Hey,
I have noticed that in my V2 Plus4 Pro i do have ADF4350 installed. Checking Analog site the slightly newer version has been improved (a bit).

Main improvements compared to the ADF4350:

* Improved 1/f in-band phase noise (5 dB)
* EVM improvement of up to 30%
* Lower PFD spurs
* Wider output range: 35 - 4400 MHz
* Small frequency/phase jumps possible without band select.

I do understand that the change probably would be unnoticeable, but do you think that maybe it is worth upgrading to fight for every 0.1dB of accuracy?

Analog: https://ez.analog.com/rf/f/q-a/76030/adf4351-vs-adf4350---what-are-the-differences

My question is, has anyone made such an improvement on their device?

On a parallel forum, - where there was a history of this topic ( three-year old theme was) , no response was received.

I have the feeling that by changing and improving the reference oscillator, we can have a greater influence on the phase noise and, of course, on the frequency stability.

Istvaan

Hi, what we know is only changing ADF4350 to ADF4351 does not make any
measurable difference in the V2 series. In VNA6000 we are using a
higher performing synthesizer circuit to achieve its lower phase
noise.
Thanks

On Thu, Aug 10, 2023 at 8:29 AM István <erretrete@freemail.hu> wrote:

>
> My question is, has anyone made such an improvement on their device?
>
> On a parallel forum, - where there was a history of this topic (three-year old theme was), no response was received.
>
>
> I have the feeling that by changing and improving the reference oscillator, we can have a greater influence on the phase noise and, of course, on the frequency stability.
>
> Istvaan
>

at first look for the clock that is driving the 4350/51 ... if you
change that for a low phase noise but high stability clock .. then you
are on the right track

eplacing that chip can be done later

i bet the main clock is what sets stability (and often also the phase noise)

dg9bfc sigi

Am 10.08.2023 um 10:06 schrieb István:

> My question is, has anyone made such an improvement on their device?
>
> On a parallel forum, - where there was a history of this topic
> (three-year old theme was), no response was received.
>
>
> I have the feeling that by changing and improving the reference
> oscillator, we can have a greater influence on the phase noise and, of
> course, on the frequency stability.
>
> Istvaan
>

On 8/10/23 4:59 AM, Siegfried Jackstien wrote:
> at first look for the clock that is driving the 4350/51 ... if you
> change that for a low phase noise but high stability clock .. then you
> are on the right track
>
> eplacing that chip can be done later
>
> i bet the main clock is what sets stability (and often also the phase noise)
>
> dg9bfc sigi
>
> Am 10.08.2023 um 10:06 schrieb István:
>> My question is, has anyone made such an improvement on their device?
>>
>> On a parallel forum, - where there was a history of this topic
>> (three-year old theme was), no response was received.
>>
>>
>> I have the feeling that by changing and improving the reference
>> oscillator, we can have a greater influence on the phase noise and, of
>> course, on the frequency stability.
>>
>> Istvaan
> _._,_._,_

I don't know that stability is a big driver - unless you're making very
narrow band measurements and a few Hz makes a difference. 1 ppm is 1 kHz
at 1 GHz, after all.

Phase noise is a bit trickier - The receive LO has a different
synthesizer than the Stimulus LO, so the noise doesn't cancel out.

But then, usually, you worry about phase noise when looking to separate
multiple signals (e.g. reciprocal mixing) - here, there's only one
signal, and what you'd care about is the noise at 5kHz out, which sets
the noise floor. I suspect other components are bigger contributors to
the noise floor.

Taking the first hit on Mouser for a 26 MHz TCXO I see phase noise at
1kHz is -134 and -146 at 10 kHz. Bumping that up to, say 520 MHz,
you'll see an increase of 20log(20) or 26 dB. So at 5 kHz, we're
looking at probably around -116 dBc.
Since the stimulus is ~0 dBm, and the dynamic range of the system is
~80dB, and allowing for ~1kHz bandwidth in the filter (30dB), the noise
power is -86 dBm. So pretty small.