Hi,
I am observing spurs in FFT of adc data and was wondering if this has been observed by TI as well. I have two sets of AWR1243 BOOST + MMWAVEDEVPACK + TSW1400 boards and my observation is present across both sets of boards.
Board Sets: AWR1243BOOST + MMWAVEDEVPACK + TSW1400.
Chip Config: Single Chip
SW: Latest Radar Studio and latest mmwave dfp (0.09.01.06)
Data Capture Config: All Tx off, All Rx ON, Fullpower ADC, LVDS 4 lanes at 600 Mbps, Data format: Complex 2x
Cont Stream: 77 GHz, Data Rate: 37.5 MHz, Rx Gain: Varied from 24 to 48 dB, 16384 samples collected, 16 averages.
After receiving the data with the TSW1400 board, I post processed the data in Matlab with appropriate scaling of FFT to convert data to dBFs. I performed a windowed 16384 pt fft and averaged the fft output of 16 data frames. I did not remove the DC for the plots that I show in this post. I matched the result with the post processing script that is provided with radar studio and they lined up.
At low gains (24 to 36 dB) I observe spurs at different IF frequencies for all the receiver channels.
The magnitude of the spurs does not change with Rx gain. As the gains are increased to 42 and 48 dB, the spurs are not observed because the noise floor is also gained up.
The spurs do not increase in magnitude with increasing gain. When I performed multiple captures, the magnitude of the spurs dont change drastically. Occasionally may be one in ten times, I do get a clean spectrum even at minimum gain but it is not repeatable.
I changed the data rate to 20 MHz, and the spurs still show up at lower gains. Changing the LO frequency from 77 to 78 and 79 GHz does not move the spurs.
I would greatly appreciate your help. Thank you !
RJ
