As my circuit is a little complicated, I don`t know how to pose the PCB here. So I just want to know the probable causes of the problem and I will test them one by one later and report back.

My problem is that:

The MCU(STM32F4) would run a function with 300Hz frequency. When the function runs, I can observe a set of noise(2100Hz,2400Hz,2700Hz,3000Hz,3300Hz...)in the analog circuit. Amplitude of the noise is about several mVpp. The observation point is the output of a 45dB-gain, 3kHz-center-frequency bandpass filter. I need to get rid of those noise as my desired signal is also 3000Hz.

So does anyone meet the same or similar problem before?

More information:

  1. I once modified the frequency of the function to 600Hz and the set of noise became 2400Hz,3000Hz,3600Hz... and the amplitude became a little larger.

  2. the digital ground and the analog ground are in the same plane.


These days I have done some experiments and found that the noise maybe comes from the ground plane.

I measured the voltage difference between the GND terminal of the power supply (connected to the main plug of the device) and the GND on my PCB. The 300Hz noise appeared. So I think the noise is from the noisy ground plane. But I am not sure my measurement is correct. Is it?

One more strange phenomenon: I measured all the MCU pin. Also found that noise. So is this phenomenon due to the noisy ground plane?

  • \$\begingroup\$ Does this digital function have any influence on ADC/DAC sampling? In other words, are you generating or reading analog samples whenever this routine is run? \$\endgroup\$ – alex.forencich May 25 '15 at 3:36
  • \$\begingroup\$ Standard medicines: separate grounds, put an LC filter in MCU power supply, shield analog part, put the cap as close as possible to MCU pins etc. What is LIA btw? \$\endgroup\$ – ilkhd May 25 '15 at 3:39
  • \$\begingroup\$ @billyzhao It may not be necessary if you can filter power noise on the digital side very well. You can try, if you have a second PCB to power only analog part from the same power supply and connect to the digital on the first PCB and see how it goes. Crude, but can be convincing. \$\endgroup\$ – ilkhd May 25 '15 at 4:23
  • \$\begingroup\$ Do you have an ADC driver amplifier to isolate the switched capacitor ADC input from the rest of the circuit? \$\endgroup\$ – alex.forencich May 25 '15 at 4:25

Maybe you need to have larger decoupling caps nearest to the microcontroller power pins, ro reduce the path length that is able to couple its signal to other PCB trace or wire.


What is most likely going on here is the switched capacitor input on the ADC is creating a measurable pulse every time it connects the sampling capacitor to take a reading. I would recommend adding an ADC driver amplifier to isolate the sample and hold from the rest of the circuitry.


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.