Tool/software:
Hello,
I am working on AWR2944P with HWA 2.1, where I want to implement weighted DC removal before VFFT. For the first 4 range bins, I want to:
-
Compute mean across all chirps for each Rx channel.
-
Subtract that mean from all chirps.
-
For the first 4 range bins, apply scaling weights
[1, 1, 0.5, 0.25]
to the DC estimate before subtraction respectively.
I understand the HWA has a built-in DC estimator (DCEST_EN=1
) that latches the mean internally, and that subtraction can then be enabled with DCSUB_EN=1, DCSUB_SELECT=1
.
My questions are:
-
Where are the outputs of the DC estimator stored? From the TRM, it seems the estimates are only latched internally and not written to memory or CPU-readable. Is that correct?
-
If I want to apply my own scaling factor (e.g., 0.5 or 0.25 for bins 2 and 3), can I:
-
Use a paramset with the Statistics block (SUM mode) to compute the per-Rx sum across chirps,
-
Divide by the number of chirps in the CPU to get the mean,
-
Multiply by my scaling weight,
-
Then load the result into the
DC_SW_I/Q
registers, -
And finally configure the next paramset with
DCSUB_EN=1
,DCSUB_SELECT=0
to use the software DC values?
-
-
Is there any recommended way to combine both approaches — e.g., using HWA for DC estimation but still being able to scale the results before subtraction?
Any clarifications on how to correctly implement this flow in HWA 2.1 would be greatly appreciated.
Thanks.