[eagle] Re: Eagle 10 MHz Clock

Tom Clark, K3IO K3IO at verizon.net
Tue Apr 10 20:40:30 PDT 2007

Bill Ress wrote:
> (snip)
> If indeed the clock is used by all LO's it is important that the LO 
> designers know the reference's characteristics so that synthesis schemes 
> and loop bandwidths can be better defined.
Assume that it is a low noise crystal. Signals will come from an N-port
driver that is a part of the USO. A good standard level will be 0 dBm
into 50 ohms (although it could be a tad higher -- like up to 1V RMS =
+13 dBm if really needed).

I would suggest that each user system could use the precision reference
if it is available, but should switch over to an internal oscillator if 
the precision source goes away. This is the way most counters, spectrum
analyzers, etc work and it permits a good, simple "fail soft" option.
> With that in mind, can you Tom, 1) determine what the specs of the 
> $ystem USO might be (or point me to a place when I might deduce them), 
> and 2) what's a realistic time frame for deciding if that $ugar 
> daddy/mommy comes through with a commitment? We not only have to get a 
> handle on the 10 MHz oscillator but the RF distribution circuitry after it.
> In regards to the RF distribution circuitry, I count 8 users of the 10 
> Mhz clock. Help me out here if I missed one....
> 1) V Transmitter
> 2) U Receiver (maybe two?)
> 3) L Receivers (two)
> 4) S Transmitter
> 5) S2 Receiver
> 6) C Transmitter
> 7) IHU
> 8) Spare (for X or K beacon or a "contributing" customer)
As Lyle mentioned, all the DSP widgets need the reference signal also.
For precise ranging it is likely that some really tight special DSP code
will be loaded into the DSP engines for maximum accuracy. This will be
to ensure the flattest possible phase response thru the system (the
group delay tau is measured as d(phase)/d(freq) over the however wide
the passband can be).

Thinking of obtaining the best possible accuracy, we may want to do U/V
and a microwave pair simultaneously to provide a calibration of the
ionosphere; the one-way group delay thru the ionosphere varies as 1/f²
and has a magnitude ~2M at S-band.

73, Tom

More information about the Eagle mailing list