Impact of Antenna switching in IQ samples

[SRINATH]

Hi ,

1. I am using nrf52833 Soc as transmitter and nrf52833 + Core-hw antenna array(4X4) as receiver. I have set the switching pattern of antenna array as round-robin pattern (7,7,8,9,10,11,12,13,14,15,0,1,2,3,4,5,6) . During this time I have received the unstable iq samples as fig_1 in which the phase difference in the reference period was not 90 degree(unstable and high fluctuations in phase difference).

2. In the second case,I have used the same setup as above.I have changed only the antenna switching pattern to return to first pattern(7,7,8,7,9,7,10,7,11,7,12,7,13,7,14,7,15,7,0,7,1,7,2,7,3,7,4,7,5,7,6).This time Iam getting a proper signal with reference period showing a regular 90 degree phase difference as shown in fig_2.

What could be the reason for unstable phase difference in case_1 and stable phase difference in case_2 (the only change - antenna switching pattern).

                                               Fig_1

                                                Fig_2

[SRINATH]

I think you didn't get my question and there is some misunderstanding.Yes,I have converted the raw iq data to phase and amplitude. In the below attached csv file you could see that the phase difference between samples are approx 90 degree.This sample is taken with the antenna switching pattern of return to first(7,7,8,7,9,7,10,7,11,7,12,7,13,7,14,7,15,7,0,7,1,7,2,7,3,7,4,7,5,7,6). Similarly when I used round robin antenna switching pattern ((7,7,8,9,10,11,12,13,14,15,0,1,2,3,4,5,6) I observed phase difference in the range of -170 to +170 considering phase difference only in the reference period. 
Amplitude - sqrt(I2+Q2)

Phase - arctan2(Q/I)

Above phase values are in degrees.

The phase difference is the difference between consecutive phase values.

TAG_778797107.csv

[PaKa]

So the phase data was not available until you sent the CSV file so not possible to comment on that.

So if the CSV file is a standard Bluetooth DF package then the 1st 8 samples are the reference period. Since these are taken 1µs apart and measuring something close to 250KHz you should see a phase shift between each of these close to 90°, data you have show ~95° so inside expectation.

Then the expected shift to the next sample (assuming you have the same antenna as the reference) will be ~95° multiplied by the sample spacing, if this is 4µs then about 20°.

Your values seems to be in the ballpark area for this.