how good should be power supply ripple for a VCO to operate properly

Hello..

I am doing research on wireless recievers. I am using a VCO from Hittite to provide LO. I have supplied power to VCO from a Rhode & Schwarz regulated power supply unit. I did not see the expected performance of VCO. Lot of phase of noise was present. We are planning to buy a SMU. I want to know how good should be the ripple from power supply unit for better performance of very sensitive RF components like VCO. I would use this info during purchase of SMU.

Kindly let me know !

Parents
  • Hello,

    Noise on the supply bias feeds, particularly the VTUNE port, will need to be as good as possible. Vcc needs to be very clean as well but but the impact isn't as severe since the modulation is a function of frequency pushing due to supply variation as opposed to directly modulating the resonant tank via the tune port. Narrow band FM modulation theory (shown below) illustrates how noise can set sidebands on our VCO and degrade the phase noise. If you know the maximum side band power that your system can tolerate (DSB below) at the various offsets (modulating frequencies) then you can derive the maximum amount of noise that you can tolerate on the power supply (assuming it doesn't pickup additional noise as it makes it's way to your device).


    To simplify all this I'd recommend that you use the lowest noise LDO's that ADI offers instead of strictly relying on the SMU. Worst case they can serve as a backup. These offer state of the art performance and should work fine for your receiver application assuming they can handle the specific bias requirements. You may also need to consider using separate LDO's for Analog, Digital & RF to improve isolation as well as additional external filtering on the IO's of the LDO's if these aren't enough but I suspect the basic apps circuit shown in the datasheets will be good enough. I would recommend the following parts: ADM7150 (best single function LDO available), ADM7170 (good enough for all but the most stringent requirements), HMC976 (higher current handling, usually used for VCO only, see attached), or HMC860LP3E or HMC1060LP3E (quad outputs in small form factor for isolating analog and digital bias feeds on PLL synthesizers).  All of these are available on the Analog Devices website.


    You can use the equations below to help you determine how much margin each will provide either relative to the SMU or the LDO you are considering. 

    As you know, narrowband FM is defined as having a Modulation index less than unity (no more than 3 significant frequencies on either side of the carrier) where...


    1) β = Modulation Index = ▲f / fm

       

              fm = Modulating Frequency and

     

              ▲f = Frequency Deviation

     

      2) ▲f = Kf * Am


              
    Kf = VCO Modulation Sensitivity (or Tuning Sensitivity) and


             
    A
    m
    = Amplitude of the Modulating Frequency (amplitude of noise on our Vtune port).


    Using properties of an nth order Bessel Function of the 1rst kind we can show that for narrow band FM where β << 1 the first side band power is...

    3) DSB = 10 * Log (β2/4) where


                    DSB = Desired First Sideband (or in our case the ‘undesired’ first sideband)


    If we re-arrange equation 3) we can derive the Modulation index (β) from our first sideband power level such that

       

    4) β = 2 * √ 10^(DSB/10)


    Finally if we substitute equation 1) into equation 2) and re-arrange we can derive the level of the modulating signal (the amplitude of the noise causing the side band).


    5) Am = ( β * fm ) / Kf


    Hope this helps,

    Marty

Reply
  • Hello,

    Noise on the supply bias feeds, particularly the VTUNE port, will need to be as good as possible. Vcc needs to be very clean as well but but the impact isn't as severe since the modulation is a function of frequency pushing due to supply variation as opposed to directly modulating the resonant tank via the tune port. Narrow band FM modulation theory (shown below) illustrates how noise can set sidebands on our VCO and degrade the phase noise. If you know the maximum side band power that your system can tolerate (DSB below) at the various offsets (modulating frequencies) then you can derive the maximum amount of noise that you can tolerate on the power supply (assuming it doesn't pickup additional noise as it makes it's way to your device).


    To simplify all this I'd recommend that you use the lowest noise LDO's that ADI offers instead of strictly relying on the SMU. Worst case they can serve as a backup. These offer state of the art performance and should work fine for your receiver application assuming they can handle the specific bias requirements. You may also need to consider using separate LDO's for Analog, Digital & RF to improve isolation as well as additional external filtering on the IO's of the LDO's if these aren't enough but I suspect the basic apps circuit shown in the datasheets will be good enough. I would recommend the following parts: ADM7150 (best single function LDO available), ADM7170 (good enough for all but the most stringent requirements), HMC976 (higher current handling, usually used for VCO only, see attached), or HMC860LP3E or HMC1060LP3E (quad outputs in small form factor for isolating analog and digital bias feeds on PLL synthesizers).  All of these are available on the Analog Devices website.


    You can use the equations below to help you determine how much margin each will provide either relative to the SMU or the LDO you are considering. 

    As you know, narrowband FM is defined as having a Modulation index less than unity (no more than 3 significant frequencies on either side of the carrier) where...


    1) β = Modulation Index = ▲f / fm

       

              fm = Modulating Frequency and

     

              ▲f = Frequency Deviation

     

      2) ▲f = Kf * Am


              
    Kf = VCO Modulation Sensitivity (or Tuning Sensitivity) and


             
    A
    m
    = Amplitude of the Modulating Frequency (amplitude of noise on our Vtune port).


    Using properties of an nth order Bessel Function of the 1rst kind we can show that for narrow band FM where β << 1 the first side band power is...

    3) DSB = 10 * Log (β2/4) where


                    DSB = Desired First Sideband (or in our case the ‘undesired’ first sideband)


    If we re-arrange equation 3) we can derive the Modulation index (β) from our first sideband power level such that

       

    4) β = 2 * √ 10^(DSB/10)


    Finally if we substitute equation 1) into equation 2) and re-arrange we can derive the level of the modulating signal (the amplitude of the noise causing the side band).


    5) Am = ( β * fm ) / Kf


    Hope this helps,

    Marty

Children
No Data