RTD temperature measurement / Calibration

[themroc]

I need to meassure the the temperature in a flow quiet accurately. In order to do this I choose 4 wire RTD precision sensor. The sensor has an accuracy of < 0.06°C.
Now my problem Normally I use a data aquisition here I can use a 4 wire connection.
But in this case I would like to use a hand held device. The Provider of the sensor only provides 3 wire RTD's handheld.
I would like to use the all the 4 wires in order to keep the accuracy as good as possible.
On the internet I do not find meassurement devices only. Normally they come as meassurement and calibration device.
Which are actually quiet expensive!

Two Questions in this context:

1. For what is the calibration offered, how is it used? Is it to simulate a different temperature and than to see whether this temperature is displayed correctly in the data aquisition?

2. Does anyone know a company which offers 4 wire RTD handheld measurment devises only?

I noticed that the accuracy of those devices is not better than 0.1°C, Is it so difficult to meassure resitance properly?
Is the assumption right that when I get here an error of +-0.1°C and with the sensor of 0.06°C that in total I have got at least an error band of +- 0.16°C?

 

[MintJulep]

The advantage of using a 4-wire RTD is that it allows you to apply a current from an external source, and simply measure the voltage across the RTD. The voltage will be proportional to the resistance.

The voltage may be measured by any volt meter, which can be obtained in whatever accuracy range in needed.

 

[IRstuff]

Resistance measurement is not the only issue. The actual calibration curve itself requires fitting measured data to the model of an RTD.

If the measured temperatures are within a relatively small range, you can specifically calibrate for that range to get improved accuracy.



TTFN

 

[themroc]

IRstuff,
does it mean that even if the sensor accuracy is stated at 0.06°C at 0° that I need to do an additional calibration if I want to use it for tenperatures 20°C to 80°C?
Is it not linear in this range?

 

[IRstuff]

You need to read the specifications carefully, as well as fully understanding the calibration method.

A typical calibration might involve several specific temperatures of calibration, and the RTD model is curve fitted to those data in a piecewise fashion. Between the calibration temperatures, the non-linearity of the device itself kicks in, as well as the curve fit error of the calibration constants, as well as the actual measurement error.

Don't forget that the RTD model is a function of absolute thermodynamic temperature, so a 0.1ºC error requires less than 0.034% uncertainty in the measurement and the calibration process.

TTFN