AWR6843AOP: One-time calibration and runtime calibration implementation

Hi, ED:

1.Based on the above e2e discussion, we should be able to increase the robustness of the equipment to temperature effects and process differences through the one-time calibration function?

(But I speculate that a complete runtime calibration will have better results.)

If temperature change, there can be a step change in TX/RX gain when running one-time or run-time calibration.  The difference is that when you run one-time, the host can stop the signal processing chain, and reset the signal processing chain after the gain change if the signal chain expect no sudden change over time.  For example, if the clutter removal algorithm designed to run cross frame boundary, then a sudden gain change will be devastated.   In such a case, the host need to reset the clutter removal estimation before apply the gain change.  This is just an example to show that the host controlled one-time calibration can have a better control over the system. 

2. "Cascade Coherency and Phase Shifter Calibration Application Note" Does the content of this document apply to the AWR6843AOP?

AOP device is not designed for cascade, so this application notes should not be applied. 

3. There is a set of default LUTs in Ti’s SDK for runtime calibration? But it can only be applied to Ti's EVB. If the user designs the PCB by himself, does it need to be calibrated to obtain the LUT with the corresponding temperature?

The TX and RX gain LUT over temperature is filled up during the boot-time RF calibration, and can be updated using run-time calibration or one-time calibration.  It is part of the calibration process and part of the firmware.   Customer is not expect to touch this portion. 

4.  The LUT structure is described in the AWR1xx Radar Interface Control document. What is the link to this document?

the newest version of ICD is in DFP 01.02.06.03, download and install it. and find ICD in C:\ti\mmwave_dfp_01_02_06_03\docs

https://www.ti.com/licreg/docs/swlicexportcontrol.tsp?form_type=2&prod_no=mmwave_dfp_01_02_06_03_win32.exe&ref_url=https://software-dl.ti.com/ra-processors/esd/MMWAVE-DFP/01_02_06_03&_ticdt=MTY0MTk3Njc5OHwwMTc5N2VhMGVhNTQwMGFiMzNhNjQzYTRmODcwMDMwODMwMDIxMDdiMDA3ZTh8R0ExLjIuMTUxNTg5OTc4My4xNjIxODI1OTQ1

5. Among the calibration list, it seems that only TX power calibration, RX gain calibration, and TX phase shifter calibration are more important? And the users only need to focus on these 3 calibrations?

Please provide more details for this questions.  Important in which way?

Best,

Zigang

Hi Zigang,

Thank you for your reply.

Regarding the reply to question 1, I still have some unclear parts.
1. Will TX gain and RX gain change due to temperature changes? Or do they change because temperature changes trigger run-time calibration?
2. Will the one-time calibration set the appropriate TX gain and RX gain based on the current temperature, and reset the signal processing chain?
3. Run-time calibration will also set the appropriate TX gain and RX gain based on the current temperature, but it will not reset the signal processing chain and cause abnormal changes in SNR or point cloud?

4. Is it enough for the reset algorithm to trigger a one-time calibration? Are other settings required?

Reply to the third question
LUT calibration is mentioned in the 3.8 TX Power Calibration and 3.9 RX Gain Calibration in the document "Self-Calibration in TI’s mmWave Radar Devices".
The file describes the following: "The LUT can also be replaced with a user-programmed LUT"

This confuses me, anyway, after your explanation, for now we will try to set up a CLI controlled one-time calibration test first.

Best Regards,

ED

HI, ED:

1. Will TX gain and RX gain change due to temperature changes? Or do they change because temperature changes trigger run-time calibration?

[ZY] The run-time calibration is periodic, and whenever it runs, it will check the temperature first.  If the temperature change reaches certain value, it will trigger the TX and RX gain change.  

2. Will the one-time calibration set the appropriate TX gain and RX gain based on the current temperature, and reset the signal processing chain?

[ZY] During one-time calibration, the host triggers the calibration, and it will again check the temperature first, and TX and RX gain will be updated if the temperature change reach a certain value.  The signal chain reset is not part of the one-time calibration, the host needs to make a decision after one-time calibration and issue the process reset.   Customer also need to add the implementation of the signal chain reset process. 

3. Run-time calibration will also set the appropriate TX gain and RX gain based on the current temperature, but it will not reset the signal processing chain and cause abnormal changes in SNR or point cloud?

[ZY] Run-time calibration is periodic and automatic.  After run-time calibration, the chirping of the next frame is done continuously.   

4. Is it enough for the reset algorithm to trigger a one-time calibration? Are other settings required?

[ZY] I do not understand your question here.  

Best,

Zigang