Reading temperature with RTD PT1000

[nickagian]

Hi all!

Attached is a circuit I have designed for measuring temperature with P1K0.232.6W.A.010 RTD sensor, connected in the 4-wire configuration.

U2, U3 and R3 produce a constant current source for the RTD. Theoretically, U2 produces 1.25V and the constant current is around 1.25/3.16k=~396uA. In reality, the current that I measure is 401uA. Anyway, the voltage developed across RTD is first amplified with U5 (with a theoretical gain of [R_gain/100k] + 1 = 2 / In reality R_gain has the value of 99.5k and thus the gain is around 1.995) and then measured with a 12-bit ADC, integrated in the EM250 ZigBee SoC. The measured value is given to me directly in mV, with help of the API offered with EM250 from Ember.

To convert the result from the ADC into Temperature I do the following process:

1) The measured voltage is divided with 1.995 (the actual gain), to find the real voltage value across the RTD.
2) Afterwards, I calculate the resistance value of RTD by dividing the above calculated voltage with the measuring current.
3) To find the temperature, I use the equation of α: α=(RT-R0)/(R0*T)=0.00385

The problem I face is that the measured temperature is quite different from the real. The error is around 6deg. Celsius and I cannot find the reason for that.

Can anyone please help me with this?

 

[ericgibbs]

hi nick.
A quick point is the AD623 drawing is wrong on the inputs, is that a drawing error is it connected that way.?

 

[nickagian]

hi nick.
A quick point is the AD623 drawing is wrong on the inputs, is that a drawing error is it connected that way.?

Hi eric. No no, it is just a drawing error. Sorry for that. RTD_OUT+ is connected through R2 to pin 3 "+IN" of AD623.

 

[ericgibbs]

hi Nick,
Farnell are showing that as a 2 wire PT, how do you have it connected.??

IST|P1K0.232.6W.A.010|SENSOR, PT1000, 600°C, CLASS A | Farnell United Kingdom

 

[nickagian]

hi Nick,
Farnell are showing that as a 2 wire PT, how do you have it connected.??

IST|P1K0.232.6W.A.010|SENSOR, PT1000, 600°C, CLASS A | Farnell United Kingdom

I have mounted the sensor on a custom PCB, that allows connecting 2 wires on each 'leg' of the RTD. Is this a problem? The wires are around 3m long.

 

[ericgibbs]

I have mounted the sensor on a custom PCB, that allows connecting 2 wires on each 'leg' of the RTD. Is this a problem? The wires are around 3m long.

hi,
What is the loop resistance of the 3mtr cable.

Is the adaptor PCB like this type.?

EDIT:
Is the 6Cdeg error a fixed error over the temperature range or does the error value change with temperature?

I guess you have measured the actual voltage output of the AD623 over the temperature range to determine if the error is in the electronics or your maths.?

 

[Ubergeek63]

Hi all!

Attached is a circuit I have designed for measuring temperature with P1K0.232.6W.A.010 RTD sensor, connected in the 4-wire configuration.

U2, U3 and R3 produce a constant current source for the RTD. Theoretically, U2 produces 1.25V and the constant current is around 1.25/3.16k=~396uA. In reality, the current that I measure is 401uA. Anyway, the voltage developed across RTD is first amplified with U5 (with a theoretical gain of [R_gain/100k] + 1 = 2 / In reality R_gain has the value of 99.5k and thus the gain is around 1.995) and then measured with a 12-bit ADC, integrated in the EM250 ZigBee SoC. The measured value is given to me directly in mV, with help of the API offered with EM250 from Ember.

To convert the result from the ADC into Temperature I do the following process:

1) The measured voltage is divided with 1.995 (the actual gain), to find the real voltage value across the RTD.
2) Afterwards, I calculate the resistance value of RTD by dividing the above calculated voltage with the measuring current.
3) To find the temperature, I use the equation of α: α=(RT-R0)/(R0*T)=0.00385

The problem I face is that the measured temperature is quite different from the real. The error is around 6deg. Celsius and I cannot find the reason for that.

