I want to use the ADF4350 as a local oscillator and I want it to sweep as fast as possible from a start frequency A to the end frequency B the issue is the span is large and for that i have to use a large loop bandwidth which should not exceed 1/10 of the phase detector or 1/5 at max for stability. Now for me to use a large bandwidth i need a good loop filter and i need your opinion.
Are those listed filters suitable for the task? And which one of them is better?
The frequency reference will be a TCXO equal to 10 time the bandwidth, and the phase detector is equal to the frequency reference. the loop bandwidth is the max bandwidth of the filter.

  1. LTC1560-1
  2. LT1567
  3. LT1568
  4. LTC6603

Edit: some members suggest to stick to a passive 3rd order filter and that an active filter is not necessary as this what is dictated in the datasheet and/or ADIsimPLL. In this document at figure 20 shows/proves the effects of high order filter were the phase margin is getting wider and wider as the bandwidth is getting larger which is my goal.
The filters i have listed promises a tightned matching phase .


1 Answer 1


Are those listed filters suitable for the task? And wich one of them is better?

The best solution would be to use a passive 3rd order filter as shown in the datasheet. An active filter is unnecessary and the listed examples are not suited anyway. The filter can be designed and simulated with "ADIsimPLL" from Analog Devices.


  • \$\begingroup\$ I don't know why you think that a 3rd order filter is the best solution except it "shown in datasheet" as all of us know going up in order filters increases the phase shift which mean it slow down the lock time and on the other hand in datasheet or with ADIsimPLL they dont go above KHz bandwidth . \$\endgroup\$
    – Tintin
    Commented Jul 10 at 10:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.