• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 52 subscribers
  • Views 168 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LP-EM-CC2745R10-Q1: multi-spikes on the 48M crystal

Albin Zhang
Expert 5910 points
Part Number: LP-EM-CC2745R10-Q1

Tool/software:

Hi, 

my customer found the multi-frequency when they probe (coupling) the 48MHz signal. it is difference with CC2642 platform (it is very clean 48MHz)

why the phenomenon is happening? what's the impact to performances potentially?  

BR. Albin

  • 0
    TI__Genius 9165 points

    Hi Albin,

    I've assigned someone from the team to support you in your question. Please expect a response tomorrow.

    Best,

    Nima Behmanesh

  • 0 in reply to Nima
    TI__Intellectual 1555 points

    Albin,

    I'll try to duplicate this in the lab with 1 of my LPs and I'll try to provide results soon. In the meantime, are you able to output an RF tone? If so, what's the frequency accuracy of the RF signal? If you adjust the cap bank setting in SmartRF Studio, does this affect the measured 48MHz coupled signal?

    Regards,

    Andrew

  • 0 in reply to Andrew Lugo
    TI__Expert 5910 points

    Hi Andrew, 

    All the RF is good and the cap tuning works well. this is customer's feedback. I don't have the coupling probe, so can't do the test in my lab. 

    so this is why I asked "potentially" ：）just worries about the PLL unlock or lock to other freq.

    BR. Albin

  • 0 in reply to Albin Zhang
    TI__Intellectual 1555 points

    Albin,

    Let me discuss this with a designer and I'll report back to see if there would be any concerns with what the customer is seeing. It's good to hear that the CC device is operating normally. I'll report back as soon as I get the feedback. 

    Regards,

    Andrew

  • +1 in reply to Andrew Lugo
    TI__Intellectual 1555 points

    Albin,

    This is not a problem.   They are probably seeing the HFXT and the HFOSC being picked up by this probe, both at very low levels.  The HFXT is sent directly to the PLL as full rail-to-rail digital signal (much higher than -90dBm).  The HFXT clock is on the crystal pins so the probe will pick it up.   The HFOSC is used for the SoC CPU and many other blocks so its frequency will show up on the supply which comes to an external cap (VDDD) due to digital switching noise.   The two signals below are ~4ppm apart.   In CC2745, the HFOSC is adjusted to be very close to HFXT using a simple frequency locked loop.   In CC2642, the HFOSC is just trimmed to be near HFXT, but due to the finite trim step size, it can easily be 0.5% (5000ppm) off, so on this same scale they wouldn’t see it.  Also, to compensate for the larger step size, the LSB of the HFOSC frequency trim in CC2642 is modulated, which will spread out the digital switching noise making it even harder to see with this probe.

    Regards,

    Andrew