• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 30 subscribers
  • Views 239 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.

AWR2944: Questions related to runtime calibration

L mx
Prodigy 90 points
Part Number: AWR2944

Tool/software:

Hi Ti Teams:

Currently, I have enabled the runtime calibration function in my code. The calibration module mask is set to: 0xF10. I understand that this configuration corresponds to enabling all runtime calibrations.

The testing method is to place the radar in an environmental chamber with temperature variations ranging from -40℃ to 85℃, and then back down to -40℃. Then continuously monitor the CR located 1.5 meters ahead, and finally extract the phase and power data of this CR.

The test results are as follows.  

The Y-axis represents power (in dB), and the X-axis represents the number of radar frames.,These two pictures show the test results of power. The left picture represents the run_time_cal is enable, while the right picture shows the run_time_cal is disable. In my opinion, these test results are in line with my expectations.

However, in the phase data of the same time period, I don't feel that the calibration effect has been reflected.

The Y-axis represents the relative error of phase between each channel(The top yellow line segment in the figure represents the radar TX temperature, while the remaining 16 line segments represent the relative phase errors between each channel.), with the unit being degrees. The X-axis represents the number of radar frames.

These two pictures show the test results of power. The left picture represents the run_time_cal is enable, while the right picture shows the run_time_cal is disable.

When comparing these two graphs, apart from the area marked by the red box in the left graph where there is a phase change, indicating that calibration has been performed, there is no significant advantage in the other data. The relative error remains very large.

Also, attach an absolute phase image of all channel setting when run_time_cal is enabled.

It can be observed that the phase changes in each channel are not consistent.

Therefore, for the run_time_cal function of the phase, is this a normal result of the chip's performance, or is there something wrong with my configuration?

Best Regards

Minxi

  • 0
    TI__Expert 4060 points

    Hi Minxi,

    Let me get back to you on this in a day's time. 

    Regards,

    Shailesh

  • 0 in reply to Shailesh Joshi
    TI__Expert 4060 points

    Hi Minxi,

    There are a few points to note on this:

    There is no internal calibration for phase mismatches across temperature as these mismatches depend on board design and other RF aspects. Hence, any specialized compensation for the same should be done by the customer in post-processing.

    As per evaluation in TI's lab, if the phases are made 0 by performing a zero-angle calibration at 25C temperature, then across temperatures a phase mismatch of +/- 5C should be observed. 

    Regards,
    Shailesh

  • 0 in reply to Shailesh Joshi
    Prodigy 90 points

    Hi Shailesh

    I have a few questions regarding your reply.

    1.I noticed that the "run_time_cal" module in the manual has "TX_PHASESHIFTER_CALIBRATION_EN". According to your explanation, isn't this used for phase calibration when the temperature changes?

    2.If I need to perform phase calibration at 25℃, could you please tell me how to do it? Are there any relevant documents available? At present, I only see two files. One is the self-calibration document within page 2944, and the other is an extremely old document regarding the calibration of the cascading chip 2243.

     

    3.If the aforementioned phase calibration method is used, is it necessary to obtain the calibration result by using “AWR_PHASE_SHIFTER_CAL_DATA_SAVE_SB” and then restore the calibration result by using “AWR_PHASE_SHIFTER_CAL_DATA_RESTORE_SB”? Also, is it necessary to disable the phase calibration function in RFinit and run_time_cal?

    Best Regards

    Minxi

  • 0 in reply to L mx
    TI__Expert 4060 points

    Hi Minxi, 

    From what you have plotted in the query, it presents the phases of the channels across temperature for a given phase shifter configuration. Whereas, the TX phase shifter calibration is for accurately achieving a desired "shift" in phase, e.g. shifting the TX phase by 90 deg. So, the TX phase shifter feature is not being used when you are measuring the phase mismatch across temperature. While there are calibrations for ensuring that TX phase outputs follow the configured phase shifts, there are no device-calibrations for correcting received phase imbalances across temperature.

    The zero-angle calibration that we expect the customers to do here is to put a CR in front (at 0 deg angle) of the radar at a few meters distance at 25C, measure and record any phase and amplitude offsets across all channels. Apply this correction to the data every time typically along with range processing. With this, we expect the deviation in the phase imbalance to be within +/- 5 deg across temperatures.

    The calibration save and restore features are for avoiding boot-time calibrations in field, as performing these calibrations in field can be risky when there is RF interference in the environment. If the device boots and performs boot-time calibration with interference present, it can degrade the performance until boot calibrations are performed again. Hence, it is recommended to perform the boot calibrations once in factory and always restore the data when in field. 

    Let me know if you have any queries on this.

    Regards,
    Shailesh

  • 0 in reply to Shailesh Joshi
    Prodigy 90 points

    Hi Shailesh

    Thanks for your reply.So now I have only one question left.

    I still don't quite understand the operation steps for phase calibration of the customer's factory.

    Now I understand how to set up the environment. But could you provide more detailed explanations regarding data collection and application of correction values?Thank you.

    Best Regards

    Minxi

  • 0 in reply to L mx
    TI__Expert 4060 points

    Hi Minxi,

    Is your question on zero-angle correction data?
    With the corner reflector at boresight, the phase/amplitude variation across channels is expected to be 0. If it is not zero (likely the case, due to any imbalances), the measured imbalance is saved in a flash memory for each channel. Then the application code should correct the phase and amplitude in each chirp per-channel. In AWR2944, HWA's complex vector multiplication feature can be used to apply this correction before or after performing range FFT. The correction would have been recorded at room temperature in factory but if the same correction is applied for all temperatures in field, an imbalance of +/- 5 deg might be seen. So, one might choose multiple temperature points to perform zero-angle calibration for better accuracy if needed.

    Regards,
    Shailesh

  • 0 in reply to Shailesh Joshi
    Prodigy 90 points

    Hi Shailesh

    Thank you,Now I get the idea.

    Then we can close this issue.

    Thanks again

    Best Regards

    Minxi