none
Questions about the implementation of BPSK demodulation RRS feed

  • Question

  • Hi,

    Theoretically, BPSK demodulation should only consider the real part of the received complex samples.

    But in the TDBPSKDemap brick, the implementation considers both the real and imaginary parts, in the "demap_dbpsk_bit" method. When I change the code to make it only judges the real part, some errors appear in the demodulated bits. 

    BTW, when I record the received samples, I find the imaginary parts are far from zero, while that in the transmitted samples are really zero.

    Would someone tell me why imaginary parts deviate so far from zero in the receiving side? and why sora boxes implement the BPSK demap in this way? Thanks very much!


    Tuesday, April 1, 2014 9:50 AM

Answers

  • Frequency offset can not be perfectly compensated using static configuration on radio like what is performed by "dut freqoffset". Finer grained compensation should be performed during demodulation, namely channel estimation and compensation.

    -Jiansong

    Friday, April 4, 2014 6:12 AM

All replies

  • Hi,

    I think I have already figured out it, the boxes implement the demap process in this way to combat the phase offset, right? I have opened the dump signal through the "sdscope-11b.exe" tool, and seen the phase rotation. 

    However, when I obtain the frequency offset through HVT and set this value through dut tool, I find the phase offset still exsits.

    Does this phase rotation induced by the frequency offset or some other reasons?  

    Thanks very much!

    Wednesday, April 2, 2014 3:45 AM
  • Hi,

    If your question is about why you can see phase difference in the receiving side.

    It could be frequency offset. Also, the distance between sender and receiver could also change the
    phase.

    Muye
    Wednesday, April 2, 2014 5:18 PM
  • Hi Muye,

    Thanks very much for your response!

    I find the frequency offset can not be set through the dut tool, because no matter what value is set, I see the same phase difference in the receiving side. 

    I set it in the receiving side, and the command is:

    dut freqoffset --value 976

    Do you know whether the operation is incorrect, or the offset really can not be set through dut tool? Do you have some other ideas to do it?

    Thanks!

    Friday, April 4, 2014 1:11 AM
  • The rotation is caused by frequency offset.

    If you carefully tune the frequency offset, you can make the rotation slower, but you cannot remove the phase offset.

    Friday, April 4, 2014 5:31 AM
    Answerer
  • Frequency offset can not be perfectly compensated using static configuration on radio like what is performed by "dut freqoffset". Finer grained compensation should be performed during demodulation, namely channel estimation and compensation.

    -Jiansong

    Friday, April 4, 2014 6:12 AM