I'm trying to design RTD measuring for industrial use which has EMI/EMC/ESD.
I have 3 wires input to the board. enter image description here

obviously 3 wires out of the board are going to 3-wire configuration PT1000
1- 3V3 (RTD 1)
2- B (RTD 2)
3- B (RTD 3)

  • Vout is going to MCU ADC
  • in second image i attached you can see i have different resistor for op amp gain.

and what I'm thinking of is the following:
1- add capacitor for (B) before the (+) of LM358 to filter noise
2- add 3 TVS diodes to the 3 wires (to GND)
3- add current limiting 10k resistors before the RTD sensor (as shown below) 4- no 10k resistors but add PTC fuse on the 3v3 rail to the sensor to protect shorting?

enter image description here

Protection Questions:

  • is protection with TVS is required actually?
  • do you see any flaw or mistake in the circuit, any suggestion is welcomed.
  • i don't know if industrial designers protect all inputs/outputs from applying certain level of voltage like +-24v? or it always depend on the project and i don't need to think about it here. so should i think about it here?

wheatstone Questions:

  • every time the wire length (going to RTD) changes will need to change software compensation variable?
  • as you can see I'm simulating the circuit but is't practically going to work efficiently?
  • \$\begingroup\$ Better to just copy one. There are lots online :^) \$\endgroup\$ Commented Feb 24, 2019 at 13:58
  • \$\begingroup\$ I see mistakes: How the wire resistance are compensated? This is the goal of three wire connection, but I can't realize what's the purpose of yor circuit. \$\endgroup\$ Commented Feb 24, 2019 at 14:15
  • \$\begingroup\$ @MarkoBuršič i cant understand what are you refering to. its wheatstone bridge \$\endgroup\$ Commented Feb 24, 2019 at 14:23

1 Answer 1


The circuit seems flawed, as there is no well controlled voltage reference. Ti has a few good ADCs dedicated to RTD readback, but if you are using a MCU ADC a circuit like AN-1559 would be recommended.

TVS devices can be used, but they are a fairly large risk to the circuit accuracy due to leakage currents (you probably have to use higher stand-off ratings then expected). I think keeping the connections to the RTD higher impedance and clamping with low-leakage diodes is a good option, but it depends on what the external threat is.

For simulation you should do a resistance sweep instead of just transient analysis.

You probably won't have to worry much about wire-length, 3- and 4- wire help with that. But it doesn't seem critical if you are starting without RTD ADCs or very precise low-drift opamps.

It is always safe to assume that the wire-harness will be erroneously wired (I think 1-2 per 100 contacts is reasonable), so if there is 24V or whatever on the connector there should ideally be no permanent damage.

  • \$\begingroup\$ hello, i've seen that TI APPNOTE, i just skipped it because i thought its complicated and for precision, i needed somthing fast i've 1 LM358 dual in my circuit and i've spare amp of it. so i thought it will work with +/- 2C. it seems yeah i ve to look for reference voltage other wise do you think lm358 will work if the 3v3 is accurate?. for the TVS, in TI appnote, they also suggest ESD protection for inputs (i assume it as TVS). \$\endgroup\$ Commented Feb 25, 2019 at 6:36
  • 1
    \$\begingroup\$ 2C on a PT1000 looks to be around 7 ohm (temperature depending, looking at a resistance table). I think your requirement is more stringent than you expect, unless you are planning to do a lot of calibration. If I had the budget for it my preference would be a more complex (or fully integrated) solution. \$\endgroup\$
    – Zekhariah
    Commented Feb 28, 2019 at 3:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.