Referring to the schematic below, is it best practice to have the 5 V supply on the right powering the MC (arduino-esque microcontroller) share a common ground with the 24 V supply ? The 24 V supply is powering a current output sensor (4-20 mA loop), and the resistor is allowing for an analogRead() of the signal as the MC has no current input pins.

I have a feeling that both power supplies and MC should share a common ground to reduce noise for the analog signal, is this reasonable?

enter image description here

  • \$\begingroup\$ Suppose you did disconnect the 24V ground from the MCU ground. How is the MCU going to measure the voltage (potential DIFFERENCE) at the top terminal of the resistor without something to measure against? Or put another way, how is the current, however, miniscule, that flows into the microcontroller pin (required the MCU to read or interact with anything) going to circulate back to the 24V supply? \$\endgroup\$ – DKNguyen Jun 19 '19 at 21:09
  • \$\begingroup\$ I'm not sure if I understand what you're trying to get at. The MCU wouldn't know what the desired potential difference would be, and could reference the 5V GND instead. Edit: Are you saying that the MCU GND and current loop GND should be separate? \$\endgroup\$ – MarkoDragic Jun 19 '19 at 21:12
  • \$\begingroup\$ "The MCU wouldn't know what potential difference would be". Correct. " and could reference the 5V GND instead." Incorrect. In the real world, currents, however small and negligible in analysis, are required to flow to develop voltages across things. How is the current from the 24V supply's positive terminal supply going to flow into the MCU pin and circulate back to the loop back to the negative terminal on the 24V supply? It can't end up at the negative terminal on the 5V supply because that's not a loop. \$\endgroup\$ – DKNguyen Jun 19 '19 at 21:15
  • \$\begingroup\$ Ok, so the 24V loop and 5V loop should be separated and not share a common ground as in keep the loops intact and separated? \$\endgroup\$ – MarkoDragic Jun 19 '19 at 21:19
  • \$\begingroup\$ Nonono. What I'm getting at is if currents from something don't have a path to circulate (aka into then out of), then out of the MCU, then that thing is essentially invisible and "not there" as far as the MCU is concerned. No current circulation into the MCU = no currents (obviously) and no voltages developed inside the MCU = no interaction with the MCU. \$\endgroup\$ – DKNguyen Jun 19 '19 at 21:21

Some microcontrollers have a sperate analog and digital ground to control the ground return currents from the microcontroller. If the microcontroller is switching large amounts of current (as it changes state) the currents exit through the ground pin. The small amounts of resistance and inductance in the pins and PCB can create small voltages that may create noise on the ADC subsystem of the microcontroller.

Make sure you have a good solid ground plane beneath the microcontroller and watch the return currents on the PCB (currents flow and create voltage even through large ground planes. Sometimes it is well to place a divider between the digital and analog sides of the ground to route large return currents away from the microcontroller. (you route currents by proper component placement, currents return to the source)

For most microcontrollers, the ADC's are 12-bits and it's easy to keep the noise below the 0.8mV per bit level (if you have a 3.3V ADC). So go ahead and tie the grounds together.


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.