Which RTD is better for noise-reduction: 100 ohm or 1k ohm?

[AAO]

Hi all,

We are planning to use RTD (Resistance temperature detectors) for our system which basically are temperature sensors that contain a resistor that changes resistance value as its temperature changes. The RTD can come in two nominal values: either 100 ohm or 1k ohm. Knowing that the RTD work by current excitation and the sensor will fed from the data acquisition by 1 mA for the 100 ohm RTD, and by 0.1 mA for the 1k ohm RTD.

My question is regarding which one might be less prone to receive noise from the environment (that may contain a motor, variable frequency drive...etc)?

Thank you.

 

[Tom.G]

Lower impedance always has less voltage noise pickup.

 

[AAO]

Thanks Tom!

Would you please explain a little bit, or maybe share a related link that may explain this?

Thank you.

 

[Tom.G]

Well, let's see.

The impedance of free space is approx. 377 Ohms (https://en.wikipedia.org/wiki/Impedance_of_free_space).

Assuming the noise is radiated energy from the source, the impedance of the receiving circuit acts as the second resistor of a voltage divider. The lower the value of this second resistor, the lower the voltage drop across it.

 

[Baluncore]

If the RTD is at the end of a long twisted pair cable, then cable resistance may be a problem. That is because the voltage is returned through the same wires that the current loop to the RTD uses. 1mA in RTD100 gives 100mV. 100uA in RTD1K0 gives 100mV. The effect of long line resistance on voltage will be bigger by a factor of 10 with the RTD100 than the RTD1K0.

The transmission line impedance of twisted pair cables is around about 100 ohms, so with the RTD100 as a load, the line will be better terminated and not ring as much, when switched or picking up noise, as would an RTD1K0.

If you use two twisted pairs in parallel at the RTD100, you can send the 1mA regulated current down one pair and return the voltage difference across the RTD on the other. That will eliminate cable resistance from the equation. With that configuration you can also use a ratiometric converter and reference resistor at the instrument end of the cables. It is a good application for Cat-5 cable.

You might consider a balanced RC filter at the received voltage end to reduce any RF noise picked up on the line. I would terminate the line by using a couple of 56R resistors from the voltage return pair to a polyester capacitor of about 100nF where the voltage will be measured.

 

[AAO]

Well, let's see.

The impedance of free space is approx. 377 Ohms (https://en.wikipedia.org/wiki/Impedance_of_free_space).

Assuming the noise is radiated energy from the source, the impedance of the receiving circuit acts as the second resistor of a voltage divider. The lower the value of this second resistor, the lower the voltage drop across it.

Many thanks Tom for the nice explanation

 

[AAO]

If the RTD is at the end of a long twisted pair cable, then cable resistance may be a problem. That is because the voltage is returned through the same wires that the current loop to the RTD uses. 1mA in RTD100 gives 100mV. 100uA in RTD1K0 gives 100mV. The effect of long line resistance on voltage will be bigger by a factor of 10 with the RTD100 than the RTD1K0.

The transmission line impedance of twisted pair cables is around about 100 ohms, so with the RTD100 as a load, the line will be better terminated and not ring as much, when switched or picking up noise, as would an RTD1K0.

If you use two twisted pairs in parallel at the RTD100, you can send the 1mA regulated current down one pair and return the voltage difference across the RTD on the other. That will eliminate cable resistance from the equation. With that configuration you can also use a ratiometric converter and reference resistor at the instrument end of the cables. It is a good application for Cat-5 cable.

You might consider a balanced RC filter at the received voltage end to reduce any RF noise picked up on the line. I would terminate the line by using a couple of 56R resistors from the voltage return pair to a polyester capacitor of about 100nF where the voltage will be measured.

Many thanks Baluncore for the nice explanation.

In fact to elimnate any error related to the lead resistance, we were planning to use the 4-wire RTD arrangement which I think correlates exactly with what you said about using pair of twisted cables. Using this 4-wire configuration (shown below), as you confirmed will increase the accuracy and reliability of the resistance being measured, and the resistance error due to lead wire resistance shall be zero.

Probably now, my concern: Which one might be less prone to receive noise from through EMI, a 4-wire 100 ohm RTD OR a 4-wire 1K ohm RTD ? According to Tom's explanation above, the 100 ohm RTD shall be less in receiving external noise. Is there any estimate of how much would the 1K ohm pick up noise more than the 100 ohm, and how can we judge if this difference can be negligible?

The solo advanatge I can see so far of the 4-wire 1K RTD over the 4-wire 100 ohm RTD is just better resolution (it can detect smaller change in temperature)?

 

[Tom.G]

@Baluncore had some good tips.

