RTD data acquisition

[fmazzelli]

Hi, 

Our lab need to conduct really accurate measurements of temperature differences (between 1 and 2 °C)

We have installed RTD sensors in 1/10 din class (accuracy lower than 0.08 °C  at our temperature conditions, between 0°c and 100°C)

At the moment we are acquiring the data through a National Instruments cFP-RTD-124 module mounted on a cFP-1804 interface. This gives typical errors of +-0.15 °C which becomes the main source of error. Hence, we would need a DAQ system which is accurate at least as much as the RTD.

Do you have anything like that?

Thank you

Best regard

Federico

[Henrik_Volkers]

For temperature differences you can use your equipment if you do a individual calibration of  your sensors. 

Calibration means, that you compare all sensors at equal temperature(s) in the range of interest, and use this result to compensate your readings.

The resolution of the RTD-124 is 0.016°C, so I would assume you could get down to 0.05°C ... and 50mK is already challenging for the rest of your setup (heat transfer, etc) 

[fmazzelli]

Hi, thanks for the answer,

The problem is that I also need the absolute value of temperature from each reading.

Do you think it would be enough to put the two RTD next to each other in a cold and hot bath (0 and 100 °C) and recalibrate in such a way that they give the same correct value? 

Do you think the RTD are linear enough to calibrate just with two temperatures?

[Henrik_Volkers]

OK, please provide more (all) information about your requirements.

-absolut temp. range and resolution

-differential range and resolution

-media (in water, air,wood,...)  mass! 

-type of sensor

-coupling!

-temp. change rate

-... 

just to name some aspects you need to consider 😉

and there is still room for a lot of sources of error 😄

And always remember, the sensor will report what it feels. That might or might not be the temperature of interest!

[fmazzelli]

ok, these are our conditions:

- absolute temperature range between 0 and 100 °C, the resolution is not so important for this measures, let's say an error of +- 0.25 °C is acceptable

- the differential range is from dT=1.5 °C to higher values, we would like an error lower than 10%, that is +-0.15°C

- the 4 wire RTD are immersed in water (the information on the DAQ and RTD are in my first post)

- The measerument are supposed to be in steady conditions but we have oscillations of the temperatures due to external conditions which are not stable (we are working inside a factory). By taking the standard deviation from several acquisitions we generally find fluctuations smaller than 1% (that is, std(T)/T<1%)

Do you think we could be able to measure with good accuracy such dT?

The resolution of our system is quite high as you said, but how could we be sure of reaching a certain level of accuracy by calibrating ourselve the system?

Hope I was clear, sorry but I'm not an expert in this field 

Thanks for helping!!!

Regards

Federico

[Henrik_Volkers]

So no problem for absolute temperature.

For differential temperature: Put all sensors as close as possible and step througth your temperature range.

STEP: until the readings are stable ,

placing them into water is fine (good coupling), using a stirred water is better 😉

All readings should be within 0.2°C (guess, detailed spec calculation needed)  but have offsets and minor different slope.

Take the mean of all sensors as a best guess and graph the differences.

Repeat the measurements to get a feeling about the repeatability.   should be in the <80mK range (guess)

Use this data to correct individual sensors and calc differences

Maybe an aluminium block with holes for each sensor placed in your water ( the sensors should be immersed as far as possible , or just like you use them)

Flutiations of the setpoint for the calibration can be minimized if you place the sensors into a thermal mass and decouple it more or less with isolation from the heater, however thermal heat flux has to kept in mind 😉

Record the values 1/s ( or faster and average)  and post some data and (a) picture(s) of your calibration setup for detailed recommendations.