Post Go back to editing

Phase Noise Floor of PFDs


I have made use of many PFD products from Hittite in various applications. In most of them, I try to push hard to get the best phase noise performance from these parts. In datasheets of various PFDs (e.g. HMC3716, HMC439 etc.), it is indicated that the PFDs have much better phase noise floor when a square wave rather than a sine wave is applied to the part. I have no experience of working with a square wave reference input.

I wonder how this phase noise floor measurement is performed? My guess is that a low phase noise OCXO or similar oscillator is used as reference in a PLL where a VCO with the same frequency is locked to this signal provided that the loop bandwidth is sufficiently wide.

Most oscillators with low phase noise characteristics provide sine wave outputs. How can we convert these signals to square waves?

  • Is saturating an amplifier an option?
  • Can a frequency divider (e.g. HMC794) be used in a divide-by-one configuration for that purpose?

I am expecting your comments/experiences on this square wave operation issue.

Kind regards.

  • Hi there,

    Yes, the high slew rate of the square wave improves the residual phase noise of these PFD devices, this is especially important when working with frequencies below 100MHz or so. 

    1) You could amplify and "clip" the signal and generate a clipped sine wave.

    2) While you can't use a /1 as that is merely a pass through you could operate at 2x the frequency then use /2 to generate a square wave with a 50% duty cycle.

    3) Could use a comparator circuit (use one of Analogs high speed comparators) or alternatively, diodes (to rectify) or an RLC circuit can be used to clip the signal also.

    Best Regards,