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.

IWR1443: minimizing power consumption - questions

Expert 6460 points
Part Number: IWR1443
Other Parts Discussed in Thread: TIDEP-0091, SYSBIOS

Team,

we are exploring various ways to lower power consumption of IWR1443.

1. During booting, we noticed there is a serial data stream sent out on UART at the end of RTOS booting. This data stream lasts for about 10 msec. Could this be somehow switched off? Could this be somehow eliminated in order to shorten the bootup time?

2. The datasheet clearly states that we have to release nRESET when all power voltages are applied and stable. Would it be possible NOT to apply power for some RF stages e.g. during system bootup, and only later on, in order to reduce consumption?

thank you.

  • Hi Bart

    Have you seen the TIDEP-0091 Design? This code set implements a hard coded chirp configuration that is set in c code so no CLI configuration is needed. Also data is output over SPI to an external MCU.

    Also the TIDEP-0091 performs power switching between PWM/PFM modes on the LP87524 -J PMIC for better efficiency in chirping/non-chirping phases.

    Let me know if you have more questions.

    Regards,

    AG

  • Hi Akash,

    yes, of course we know the official power optimization design. This is implemented already on our end, but further reduction of power consumption would be needed in our application.

    Could you advise on the questions 1 and 2?

    Thank you.

  • Hi Bart

    For question 1, the UART stream should only be coming from the external MCU not from the IWR1443 device. Can you elaborate more on what you are seeing?

    For question 2, this isn't possible. Rather the solution is the power switching mechanism discussed in my post above. This is implemented in the TIDEP-0091 but programming the PMIC is performed by the external MCU (MSP432), so customer may not have implemented it yet.

    Regards,

    AG

  • Hi Akash,

    My name is Gabor Benyhe. I am working on the same project what Bart mentioned. Thanks for your reply, of course we checked the TIDEP0091 design guide, and could get under 20 mJ in our design, but still seeking for further possibilities of energy reduction. One idea is to check whether the boot time could be somehow shortened. I have noticed, that for example if I do not use the CLI it also shortens the system boot time by a few hundreds of microsecs. I also noticed, that there is a data stream appearing on the UART at 115kbps during booting. Do you know the purpose of this data stream? It appears to have some short header-like part, than blocks containing \x00\xfa\x71 come periodically, with almost exactly 424 microseconds between each block, then some closing block at the end of booting. Could this data sending be somehow switched off? Of course switching it off would not necessarily shorten the boot time, but I would like to know at least the purpose of this serial data and the possibilities around it.

    Best Regards

    Gabor

  • Hi Akash,

    Regarding question 1: I am entirely sure that it comes from the IWR. We are not using the demo board, we have our own PCB design. The MCU sits on a separate PCB. Even if the UART lines are disconnected from the MCU, the data is still there on pin R5 during bootup. It can be some debug firmware trace or something similar, I can only guess.

    Regards:

    Gabor

  • Hello Akash,

    do you have some feedback for us based on the above details from Gabor?

    Thank you.

  • Thank you Bart and Gabor

    I am still looking into this and will need a couple of days to provide a recommendation. Aiming to have an update End of Thursday.

    Regards,

    AG

  • Hi Akash,

    No problem, thanks. Let me tell you my further findings also. I noticed, that the length of this data stream during boot (let's call it as debugtrace for simplicity) varies depending on which library components does the application use. That's sort of logical. But I also noticed, that for example if I use SPI for interfacing, the length of the debugtrace is 13.2 msecs, if I use UART it becomes 13.05 msecs, and the surprising thing comes now: if I involve both SPI and UART the length of the debugtrace shortens to 10.6 msecs, which makes a significantly (about 10%) shorter total booting time.

    Question 1: Is there any document dealing with this behaviour of SYSBIOS booting process?

    Question 2: if this serial data is sort of a debug trace indeed, does this mean that some components include debug functionality? Maybe a release version also exists? If yes, can that be faster due to the lack of debug functionality?

    Best regards

    Gabor

  • Hello Akash,

    do you have some updates for us?

    Thank you.