MMWCAS-RF-EVM: MMWCAS-RF-EVM: azimuth range is not as expected in TXBF output result by processor sdk

Could you share the link to the previous post regarding the updated release?

Hi Stanley,

Here's original question link:  e2e.ti.com/.../908543

Hi Wei03, 

Please see the attached draft of the Cascade TX phase shifter calibration example which will be released in the next mmWave Studio release. This shows a general TX phase shifter calibration flow as well as example mmWave Studio scripts to be run with the MMWCAS-RF-EVM and Matlab. This flow will provide the TX calibration data you need for setting up a calibrated TX beam-forming/steering example. However, additional work will still be needed to port that into the TDA2 Processors SDK radar TX-BF example.

4505.AWRx_TX_Phase_Cal_Scripts_20200804_rev4.zip

Please let me know if that helps. 

Thank you,

-Randy

hi Randy,

1. while running "cascade_TX_Phase_Calibration.m", it promps error at "

Hi Wei03, 

Were you able to generate an appropriate TX Phase Shifter lookup table? 

Thank you,

-Randy

Hi Wei03, 

The process you are outlining looks roughly correct, but I would still recommend following the whole procedure I outlined. In my rev4 release there is also a set of mmWave Studio Lua scripts and Matlab scripts for verifying that the TX phase shifter calibration was effective, by setting up a simple TX-BF example configuration with and without the calibration data applied. Have you given that a try as well?

Also, are you saying that you used the phase offset data in the calibrateTXPhaseResult.mat directly? Or did you do the necessary, additional processing on this data to convert it into a beam-angle vs. phase shifter programming lookup table as I did in my examples?

The rev3 version of that TX phase cal tool had a few problems and calibration data capture took too long. That is why I posted the rev4 version. However, rev4 was only tested with mmWave Studio 3.0.0.7 and DFP 2.1.5.3. 

You need to move to the mmWave Studio 3.0.0.7 anyway since it is a release that is fully compatible with the AWR2243 and DFP2.x releases. It is available here as part of the MMWAVE-STUDIO-2G package: https://www.ti.com/tool/MMWAVE-STUDIO

DFP 2.1.5.3 is availble here as part of the MMWAVE-DFP-2G package: https://www.ti.com/tool/MMWAVE-DFP

Thank you,

-Randy

Randy,

There is a bug in the Cascade_Phase_Shifter_Calibration_AWRx.lua you shared in rev4 package.

The number of parameters of ar1.FrameConfig_mult in mmwave studio 3.0.0.7 is 8, instead of 9. Pls find my correction below and kindly update it in your code base.

--   if (0 == ar1.FrameConfig_mult(1,start_chirp_tx,end_chirp_tx,nframes_master, nchirp_loops, Inter_Frame_Interval, 0, 0, 1)) then -- mmwave studio 2.x
   if (0 == ar1.FrameConfig_mult(1,start_chirp_tx,end_chirp_tx,nframes_master, nchirp_loops, Inter_Frame_Interval, 0, 1)) then  -- mmwave studio 3.0.0.7
...
   -- Slaves
--   if (0 == ar1.FrameConfig_mult(deviceMapSlaves,start_chirp_tx,end_chirp_tx,nframes_slave, nchirp_loops, Inter_Frame_Interval, 0, 0, 2)) then -- mmwave studio 2.x
   if (0 == ar1.FrameConfig_mult(deviceMapSlaves,start_chirp_tx,end_chirp_tx,nframes_slave, nchirp_loops, Inter_Frame_Interval, 0, 2)) then -- mmwave studio 3.0.0.7

Randy,

1. I compared the Cascade_Phase_Shifter_Calibration_AWRx.lua in rev3 and rev4 package and they are same. I got similar error as Wei when I ran cascade_TX_Phase_Calibration.m in rev4 package. I am using mmWave Studio 3.0.0.7 and DFP 2.1.5.3

Would you pls kindly check if you have updated lua to capture the calibration data for TXBF?

dataFolder_calib_data_info(folderIdx).name = device1_txChannel1_phaseShiftValue00
paramFile= C:\ti\mmwave_studio_03_00_00_07\mmWaveStudio\PostProc\device1_txChannel1_phaseShiftValue00\device1_txChannel1_phaseShiftValue00.mmwave.json
Index exceeds matrix dimensions.

Error in cascade_TX_Phase_Calibration (line 150)
            calData_1DFFT(:, :, idxTX, idxRX) = fftshift(fft(calData(:, :, idxRX, idxTX),
            genCalibrationMatrixObj.numSamplePerChirp, 1));

2. I can't found parameter_file_gen_TXPS_VerificationCalib_json.m in rev4 package which you mentioned in the user guide. Would you pls help to share？

\mmWaveStudio\MatlabExamples\4chip_cascade_MIMO_example\main\cascade\paramGen
 parameter_file_gen_TXPS_VerificationCalib_json.m 

Randy,

Would you pls help again?