You are certainly on the right track for accuracy by using 4-wire RTDs.
Other things to consider for high accuracy are:

• Operating Temperature - High temps need Kapton or Teflon insulation.
• Damp or Wet environments (especially over a few tens of feet) - Generally need Teflon insulation for 1% or better accuracy (PVC absorbs too much moisture).
• Use twisted, shielded cable for maximum noise immunity. Ground the shield at only one end, usually at the electronics end.
• Use a continuous length of wire if at all possible. If you MUST splice, use screw terminal blocks and tighten securely.
• Put all wire runs in metallic conduit, preferably steel if in a high magnetic field environment, i.e. near high HP motors, transformers, or power lines supplying them.
• Use RTDs that are guaranteed interchangeable to the desired accuracy - for when you have to replace one.

There may be other gotcha's depending on your specific application, which you haven't yet shared.

 

[Baluncore]

The solo advanatge I can see so far of the 4-wire 1K RTD over the 4-wire 100 ohm RTD is just better resolution (it can detect smaller change in temperature)?

Since both devices will have the same output voltage, neither will have an accuracy advantage over the other. Both have a slope that passes through absolute zero, so accuracy will be determined by the calibration accuracy and the 0°C offset reference stability.

You have not specified your application or the accuracy you require. If you are multiplexing or switching the current through the RTD then self heating may be a problem. Self heating will be 10 times greater in the RTD100. The W=I²R effect will be small. For an output of 100 mV;
RTD100 @ 1mA dissipates 100 uW. RTD1K0 @ 100uA dissipates 10 uW. The thermal coupling of the RTD to the thermal mass and the thermal capacity may be important.

Excitation with 1 mA is not as efficient as with 100uA since current will probably come from a linear regulator and generate 100 mV in both cases. To compete, instruments require progressively lower power consumption for longer battery life.That is the primary incentive to use RTD1K0.

The PT100 has always been the standard RTD.
The RTD100 will better match the transmission lines and have a lower voltage noise.
Use the RTD100 unless you are powering the unit from a limited power supply.

 

[AAO]

@Baluncore had some good tips.

You are certainly on the right track for accuracy by using 4-wire RTDs.
Other things to consider for high accuracy are:

• Operating Temperature - High temps need Kapton or Teflon insulation.
• Damp or Wet environments (especially over a few tens of feet) - Generally need Teflon insulation for 1% or better accuracy (PVC absorbs too much moisture).
• Use twisted, shielded cable for maximum noise immunity. Ground the shield at only one end, usually at the electronics end.
• Use a continuous length of wire if at all possible. If you MUST splice, use screw terminal blocks and tighten securely.
• Put all wire runs in metallic conduit, preferably steel if in a high magnetic field environment, i.e. near high HP motors, transformers, or power lines supplying them.
• Use RTDs that are guaranteed interchangeable to the desired accuracy - for when you have to replace one.

There may be other gotcha's depending on your specific application, which you haven't yet shared.

Many thanks Tom for the valuable tips! Our application is a vapor compression refrigeration cycle that involves measuring temperatures inside the pipes. The temperature media is carbon dioxide. The temperatures are relatively high (~ 200 °C). The test rig located inside a building within a research lab (So, I don't think our environment is wet). The test rig has a compressor that is run by 230 V AC compressor, and variable frequency drive, plus other components. There might be also two portable chillers that may run at 480 V, within their enclosure there is of course compressor that run with AC motor. Based on that, do you think this environment is a high magnetic field one?

 

[AAO]

Since both devices will have the same output voltage, neither will have an accuracy advantage over the other. Both have a slope that passes through absolute zero, so accuracy will be determined by the calibration accuracy and the 0°C offset reference stability.

You have not specified your application or the accuracy you require. If you are multiplexing or switching the current through the RTD then self heating may be a problem. Self heating will be 10 times greater in the RTD100. The W=I²R effect will be small. For an output of 100 mV;
RTD100 @ 1mA dissipates 100 uW. RTD1K0 @ 100uA dissipates 10 uW. The thermal coupling of the RTD to the thermal mass and the thermal capacity may be important.

Excitation with 1 mA is not as efficient as with 100uA since current will probably come from a linear regulator and generate 100 mV in both cases. To compete, instruments require progressively lower power consumption for longer battery life.That is the primary incentive to use RTD1K0.

The PT100 has always been the standard RTD.
The RTD100 will better match the transmission lines and have a lower voltage noise.
Use the RTD100 unless you are powering the unit from a limited power supply.

Many thanks Baluncore for your feedback!

