CC2652P7: How to achieve the lowest power consumption?

Part Number: CC2652P7
Other Parts Discussed in Thread: SYSCONFIG, CC1352P7

Dear sir,

Now we are trying to use CC2652P7 with zigbee mode. And the device is a battery-powered end device.

We cannot achieve the goals stated in the manual at the lowest power consumption.

According the datasheet, the lowest power consumption should be 0.9uA, but our lowest power consumption is 200uA.

bae32059-1c5a-4f69-9c3f-4f39bf40d6b6.png

Please let me know which aspects we should check?

If you need more information, also please let me know.

Thanks

Chad

  • Hi Chad,

    Here are a few resources:

    https://dev.ti.com/tirex/content/simplelink_cc13xx_cc26xx_sdk_8_30_01_01/docs/zigbee/html/zigbee/power_configuration.html 
    https://www.ti.com/lit/swra478 
    https://www.ti.com/lit/swra625 

    Make sure you have set CUI_DISABLE and removed BOARD_DISPLAY_USE_UART from the pre-definitions.  You will also want to increase the poll period in SysConfig and plausibly decrease the output TX power.  Keep in mind that 1 uA is only obtainable without any radio activity (i.e. no TX/RX from data polling or otherwise).  You will want to evaluate a TI LaunchPad EVM to confirm the power measurement setup before considering your own custom hardware and software.

    Regards,
    Ryan

  • Hi Ryan,

    We had tested this case:

    I think the pre-definitions should be closed in this case, but our bottom current still have 200uA。And for the other URL, we also checked, but nothing is discovered.

    Do you have any other suggestion?

    Thanks

    Chad

  • It's good that you are starting with TI Drivers examples.  My original suggestions apply to Zigbee projects and should be referenced in the future.  The gpiointerrupt example is the correct example for Standby low-power mode, as gpioshutdown is specific to the Shutdown low-power mode which is not utilized by Zigbee End Devices, but either should demonstrate operating at less than 1 uA by default.  This is verified by TI before a SDK is released.  Therefore I believe there is an incorrect configuration with your hardware setup or power measurement tools.  Can you please further describe these details?

    Regards,
    Ryan

  • Hi Ryan,

    Thanks for your suggestion. 

    We used a power monitor to supply power。

    I couldn't find the manual of our monitor on their website. Here is a similar one on their website:bluebird-labs.com/.../ 

    For our model, the label indicates that it can measure 100nA。

    About the hardware configuration, I am not sure what are you mean. Now we used a module: https://www.szrfstar.com/product/cc2652p7_matter_zigBee_thread_ble_rf_ipex_module-en.html, We have disconnected the other circuits, and only this module is working. We also tested directly powering the module, same result.

    Thanks

    Chad

  • Hi Ryan,

    Another problem, we use the example of zed_genericapp_LP_CC1352P7_4_tirtos7_ticlang as our basic project.

    And for the problem, now we are trying to shutdown mode to confirm the bottom curent, but it is failed. As you know, at the end of shutdown, there is a while(1), usually, this while should not run. but it stay in the while(1) when we debug.

    Even we move the Power_shutdown to the beginning of main function, it still does not work.

    Did we do something wrong somewhere?

    Thanks

    Chad

  • Since you are using a non-LaunchPad hardware module you will need to select "USE CUSTOM BOARD" in the Board View of SysConfig.  You should also remove all non-existent LEDs and pushbuttons that do not exist on your board, except perhaps for a GPIO that can wake the device to restart from shutdown.  You will need to make sure that GPIO is not already at the shutdown wakeup level when you are attempting to enter shutdown.  I would also appreciate if you could report on your evaluation of the gpiointerrupt example as it should be capable of achieving 1 uA which I would like to compare against your gpioshutdown measurements.

    Confirm sure you have installed and are using the correct SimpleLink F2 SDK Dependencies in your project.  

    You can copy simplelink_cc13xx_cc26xx_sdk_8_30_01_01\source\ti\drivers\power\PowerCC26X2.c directly into your zed_genericapp project to determine why Power_shutdown fails.  Shutdown is typically not considered for ZED as it required a restart of the device for which it must then rejoin the previous Zigbee network.  This is why standby mode is commonly acceptable, although it should not be preventing you from using shutdown mode.

    Have you considered contacting RF Star for more advice from the manufacturer concerning this matter?

    Regards,
    Ryan

  • Hi Ryan,

    Sorry for the later reply.

    We measured a current of 1uA using the shutdown example. It should be a problem with the clock source. Do you have any examples of switching between LF and HF?

    Beside, for the 100nA-reset and shutdown mode, can it only be exited through an IO interrupt? Just to confirm.

    Thanks

    Chad

  • Hi Chad,

    The Power TI Driver is fully responsible for switching clocks in the back end during power state transitions, however be sure to select the correct clock source in the SysConfig Device Configuration Module.

    Shutdown can only be exited through an IO interrupt or technically a toggle of the RST pin without power cycling the device.

    Regards,
    Ryan

  • Hi Ryan,

    About the standby mode. Datasheet and driver are different about the crystal oscillator。

    About the 32.768kHz crystal oscillator, how to understand it?

    Thanks

    Chad

  • Once the LF XOSC is qualified and switched to then standby low power mode is enabled.

    Regards,
    Ryan

  • Hi  Ryan,

    Another problem, when we use the XDS110 debugger, where should I set the break-point so that I can monitor the switch to Standby mode?

    Thanks

    Chad

  • PowerCC26XX_standbyPolicy from <sdk_install_directory>\kernel\tirtos7\packages\ti\dpl\PowerCC26X2_tirtos.c (assuming TI-RTOS) calls Power_sleep in PowerCC26X2.c if the entry conditions are correct.  You should be able to bring these files into your workspace for debugging if warranted.

    Regards,
    Ryan

  • Hi Ryan,

    Good news, We confirm that it is caused by the hardware. After hardware improvements, it has been significantly reduced to 20uA and there is still room for improvement.

    I will close this case now.

    Thanks

    Chad