In a board is it okay to just join low frequency power(>10 Hz, 1-15 A, 24 V) with “high” frequencies(~100-480 MHz), being supplied by “different” power supplies but referenced with the same ground without any isolation whatsoever?

Long one

I started designing the board around a 480 MHz microcontroller and to my luck after a week or so of troubleshooting the board came to life (which is really nice considering I only designed a few simple sensor boards and nothing relating to microcontrollers)

Then, as usual, scopes changed, and now I had to include some solenoid control, and other stuff, but since I'm from computer engineering, I started to see “updates” to requirements as something annoying to deal with.

So I decided to just keep the design flexible by adding some H-bridges for power in case the project quickly changed+ I added some power MOSFETs in case the aforementioned design fails. The H-Bridge itself drives up to 20 kHz although for the use-case I'm probably not gonna use it, and the solenoid itself can only achieve up to 10 Hz.

About power, the two powers are separate due to fusing although before the wiring, fuses etc… it all goes to two 12V batteries stacked in series so we can get 24V. The ground however is separated by wires, until they all go to a single point in the chassis to avoid ground loops.

My colleague says there's virtually no issue. However, I'm afraid of noise, sudden manufacturing, transients, EMI transmissions etc., not to mention that from the team's “perspective” money isn't an “issue” per se, as we're sponsored. So it's even if it's to squeeze some performance, and it's somewhat justifiable. It's worth to place it on the board (plus I'm biased and I really wished to try out some digital isolators from Wurth),

  • \$\begingroup\$ What is your question? Please be specific. \$\endgroup\$
    – Voltage Spike
    Commented 2 days ago
  • \$\begingroup\$ I mean is there any issue in not isolating it with optocouplers/digital isolaters? what problems can arise? \$\endgroup\$
    – Imeguras
    Commented yesterday
  • 1
    \$\begingroup\$ So the Question is:"What kind of precautions do I need to take to prevent the high power part disturbing or destroying the microcontroller side of the board?", right? \$\endgroup\$
    – kruemi
    Commented yesterday
  • 1
    \$\begingroup\$ @kruemi or vice versa, yes but im also unsure about the problems that can appear \$\endgroup\$
    – Imeguras
    Commented yesterday
  • \$\begingroup\$ When working at (relatively) low voltages like 24V and high currents voltage drops might hurt you pretty quickly. Take care when selecting connectors, make traces plenty wide (don't go with minimum requirements) and don't forget to use fuses to prevent catastophic failures. \$\endgroup\$
    – kruemi
    Commented 23 hours ago

1 Answer 1


Short: Mixing 24V electronics with low power, high frequency microcontroller circuits can be done in is routinely done in many control applications.


Usually, the following things are to consider:

  • Signal levels must be converted according to the requirements of the components in both directions
  • Transients in current / voltage from the high power side should not influence the function of the uC side.
  • Faults on the power side should not destroy the uC and components

The low power, high frequency stuff influencing the high power side negatively is normally no big concern.

As you're using the same ground anyway, optocouplers are of limited use (they are not useless but often just not worth the effort).

So what to do?

  • Separate the parts on the board physically (keep stuff of the same power level together)
  • Take great care with the power rail design (use ground planes or really thick ground tracks, keep the area between power/signal traces and the gnd return as small as possible), use deblocking capacitors really close to each Vcc/GND pin pair
  • When using separate supplies for different parts of the board, spend some thought on the power-up and power down sequence. You want to prevent cross-feeding (signals from one part powering the other part via signal pins) at any moment
  • spend a lot of time to think about the GND design! Where does current flow? Where do you have voltage drops and what could be the influence on your signals? Where could noise and other distortions enter your design. (I see problem arising from the Idea of connecting the grounds externally to the board, I'd suggest bonding them on the board and getting rid of the external duble connection).

It's solvable but you will have to find the right mindset. Always keep in mind that current only flows in a closed loop. So if you have a signal there is always a ground path that acts as return. Consider the whole path, not just the signal.

Everything I write here can be applied to every design you do. Do it that way and it will save you a lot of headache in the future.


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.