[SOLVED] how to remove the lead resistance error for RTD measurement

[ud23]

I have built the circuit for temperature measurement using the constant current source.Problem is that if lead resistance comes in the circuit then it gives the error in the temperature measurement.I want to achieve a 0.1 deg Celsius accuracy.

I am trying to do is that measure the voltage across the RTD and then Voltage across the RTD 2 and 3 terminal.But for higher temperature values its measurement not comes accurate.

Please tell me what am I doing wrong here?

https://obrazki.elektroda.pl/1369039500_1429016673.jpg

 

[KlausST]

Hi,

look at the first OPAMP in the circuit of this thread:

https://www.edaboard.com/threads/335060/

Klaus

 

[ud23]

I think you are pointing out at POST #1 circuit I have used the Same circuit before the Filtering stage opamp.

 

[Vbase]

In your circuit the positive side of the RTD has twice the error caused by the wires and it goes into the none-inverting input which has gain of 2 that means you get 4 times the error at the output of the amp.
You need to connect the positive of the RTD to the inverting input of your differential amp and the errors will cancel out. You will get negative voltage output relating to RTD value.
The alternative circuit to get positive output is to use 2 current sources.

 

[KlausST]

Hi,

the connection to inverting and non inverting inputs are mixed up.

Klaus

 

[Vbase]

Another way is to increase R6 to 300K and to add 100K from none inverting input to ground. This will cancel the error but will reduce the gain by 4.

 

[crutschow]

For accurate 4-wire measurement you should use an instrumentation amp IC which has a true differential input with selectable gain and high common-mode rejection.
You could build your own using op amps but they suffer from offset problems due to resistor tolerance mismatch.

 

[ud23]

I have made changes as you all suggested but at higher temperature it still have error of .2 or .1 deg error in the measurement.Is there any other way to remove this error.

Thanks for your help all of you.

 

[KlausST]

Hi,

Error sources i see:
* current is not exactely 1.0000mA
* input impedance of opamp circuit
* offset voltage of opamp
* bias current of opamp ( could be improved with 50kOhms at vin+ instead of 115k)
* offset current of opamp input
( i don't have the opamp parameters in mind to check the magnitude of errors)

But an error of 0.2 °C at 850 °C is not bad.
If you need better precision, then use a true instrumentation amp, as crutschow suggested. And mind to use 0.01% tolerated resistors.
Are you sure that the RTD has that good precision?

Klaus

 

[ravindragudi]

The self heating of the RTDs will also be a factor of inaccuracy. Try to reduce the the source current to it from 1mA to 0.1mA if feasible.
Also as suggested by others, use of INA (from TI or Analog) would help reduce error contribution from signal conditioning circuits.
Proper isolation of the PCBA (circuit) from the measurement environment. Only the RTD with the probe should be exposed to the measurement environment - if possible.

 

[Vbase]

With the results you get you can apply for a job with NASA.

 

[ud23]

No. That is just results of simulations and I am sure that during my real test this results would be different and I will get more error due to the facts mentioned here.

So I am not that talented to have JOB in NASA.Many more knowledgeable and talented Guys in this forum.

 

[FvM]

You claim to observe a temperature error without showing any results. How can we know that you are applying the correct RTD curve? It's not plausible why the lead resistance effect should increase at higher RTD temperatures.

 

[ud23]

Yes respected sir FvM, you are right but from the simulation results I get and observed with Proteus simulation I would assume their will be some more error in real world.As you said no error in the simulation as you can check the attached result file.

But considering the ADC gain error,Op amp offset error and resistor tolerance mismatch error their should some difference from simulation and real world circuit results.

So I try to get more accurate result in simulation.

 

[FvM]

After reviewing the table, I don't understand which errors you are talking about.

There's however a serious problem of your circuit that didn't come into view yet. LM134 is a PTAT (proportional to absolute temperature) current source, it shouldn't be used for this application. Apart from this point, the finite current source output impedance is generating the largest error contribution, but the current measurement resolution is too low to recognize it.

 

[ud23]

thank you Sir,I was talking about the error which will come when I'll test the circuit with real world parameters.

And what is the best replacement for the lm134 ?Is there any IC which have less temperature?

 

[fouwad]

See microchip application note AN687