IWR1443BOOST: CFAR Threshold format clarification

Hi Predrag,

 I want to clarify/correct the last paragraph in Wesley's response (I provided him that information, it is not his fault, I later realize it was not quite right) about 14xx usage of FFT to calculate sum as DC bin for your benefit and also for benefit of everyone else who may follow this thread in future. The FFT by definition does not produce 1/N * sum(), by definition of FFT, it will only produce the sum() because X(0) = sum[x(n)*e^(-j2*pi/N*0*n)] = sum[x(n)]. We set the size of the FFT to be the next power of 2 of the virtual antennas i.e if N = number of virtual antennas, then fft size is N' = 2^ceil(log2(N)) and we enable butterfly scaling for all stages because of which the output is scaled to 1/N', not because of the definition of FFT but because of our configuration of using the HWA. So the output of the FFT = 1/N' sum(.). So we nearly get the average but it is not exact if N is not a power of 2. e.g if you have all three transmit involved in the TDM chirping (i.e you have elevation along with azimuth), then N = 12 and N' = 16. So an additional correction needs to be applied when computing the threshold from dB to Tcli. This is already taken care of in the visualizer when user fed value in dB is translated to the CLI command format. But it is not shown in the UG which has the formula Tcli = 512 * dB / 6. The correct formula is Tcli = 512 * dB / 6 * N/N', this you will see in the UG of the next (post 1.2) SDK release. So the 14xx does have dependency on number of virtual antennas but it is in the sense of this correction factor, not like the 16xx case. When number of virtual antennas is already power of 2 (e.g 4, 8), then N = N' so no correction is required.

thanks