AoA IQ data seem NOT uncorrect, when used one antenna?

[upupsky]

software: ncs v1.8.0

hardware: my test nrf52833 board, with 9 antennas.

I do experiment like this:

aoa rx:  in "nrf52833dk_nrf52833.overlay" file, I set  "dfe-antenna-num = <2>;" and "dfe-pdu-antenna = <0x4>;"

             in "main.c" file, I set "const static uint8_t ant_patterns[] = {0x4, 0xB};"

this means, I use only 2 antennas(1 antenna 0x4 receive pdu and 8us ref period samples; 2 antenna 0xB receive 148us cte signal[will not switch to 1 antenna]), the 45 samples like this:

QUESTION:

    cte signal f=250KHz(T=4us), slot 4us.  after 1 antenna samples 8us, then switch to 2 antenna samples 142us(37 samples). the phase of the 37 samples  should  be permanent(because aoa tx tag not move all the time), the right picture, it seems cte signal is not 250KHz(4us){the cte almost T=4us*24=96us},  I wanted like this(red points):

where am i wrong?

thanks.

 

[PaKa]

All the frequencies in transmitter and the receiver have inaccuracies so you will never ever see a fixed frequency in the system. The algorithm used to process the data should assume the frequencies will be fixed over the period of one CTE event but the absolute frequency value can be inside a fairly wide range so you will need to calculate this out of incoming data for each CTE event. This is what you use the reference period for.