• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 1364 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

IWR6843: Calibration best practice

Aaron Fulton
Prodigy 50 points
Part Number: IWR6843

Can you advise the best setup for calibration. I have designed a product using the IWR6843 module and am now working through quality checks and was wondering what the best setup for calibration is.

I have been using the calibration procedure outlined in the documentation (as I understand it) but I'm seeing some very large calibration values which I need to get to the bottom of. I am not sure if this because of a manufacturing tolerance issue or if my calibration setup is the problem, or if this normal. I will also need to build a test rig for the manufacturer in due course so need something that will work in a factory setting. I am using the reference design for the IWR6843ISK antenna.

For testing, I have a 1200mmx1200mm sheet of metal which is placed 1.000m away from the sensor and I then run the calibration routine as explained in the documentation. I have also found the results to be sensitive to the exact setup (moving things a few mm can make a noticeable difference).

Typical calibration results:
Range bias:0.08221510052680969,
Phase bias real 1:-0.6945499777793884, Phase bias complex 1:-0.01834000088274479,
Phase bias real 2:-0.4772599935531616, Phase bias complex 2:-0.6360200047492981,
Phase bias real 3:-0.6244800090789795, Phase bias complex 3:-0.5064399838447571,
Phase bias real 4:-0.12756000459194183, Phase bias complex 4:-0.6656500101089478,
Phase bias real 5:-0.7667499780654907, Phase bias complex 5:-0.23654000461101532,
Phase bias real 6:-0.32686999440193176, Phase bias complex 6:-0.8762199878692627,
Phase bias real 7:-0.5512099862098694, Phase bias complex 7:-0.76569002866745,
Phase bias real 8:0.059450000524520874, Phase bias complex 8:-0.7704799771308899,
Phase bias real 9:-0.8706099987030029, Phase bias complex 9:-0.021940000355243683,
Phase bias real 10:-0.5823400020599365, Phase bias complex 10:-0.7919600009918213,
Phase bias real 11:-0.7581800222396851, Phase bias complex 11:-0.6520699858665466,
Phase bias real 12:-0.12848000228405, Phase bias complex 12:-0.8274800181388855

Questions:
1. What does the ideal calibration 'test rig' look like?
2. If I were to write a testing specification for the manufacturer, what would sensible tolerances be?
3. What could be causing such large phase bias values and should I be concerned?

  • 0
    TI__Guru 50240 points

    Hi Aaron,

    Which calibration document are you referring to?

    We'll need a few days to look into this.

    Regards,

    AG

  • 0 in reply to Akash Gondalia
    Prodigy 50 points

    Thanks AG,

    For the calibration documentation, I'm referring to chapter 6.8 "Range Bias and Rx Channel Gain/Offset Measurement and Compensation" of the MMWAVE SDK User Guide (Document Version: 1.0) and the mmwave_sdk_<ver>\packages\ti\datapath\dpc\objectdetection\<chain_type>\docs\doxygen\html\index.html document. We don't use the mmwave Visualiser to run the calibration but do use the same profile_calibration.cfg and commands as per the documentation.

    regards

    Aaron

  • 0 in reply to Aaron Fulton
    TI__Mastermind 25795 points

    Aaron,

         Ideal calibration setup would be similar to the setup shown below. 

    This would need controlled environment, where radar reflection should be from known reflection source. For this mmWave absorbers could be used to suppress the clutter.

    If there are clutter it might results in wrong calibration values. 

    Known RCS corner reflector is a better source, it provides point target, as compared to metal plate, it may results in multiple objects reflecting from multiple places making calibration harder.  

         

    You could refer to the Antenna radiation pattern measurements depending upon type of design chosen (ISK, ODS, AOP) for various antenna. It has antenna pattern for various antenna pairs used based on that you could arrive at the tolerance that your system desires. 

    Typically phase will have large dynamic range i.e. Receiver antennas are kept at lamda/2 spacing at 60GHz phase of this signal varies rapidly depending upon angle of arrival. In this case, I believe problem is due to multiple reflections that you are picking up from multiple range bins (Probably due to uncontrolled setup and lack of point target).  And in each range bins phases could be different. You need to ensure setup is cleaned up for this. 

    Thanks and regrds,

    CHETHAN KUMAR Y.B.