[HMC3716]

I have a simple question:

If the frequencie of the input and the LO are diferent, how will the output voltage be? Is the one port sent out a votage meant the difference and the other sent out the phase difference? or what?

Thanks

Parents
  • 0
    •  Analog Employees 
    on Sep 6, 2017 12:17 AM

    Hi there,

    The output of NU & ND function identically however, when the delta in frequency between the VCO and REF is great, only one output is active at a time. Refer to the last plot in the datasheet comparing LD, NU and ND where the VCO frequency is being held 11Hz higher than the REF. This is why ND appears to be constant and only NU is varying. If the VCO frequency were 11Hz below the REF frequency, the traces for ND & NU would be reversed in this plot.  Notice that the period of the saw tooth voltage waveform is 90.9mS which is the period we'd expect for an 11Hz difference in frequency. As the difference between the VCO & REF deceases and the delta approaches a limit of 0hz for the first time ( Note that 0hz is really never achieved for any real length of time as the VCO is always moving) the phase detector output pulses become very narrow and the duty cycle reduces to something much less than 50%. Now that they are close enough to declare "lock" the VCO frequency may exceed the reference frequency in one instant and in the next instant be less than the REF so both outputs will be active as they constantly work to maintain phase lock. This region is called the "dead zone" and results in increased spurious in fractional PLL's. Fortunately modern fractional PLL's can force the VCO to operate just slightly to one side of the dead zone which minimizes the fractional spurs.

    hope this helps,

    Marty

Reply
  • 0
    •  Analog Employees 
    on Sep 6, 2017 12:17 AM

    Hi there,

    The output of NU & ND function identically however, when the delta in frequency between the VCO and REF is great, only one output is active at a time. Refer to the last plot in the datasheet comparing LD, NU and ND where the VCO frequency is being held 11Hz higher than the REF. This is why ND appears to be constant and only NU is varying. If the VCO frequency were 11Hz below the REF frequency, the traces for ND & NU would be reversed in this plot.  Notice that the period of the saw tooth voltage waveform is 90.9mS which is the period we'd expect for an 11Hz difference in frequency. As the difference between the VCO & REF deceases and the delta approaches a limit of 0hz for the first time ( Note that 0hz is really never achieved for any real length of time as the VCO is always moving) the phase detector output pulses become very narrow and the duty cycle reduces to something much less than 50%. Now that they are close enough to declare "lock" the VCO frequency may exceed the reference frequency in one instant and in the next instant be less than the REF so both outputs will be active as they constantly work to maintain phase lock. This region is called the "dead zone" and results in increased spurious in fractional PLL's. Fortunately modern fractional PLL's can force the VCO to operate just slightly to one side of the dead zone which minimizes the fractional spurs.

    hope this helps,

    Marty

Children
No Data