How can I calculate the effect of accuracy on voltage for a 4-20 mA current loop for a 500 Ohm 0.1 % shunt resistor?

How much is the accuracy affected by this resistor, I mean.


The tolerance of the resistor directly affects the voltage across it, all other things being equal.

So at 20mA, a 0.1% error in resistance will affect the voltage across the resistor by 0.1%, which is 0.125% of span at 20mA (worst case). Span is 20mA - 4mA = 16mA.

At 4mA it's 0.1% of 4mA, or 0.025% of span.

Real current sources also have some output resistance (effectively in parallel) so there will be an effect for a non-zero (or differing from calibration value) load resistance. Usually that output resistance is positive so the voltage across the resistor is a bit less than would be otherwise expected. It is caused by the finite gain of the feedback loop.

Both of those errors are added to whatever error is in the signal to begin with, and may increase or decrease the net error. Worst-case they make it worse, of course.

  • \$\begingroup\$ The reason I asked because I have this current output sensor: crlsensors.com/prodDocs/sa-107aic.pdf And its accuracy is given as +0.01% F.R under resolution section. They didn't mention the type of resistor they used to measure the accuracy on the data sheet. For total accuracy/uncertainty should I add the accuracy you calculated to the accuracy they give in the datasheet? \$\endgroup\$ – panic attack May 4 '18 at 9:23
  • \$\begingroup\$ No it's not 0.01%. The gain accuracy is +/-1% of reading and the zero accuracy is an additional +/-0.31% of span , with a temperature drift (maybe an additional +/-1% over the operating range), so a 0.1% resistor isn't adding much error. You may be able to calibrate the resistor and the full-scale sensitivity error out if you know g and fiddle with it to get the maximum reading. For data sheet purposes the resistor is assumed to be perfect. The output resistance I mentioned is 50M. It's probably calibrated with something like your 500 ohms. Also frequency roll-off/peaking: +/-5% error to 50Hz \$\endgroup\$ – Spehro Pefhany May 4 '18 at 9:31
  • \$\begingroup\$ The +/-200ppm/K bias drift spec means you won't gain much from using a $20 0.01% resistor vs. a $0.25 0.1% 10ppm resistor. \$\endgroup\$ – Spehro Pefhany May 4 '18 at 9:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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