• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 2 answers
  • Subscribers 47 subscribers
  • Views 553 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

ADS1298 modulator noise

Matteo Cardani
Prodigy 60 points

Other Parts Discussed in Thread: ADS1298

Hello, I'm using an ADS1298 to implement a 12 lead electrocardiograph. However I'm having some issues do to the noise introduced by the sigma-delta modulator. Sampling the signals at 500Hz in fact introduce a sinusoidal noise with frequency of 62.5Hz due to fact the the high  modulator frequency is folded into the pass band. I tried to filter the input signals (cut off frequency 1khz), however that not also did not resolve the issue, but worsened it as the ECG signal amplitude was reduced. I also tried to use your evaluation board in which the input signals are filtered with a second order RC filter, but I had the same problem. The only thing that worked for me was raising the sampling frequency till this noise is not bent anymore in the pass band.

As this solution is just a workaround, can you give me some information on how I can remove this noise? 

Thank you,

Matteo Cardani

  • 0
    TI__Mastermind 38745 points
    Hey Matteo,

    I doubt that the modulator frequency is aliasing to 62.5 Hz. The modulator frequency is either 256 kHz or 512 kHz depending on whether you are in low-power or high-resolution mode, respectively. The digital filter acts on the data sampled at that frequency so if that specific frequency were aliasing, it would alias at 0 Hz. Even if an alias was occurring about the 500 SPS data rate, the modulator frequency aliases at integer multiples of 500 Hz; all of which fall at a zero for the sinc filter which leaves the only possible alias at 0 Hz again. The modulator frequency aliases falling exactly in these zeros is guaranteed by the fact that the data is being sampled with that clock.

    In addition, the amount of anti-aliasing in the scenarios you've described seems sufficient for a signal that was stationed 62.5 Hz away in either direction from the modulator frequency. In addition, the passive anti-aliasing filters, the PGA on the ADS1298 also has limited bandwidth. That means that in the case with the EVM, you can model the anti-aliasing as a third order system. These factors lead me to believe that the signal is in the passband for the system. A common source of parasitic signals in mixed signal systems is the power line signal. What frequency is used for power delivery in your region?

    Regards,
    Brian Pisani
  • 0 in reply to Brian Pisani
    Prodigy 60 points

    Hi Brian,

    Thank you for your answer! I live in Italy, so the frequency of the power delivery system in my region is 50Hz, which I filter afterwards with a notch filter. In the FFT of the raw signals in fact I see three peaks: one at 50 hz, one at 62,5hz and one at 187,5hz (multiple of 62,5hz). I thought that the modulator was aliasing at 62.5 hz because this peak changes only if I change the sampling frequency of the ADS1298. Changing frequency or turning of other possible noise source (SPI; UART, oscillator of the micro-controller)  seems to not affect the noise. 

    Best Regards,

    Matteo Cardani

  • 0 in reply to Matteo Cardani
    TI__Mastermind 38745 points
    Hey Matteo,

    Changing the data rate will change the response of the on-chip digital filter so I would expect the magnitude of the noisy signal to change even if it was occurring "un-aliased" in the passband. Have you tried moving to a different bench in your lab or perhaps to an entirely different room? There could be some electric motor or something nearby that is causing the noise.

    Brian