[ale] Re: WWVB to computer

[deritchie at earthlink.net]

Alfo, you might want to look at www.tapr.org and look at the docs for TAC2 (no longer produced, but it may give you some ideas).

How much accuracy does someone really need that isn't already handled by ntp?

-- Dave Ritchie
Sent via BlackBerry from T-Mobile

-----Original Message-----
From: John Mills <johnmills at speakeasy.net>

Date: Tue, 18 Dec 2007 16:43:03 
To:Atlanta Linux Enthusiasts <ale at ale.org>
Cc:"Michael H. Warfield" <mhw at WittsEnd.com>
Subject: Re: [ale] Re: WWVB to computer


Christopher -

I enthusiastically second Mike's recommendation.

I looked into this many(!) years before GPS for a satellite ground station 
that had to get time references from various international standards 
bureaus, including WWV (but not WWVB). Our standards lab at S-A _did_ 
synch from WWVB. There are several factors that complicate using WWV[B], 
render it expensive, and/or limit its accuracy.

You don't say anything about the accuracy you need, but ...

First, the accuracy: "shortwave" WWV gives you a "beep" tone every minute,
which can be detected with an accuracy around 60 msec (to drive - say - a
counter or similar application). Not even close to GPS accuracy! This
turned out to be our only global option c.1976, and the 60 msec. was a
limiting factor for some orbital computations: good enough, just.

WWV signals also include modulation to drive time-code generators to much
better accuracy, but this requires a special demod and the time-code
generator to follow it.  I have also been told (but did not verify) this
approach is not reliable very far from Colorado Springs!

WWVB _can_ be demodulated to very good accuracy over the continental US,
but you have to phase-lock your time generator to the 60KHz _carrier_. S-A
used a large, multi-turn loop antenna on the roof, a high-grade PLL
filter, and (again) one or more time-code generators and/or converters. 
Rate information comes off the 60Khz and (IIRC) phase demodulation gives 
you an encoded absolute time.

Systems I delivered more recently had a remote "hockey-puck" GPS antenna,
and a receiver and decoding logic that plugged into the time-code
generator. You can't believe what a relief, economy, and improvement this
was!

If you only need a serial or USB computer interface (and accuracy), you
may find a "consumer oriented" GPS that has a remote antenna.

(Interesting side note: GPS time may carry a fixed offset from UTC - I
don't remember exactly how that works, but it bit us when we were
recording time and position from a simple GPS into a laptop computer in a
small aircraft, and trying to correlate it with our radar track of the
same airplane, using time derived from USNO. #8-)

NIST used to have an excellent, free paperback reference on timekeeping.  
I expect Google on 'time-code generator' or a visit to [www.usno.navy.mil]
would give you a lot of data to sort through. If not, drop me a line and
I'll see how many old vendors have survived or adapted to GPS.

Finally, have you checked into 'ntp' referencing? I think it even works
effectively over a dialup, by using multiple round trips and [optionally]
multiple servers to sync-up. The full ntp software download also includes
a server you can run on your LAN that is corrected for offset and average
drift rate each time you re-sync. Again, USNO (and an interested faculty
member at U.Del. -- follow USNO's links) have a wealth of info.

Just some thoughts.

 - Mills

On Tue, 18 Dec 2007, Christopher Fowler wrote:

> On Tue, 2007-12-18 at 13:09 -0500, Michael H. Warfield wrote:
> > In this day and age of cheap GPS modules, I wouldn't even consider
> > going the WWVB route. 
> 
> GPS propagation inside Quality Services in Suwanee?  Feasible?

_______________________________________________
Ale mailing list
Ale at ale.org
http://www.ale.org/mailman/listinfo/ale