Thermocouple measurements require two temperature measurements - the thermocouple gradient measurement and the cold junction measurement.

a) The error in cold junction measurement is usually greater than the 0.2 - 0.3 Deg C error you're seeing.

b) Even the special limit of error on the standardized ANSI tables has errors on the magnitude of plus/minus 1.1 Deg C for J/K elements

RTD's have much lower inherent uncertainty than thermocouples for that level of accuracy.


B) What are comparing with/to?
> I directly measure at the point where the thermocouple is connected
> I measure the temp value by a multimeter on the point where the thermocouple is connected I get 37.9degreeC

A 0.2-0.3 Deg C difference between two measurements by different devices and different sensors is not uncommon because you are probably measuring 2 different physical points and seein the thermal gradient between them or the thermal gradient induced by making a measurement where heat is drawn away adding a cooler mass to the surface.

How are the permanent T/Cs installed? threaded into metal? Inserted into a chamber?

How is your mulitmeter measurement made? Which element? T/C? RTD? What style? Surface contact? pointed tip pressed against the object?

Do you let your multimeter equilibrate so the terminal temperatures are read correctly by the cold junction?

Was the Multimeter's cold junction measured and calibrated at the last calibration?


C) At a minimum, you might consider using dual element thermocouples and using one for the system and the other for your multimeter reading, but the cold junction error on the Multimeter might well show an error like you're seeing. Again, think RTD.

 

[danw]

If
- your multimeter is a traceable, certified and calibrated
- you are truly measuring the same point with the multimeter as the embedded thermocouple

then consider the measurement an in-situ calibration of the embedded thermocouple and apply the complement of the measured difference to the PLC measurement as a calibration correction factor.

If the calibrator measurement is 0.3 Deg C low, add 0.3 Deg C.

If the calibrator measurement is 0.3 Deg C high, subtract 0.3 Deg C.

But be prepared to defend the assertion of "making the same measurement". I'm not sure that you are. But if you are, applying a calibration correction factor to compensate for either sensor error or AI error is legitimate.