[time-nuts] Timing Distribution in Mountainous Terrain

[Oz-in-DFW]

On 9/10/2010 4:04 PM, Ralph Smith wrote:
> OK, stop me if this is really stupid. The initial site is in Colorado.
> Would it be possible to use WWV? In particular:
>
> 1) Lock a reference to the carrier of one of the WWV signals
> 2) Generate PPS off of WWV-locked reference
> 3) Periodically send difference of GPSDO PPS and WWV-locked PPS home,
> along with GPS lock indication
> 4) When GPS goes away do the math at home and correct for the timing drift
> of the GPSDO compared to WWV-locked reference
>
> There are probably several fatal flaws with this approach. In particular,
> the following are required:
> 1) Ability to maintain constant lock to WWV
> 2) Common-mode error. Will the propagation from WWV be similar enough for
> all stations to it be a practical common reference.
> 3) Adequate resolution. Even if, for some reason 1 and 2 are possible,
> would the result be good enough to use.
>
> Like I say, probably completely unworkable, but what are your thoughts?
>
> Ralph
>
There are a number of problems I see with this, each of which is
sufficient to eliminate this as an option. 

   1. Propagation varies with temperature, humidity, and other
      atmospheric effects like dust content.  Over a several hundred
      mile path this is going to be at least an order of magnitude
      greater error term than the overall budget. 
   2. This is HF and skip paths are common.  The WWV wave form provides
      no mechanism for discriminating between ground and skywave.  LORAN
      accomplished this by using a short pulse.  The Skywave path was
      always later than the groundwave, and for most cases was
      completely distinct.  This can be 100s of microseconds.
   3. It's conceivable that you could have long-path propagation that
      makes a complete circuit of the earth - really long delays and no
      good way to know in the absence of a stable reference.  About 150
      milliseconds...
   4. Multipath with sky and ground waves adds yet another variable. 

And WWVB doesn't really fix much of this. 

-- 
mailto:oz at ozindfw.net    
Oz
POB 93167 
Southlake, TX 76092 (Near DFW Airport)