How to perform temperature sensor calibration at site ?

[syne]

How to perform Temperature Transmitter and temperature gauge Calibration at Site ? What are the tools required to perform calibration of RTD OR Thermocouple ?

 

[btrueblood]

To what level of precision?

Google "dry block calibrator" - probably one of the more portable methods. Precision will not be comparable with a lab-grade platinum RTD, but you should be able to find calibrators to get within 1 deg. C of error or less.

 

[DanEE]

As mentioned depends on what level of precision, how many points and perhaps what's available/conditions at the test site.....

In my past career with the thermocouple system we used to instrument equipment under test with a number of test points of interest, while undergoing a THA test cycle, an initial reference check with all thermocouples in an ice/slury mix to check calibration at 32F was sufficient. e.g.

http://www.thermoworks.com/learning/thermapen101_creating_an_icebath

 

[BrianE22]

Ice slurry as Dan says and boiling water for another point.

 

[ScotchWhisky]

If you have a thermowell adjacent to the RTD you can insert a (calibrated) temperature probe and check they are reading within 0.5 degC of one another. As for the transmitter you can inject a resistance at 5 points across the range of the transmitter and measure the output which you compare against the expected output ensuring the %error is within spec.

 

[roydm]

I have never seen a dry block calibrator used on site, it would be too slow in fact I'm not even sure I've seen one in an instrument shop, more of a teaching tool.
A decade box is used for RTD, you cannot adjust the probe only the transmitter.

A thermocouple calibrator for thermocouple transmitters, these used to be based on a Wheatstone bridge and a standard cell with Cold Junction Compensation but I'm surely way out of date.

Ice or boiling water as a quick and dirty check.

 

[UniCO2]

I've used ice bath and boiling water (need to know altitude/pressure for the boiling water point). For NIST traceable need a calibrated thermocouple as reference and also a calibrator like a FLUKE 726 so you get the reference TC hot along with your target (in a dry block/furnace sand) and then check both output to see if it is still in spec. They also have small sand fluid bed's with heaters etc.

http://www.omega.com/pptst/FSB.html

 

[roydm]

I can see using a fluidized bed to calibrate a number of instruments in a fixed location but you would hardly lug it up the side of a column to check an instrument in a Hazardous location

You assume your Platinum wire RTD is accurate, if not don't fudge the transmitter to make up for it.
Substitute a precision decade box and set the transmitter to give the correct output.
You cannot always get access to the temperature element for example the 10 Ohm Copper or thermistors in a motor but the Decade box is universal.

Sometimes you encounter American Curve .00392 when expecting European curve .00385 RTDs, I can see a fluidized bed being useful for that but boiling water will do just as well to identify the type.

Thermocouples don't drift either, use hot water and a thermometer to prove it's working but a calibrator for setting the transmitter.

 

[arslansharif]

For RTD use decade box and set your resistance according to temperature.Fotr thermocouple use mv source .

 

[John2025]

We use an Omega CL355A and send it out to be calibrated to maintain the calibration chain