Hi Randy,

I tried the new scripts in the office, they could run with no error. I will try again next week in mmwave chamber to compare the results. Many thanks.

Could you please provide how to port the calibration data to TDA2 Prossessor_SDK_Radar?

Thanks,

Wei

Hi Randy,

I tested AWRx_TX_Phase_Cal_Scripts_20200811_rev4.zip in a mmwave chamber, but the result is not expected like the doc.
When running Cascade_Phase_Shifter_Calibration_AWRx.lua, it occurs error. The complete 64 phase is complete in 3 times by changing the phase start index.

Then LUT generated as follows, it looks abnormal. The data in red is not close to zero phase.

0,6,37,32,34,3,59,63,29
0,8,42,35,42,13,7,14,50
0,10,47,42,52,24,20,27,2
0,12,52,50,60,35,34,45,20
0,14,56,56,5,46,46,60,36
0,17,60,63,14,58,60,11,55
0,19,1,5,22,4,9,25,8
0,21,5,11,28,15,22,41,25
0,22,9,18,40,25,35,58,43
0,24,13,24,50,39,50,9,62
0,26,18,29,58,50,62,24,15
0,32,22,38,3,60,11,40,34
0,30,26,45,12,7,24,56,51
0,34,30,52,21,18,38,8,5
0,35,35,59,32,32,52,22,22
0,0,0,0,0,0,0,0,0
0,40,46,8,47,52,14,54,59
0,42,50,14,57,63,27,6,12
0,46,54,21,2,10,41,21,32
0,47,59,26,10,21,55,36,49
0,51,63,35,19,33,4,53,2
0,52,3,41,27,44,17,5,19
0,54,8,49,36,55,32,20,36
0,57,12,55,46,1,43,35,55
0,59,17,62,55,12,58,51,8
0,61,21,4,63,23,6,2,25
0,63,25,10,9,35,19,18,45
0,2,32,17,18,46,34,34,62
0,4,34,23,25,57,46,49,15
0,6,37,32,35,3,59,63,33
0,8,42,35,42,14,8,15,50

After modify LUT and run Cascade_Configuration_TXBF_AngleSweep.lua script, the uncalibrated and calibrated beam angle looks the same.

Hi Randy, 

Could you confirm following modifications are correct in TXBF_PS_LUT_Generate.m

psOffsetCal(antIdx,1)= 0;

for psIdx = 2:64

psOffsetCal(antIdx, psIdx) = phaseOffsetValuesUnwrap(txIdx, SelectedRx, psIdx-1) 

==> 

%%% psOffsetCal(antIdx,1)= 0;  %phase shift should not be zero for boresight

for psIdx = 1:64

psOffsetCal(antIdx, psIdx) = phaseOffsetValuesUnwrap(txIdx + (devIdx-1)*3, SelectedRx, psIdx) %txIdx error here

Then the LUT are as follows, there are phase shift at boresight

0,5,36,14,44,9,52,14,48
0,7,41,21,53,20,0,32,0
0,9,46,26,61,32,12,46,17
0,11,51,34,5,43,25,62,35
0,13,55,40,14,54,39,11,53
0,16,59,49,22,0,53,26,6
0,18,63,55,33,11,1,44,23
0,20,4,61,40,22,14,59,41
0,21,8,3,50,34,26,9,60
0,23,12,10,59,45,41,24,12
0,25,17,17,3,56,56,41,32
0,31,21,23,12,2,4,58,50
0,29,25,28,21,14,17,7,2
0,33,29,38,31,24,30,23,20
0,34,34,44,39,35,45,39,36
0,36,40,51,50,49,59,57,57
0,39,45,57,57,59,7,6,9
0,41,49,0,2,6,20,22,26
0,45,53,6,10,17,34,39,46
0,46,58,12,20,26,49,55,63
0,50,62,20,27,39,62,5,17
0,51,2,25,36,51,9,20,35
0,53,7,33,46,62,22,38,53
0,56,11,39,56,8,37,53,6
0,58,16,46,0,19,51,3,23
0,60,20,53,9,30,63,18,41
0,62,24,60,18,41,11,34,60
0,1,31,2,25,53,24,50,12
0,3,33,8,34,63,39,0,30
0,5,36,15,44,10,53,15,49
0,7,41,21,53,21,0,30,0

And I don't find "cascade_TxBF_dataCapture.m" in mmwave studio3.0.0.7, and I use the one in version 2.1.0.0 to do txbf and then  "cascade_TxBF_signalProcessing.m"?

Thanks,

Davy