I transmit OFDM signal with bandwidth 20 MHz via ADALM-Pluto SDR device. To study the effect of the wireless channel, I used the same device for transmission and reception.
The maximum bandwidth of my ADALM-Plut SDR device is 20 MHz. I have read that due to the difference between the Local Oscillators, there will be some frequency offset of delta which will cause phase rotation of samples in time domain and there will be also a frequency sampling offset which will cause a phase rotation of samples after fft-output at the Receiver in frequency domain.
As you can see in figure(24), the samples after fft-output at the receiver show amplitude compression towards the zero when transmitting the signal at the maximum bandwidth of the SDR (20 MHz). This issue also happens when I use the Automatic Gain control at the receiver setting.
In contrast, when I make transmission with a bandwidth of 2 MHz (sampling frequency of 2 MHz) via the same ADALM-Pluto device, I obtain figure (17) which shows samples after fft-output rotating linearly in frequency domain.
In both cases above, the frequency offset was corrected using coarse and fine correction with short and long training sequences respectively.
When I perform equalization to both cases, the first case shows low performance after equalization in which some samples spread away from the +1 and -1 as I am using BPSK baseband modulation.
For the second case, it shows good performance after equalization.
My question for now are:
Which case is the normal state for ofdm-samples after the fft (the one which shows compression of amplitude towards the zero (fig. 24) or the case which shows linear rotation of samples (figure. 17) ).
What is the cause of the amplitude compression shown in the first case and why this issue vanish when I use low bandwidth like 2 MHz
Read more here: Source link