Re: [time-nuts] Digital Phase Lock Loops

[Jim Harman]

Martyn,

Are you willing/able to write or modify some code to implement the loop?
GPSDOs typically use a digital PLL which can have a time constant of 10,000
sec or more. The biggest challenge with very long time constants is
avoiding overflow and roundoff issues in the calculations.

You might want to study Lars Walenius' GPSDO design as described here
https://www.eevblog.com/forum/projects/lars-diy-gpsdo-with-arduino-and-1ns-resolution-tic/


Unfortunately Lars has passed away but I have modified and extended his
system and can provide some help if you are interested.

On Tue, Nov 19, 2019 at 6:38 AM  wrote:

>
>
> Hello,
>
>
>
> Have anyone had any luck designing a phase lock loop with very small
> bandwidths, e.g. less than 0.2 Hz.
>
>
>
> I need to lock a 100 MHz Ultra low oscillator to a 10 MHz ultra low
> oscillator and I need a loop bandwidth less than 0.2 Hz.
>
>
>
> Quartzlock do a digital PLL board with bandwidths to 1 mHz.  However I
> never
> got it to work.
>
>
>
> The 10 MHz reference has phase noise of -116 dBc/Hz at 1Hz offset.  The 100
> MHz VCO make -104 at 1 Hz.
>
>
>
> Obviously I'm trying to preserve the 100 MHz phase noise.
>
>
>
> Any advice would be appreciated.
>
>
>
> Best Regards
>
> Martyn
>
>
--Jim Harman
___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Digital Phase Lock Loops

[Bob kb8tq]

Hi

Unless the 100 MHz “VCO” is an OCXO, it is going to be tough 
to keep things locked at a fractional Hz sort of bandwidth. It’s simply a matter
of how much the phase / frequency is likely to change while you are trying
to get it to lock / keep it locked. Things like shock and vibration get into it 
along 
with temperature. 

There’s nothing “magic” about digital vs analog. You do the same things 
with both. Gain margin / phase margin / damping matter in any control loop. 
Both approaches run into side issues with things like divider and phase 
detector 
noise floors. 

You may need a wide bandwidth loop to get things lined up and a soft 
switchover ( = no spikes ) to a narrower bandwidth. You see that done in GPSDO’s
a lot.  Plenty of grubby details to deal with ….. 

Bob

> On Nov 19, 2019, at 4:48 AM, mar...@ptsyst.com wrote:
> 
> 
> 
> Hello,
> 
> 
> 
> Have anyone had any luck designing a phase lock loop with very small
> bandwidths, e.g. less than 0.2 Hz.
> 
> 
> 
> I need to lock a 100 MHz Ultra low oscillator to a 10 MHz ultra low
> oscillator and I need a loop bandwidth less than 0.2 Hz.
> 
> 
> 
> Quartzlock do a digital PLL board with bandwidths to 1 mHz.  However I never
> got it to work.
> 
> 
> 
> The 10 MHz reference has phase noise of -116 dBc/Hz at 1Hz offset.  The 100
> MHz VCO make -104 at 1 Hz.  
> 
> 
> 
> Obviously I'm trying to preserve the 100 MHz phase noise.
> 
> 
> 
> Any advice would be appreciated.
> 
> 
> 
> Best Regards
> 
> Martyn 
> 
> 
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.