• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 32 subscribers
  • Views 464 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

IWR1642: Eval Board and Demo Questions

Ryan.Bishop
Expert 3900 points
Part Number: IWR1642
Other Parts Discussed in Thread: DCA1000EVM

Hello,

I am assisting one of our customers as they evaluate the IWR1642 eval board from TI. They had the following questions, sorry it's a long list but hopefully some of them are easy answers. Let me know if you have any questions, I can be contacted via email if so:

1. The overall goal of our current experiments is to pull the raw data off of the sensor for analysis via an external program to evaluate the radar. Is this possible ? What is the best way to extract the data?

2. Has TI experienced any IC over heating issues when trying to use more memory or increase # of chirps over UART? We increased the chirps and the card stopped working (and got very hot) and is now DOA.

3. Why in the Vital Sign Demo only 2 chirps are collected and only one is processed? Does it take too much time to compute heart rate and respiration from more chirps per frame to improve SNR?

4. Why is there a limitation in the Frame duration (~30 msec) in the Demo? Is just for the Demo or a general guideline?

5. What would be the best way to collect just the fft of the chirps (no processing for heart rate and respiration) and what would be the max rate of chirp we can collect?

6. Can we go above 1Mbps transfer rate from UART?

7. What are the advantages/disadvantages to using more than one antenna ?

Thank you!
Ryan B.

  • 0
    TI__Guru 50240 points
    Hi Ryan,

    I can answer question 1 right now, but questions we'll need some time for questions 2-7.

    1. The overall goal of our current experiments is to pull the raw data off of the sensor for analysis via an external program to evaluate the radar. Is this possible ? What is the best way to extract the data?

    [AG] The new DCA1000EVM is specifically design for raw data capture over Ethernet. This is going to be by far the best way to proceed with Raw Data Capture evaluation.


    Cheers,
    Akash
  • 0 in reply to Akash Gondalia
    TI__Guru 50240 points
    Hi Ryan,

    2. Simple answer, yes in certain cases changing the mmwave parameters changes the heat generated. There is an area on the back
    Of the board where thermal putty can reject heat to a metal plate (must be electrically insulated).

    The Profile – Chirp – Frame
    Has to be tuned to send the data over the UART.
    Typically the idle time is increased to allow the DMAed data to be exported before the next chirp or frame.
    (depends on use case).
    Increasing the idle time, also lowers the average heat from the board.

    If the mmwave sensor is DOA, measure the power rails, and LDOs. Use CCS and follow the datacapture demo to see
    If the device boots, (check Warm Reset release, and NERROR Out).


    7. The number of antennas changes the angular resolution (Rx, or MIMO Tx). Placing the antennas in a different orientation than the EVM, can provide for more elevation coverage if needed.

    Use the Sensing Estimator, for the system use case, change the number of Rx or Tx antennas to experiment.




    Cheers,
    Akash
  • 0 in reply to Akash Gondalia
    TI__Mastermind 21295 points
    For question 6, the IWR16xx/14xx Industrial Family TRM (www.ti.com/.../swru522b.pdf) lists a max UART (a.k.a. SCI) baud rate of 3.125 Mbps.

    Regards,
    John
  • 0 in reply to JohnH
    TI__Guru 50240 points
    Hi Ryan,

    3. Yes, in principle more chirps can be collected per frame and these chirps can be averaged to improve the SNR. The next version of the Vital Signs Demo will include this option. The only tricky part is to figure out how many chirps to average as averaging would only improve the SNR if the signal is “constant” and the noise is random and in the vital signs use-case we have a fluctuating signal. My guess is averaging over 8-16 chirps should be fine.

    4. Not sure where the 30 ms number is coming from. The frame Periodicity set in the current demo is 50 ms and that’s just because of the limited throughput through the UART.
    Based on the Nyquist criteria a sampling rate of 20 fps (or 50 ms) should be enough to capture vital signs signature which can have a max frequency of up to 4 Hz. However, higher sampling rates will be desirable when a person is moving (to prevent phase wrapping) or if we have noise sources in the scene vibrating greater than the Nyquist limit (for example Car engine vibrations) to prevent frequency aliasing.

    5. The GUI has a “Save Data” option. That would save the complex FFT of the chirps and you can use your own processing methods to estimate the heart-rate and breathing-rate. If you are using the Matlab version of the GUI then please make sure that the plots are disabled (through a CheckBox) before saving the data. The max rate would again be limited by the speed of the UART , computer specs etc. On my PC I can get up to 20 chirps per seconds


    Cheers,
    Akash