• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 21 subscribers
  • Views 229 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Clock ouput with reference input

Zain Alishammat
Prodigy 220 points
Part Number: LMK05318B


Tool/software:

Hi everyone,

I have the LMK05318B EVM with the following settings in TICS Pro:

  1. ZDM-Mode is on
  2. Reference Input with 1 kHz (DPLL Mode is enabled)
  3. Output Frequency set to 98.304 MHz
  4. Output Channel 7 
  5. Output format LVCMOS

The module cannot output 98.304 MHz exactly. The value fluctuates around 98.304 MHz.

I noticed this when I entered the frequency in "Set output"  APLL1 or APLL2 is assigned as the source — in my case, APLL2. I can't change the source anymore; it's fixed.

Because the TICS Pro assigns APLL2 to output 7 at this frequency, Cascaded Mode is enabled.

When Cascaded Mode is enabled, VCO1 is fixed to 2.5 GHz (see datasheet, page 24).

I tried this with other channels. I got the same results!

So, I don't want to enable Cascaded Mode. I just need to output the frequency from APLL1, not APLL2.

Furthermore, I calculated the jitter from Channel 7. It is 33 ps. However, the datasheet mentions 125 fs. Why is that?

BR

Zain

  • 0
    TI__Genius 17045 points

    Hi Zain,

    It is not possible to output 98.304 MHz from VCO1.

    VCO1 is centered at 2500 MHz and can only be pulled +/- 100 pm. This is always true and is regardless of whether cascaded mode is enabled or disabled.

    In contrast, VCO2 has a much wider pulling range which allows us to pick from a variety of VCO2 options. The GUI helps to visualize this.

    VCO1 is based on the BAW technology which gets us ultra low jitter clocks that can be divided from 2500 MHz, such as 156.25 MHz, 100 MHz, 125 MHz, 312.5 MHz....

    For any other frequency outside that domain, such as  98.304 MHz, VCO2 must be used.

    Also, how are you calculating the jitter? Is this using a phase noise analyzer? What is the full frequency plan and RMS jitter integration range?

    Regards,

    Jennifer

  • 0 in reply to Jennifer Bernal
    Prodigy 220 points

    Hello Jennifer,

    Jennifer Bernal said:
    For any other frequency outside that domain, such as  98.304 MHz, VCO2 must be used

    However, using VCO2 in cascaded mode does not solve the problem. The clock generator cannot generate 98.304 MHz and remain unstable. I tried everything.

    This is a huge problem!

    So how can this be accomplished, i.e. how can 98.304 MHz be generated and maintained stable?

    For measuring jitter, I used the jitter analyser in the oscilloscope.

    Even the jitter remains unstable. After measuring thousands of samples, the jitter value fluctuates. But this makes sense, since the frequency remains unstable.

    BR

    Zain

  • 0 in reply to Zain Alishammat
    TI__Genius 17045 points

    Hi Zain,

    It's not recommended to use the scope to calculate jitter because there is inherent scope noise which impacts the actual output jitter measurement.

    Instead, a phase noise analyzer should be used to calculate jitter.

    Please review this app note for more information on scope vs a phase noise analyzer: https://www.ti.com/lit/an/scaa120b/scaa120b.pdf .

    Here are phase noise plots showing the output RMS jitter performance, for 12 kHz to 20 MHz integration.

    DPLL LBW = 10 Hz

    LMK05318B_XO=12.8M_OUT=98.304M_DPLLLBW=10Hz.tcs

    DPLL LBW = 0.1 Hz

    LMK05318B_XO=12.8M_OUT=98.304M_DPLLLBW=0.1Hz.tcs

    The DPLL LBW should be at least 1/10 of the TDC frequency for proper DPLL operation. A more narrow DPLL LBW, such as 0.1 Hz, means less PRIREF noise impact on the outputs, but longer time for the DPLL to phase lock (LOPL flag to clear).

    Additionally, I recommend using a lower XO input frequency such as 12.8 MHz since you must use the PPS phase validation detector to lock to the low REF input frequency of 1 kHz input.

    Regards,

    Jennifer

  • 0 in reply to Jennifer Bernal
    Prodigy 220 points

    Hi Jinnifer,

    thank you for the config file.

    It did not work!

    I still get an unstable frequency of around 98.304 MHz, but never exactly 98.304 MHz. It's exactly the same as before: changing the XO frequency to 12.8 MHz doesn't really affect the frequency stability or make it exact.

    I tried all the VCO2 Freq plan in TICS Pro and no one works!

    So, there are two major problems here:

    1. I never get 98,304 MHz out

    2. The frequency remains unstable

    can you help please?

    BR

    Zain

  • 0 in reply to Zain Alishammat
    TI__Genius 17045 points

    Hi Zain,

    As mentioned, a low noise instrument such as a frequency counter or phase noise analyzer should be used to get an accurate frequency measurement. A scope must not be used to prove if the frequency is exactly 98.304 MHz.

    Can you please check with a frequency counter or phase noise analyzer?

    Regards,

    Jennifer