Hi, everyone!

Here I have a question about the reference crystal oscillator for frequency synthesizers, such as ADF4351. The output frequency range is usually wide, for example the output range of ADF4351 is 35 MHz to 4400 MHz. The crystal oscillator has a decisive impact on the stability of the output frequency, especially for high freqency. Which parameters of the crystal oscillator should we focus on, the short term frequency stability or the the frequency stability over temperature?

Take the TCXO (Model TXO200B 25MHz) on EVAL-ADF4351 for example, its frequency stability over temperature is +/-2.5 ppm@25 deg C, and its root allan variance (1 second Tau) is 1.0 ppb. And the output freqency is 1422 MHz. What about the frequency stability of 1422 MHz? In my view of point, we should pay attention to the allan variance, that is there exist a 1.422 Hz jitter around center frequency 1422 MHz. And it is based on the assumption that there's on increased phase noise after the upconversion.

Thanks!

Best Regards!

Any offset at the reference will appear at the ADF4351 output multiplied by N. N is the ADF4351 N divider value. So, a 100 Hz offset on a 10 MHz reference signal, will be a 40 kHz offset on a 4 GHz output (N = 400).

The frequency stability is how close the XO output is to it's specified output frequency; i.e. what's the max offset. Allan variance relates to the jitter on the signal; this is more to describe the phase noise of the XO output.

Focusing on short term frequency stability vs. frequency stability over temperature depends on the application and your needs.

If you want something very stable, look at an oven controlled XO (OCXO); however, they draw a lot of current and aren't very robust; i.e. liable to fail.