Post Go back to editing

AD9517: Additional spurs when using REF1 and REF2 Simultaneously

I use AD9517-1 to generate DAC reference clock, REF1 and REF2 are 10MHz, AC-Coupled, LVCMOS, both about 100Hz deviation, the clock can be locked, VCO 2400MHz. If I use any of the references, the clock output spectrum is very good, but if I enter two clocks at the same time and prefer to use REF1, the clock output near the spectrum(in 10kHz range) will appear some spurs, the range of about -50dBc. I analyzed REF2 for REF1 interference, but the actual working conditions REF1 and REF2 must exist at the same time, in this case the spurs can be eliminated? Or what other ways to reduce these spurs?
Looking forward to reply, thank you very much!

  • Dear Zhang,

    Unfortunately, putting two CMOS inputs on adjacent pins is the source of the coupling. You could reduce the coupling by reducing the amplitude or input slew rate, but you might make the close in phase noise worse on the AD9517 output. There is a very small chance the turning off the unused input receiver may help a little bit, but I suspect that the coupling is on the bond wire.

    I wish I could tell you a perfect way to remove the spurs, but the only other way to fix this is to shut off the unused input signal at the source.

    -Paul Kern

  • Dear ,

    Thank you very much for your timely response. As you said, closing the clock receiver does not remove the spurs. I will seriously consider the future design of this problem.