2
\$\begingroup\$

I recently made a low side current sense circuit to test a power supply. It works well, but I am trying to understand the strange behavior that happens when I apply the negative lead of my probe to different places on ground.

When I place the positive and negative leads directly across the .1 ohm sense resistor, I get a very precise voltage measurement that I can then use to calculate the current.

If I move the negative lead further away from the negative side of the resistor terminal, yet still connected to ground, I start to get a voltage that does not accurately correspond to the current flowing through the resistor.

Can someone explain what is going on? I have been reading some Microchip articles on low-side sensing and they briefly mention that ground references can become skewed, and they use the term "ground loops" which I am not sure I completely understand.

This is a crude drawing to illustrate the problem. As you can see, when I leave the positive lead on the positive side of the resistor, but move the negative one from directly touching the negative side of the resistor to a different spot on ground, represented by the blue pen, I get discrepant results.

enter image description here

\$\endgroup\$
7
  • \$\begingroup\$ Which frequencies do you expect on your board? How big is Rsense? How much resistance do you add, when moving the ground probe away? \$\endgroup\$
    – Christian
    Commented Jun 30, 2020 at 18:27
  • \$\begingroup\$ You are only supposed to measure across Rsense and not anywhere else. \$\endgroup\$
    – user105652
    Commented Jun 30, 2020 at 18:39
  • \$\begingroup\$ @jm567 Obviously in the sketch there is no difference. So more information is needed. Perhaps you would take a photo showing both probes in the first case and a second photo showing the second case. That might be a good start. \$\endgroup\$
    – scorpdaddy
    Commented Jun 30, 2020 at 18:40
  • 2
    \$\begingroup\$ This makes perfect sense. To layout this board properly, lay down +sense and -sense as dedicated tracks. Using GND instead of a -sense track means that all currents flowing on GND inbetween the resistor and your sense input will alter that signal. GND is not 0Ω - all traces have a small resistance - so (big) currents will always skew (tiny) analog signals, even on ground tracks. \$\endgroup\$
    – rdtsc
    Commented Jun 30, 2020 at 18:56
  • \$\begingroup\$ Does the measured voltage increase or decrease when you measure farther from the resistor? \$\endgroup\$ Commented Jun 30, 2020 at 19:05

2 Answers 2

3
\$\begingroup\$

Your wires or PCB is not made of superconductor.

Real world conductors have resistance.

So when the currents are flowing in conductor with some resistance, there will be voltage difference.

\$\endgroup\$
0
\$\begingroup\$

You can have other circuits which return currents are passing through the area where you place the negative probe. This will increase the voltage or decrease depending on the direction of current flow.

To measure precisely and avoid the extra resistance from the ground which is not ideal zero impedance, always route a dedicated trace from both ends of a sense resistor as also mentioned in the comments, or use a Kelvin connection to do the measurements.

\$\endgroup\$

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.