[time-nuts] WWVB format change in 2012

Paul tic-toc at bodosom.net
Mon Feb 29 13:39:19 EST 2016


On Mon, Feb 29, 2016 at 6:35 AM, Hal Murray <hmurray at megapathdsl.net> wrote:

> The audio stuff seems particularly ugly.
>
> Are there any good non-GPS options these days?  In this context, "good" is
> a
> bit hard to pin down.  My straw man is either something in production or
> something like a TBolt or Z38xx that is available surplus at hobbyist
> prices.
>

On Mon, Feb 29, 2016 at 6:39 AM, Hal Murray <hmurray at megapathdsl.net> wrote:
tic-toc at bodosom.net said:
>> If you can get audio they work.  The documented accuracy for WWV is
within 1
>> millisecond of the time pulse.

>Where is that documented and/or can anybody verify that they get results in
>that range?

First off this is a bit time-nuts relevant because RF timing propagation is
germane but the basic pulse extraction is low accuracy and hence not so
much time-nut relevant.  That also informs the forest for the trees issue
here: the original intent (as in Dr. Mills PDP-11 assembler coded
demodulator) is a periodic (i.e. daily) clock update.  So despite the
predicted 1ms accuracy a .1s system error was probably considered
acceptable.  The WWV refclock shares this operational sense with the ACTS
driver.  You phone home maybe once a day and, in the case of the ACTS
driver, NTPD won't use it unless you're orphaned.

Regarding non-GPS.  Do you consider a CDMA receiver non-GPS?  How about a
disciplined oscillator in hold-over?  In consideration of the recent GPS
glitch is GLONASS non-GPS or are you speaking more broadly (ie. GPS as a
synonym for GNSS).

And lastly despite the politics it would seem the folks on the
ntp-questions and ntp-hackers are a better resource.  Unless they're all
here.


More information about the time-nuts mailing list