This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

PCM2901: 1kHz noise removal

Part Number: PCM2901

We are designing a low-cost two channels data acquisition system based on PCM2901. The device acquires the signals form IEPE accelerometers.

 I attach the schematics. of the device. The USB interface is isolated by the use of the chip ADUM4160BRWZ

The prototype has been built on a 4 layer PCB The outer layers have signals and power-supply traces. The inner layers have groud planes. There is a clear separation between the two sides of the isolation

When observing the data acquired, harmonics at 1,2,3kHz are clearly visible even when no load is applied. The signals at the two opamp outputs (looking at pin 7 of  U12 and U13) shows pulses with a period of 1ms when the software is acquiring the data.

By making some research, I notice that 1kHz noise is quite a common issue in USB audio device. 

I have already made some tests by connecting a 100uF tantalum capacitor close to the USB connector and close to the 3.3V regulator output but the behaviour does not change.

I have also built on a breadboard a common mode filter to be placed near the USB connector. The filter was made by the common mode  choke 744226 and two 100uF 10V capacitors. In this case the device is not recognized by Windows.

Below are attached the time and frequencies domain acquisitions by connecting to one channel a 6MOhm resistor and to the other channel a 20MOhm resistor

Which kind of filter do you suggest in order to reduce the harmonics multiple of 1kHz?

Thanks in advance

  • Hello,

    The 1kzh (and harmonic) content usually usually couples through the power-supply connections since the CODEC is directly powered from the USB power-supply connection.  Proper decoupling, very little distance between the CODEC and incoming USB power connections will help with this behavior. I have a feeling this has to do with issues with the "+5Vi" net and subsequent "+3.3Vi" nets.  If these power connections/ GND connections end up having the 1kHz signal on them there's not going to be much we can do as it will be up to the internal PSRR/CMRR of the products to reject the noise.  Would you share a layout for review?

  • Thanks for the reply. In the zip folder there are the gerber files and the project created in Cadence 16.6Project DAQ Sound card v1.zip