1 Reply Latest reply on Feb 14, 2019 10:01 PM by PradiptaB_11

    CY29430 Fractional-N Boundary Spurs; "Fout may have [m^Fout-nFref] spurs, choose different reference input clock frequency for better jitter performance-

    brcuc_2268776

      I understand the concept of fractional boundary spurs being inherent to fractional-N PLL synthesizer technology; I am not worried about the below warning returned by the ClockWizard software, but I have a few questions in regards to this subject topic.

       

      Capture.PNG

      Below definition for where you can expect to see fractional boundary spurs/Clockwizard software doesn't allow you to program the part (copied/pasted directly from Cypress AppNote "AN210253", pg. 10):

      ==================================================

      "The output programmed frequency (         ) and the input reference frequency (         )

      holds a relationship of 0.01 ≤ |  ×         −  ×         | ≤ 1.5,

      the output phase noise plot will show spur within the integration bandwidth limit (12 kHz to 20 MHz)"

      ==================================================

       

      I have the Cypress Programming Kit for the CY29430 device; my "Fref" value for the CY29430 = 122.88MHz; (XO version, not VCXO version, 5.0x3.2mm)

       

      I'm sure Cypress has developed a list of "Fout" frequencies that will exhibit fractional boundary spurs and return this Warning Message (based off "Fref" = 122.88MHz). It would be extremely helpful to know ahead of time which "Fout" carrier frequencies I will experience the above situation with (fractional boundary spur occurence+inability to program the part) and have to obtain a completely different CY29430 device at a different "Fref" base frequency, so that I can program the part to the expected target "Fout" frequency. If not a list of frequencies, then you should at least be able to provide which fraction ratios (M/N) from equation above you've detected fractional boundary spurs at.

       

      I've tried doing a ton of calculations on my own, but I have too many unknown variables to reach a solid answer to this question. To put it simple, which values for "Fout" @ "Fref"=122.88MHz yield the above Warning Message due to a boundary spur??

       

      • The value for "Fout" that I just ran into this issue with was "Fout"=266MHz. I cannot program this device @ 266MHz (when "Fref"=122.88MHz), which is honestly a very common frequency and used in a lot of applications.
        • Also, I'm almost 100% certain I've hit fractional boundary spurs when I program the device to "Fout"=153.6MHz & "Fout"=102.4MHz ("Fref"=122.88MHz)
        • I have an E5052B Signal Source Analyzer that I extract phase noise plots on with these devices and they seem to definitely exhibit fractional spurs on the plots
        • And these values for "Fout" exhibiting spurious activity actually made sense to me ("Fout"=153.6MHz @ M/N=5/4 & "Fout"=102.4MHz @ M/N=5/6)
        • However, the ClockWizard program definitely doesn't return the above Warning Message indicating there are spurs at these (2) "Fout" carrier frequencies... it still lets me program them.

       

      So now I'm pretty confused as to what methodology Cypress took to detect and define these boundary spurs. I'm having a hard time calculating with confidence at what other "Fout" frequencies I will likely see fractional spurs at and will not be able to program these devices to ("Fref"=122.88MHz; XO; 5.0x3.2mm).

       

      I could just input infinite amounts of frequencies into the software and see what returns the warning message, buuut not quite the most efficient way to do this. I am very much hoping you have some information to provide me on this.

       

      Thank you-