The RTD that we plan to use belongs to Class A, which has accuracy defined as: Class A = ±(0.15 + 0.002* t) [-30 to 300°C] (Where t is temperature). The data acquisition module is either NI 9216 for 100 RTD OR NI 9226 for 1K RTD. Both NI devices have similar temperature accuracy, but the offset error for 1K RTD is of factor of 10 compared to 100 RTD (as shown Page 5).

We won't be switching or multiplexing, these NI modules provide the excitation current (1 mA or 0.1 mA) per channel (or per RTD). But I agree with you that this is another advantage for the 1K RTD is that it has less self heating power. Here the self heating index (SHI) comes into play, which is the ratio of resistance changes (in ohms) to unit electric power generated in the RTD sensing element (in mW) as the result of application of electric current (EX: SHI = 25 mW/ °C; thus if you have self heating power=25 mW, this will give you 1 °C error). The SHI should be provided by the RTD manufacturer, but sometimes they don't do so. If we assume same SHI for both RTDs, then I think that means that 1K RTD might have better accuracy over 100 RTD in that regard.

But still as you and Tom confirmed, the 100 RTD will have lower voltage noise.

If you have further thoughts based on our application, please let me know.

 

[jim hardy]

but the offset error for 1K RTD is of factor of 10 compared to 100 RTD (as shown Page 6).

Observe:
1. They give that error in OHMS not degrees and your Pt1000 has tenX the ohms per degree,
...so i think they're equivalent.
2. Offset Error shown is for 3 wire measurement and you're using 4 wire ?

old jim

 

[AAO]

Observe:
1. They give that error in OHMS not degrees and your Pt1000 has tenX the ohms per degree,
...so i think they're equivalent.
2. Offset Error shown is for 3 wire measurement and you're using 4 wire ?

View attachment 209659

old jim

Thanks Jim!

1- Yes, you are correct! they are equivalent.
2- I was mistaken in the page number, it should be page 5 (for 4-wire). I have corrected that on my previous reply too. Thanks for the catch!

 

[Tom.G]

The test rig has a compressor that is run by 230 V AC compressor, and variable frequency drive, plus other components. There might be also two portable chillers that may run at 480 V, within their enclosure there is of course compressor that run with AC motor. Based on that, do you think this environment is a high magnetic field one?

Depends on what conduit material the power wiring is in and the proximity to the RTD wiring. If the RTD wiring is parallel to and inches from the power wiring, that's definitely a problem. I have successfully put a large VFD and RTD electronics in opposite ends of a five foot wide enclosure. The power stuff went in and out one end and the instrumentation out the other. If your control enclosure doesn't allow such separation, consider a second enclosure for the instrumentation.

Other tips that can help are to twist each set of power conductors so the fields will partially cancel each other; this may require the next larger size conduit. And wherever possible, route the RTD wiring perpendicular to the the power wiring.

If the rig is assembled enough to power-up, you can run a pair of wires, without conduit or shielding, using the proposed routing. Connect a resistor to simulate the RTD at the far end, and an Oscilloscope, headphones, or audio amplifier and speaker, to the near end. Then change the routing around for minimum pickup.

p.s. I just looked at the 9216 datasheet. Using the High Resolution mode, the built-in filtering provides at least 85db suppression of power line frequency and harmonics. That should make the above precautions largely unneeded; I would still stick with the steel conduit though. The 9216 High Speed mode on the other hand, has no filtering. If using High Speed mode, pay attention to all the above.

 

[AAO]

Depends on what conduit material the power wiring is in and the proximity to the RTD wiring. If the RTD wiring is parallel to and inches from the power wiring, that's definitely a problem. I have successfully put a large VFD and RTD electronics in opposite ends of a five foot wide enclosure. The power stuff went in and out one end and the instrumentation out the other. If your control enclosure doesn't allow such separation, consider a second enclosure for the instrumentation.

Other tips that can help are to twist each set of power conductors so the fields will partially cancel each other; this may require the next larger size conduit. And wherever possible, route the RTD wiring perpendicular to the the power wiring.

If the rig is assembled enough to power-up, you can run a pair of wires, without conduit or shielding, using the proposed routing. Connect a resistor to simulate the RTD at the far end, and an Oscilloscope, headphones, or audio amplifier and speaker, to the near end. Then change the routing around for minimum pickup.

p.s. I just looked at the 9216 datasheet. Using the High Resolution mode, the built-in filtering provides at least 85db suppression of power line frequency and harmonics. That should make the above precautions largely unneeded; I would still stick with the steel conduit though. The 9216 High Speed mode on the other hand, has no filtering. If using High Speed mode, pay attention to all the above.

Many thanks Tom for these valuable tips! It will definitely come into play when I begin assemble things. Thanks again!