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: Crystal or oscillator as reference

Part Number: IWR1443

Hi,

in the datasheet of the IWR1443 (p. 26 in Rev. C) figure 5-2 is showing the wake-up sequence and its timings.

This figure shows that the wake-up is about 6.5ms faster when using an external oscillator instead of a crystal at the reference input.

During our measurements we didn't see such an improvement. Do we need to do any further software modification to see this timing improvement?

Thanks and best regards,

Robert Storch

  • Hello

    The referenced diagram is for Crystal input, please let us know where you are referencing the ext clock based timing from.

    Thank you,

    Vaibhav

  • Hi Viabhav,

    the figure I'm referencing to is the following

    I've marked the point where TI mentions the distinction between XTAL mode and REFCLK mode is shown.

    If XTAL and REFCLK is not referring to the reference input - what is meant here else?

    Thanks and regards,

    Robert

  • Robert,

         Could you please elaborate more on "During our measurements we didn't see such an improvement". Did you not see any improvement at all or how much improvement did you see?  

    Also could you please give more details on your measurement scheme. 

    I guess you have done the modification in the hardware for feeding the external clock option (CLKM  to ground) and recommendations as per Table 5-7 datasheet specifications.  Please note that, when VIO supply is not present, external clock should not be driving the CLKP pins. 

    Thanks and regards,

    CHETHAN KUMAR Y.B.

  • Hi Chethan,

    thanks for your answer.

    Regarding the measurements we did:

    We measured the time between enabling the DCDC for the IWR until we saw activities on the QSPI, i.e. flash download begins.

    This time is about 5.8ms and it was almost the same if using an oscillator or a crystal.

    Regarding our configuration: yes, we tied the CLKM to gnd and use an oscillator fulfilling the specs of table 5-7. Although we took care about the VIO and oscillator activity.

    Regards,

    Robert

  • Robert,

        Thank you for the confirmation. Startup time is function of multiple elements, i.e. Power supply stabilization--> SOP assertion --> nReset release-->Boot mode detection--> Crystal oscillator build up and stabilization --> APLL lock and clock stabilization, Clock good check release-->Boot ROM self test check-->MSS application download from external QSPI flash.

    This entire sequence worst case scenario would be 7msec. Typically a component of the startup time is crystal oscillator wakeup and stabilization and it's not constant, It's stabilization varies depending upon crystal property, ESR, -ve resistance, drive strength,  Cap-loading etc. can have large variability depending upon the design. In your application, it may be possible that crystal oscillator is coming up and settling to desired amplitude and frequency faster. Hence you may not see significant improvement w.r.t to external clock option. However, I expected atleast some improvement w.r.t. to crystal option. Probably you have better design that makes crystal startup faster!. You get benefit only to the extent of crystal startup time reduction, If that is small portion of the entire wakeup period, you may not get significant reduction in startup time with an external clock.   

    Thanks and regards,

    CHETHAN KUMAR Y.B.

     

  • Hi Chethan,

    thanks for your answer.

    Of course, I understand that the startup time is dependent on a variety of parameters.

    Nevertheless, if I got you right, there is nothing to be done in software to speed this procedure up?

    Regards,

    Robert

  • Robert,

    Yes, that's right. 

    Thanks and regards,

    CHETHAN KUMAR Y.B.