• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 49 subscribers
  • Views 806 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.

EEG Frontend

Brian Anderson
Prodigy 175 points
Other Parts Discussed in Thread: ADS1282, ADS1298, PGA103

The ADC1298 seems to be a perfect solution for ECG and similar applications where we expect large signals in the mV region. How about advanced EEG applications? There are quite a lot posts expressing concerns regarding the 2..5uVpp input noise level of the ADC1298. Old-fashioned approaches such as low-noise input amps don't seem to be very elegant. I'm wondering if TI is preparing devices similar to the 1298, just with lower input noise and/or higher resolution? We are considering the ADS1282, which, however, is quite expensive...

Any ideas, comments and suggestions regarding a high-performance EEG frontend are very welcome :)

Brian

  • 0
    TI__Guru**** 182191 points

    Hi Brian,

    Thank you for your comments on the ADS1298!  For ECG/EKG, the part is very well suited to the higher voltages as you noted.  For the EEG side of things, there are a few options - you can use a low noise PGA ahead of the ADS1298.  We are working on a few app notes/solutions to address this application and we'll keep the forum posted on our progress.

  • 0 in reply to Tom Hendrick
    Prodigy 175 points

    Hi Tom,

    looking forward to this! For us, very low 1/f noise performance is also very important. I'm checking out chopped amplifiers as well as the ADS1282. May I ask (a probably silly) question regarding the ADS1282: with a PGA setting of 4 in normal mode, 250SPS, the input related SNR is -129dB, which equals a rms noise of 0.16uV. At 31bits, the LSB resolution is 0.6nV - so, I have like 8bits to show the noise. Am I making a mistake here?

    In the meantime, my test circuit works and I'm checking out the ADS1282.
    PGA=64, Chop ON, 250SPS ==> approx. 66nVrms noise with internal short to 400 Ohms. I was expecting 60nVrms- this looks great!. However, with some external protection circuit, noise goes up to 180nVrms- I think this is mainly due my resistors (never measured resistor noise before that directly ;-) So, that circuit needs some thinking... However, although not expecting miracles I'm feeling bad to throw away at least 7 bits..

    Nevertheless, this part rocks- only, its expensive...

    Brian

  • 0 in reply to Tom Hendrick
    Prodigy 175 points

    Hi Tom,

    which low-noise PGA could you recommend? I was browsing for such devices but didn't find any with noise lower than 10nV/rt(Hz). Also, are there multi-channel PGAs (not multiplexed- I need simultaneous A/D conversion)?

    Thank you,

    Brian

  • 0 in reply to Brian Anderson
    TI__Expert 3260 points

    Hi Brian,

    Unfortunately, the lowest noise PGA we have currently is the PGA103 (with 11nV/Hz), and you would still need multiple, since we do not currently have multichannel PGAs.

    Best Regards,

    Michael

  • 0 in reply to Michael W
    Prodigy 140 points

    Be careful here - EEG applications are much more sensitive to the 0.1-10Hz p-p noise. 

    This is 1/f noise territory - the nV/Hz numbers quoted, typically at 1kHz to 100kHz don't mean much.  nv/Hz at 1Hz would work too, but that's not something that shows up without digging or measuring.


    Regards,

    Gary