[hpsdr] GPS 1PPS board

[John Ackermann N8UR]

If I did the math correctly, that value is 1x10e-10.  That's actually a 
couple of orders of magnitude less than the accuracy of the GPS 
constellation; it's generally accepted that averaging over a day or so, 
you can get to within parts in 10e-13.

1x10e-10 is a fairly common performance spec for simple GPSDOs.  Better 
ones, like the surplus HP and Trimble units many hams have, are more 
like 1x10e-12.  But both are way better than any HF or VHF work could 
ever require, so most of this discussion is in "time-nuts" territory. :-)

But the key thing is that any frequency system based on GPS has to 
average over a long time period, usually 1000 seconds or more, because 
the jitter on the GPS signal is much worse than 1x10e-10.  Even really 
good GPS timing receivers like the Motorola M12+ have about 30 
nanoseconds of jitter.  A unit that has 1 microsecond of jitter bounces 
around by 1x10e-6, or 1 Hz at 1 Mhz, potentially every second.

So, you need to average the signal using a PLL with a very long time 
constant, and use that to steer a crystal oscillator that is stable over 
shorter time periods, to make a workable frequency standard out of GPS.

And there the key is that a 10 kHz pulse rate makes the PLL easier to 
implement -- but in the end it won't be any more accurate than the 1 PPS 
signal because the 10 kHz pulses are themselves only "disciplined" once 
per second by the GPS engine.

73,
John
----

Chris Salinas wrote:
> John,
> I just read a website on this TU00 model and that they measured the 
> accuracy of +/- .000001 Hz at the 10 Khz pin. Too me it sounds stable 
> enough for a majority of work.
>  
> I ordered one yesterday just too play with and do some measurements. Can 
> always use it for APRS if it doesn't hold water.
>  
> I'll let you all know what I find.
>  
> Chris N0TTW
> 
> --- On *Mon, 4/6/09, John Ackermann N8UR /<jra at febo.com>/* wrote:
> 
> 
>     From: John Ackermann N8UR <jra at febo.com>
>     Subject: Re: [hpsdr] GPS 1PPS board
>     To: "Henry Vredegoor" <henry.vredegoor at gmail.com>
>     Cc: "'Graham / KE9H'" <KE9H at austin.rr.com>, "'Chris Salinas'"
>     <n0ttw at yahoo.com>, "'HPSDR discussion list'" <hpsdr at lists.hpsdr.org>
>     Date: Monday, April 6, 2009, 3:09 PM
> 
>     I'd put it a little differently:  the 10kHz signal does not make it
>     possible to get *better* performance than a GPS with a 1 PPS signal,
>     but it allows much simpler circuitry since filtering a 10 kHz signal
>     is much easier than filtering 1 PPS, particularly if using analog
>     components.
> 
>     However it's important to realize that the GPS part of any GPSDO
>     affects only the long term frequency stability.  When measuring over
>     time periods shorter than around 1000 seconds, the performance of
>     the XO is paramount.  One of the problems with building a cheap
>     GPSDO is that good XOs are expensive, and therefore the short term
>     stability is pretty much  tied to the builder's budget or scrounging
>     skills.
> 
>     The control loop should (a) optimize the "cross-over" point where
>     the XO stability matches the GPS system stability; and (b) avoid
>     messing up the short term stability through excess noise applied to
>     the XO's tuning voltage.
> 
>     John
>     ----
> 
> 


 1239051722.0