[time-nuts] GPS first LO need to be locked?

[jimlux]

On 3/30/17 11:06 AM, Peter Monta wrote:
>> I am curious if the first local oscillator on a GPS receiver must actually
>> be locked or coherent to the reference oscillator in the GPS receiver
>> typically running at some 10 MHz approximately. Or as long as the first LO
>> is quite stable it doesn't matter because the receiver can track the code.
>>
>
> It doesn't matter, so long as the first LO is in the ballpark so that the
> Doppler search is not needlessly large.  I'm not so familiar with the early
> receivers, but I imagine a single reference oscillator serves for
> everything---there would seem to be no reason to have more than one unless
> the antenna/downconverter were physically separate from the rest of the
> receiver.  If an older receiver used a physical source at 10.23 MHz, it
> would still need to be offset slightly for each satellite because of "code
> doppler", but this choice of frequency might slightly simplify the
> circuitry.  Current receivers would use any convenient physical rate, then
> synthesize the code rates.
>

BTW a lot of GPS receivers don't have a "first LO".. they are more like 
a Tuned RF receiver - an input BPF for L1, L2, or L5, then direct 
sampling at around 30-40 MHz -  something that makes the GPS signals 
alias down somewhere convenient (and always have positive frequency 
offset from zero, even at max negative Doppler)