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.

ADS1299: Artifacts in frequency spectrum

Part Number: ADS1299


We are seeing noise in the spectra of our recordings that we suspect is due to the sinc filter referenced on p25/26 of the ADS1299 datasheet, and would like to know (1) does the behavior shown below match what you would expect to see if the sinc filter was the cause, and (2) do you have any suggestions on how to reduce or eliminate this artifact (depicted below).

We're running the ADS1299 to collect EEG signals from a single channel input. In all recordings we are seeing addition peaks in the power spectra of our recordings that occur ~4 Hz above and below 60 Hz with additional smaller peaks every 8 Hz up and down the spectrum. To rule out an environmental source, we placed our recording device inside a Faraday cage which eliminated the 60 Hz peak, but the side band peaks remained.


The artifact resembles the shape of figure 31 on p. 26 of the ADS1299 datasheet. This lead us to suspect that it may be related to our sampling rate which we set to 250 Hz. We increased it to 1000 Hz, but evidence of these peaks still remained. We are hoping to get advice on possible sources of this noise and ways to remedy it.

I've attached an image displaying recordings we collected outside of the Faraday cage at 250 Hz (Left column), inside the Faraday cage at 250 Hz (middle column), and inside the Faraday cage at 1000 Hz (right column). The top two rows are example traces of 2 different styles of recordings. The gray background tracing has no filters applied. The blue and red tracings are notch filtered at 58-62 Hz then bandpass filtered at 1-50 Hz. These colors correspond to the 3rd row which shows the power spectra of these time series. The filtered spectra are shown in solid lines. The spectra of the raw data are in dotted lines. The triangles indicate the peaks that are in question.

  • Hi Chris,

    Welcome to the E2E forum !!! 

    Thanks for the post. Please let me check the team and get back to you. 

    Thanks

    -TC

  • Hi Chris,

    Thanks for the patience.

    I don't think the artifacts in the diagrams are the reason caused by the Sync filter. Figure 31 of the datasheet shows the transfer function of the Sync filter at multiple fMOD frequencies. This plot shows that the passband repeats itself at every fMOD, and the peaks remain the same for different data rate (fDR). Any interference signals within the passband will be aliasing back to the baseband, causing the artifacts. Thus, there is a need for an antialiasing filters in the system to attenuate any interference in frequencies around multiple fMOD.

    Some experiment that you try is to change the clock frequency of the device and see if the artifacts remain at the same frequencies. In addition, you may want to look at the hardware design and see if you have any interference signals within the multiple fMOD frequencies that may be aliasing back to the baseband and cause the artifacts.

    Thanks

    -TC