IWR1642: The IWR1642 can not transmit at low temperature.

Hello,

Things we expect may work differently at temperature difference are calibration.  

Looking at the mmwave User Guide, we need to trace the execution of each API, looking for the first error code.

a) you have commented the power supplies are OK, please check them in the -20C environment

b) if you enable the PMIC_CLKOUT, or the MCU_CLKOUT, following the TRM register programming, you can get a divided PLL output - check for desired frequency

c) using the section of the mmwave sdk user guide, 5.3.1.2 for xWR1642, you can check each API call, and inspect the return status code, or the check status.  RF calibration (rlRfinitm, RK_RF_AE_INITCALIBSTATUS)

The status at room temperature, when it works, should match the status at -20C, given the power supply, the 40Mhz clock are OK.

Regardsl

Joe Quintal

Hello,

This indicates you don't have a clock problem.

I sent the error information, you have a receiver gain error.   The bit field for this is shown in the initial and run time calibration mask, and the error code indicates if its not a '1' for the status bit that this bit is in error.

Please download the "dfp" radar studio package.  In the docs folder there is an AWR1xx_Radar Interface_Control.pdf.

The calibration masks (what tests are run) is discussed in 5.2.9

In the CLI commands the Initial and RunTime calibration mask, can be set to 0 for bit 10.  You should then get the Tx output, if there are no other errors.

The calibration status is discussed in section 5.13.5.  The masked calibrations have to return a '1' to indicated PASS status.

The designer indicates that downloading the latest DFP package, the new BSS firmware file needs to be compiled and used for your test.

Regards,

Joe Quintal

Hello,

The Rx Gain calibration is used to adjust for part variation, aging, and device temperature by looping back a Tx internal signal to the Rx.

There is an internal adjustment for the 2 variable Gain sections covering the LNA gain from 24-48db.

As suggested in another E2E post, the design engineer indicates if you use a later version of DFP this is supposed to be fixed.

If you operate without Rx gain calibration, the LNA Rx Gain setting may vary over temperature and device.

Please try to use a later dfp BSS package and compile and test.

Regards,

Joe Quintal

AWR1xx_Radar_Interface_Control.pdfHello,

In the radar Studio documentation, there is an ICD document, I will attach it this time.   Under the Initial Self test, and the Repetitive testing there is a bit mask for each test.   Once you know which bit field is failing you can mask it off for temporary testing.   In most cases Radar Subsystem Test errors are handled through BSS firmware upgrades, so downloading the latest dfp package from the TI website , and using the BSS image for your device is needed.

I have tried to answer this emails 2x already, the answer does not seem to be going through E2E.

Please look up the Initial calibration command and bit field in the ICD document attached.

Regards,

Joe Quintal