Can anyone please help me with this?

ah yes a platinum RTD, brings back memories... the platinum RTD is NOT linear, it is in fact represented by a VERY complex equation. https://zone.ni.com/devzone/cda/tut/p/id/3643

That said if you add some positive feedback you can linearize it. the positive feedback increases the "reference" to the current source slightly as the temperature goes up to compensate for the "lazy" hot end of the temperature curve.

dan

 

[nickagian]

What is the loop resistance of the 3mtr cable.

I'm sorry, but I do not know how to compute that? Can you give me a hint, please?

Is the adaptor PCB like this type.?

Well, yes, something like this. Actually the adaptor PCB I'm talking about includes only the RTD. It has one through-hole position for the RTD and 4 pins for connecting the external wires. The connections between the pins and the through-hole position for the RTD are implemented on the PCB. This adaptor PCB is located around 3m from the central processing unit (implemented on another custom PCB, too).

EDIT:
Is the 6Cdeg error a fixed error over the temperature range or does the error value change with temperature?

It is not exactly the same. But more or less, the error value stays about the same with temperature range.

I guess you have measured the actual voltage output of the AD623 over the temperature range to determine if the error is in the electronics or your maths.?

Yes, but in fact I want to be sure. So tomorrow morning I will do some more measurements on this, just to be sure that it is not a matter of maths and I will post the results here.

 

[nickagian]

ah yes a platinum RTD, brings back memories... the platinum RTD is NOT linear, it is in fact represented by a VERY complex equation. Measuring Temperature with an RTD or Thermistor - Developer Zone - National Instruments

That said if you add some positive feedback you can linearize it. the positive feedback increases the "reference" to the current source slightly as the temperature goes up to compensate for the "lazy" hot end of the temperature curve.

dan

dan,

Can you please explain me a little bit what do you mean with this? Actually I have also used the "more complex", non-linear equation and the difference between the result of the linear equation (at least for the room temperature range) is very little. Thus, the measurement error that I have still exists.

 

[Ubergeek63]

dan,

Can you please explain me a little bit what do you mean with this? Actually I have also used the "more complex", non-linear equation and the difference between the result of the linear equation (at least for the room temperature range) is very little. Thus, the measurement error that I have still exists.

well at room there is very little difference. you probably have tolerance issues.

start by drawing out what you have... the schematic provided is in error showing one of the inputs from the RTD to ground... I am assuming you are using a thevenin connection to the sensor but you are not showing that eitther

Dan

 

[nickagian]

dan,

ah yes a platinum RTD, brings back memories... the platinum RTD is NOT linear, it is in fact represented by a VERY complex equation. Measuring Temperature with an RTD or Thermistor - Developer Zone - National Instruments

That said if you add some positive feedback you can linearize it. the positive feedback increases the "reference" to the current source slightly as the temperature goes up to compensate for the "lazy" hot end of the temperature curve.

dan

can you please explain me how can I "add some positive feedback" in order to linearize the RTD?

well at room there is very little difference.

Yes, I understand that and finally I intend to use the second-order equation for the maths and do some linearization (if required), like you advised me in your own post.

you probably have tolerance issues.

Yes, I agree. And I try to understand the source of these errors, so that I can eliminate them. Is it acceptable to remove this offset from the final result of the computations? Of course, this may not be a very good solution, since this error is probably not linear.

start by drawing out what you have... the schematic provided is in error showing one of the inputs from the RTD to ground... I am assuming you are using a thevenin connection to the sensor but you are not showing that eitther

Dan

Ok, yes. Tomorrow I will post a more "proper" drawing. However, it is true that I use a thevenin connection. This is what I mean by saying "RTD in 4-wire configuration". Otherwise, this is all I have. All analog front-end electronics are shown in the drawing above. Anyway, I will also need to verify that the voltage measured at the output of AD623 is what I get after the ADC.

 

[Ubergeek63]

can you please explain me how can I "add some positive feedback" in order to linearize the RTD

put a resistor between the reference and the current source and another from the output of the current source back to the + input. as the output voltage goes up with temperature, so will the current supplied to the RTD.

Dan

 

[nickagian]

Ok now. I have posted the corrected schematic (rtd3.jpg), also showing the PT1000. Have in mind that the PT1000 is around 3m away from the analog front-end.

hi,
What is the loop resistance of the 3mtr cable.

