Hello,
I have a setup comprising a TI AWR1243 eval board + DCA 1000. I acquire data using mmWave Studio and process the data directly from the *.bin files. My acquisition is pretty simple: I send a single frame with a specific number of chirps (i.e.: 128), by arming the DCA, triggering the start, saving the data and then repeating this operation a few times. My scene is static, has a simple, reflective target (not too close to the radar) and there is virtually no clutter. Note: I process the data correctly, because there is no difference between the plots I obtain vs the ones I see using the post-processing tools.
Let's say I am looking at frame #1, chirp #1. I calculate the 1D FFT over the range, find the highest peak and determine the magnitude and phase of the FFT bin containing the target. If I repeat this operation for frame #1, chirp #2, 3, 4, etc... the target response phase doesn't change much. However, if I compare the phase of (frame #1, chirp #1) and (frame #2, chirp #1), there is a difference.
I am assuming that this phase difference is caused by the async trigger start. Correct? However, I do not understand why this is happening. I am intuitively assuming that the 40 MHz chip clock starts as soon as the AWR1243 is physically powered on, so all frames always start with whatever starting phase is associated with the 40 MHz clock. Regardless, is there a way to calibrate this inter-frame (inter CPI) phase offset?
Thanks in advance!
