[time-nuts] OT: xtal osc PN
jimlux at earthlink.net
Sun Sep 19 18:44:47 UTC 2010
francesco messineo wrote:
> On 9/19/10, jimlux <jimlux at earthlink.net> wrote:
>> francesco messineo wrote:
>>>>> It's hard to explain why to ones not familiar with weak signal
>>>>> operation between broadcasting signals, but really the noise floor
>>>>> raise a lot when you have some 5 or 6 broadcasts signals in 500 KHz of
>>>>> band (all with power levels of at least 10 dB more than the levels
>>>>> used in amateur radio, often +20 dB more)
>>>> I would need some more fundamental understanding of the system and needs
>>>> to be able to understand how you come up with the above noise level at
>>>> 100 Hz.
>>> as I said, if it's not possible or not practical, of course I'll take
>>> what I can get. The receiver will be limited by its phase noise and
>>> not for example by its IMD3.
>>> I think already -110 dBc/Hz at 100 Hz is better than any LO in
>>> commercial receivers (for ham radio at least).
>> But why at 100 Hz offset. Are you looking for weak signals 100 Hz away
>> from the strong interferer? (and are worried about reciprocal mixing
>> from the LO) That would imply that the interferer has equally good
>> phase noise, and that's not particularly likely?
> the first scenario. Even if the interferer has poor phase noise, why
> folding back another share of interference on my side?
> As for the 100 Hz offset, it's just a practical measure, if @100 Hz
> things are good, @ 1KHz or more they must be even better, right?
> In real life though, interferers are often much closer than 100 Hz to
> the wanted signal and this often means no contact made (it happened
> two times to me this year already).
> On the other side of the contact however, the band was much more clean.
Sure.. but you might find yourself struggling to get good 100Hz
performance, and if your real need is at 1000Hz offset, it might be
Consider this..a PLL based LO and you have a fairly quiet VCO and you're
locking it to a quiet quartz oscillator to clean up the close in noise.
If the loop bandwidth is 500 Hz, then it's the VCO that's setting the
1kHz offset noise and the XO that's setting the 100 Hz noise.
In very narrow band applications (e.g. coherent transponders used in
deep space), we might have a crystal locking a DRO (or, these days, a
GaAs VCO), but the crystal is locked to the received signal. The
crystal to DRO loop might have a BW of a few tens of kHz, but the
receiver to crystal loop might have a BW of 1 Hz. Since the measurement
of the round trip signal is made by averaging over many seconds, we
aren't as worried about the 100Hz and farther out noise, except insofar
as it adds a bit of noise to the received signal.
On the other hand, in this sort of application, there's no interfering
More information about the time-nuts