I guess you have measured the actual voltage output of the AD623 over the temperature range to determine if the error is in the electronics or your maths.?

eric,

I made some more tests today and I discovered that probably the error that I notice in the final result is caused by the long wires. For example, I measure the voltage directly at the two legs of the PT1000 (at the remote PCB) to be around 448mV, which means around 29 deg. C, whereas the ADC measures 438mV (after dividing the true result with the gain of AD623), which means around 23 deg. C.

Do you have any good idea on how could I eliminate this error caused by the wires? Obviously the resistance of the wires plays an important role here. However it is quite interesting. This 10mV drop is caused by the input bias current of the AD623, isn't that correct? From the data sheet of the part I see that the input bias current is around 20nA. Thus, the resistance of the wire is calculated to be around (10mV/20nA/2=500k/2=)250k Ohm? Is this a rational value? I know nothing on this issue.
EDIT: But is seems to me that 250k ohm is a quite big value for the resistance of a 3m wire.

put a resistor between the reference and the current source and another from the output of the current source back to the + input. as the output voltage goes up with temperature, so will the current supplied to the RTD.

Dan

Dan, can you please look the "rtd2.jpg" attachment? Is this what you mean? And what about the resistor values? I guess I should calculate the change in the output reference voltage and try to eliminate this change with the correct resistors?

 

[ericgibbs]

hi nick,
Are you following the 'layout', wiring and decoupling as recommended on page #13 of the AD623 datasheet.?

I would expect the 3mtr cable resistance to be in the order of tens of milli-ohms, are you using screened cable, page 13 also covers the connection of the screen.

 

[Ubergeek63]

Dan, can you please look the "rtd2.jpg" attachment? Is this what you mean? And what about the resistor values? I guess I should calculate the change in the output reference voltage and try to eliminate this change with the correct resistors?

precisely. a 1% resistor with a 1% RTD is enough to give you a 6C offset.

 

[nickagian]

hi nick,
Are you following the 'layout', wiring and decoupling as recommended on page #13 of the AD623 datasheet.?

I would expect the 3mtr cable resistance to be in the order of tens of milli-ohms, are you using screened cable, page 13 also covers the connection of the screen.

I'm sorry eric, I cannot find any information regarding the layout at page #13 of the opamp. And moreover, I cannot understand what you mean by "screened cable". Do you mean shielded? Actually it is a twisted pair cable, part number is SMLB26A(2)S24-12814. At least this is what is written on the cable. Unfortunately, I haven't bought it myself and I do not know if this is the ordering code or not.

 

[ericgibbs]

I'm sorry eric, I cannot find any information regarding the layout at page #13 of the opamp. And moreover, I cannot understand what you mean by "screened cable". Do you mean shielded? Actually it is a twisted pair cable, part number is SMLB26A(2)S24-12814. At least this is what is written on the cable. Unfortunately, I haven't bought it myself and I do not know if this is the ordering code or not.

hi Nick,
Extract from the d/s, marked with boxes...
Screened/Shielded.

Also note what Ubergeek aka D, said about the resistor tolerances.

 

[nickagian]

Hi again after so much time.

I didn't have time to get involved with the circuit again, but I have studied it. However, I couldn't find a solution to my problem. What would you advice me to do in order to eliminate this error in my measurements? I would appreciate any help, because I am a little bit confused.

Dan, the thing about the resistor's tolerances that you said. That 1% resistors and RTD 1% can cause this error. For which specific resistor in my circuit where you talking about? For what tolerance you where talking about? At the amplification stage? Or at some other stage?

Thanks,
Nikos

 

[Ubergeek63]

Hi again after so much time.

I didn't have time to get involved with the circuit again, but I have studied it. However, I couldn't find a solution to my problem. What would you advice me to do in order to eliminate this error in my measurements? I would appreciate any help, because I am a little bit confused.

Dan, the thing about the resistor's tolerances that you said. That 1% resistors and RTD 1% can cause this error. For which specific resistor in my circuit where you talking about? For what tolerance you where talking about? At the amplification stage? Or at some other stage?

Thanks,
Nikos

at all stages... the current source, the amp, the ADC reference, everything.

you might be better off going ratio-metric it is certainly easier...