Data downconversion on X2 and X4

Discussion in 'X2 & X4 SoC' started by Davide Bongiovanni, Mar 16, 2017.

  1. Davide Bongiovanni

    Davide Bongiovanni New Member

    From what I understand, the X4 chip is capable of outputting data that is already downconverted. Does the X2 chip do so as well? If not, is this taken care of by the on-board microcontroller - say, on the X2M200 module?
    Follow-up to the follow-up: how is the downconversion performed? Is it just regular pulse-compression or is there something more going on?
     
  2. Jan Roar Pleym

    Jan Roar Pleym New Member Staff Member

    Hi Davide,
    both the X2 and X4 chip sample the signal at RF.

    The X2 chip outputs RF data and the down-conversion to complex baseband is done in the on-board microcontroller (e.g. X2M200).

    The X4 chip has a built-in down-conversion chain and then you can select RF data or complex baseband data output.

    Since the signal is digitized at RF, the down-conversion from RF to BB is implemented digitally. This is done
    by multiplying every frame with a complex down-conversion vector and subsequent anti-aliasing LP filtering and decimation by a factor of 8.

    This is equivalent to an analogue down-conversion chain (I/Q mixers and LP filters). However the digital process guarantees perfect amplitude/phase balance.

    Regards,
    Jan Roar
     
  3. Davide Bongiovanni

    Davide Bongiovanni New Member

    Thank you very much for the detailed answer
     
  4. David HU

    David HU New Member

  5. David HU

    David HU New Member

  6. NSANGOU CHOUAIBOU

    NSANGOU CHOUAIBOU Active Member

    Hi David,
    Contact NOVELDA team for X4 datasheet.

    upload_2017-3-20_17-13-4.png
    B.R
    Chouaibou NSANGOU
     
  7. Jan Roar Pleym

    Jan Roar Pleym New Member Staff Member

    Hi Chouaibou, we need an anti-aliasing filter because we reduce the sampling rate by a factor of 8 when down-converting the RF signal to baseband.

    Regards Jan Roar
     
  8. NSANGOU CHOUAIBOU

    NSANGOU CHOUAIBOU Active Member

    Hi Jan,
    You are right. We absolutely need to reduce high-frequency signal components in our BB after decimation.
    Thank you.

    B.R,
    Chouaibou.
     
  9. Davide Bongiovanni

    Davide Bongiovanni New Member

    Actully, I have been experimenting with thwe X2M200 sensor and I noticed that when I activate amplitude/phase output, place a planar reflector in front of the sensor and slowly move the sensor towards the obstacle, I get some periodic oscillations in the amplitude, with a period of half lambda.
    If I switch to I/Q output, with the same reflector as before and moving the sensor in the same way, I would expect the I and Q signals to oscillate in quadrature with a period of half lambda. What I actually see is that the I signal doesn't seem to oscillate as much in the positive wave of the oscillations, effectively causing the amplitude that is calculated from the two signals to not be constant.

    Does anybody know why this might happen? If what Jan said is true, then does that mean that the problem resides in the analog frontend? Or is it caused by the transmission medium, or maybe some properties of the reflector?
     
  10. Davide Bongiovanni

    Davide Bongiovanni New Member

    I conducted some experiments to better show what I mean. Here is a graph of maximum amplitude vs distance with the radar facing a big metallic reflector.
    distanceorthomm.png

    Any ideas why this is happening?