[time-nuts] Generating a stable 26MHz and 19.2MHz from 10MHz

Magnus Danielson magnus at rubidium.dyndns.org
Sat Jul 2 21:11:25 UTC 2011


Don,

On 07/02/2011 09:48 PM, Don Latham wrote:
> Hi Anders:
>
> I need something to give 2.048 MHz from a 10 MHz source, and have not
> too aggressively tried to find a way; have followed with interest many
> suggestions from TN's.  I found a company called Vectron that makes,
> ostensibly, very nice frequency translators, with an fin and an fout,
> with reasonable phase noise characteristics (?), the data sheets are
> available and intriguing, but I spent an hour or more trying to find a
> way to actually buy one.

Getting 10 MHz to 2,048 MHz isn't all that hard. You can do it in good 
old 4000 CMOS series if you need to. What application to you need it 
for? I recognice it as a standard telecom frequency and well have in 
fact implemented 10 MHz to 2,048 MHz conversion (even if that was not 
the primary target). The wideband noise needs isn't very daunting in 
telecom applications, but long-term stability is another thing.

If you really do not need to have very good wideband phase-noise, a 4046 
and a few dividers will help you.

10 MHz and 2,048 MHz has the highest common frequency in 16 kHz, so you 
need to divide by 625 on the 10 MHz and 128 on the 2,048 MHz. If you 
settle with the 4046 oscillator or chooses an off the shelf VCXO depends 
on what you need, but it is not too hard and too expensive to come by.

Cheers,
Magnus



More information about the time-nuts mailing list