[time-nuts] Question on GPS and reference standards

Magnus Danielson magnus at rubidium.dyndns.org
Thu Dec 23 23:35:34 UTC 2010


On 12/23/2010 09:52 PM, Chris Albertson wrote:
> I just looked up what it would take to build a 10:1 conjugate
> regenerative divider.  This is not easy nor cheap.  The conjugate
> number turn out to be f/10 and 9f/10.  9f/10 is close to f.

Filters for 1 MHz and 9 MHz with sufficient Q-value is cheap an 
straight-forward to design.

The amplifier does not have to be spectacular.

The part-count should be fairly low to meet your needs.

> Question for the group?  What phase noise should be expected if a
> conventional PPL was used?
>
> I can't believe that a PLL derived 1Mhz reference is not good enough
> for a 30Mhz HF transmitter.  Maybe different answer if the transmitter
> were microwave
>
> Not arguing, this is a real question.  I'm wondering how good a simple
> and convention PLL is.      My problem is that I do the calculations
> and it's not "real world" and I get perfect result.  Just wondering
> how and what cause PPL to be not perfect and by how much.

Look at the Wentzel dividers, it is essentially a divider with related 
signal treatment. You rarely use PLLs for division but for 
multiplications, which is not the same as saying you can't. It would 
just be overkill most of the times. A PLL without a divider but using a 
S/H phase-detector would be possible thought.

Cheers,
Magnus




More information about the Time-nuts_lists.febo.com mailing list