You are not logged in.
Pages: 1
Hello all:
I was recently puzzled by line 27 of wl_example_siso_ofdm_txrx.m, specifically the comment that reads "must be 2".
INTERP_RATE = 2; % Interpolation rate (must be 2)
Indeed the interpolation rate must be 2 when enabling WARPLab-in-the-loop; otherwise the example is not able to receive the waveform satisfactorily. However, INTERP_RATE may be 1 when using a sim-only approach (the example is able to receive the waveform in a pure MATLAB approach).
What is the actual mathematical or hardware reason (I guess hardware) for this behavior? I may have misunderstood something... but could you please help enlightening me? :)
As usual, thanks a lot for your time. :)
Offline
There isn't any hardware reason for that requirement. An earlier version of the code actually supported an INTERP_RATE of 1 (40 MHz bandwidth) or 2 (20 MHz bandwidth) in hardware, but it made the code messy having to deal with switching on that parameter to apply different processing (e.g. selectively applying the interpolation/decimation filter).
We ultimately decided that the INTERP_RATE of 1 not only made the example messy, but also was more generally a bad example of how to do 40 MHz OFDM since the underlying implementation in that example remains a 64-point FFT. Normally, 40 MHz OFDM (like 802.11n/ac) keep the same subcarrier width and increase the FFT size to 128-point. Doing this would be a straightforward extension of our existing example.
Offline
Oh, now I get it. Thank you so much for the speedy reply. :)
Offline
Pages: 1