You are not logged in.
I have modified the Wl_example_siso_ofdm in WarpLab reference design so that only one warp could use the RF_A to transmit the data and RF_D to receive the data. The question here is do these two RF chains share the same sampling clock?
Offline
do these two RF chains share the same sampling clock
Yes - all RF interfaces on one node share sampling and RF reference clocks.
Offline
This is the linear relationship of CSI phase in different OFDM subcarriers.
Last edited by sarah (2018-Nov-23 22:43:26)
Offline
murphpo wrote:
do these two RF chains share the same sampling clock
Yes - all RF interfaces on one node share sampling and RF reference clocks.
When I test the WarpLab SISO OFDM design using one Warp node. One antenna transmits signal(RF D) and the other one receives the signal (RF A) . I got the Channel state information for different OFDM subcarriers. However, When I look into the phase of Channel state information in different subcarriers. There is a strong linear relationship of phase with different subcarriers which looks like a phase offset caused by sample frequency shift. But if all RF interfaces share the same sample clock. There shouldn't be sample frequency shift. Do you have any idea about this strange frequency shift caused a linear relationship of phase in different subcarriers?
About the linear relationship what I say, Please refer the figure above
Last edited by sarah (2018-Nov-23 22:45:15)
Offline
There is a strong linear relationship of phase with different subcarriers which looks like a phase offset caused by sample frequency shift
You're seeing the effects of delay. A delay is a linear phase vs frequency, where the slope of the linear phase is proportional to the delay. Try adjusting the FFT_OFFSET parameter. This value controls (on average) how many samples of cyclic prefix are consumed by the receiver's FFT. In a real system this parameter adjusts the tradeoff between synchronization error tolerance (i.e. use many cyclic prefix samples) and multipath tolerance (i.e. use no cyclic prefix samples). You will observe the same behavior in simulation.
Offline