Home Forums SYSTEM Dispersion compensation using Matlab

Viewing 3 reply threads
  • Author
    Posts
    • #24092
      mai fouad
      Participant

      I try to compensate second and third order fiber dispersion in the frequency domain, the frequency domain filter works well at 25 Gbaud/s and gives results close to compensation using DCF but when increasing symbol rate to 50 Gbaud/s there is a clear difference between compensation using DSP and DCF, is there a problem in this method?

    • #24096
      alistu
      Participant

      Hi Fouad, When I ran your simulation, I encountered an error stating that it couldn’t load(‘D:\qpsk data\I_comp1’). If there is another Matlab file that you have not attached, please attach it. Otherwise. could you give an explanation about this? Thank you.

    • #24099
      mai fouad
      Participant

      I attached these files, that load In-phase and q-phase components from Matlab to optisystem also the binary data, QPSK symbols for calculating BER if needed and you may need (Inphase_comp_tx_for_Q-factor)that contains In-phase component with no biasing or amplification (i.e. constellation points at [-1+1j 1+1j -1-1j 1-1j]). you could use these files and change the paths or use your own QPSK TX and RX but you will need the last two matlab component(specially matlab component 2) to compensate the dispersion.

      Attachments:
    • #24141
      mai fouad
      Participant

      attached here a file comparing the performance of DSP and DCF, the difference in the performance can be seen from the constellation visualizer.
      i don’t know where is the problem, the frequency domain filter just multiply the signal by the conjugate of {exp(j*[(beta2*omega^2/2)+(beta3*omega^3/6)]), beta2,3: second and third order dispersion, omega: angular frequency} to compensate dispersion.

Viewing 3 reply threads
  • You must be logged in to reply to this topic.