• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 46 subscribers
  • Views 489 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.

ADS1255 alias

Paolo Geriniz
Prodigy 120 points
Other Parts Discussed in Thread: ADS1255

Dear Sirs,

our board is a multichannel AD converter with an ADS1255 for each channel.

We leave to the user the possibility to change the sample rate from 10 up to 1000 sps.

Since the analog modulator is working at 7.68 / 4 = 1.92Mhz we placed the input antialias RC filter corner freq at around 400Hz, so well below the Nyquist freq.

Nevertheless we notice alias phenomenon when we collect data.

For example at 100sps, frequencies above the Nyquist (50Hz) are spectrally mirrored, for example with 70Hz we get a mirror around 45Hz, with 80Hz at 20Hz and so on.

What can we do to get rid of the aliasing problem? 

Regards,

Paolo

  • 0
    TI__Mastermind 44680 points
    Hi Paolo,

    That is a challenge! If the signal is within the passband of interest, there's not much you can do on the pre-filtering side to remove the issue. Where does the aliasing signal come from, can you prevent it from getting into the system before the inputs?... it sounds like it is being applied intentionally, is that the case?

    The only other way I can think of removing the alias would be to implement some sort of post processing filter... You would run the ADC at a faster data rate and use a digital filter with a steep and narrow transition band to remove the aliasing frequencies. The digital filter could control the "effective output data rate" to create the appearance of 10-1000 SPS data rates. The main problems with this implementation would be that: 1) the ADC output would be nosier at the higher data rate and 2) the digital filter would likely have a long settling time (since filters will always have some trade-off between the frequency- and time-domain responses....a steep cutoff usually implies a long settling time.)

    Best Regards,
    Chris
  • 0 in reply to Christopher Hall
    Prodigy 120 points

    Hi Chris,

    thanks for your prompt reply.

    Yes we are intentionally appling a test signal on the inputs to evaluate the complete behaviour of the board.

    We will try to implement a filter/decimation on post-process.

    Thank you very much.

    Regards,

    Paolo