Post Go back to editing

HMC704+HMC586 generate spurious signals at certain frequency

Hi,our customer designing HMC704+HMC586+HMC1060  LO circuits,fref=320MHz,fpd=80MHz,LBW=300kHz,Frac B Mode,fout=4-8GHz,fstep= 1kHz.Spurious signals ( -60dBc@ 800kHz offset) can be found at 4019.999MHz,4020.001MHz,4059.999MHz,4060.001MHz,....and so on.We can see the spurious signal neither ref spur nor channel spur/int boundary spur.Can you help to analyze the causes and solutions of spur?Thanks.

  • Hi there,

    Sorry to hear your seeing elevated spurious in your design. Spurs at these levels can easily be due to coupling at the layout level, inadequate supply filtering (be sure you're using High PSRR LDO's), inadequate isolation of PLL / VCO and their supplies and RF paths and more. Unfortunately you didn't provide a schematic, register files, layout or application information so there's only so much I can do here.

    It sounds like you may be using the HMC1031 for jitter cleanup (to sync a lower quality reference to a higher quality reference in a similar manner as is done on the HMC704 evaluation board). This can improve overall spurious performance but it also is another path with dividers that can generate more spurs albeit at lower levels, especially if the RF isolation is insufficient.

    Due to the internal prescaler on the HMC704 that must be enabled when the RF input > 4GHz (REG 9 [21]) additional spurs fall at PFD /2 or at 40MHz intervals when operating above 4GHz. The prescaler divides your 4020.001MHz down to 2010.0005MHz so I would expect spurs at 10.0005MHz away and 20.001MHz away from the carrier. Since you also have RDIV = 4  to divide down the reference you may see additional spurs that are related to the reference (or a harmonic of the reference) or the PFD frequency or a harmonic of the PFD frequency.

    You can isolate where the spurs are coming from by setting the REF to 80MHz and RDIV = 1, REF to 160MHz, RDIV = 2, etc. Also try eliminating the HMC1031 and power down its high quality reference and see what happens to the 800kHz spur.

    1) Since your using a high frequency reference be sure you have enabled the High Frequency Mode (REG 8[21] since your reference is >200MHz.

    2) SDM Mode A generally provides better in-band spurious performance than Mode B - I would try using this and see what you get,  refer to REG 6 [3:2].

    3) Varying Icp Offset current from it's calculated value may help for particular spurs (REG 9[20:14]) however this will take some experimentation and may not be as consistent as the calculated value from unit to unit.

    4) Seed select (REG 6[1:0] - varying this may help.

    Best Regards,