[time-nuts] Digital PLL ICs,alternatives and digital loop filters

[Stephan Sandenbergh]

Hi all,

 

Thanks for all of the very helpful responses, links and advice. I'm very
sorry that I couldn't participate in the discussion at the time but I was a
bit under resourced on e-mail facilities during the festive season.

 

The trend that I noted from your replies is that the different techniques
for up-converting from 10MHz to 100MHz can be ranked as follows: (best noise
performance first)

 

1.	Pure Analogue PLL where two high quality OCXOs are slaved onto
another. 
2.	Analog frequency multipliers (if the reference is of low enough
phase noise)
3.	Digital PLL

 

This list is also, as expected, ranked in order of associated cost. 

 

My GPS disciplined 10MHz OCXO, an Oscilloquartz model 8788 has the following
phase noise:

 

1Hz      -100 dBc

10Hz     -130

100Hz   -152

1kHz     -160

10kHz   -165

100kHz -165

1MHz    -165

 

>From your replies I take that I'll probably ruin this clean phase noise
characteristic by trying to lock it digitally to another high quality 100MHz
OCXO, since this kind of oscillator exceeds the performance of what is
achievable with a digital PLL IC.  In this case I would also have wasted
good money on the 100MHz OCXO.

 

It seems that the second best option (cost wise) would be to analog multiply
the 8788's output to 100MHz. Theoretically, I would then sit with the
original phase noise curved that is lifted by 20logN, where N=10, which is
about 20dBs. Thus, I would start of at -80dBc @ 1Hz and end off with a
-145dBc noise floor. (in real life I guess one would add another few dBs for
noise added by the individual components)

 

It also seems that the needed analog multipliers are a good start on one's
way to a fully analog PLL - the mixer and loop filter could be added later
if required. 

 

Could anyone point me to some analog frequency multiplier manufacturers?
(Currently I am only familiar with Wenzel (thanks Bruce)) 

 

In the mean while, it is back to the books for me - I seriously need to
catch up on some PLL theory. 

 

 

Regards,

 

Stephan Sandenbergh.