I’m using a custom PCB with the Qorvo DW3220 UWB transceiver (Channel 5, 6.5 GHz) to measure Phase Difference of Arrival (PDoA) for AoA estimation. I designed the board with an antenna spacing of lambda / 2 = 2.3 cm to avoid phase ambiguity. After reading the PDoA register and converting the raw phase value by dividing with 1 << 11
(to radians), I observe phase wrapping/ambiguity when angles exceed 60°. Why does this happen despite proper antenna spacing? Do I need code changes? Other than this what else can reduce NLOS susceptibility?
Hi @olirex99 ,
We typically recommend using 0.45 Lambda spacing between antenna elements to avoid any phase wrapping.
In your custom PCB, I assume you performed the following.
- Calibrated your PDoA offset at 0 degrees
- Obtained your PDoA to AoA LUT
- Then you observed the phase wrapping at 60 degrees right?
A very high standard deviation at 60 degrees might also cause a phase wrapping issue. Do you also have any statistical data on your PDoA measurements?
Kind regards,
Emre
Hi, thanks for the answer! I am using the LUT, but I am directly computing the PDoA using the PDoA to AoA formula that involve the sin function.
I know that this is not optimal, but I want to try the physic implementation. Regarding the phase ambiguity I am pretty sure that the problem is related to the distance, since I am using exactly 0.5 lambda.
I also want to ask if you have any suggestion for the post processing of the signal. For example do you use suggest any type of filtering process? I saw that the sensors reading are seems to have noise at high frequencies. Do you use some other calibration methods other than the classic offset correction? For example polynomial calibration ecc.
Thanks!