Tool/software:
Hi expert,
We are currently investigating an unexplained phenomenon related to harmonic generation and unexpectedly low ADC amplitudes when using the AWR2944 with our designed antenna compared with the AWR2944 EVM board. Below is a summary of our observations and questions:
Test Setup
- Boards:AWR2944 with our designed antenna (AWR2944 custom board) and AWR2944 EVM board
- SDK: mmwave_mcuplus_sdk_04_04_01_02
- Firmware: mmw_Demo_DDM
- Data capture: DCA1000
Observation
I used the DCA1000 together with the custom board and EVM board and collected ADC data with mmw_Demo_DDM, without placing a CR in front of the radar.
We found that in the gym environment (a relatively clean indoor environment with fewer objects), when using the custom board, the ADC data amplitude was very small, about +/-200, whereas with the EVM board, the ADC data amplitude was about +/-4000.
In addition, to verify whether the custom board would still show an amplitude of only +/-200, I set up a simple experiment at my office desk.
In this experiment, the amplitude was around +/-5000 in the office.
Next, I placed a static target (30 dBsm CR) in front of the radar at 5 m.
The amplitude of the ADC data received by the custom board was unexpectedly small (about +/-200), even with the 30 dBsm CR present in gym.
However, in the office, the amplitude of the ADC data received by the custom board matched expectations and became significantly larger compared to the free space case.
I further processed the collected ADC data using 2D-FFT in MATLAB and observed the R-V heatmap, focusing on the range domain at zero velocity.
Focusing on the first 10 m (since beyond 10 m multipath reflections appear in indoor space), when comparing the free space case with the 30 dBsm CR case:
- With the custom board in the gym, a 30 dBsm CR appears at 5 m, but a harmonic appears at 10 m with lower ADC amp..
- With the EVM board in the gym, a 30 dBsm CR appears at 5 m, but no harmonic appears at 10 m with higher ADC amp..
- With the custom board in theoffice, a 30 dBsm CR appears at 5 m, but no harmonic appears at 10 m with higher ADC amp..
This makes me very curious why, with the custom board and a 30 dBsm CR in the gym, the ADC amplitude is very small and harmonics appear, while with the EVM board under the same condition, the ADC amplitude is larger and no harmonics are observed.
Questions
- Do you have any ideas about observation above ? Could it be caused by some component in the analog chain, or does the 2944 itself have a self-calibration mechanism (e.g., automatically adjusting TX gain or TX power) that affects the amplitude?
- In addition, I would like to further understand which observable parameters exist in the analog chain that can help identify where the issue originates when harmonic and low amplitude issues occur, which parameters do you think we can monitor, such as IF stage, etc.?
- And how can we read these parameters? As far as I know, these parameters can be read using mmWave Studio GUI, but currently we are facing an issue that the custom board cannot connect to the GUI, is there any other way to read the analog chain-related parameters?
Best Regards,
JiaQuan Zhao
