[time-nuts] Using GPSDO as a Refrence for Protable Amateur Radio Microwave Operations
jimlux at earthlink.net
Thu Dec 22 10:12:17 EST 2016
On 12/21/16 9:08 PM, John Hawkinson wrote:
> Chris Albertson <albertson.chris at gmail.com> (and Bob Camp <kb8tq at n1k.org>):
>> Why to people always build 10MHz GPSDOs?
> Because "a lot" (...) of amateur radio microwave equipment is designed
> off the shelf to accept an external 10 MHz input. [And other kinds of
> equipment, too.] If you're not designing from the ground-up, then it makes
> a lot of sense.
In most cases, folks want good close in phase noise - 5 and 10 MHz are
in the sweet spot of having good close in phase noise even after the
20log(N) bump from multiplying it up. A 100 MHz oscillator takes less
multiplication, but because the crystal is physically smaller, it
probably doesn't have phase noise that is lower than a 10 MHz multiplied up.
if you're multiplying up to 8 or 10 or 32 GHz, whether you start at 10
or 100, it's still a lot of multiplying. If you're using a PLL, the
frequency gets divided down into the phase/frequency comparator, and
dividing down from 10 is no different than dividing down from 100.
If you're doing chains of multipliers up, then starting a 10 gives you a
bit more flexibility to design the multiplier chain in terms of where
the various frequencies wind up, so as to avoid "inband" spurs later on.
Most ham radio designs tend to be mixes and matches of previous widgets.
The whole transverter model, starting with 10 meters or 2 meters, and
then mixing, converting, multiplying, etc. is great if you're trying to
re-use stuff you already have, gradually adding bands, or trying to use
30-40 year old surplus microwave gear. It's not necessarily a good
approach if you were starting from scratch and using modern components.
More information about the time-nuts