Re: LMX2541


Dear Andy,

YES definitely YES. The point is that the 100MHz OT oscillator has better phase noise than the 10MHz reference source used to adjust its mean frequency. If the loop bandwidth is too wide the benefit of the better phase noise performance of the 100MHz OT oscillator is lost.

It is the loop bandwidth of the loops locking the microwave output to the 100MHz OT oscillator that have to be wide enough to ensure that the noisy resonator tuned oscillators, later in the chain, take on the phase noise characteristic of the multiplied 100MHz OT oscillator.

If you doubt this, try making some phase noise measurements at, say, 50 GHz on the output of a high quality (eg R&S SMA100b or similar) signal generator with deliberate phase modulation of the 10MHz reference. You can plot out the transfer function of the control loop for yourself.


Alwyn G8DOH

Re: LMX2541
From: Andy G4JNT
Date: Sat, 11 Jul 2020 08:47:56 PDT 

On Sat, 11 Jul 2020 at 15:35, alwyn.seeds1 <a.seeds@...> wrote:
The PLL to lock the frequency to 10MHz reference has to be fairly narrow bandwidth so as not to spoil the phase noise performance.
The block diagram of a good signal generator shows the approach, though some synthesis schemes are of head-aching complexity to obtain continuous low-spur coverage.

NO, definitely NO

Inside the loop bandwidth, phase noise is directly related to that of the reference
Outside, far-out,  it is a function of the VCO, phase detector etc

So unless you've got a really crappy reference you wan't your loop bandwidth to be as wide as possible so you can make the most of that top-spec 10MHz or 100MHz or whatever reference.
You are probably thinking of the need to use a narrow PLL bandwidth to clean-up a noisy clock perhaps ?   A Totally different requirement


Alwyn Seeds, Director
SynOptika Ltd.,
114 Beaufort Street,
SW3 6BU,

SynOptika Ltd., Registered in England and Wales: No. 04606737
Registered Office: 114 Beaufort Street, London, SW3 6BU, United Kingdom.

Join to automatically receive all group messages.