Post Go back to editing

ADF4401A VCO stability while switching from internal to external loop

Category: Hardware
Product Number: ADF4401A

With reference to the Figure 32 (ADF4401A Evaluation Setup Block Diagram) on page 19 of ADF4401A (Rev. 0) datasheet, my queries are that when VTUNE input is switched from internal PLL to the external PFD using the switch ADG1219/ ADG1609,

1. Will the VCO frequency randomly deviate from the locked frequency (using internal PLL) during the switch transition time (~150ns for ADG1609) due to break-before-make switching or will the PLL remain in lock? There is no capacitor at the ADG1609 output to hold the VTUNE till the switching is completed (Figure 9 on Page 10 of UG-1922).

2. Will the VCO sensitivity (Kvco) determine how much the VCO will deviate during this switching period? For example in ADF4401A the VCO is multicore, so the Kvco is small, but suppose I am using HMC586 wideband VCO (4G-8G) whose sensitivity is more than 200 MHz/V will the deviation be more?

3. Is there a possibility that the external PFD will lock to an image frequency from the mixer if this momentary deviation in VCO frequency is significant? How to prevent in that case?

Please suggest


  • Hi  ,

    We're currently looking into this and will get back to you.

    Best regards,

  • Hi 

    Your question has been brought to the attention of the support engineer for the ADF4401A engr. Apologies for any delays caused. Thanks.



  • Hi. Is there any  clarifications regarding my queries. Please provide some insight on this switching between internal and external loops.

  • Hi,

    Firstly, the switching interval is not tested in detail.

    1. Because the Vtune is floating, I expect the VCO frequency to drift lower due to the leakage current. I agree, adding a capacitor at the Vtune input after the switch may help to reduce the drift.

    2. I think both Kvco and the leakage current on Vtune have an impact on the frequency drift.

    3. It is difficult to comment without testing with HMC586. Here are my thoughts:
    The image frequency is not a stable point. We may call it quasi-stable where a little deviation from the image frequency will lead the PFD to get away from the image point because it works in the opposite direction. However, that may cause either a lock to correct frequency or to an unlock state. Note that, the INV pin in HMC3716 let us mixing both from high side and low side.

    This does not look very likely for ADF4401A because the frequency band is narrow as you noted.