[hpsdr] Precise receive frequency measurement

[John Sager]

Brian, thanks for that. I'm seriously thinking of building a GPS-locked
10MHz reference anyway - plenty of designs around.

73 John G8ONH

On 08/03/13 15:15, Brian Lloyd wrote:
> ***** High Performance Software Defined Radio Discussion List *****
> 
> 
> 
> 
> 
> On Fri, Mar 8, 2013 at 1:08 AM, John Sager <john at sager.me.uk
> <mailto:john at sager.me.uk>> wrote:
> 
>     ***** High Performance Software Defined Radio Discussion List *****
> 
>     Or measure the actual error in the onboard 10MHz reference. I did that using
>     a standard frequency broadcast (RWL), and the 'fmt' program in Joe Taylor's
>     WSPR software. Mine is about 1.5Hz high after it warms up.
> 
> 
> The built-in 10MHz reference is (if I recall properly) rated at 1ppm
> accuracy. So your 1.5 Hz error at 10MHz is well within spec. 
> 
> But the thing I noticed about the built-in reference, at least on the Hermes
> board sold by Apache Labs, is that the internal temperature correction is
> not smooth but corrects in jumps. Phil figured that one out after I reported
> frequency instability. I had hoped to be able to quickly calibrate the
> Hermes board to WWV and then be able to do non-critical frequency
> measurements, i.e. under 1Hz but not quite down to 1mHz, but that appears
> not to be possible using the built-in reference. Of course, for people using
> the radio for "normal" operations, they wouldn't notice a thing. You just
> have to be anal-retentive like me to catch it. :-) 
> 
> (And I really appreciated Phil's and the folks at Apache Labs' efforts to
> figure out what was going on.)
> 

 1362824393.0