• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 22 subscribers
  • Views 540 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.

LMX2492: Sub Fractional Spurious Level

user6249907
Intellectual 295 points
Part Number: LMX2492

I was investigating the fractional spurious level in PLL PlatinumSim, and noticed a correlation between the spurious level and the Fnum/Fden of the N Divider. It seems that the Sub Fractional Spurious level caused by MASH_Nonlinear becomes lower around Fnum/Fden=0.5. What is the reason for this? Should we choose the comparison frequency so that Fnum/Fden = around 0.5?. In our application, we generate high speed sawtooth waveform, so we are concerned about the spuriousness when the output frequency is a fraction.

  • 0
    TI__Mastermind 28550 points

    Hello,

    There is an integer boundary spur that can occur when you use num/den close to 0 or 1.  So moving away from this can help reduce spurs.

    Please also see the app note on Fractional N Frequency Synthesis, integer boundary spurs are discussed in section 3.4.1.  A few pages later table 3-6 discusses worst case for this spur.  And theoretically num/den = 0.5 is the second worst case, so you may be able to improve yet further.

    Hope this helps.

    73,
    Timothy

  • 0 in reply to Timothy T
    Intellectual 295 points

    In the PLL PlatinumSim display, can I assume that Fvco%Fosc is the boundary spur?.  Fvco%Fosc is about -110dBc even when the N-Divider setting is 1/400 or 399/400, and it gets worse at 200/400, etc., but it can be avoided as you answered.

    What I am concerned about is the sub fractional spurious that shows up as ”MASH_Nonlinear” in spur decomposition. I fix the denominator to 400 and change the numerator, it tends to be lower between 101/400 and 299/400. Is there any solution for sub fractional spurious other than narrowing the loop bandwidth?

  • 0
    TI__Mastermind 28001 points

    Hi,

    PLLatinum Sim calculates the sub-fractional spurs based on a numerical sequence from the MASH;  it's pure math and not part specific.

    I think that what you might be seeing is that near the fraction of 1/2, the sub-fractional spurs might be lower, but the primary fractional spurs are higher.

    Don't have a pure theoretical answer, but I believe PLLatinum Sim's result on this.

    On the topic of sub-fractional spurs, here are some tricks:

    1. Increase dithering.  For instance, instead of the fraction 201/400, you  can use a fraction of 201000000/400000001

    2. Fractional denominator choice.   The following will have no sub-fractional choice

    a. Any denominator that is not divisible by 2 or 3

    b. 2nd order modulator if denominator is not divisible by 2

    c. 3rd & 4th order modulator if denominator is not divisible by 2 or 3

    3. MASH_SEED

    a. For fraction 201/400 & 3rd order modulator, try MASH_SEED=234

    b. For fraction of 201/400 & 4th order modulator, try MASH_SEED=289

    c.  PLLatinum Sim calculates MASH_SEED for you.

    Regards,

    Dean