Symmetricom makes GPS based NTP time servers with excellent holdover
capability.  I think our "SyncServer" with an OXCO is good for +/- 0.5
second holdover over something like 60 days.  They have options for Rb
oscillators installed that will make that much much better and it might fall
inside of your requirements.  Give them a look.  You're welcome to mail me
directly with questions about mine if you like.

-Bob

On Thu, Sep 9, 2010 at 9:37 AM, Ralph Smith ralph@ralphsmith.org wrote:

The requirement is 30 nanoseconds, so individual rubidium holdover over
six days won't cut it.

Ralph

On Thu, September 9, 2010 11:58 am, Robert Darlington wrote:

On 9/9/2010 8:37 AM, Ralph Smith wrote:

30 ns seems a little closer than most radio site applications need...
what drives this requirement?

Wow, ok, and what drives that requirement? Can you use any other
mutually visible thing, or do we assume all satellites have vanished
from orbit?

Matthew Kaufman

Does the GPS backup include other sats ? As long as all sites could see the same
sat then using it as a standard they would drift together.

Stanley

----- Original Message ----
From: Ralph Smith ralph@ralphsmith.org
To: Discussion of precise time and frequency measurement time-nuts@febo.com
Sent: Thu, September 9, 2010 11:45:04 AM
Subject: Re: [time-nuts] Timing Distribution in Mountainous Terrain

The requirement is 30 nanoseconds, so individual rubidium holdover over
six days won't cut it.

Ralph

On Thu, September 9, 2010 11:58 am, Robert Darlington wrote:

time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

On Thu, September 9, 2010 1:10 pm, Matthew Kaufman wrote:

Aircraft surveillance using multilateration.

Paranoia. People making the requirements are concerned with GPS going away
due to solar flare or some other reason.

No satellites.

Thanks,
Ralph

On 9/9/2010 10:42 AM, Ralph Smith wrote:

So timing errors just become position errors. How do the sites talk back
to the display? Can't you null out position errors if enough sites can
see a single plane, and thus learn the timing error of the drifting
(relative to other) site?

Once everyone relies on GPS approaches and ADS-B, the planes will be
grounded long before 6 whole days of GPS outage anyway.

Ok then. My best answer is to use the planes themselves as the common
reference, at least the ones high enough that enough sites can see them.

Also consider that you might be able to find additional mountaintop
sites to plant fixed squitter-emitter transponders at that can be seen
by 2 (or more) sites.

Matthew Kaufman

On Sep 9, 2010, at 10:42, "Ralph Smith" ralph@ralphsmith.org wrote:

Hmm... So the decision makers think that after a solar flare or "some other reason" (hostile destruction of the birds, perhaps?) takes out the GPS system, it'll be back up and running within six days? That sounds optimistic to me.

It sounds to me more like you would need to function indefinitely without GPS, and just use GPS for initial position survey and as a convenient way to synchronize to external time reference (with ground-based backup methods in place and periodically tested).

On Thu, September 9, 2010 1:55 pm, Matthew Kaufman wrote:

Sites communicate via landline telco. If there are sufficient mutually
visible networked sites to form a solution on an aircraft visible to
stations not in the timing network that would work, and is one of the
options we are studying.

You're making the mistake of applying logic. ;) Actually, aircraft can
continue to fly VFR or navigate using VOR/DME and inertial navigation. The
radios are part of an ADS-B installation.

Thanks, all of these are various options we are considering, considering
all of the engineering trade-offs.

Ralph

On 9/9/2010 11:57 AM, Ralph Smith wrote:

Yes, and they can make routine position reports and ATC can apply
time-based sequencing for IFR traffic... see all sorts of past solutions
to this problem.

VFR aircraft don't need anyone to know where they are anyway, and IFR
only need it if they're being provided radar (or equivalent) separation
services.

Things would work just fine after a period of adjustment, and during
that period of adjustment the planes that need this kind of position
determination would be grounded.

I run GPS-disciplined oscillators at mountaintop radio sites and am a
pilot, so this thread is particularly interesting to me.

Matthew Kaufman

Ralph,

so you're talking about 5.8E-14, right?
I'd think no off the shelf caesium, even when run in a temperature
controlled environment, will get you there.
Well, at a first glance, a 5071A with high performance tube would, if
you keep any environmental effects out.
So you'll likely need to sync them through a reliable and reproduceable
link.
Just a quick shot though.

Adrian

Ralph Smith schrieb:

Adrian wrote:

I would like to point out that the environmental sensitivities of
the 5071A are unmeasureable, and the measurement threshold is
far below 5.8E-14.  I would estimate that the 5071A (and ONLY the 5071A
among commercial clocks) could get the job done provided that you could
compare its frequency to GPS to the stated accuracy.  This would
be using the 5071 as a secondary standard.  You still need to
deal with the short term stability of the 5071A, depending on
your system needs.  JPL uses H masers as flywheels.

Rick Karlquist N6RK
member 5071A design team

See TVB's site for an experiment with moving CS standards around...

Adrian

--
"Neither the voice of authority nor the weight of reason and argument are
as significant as experiment, for thence comes quiet to the mind."
R. Bacon

Dr. Don Latham AJ7LL
Six Mile Systems LLP
17850 Six Mile Road
POB 134
Huson, MT, 59846
VOX 406-626-4304
www.lightningforensics.com
www.sixmilesystems.com

Hi

If 30 ns is the system goal, then you have a lot more to budget for than simple clock error. You could easily be below 10 ns for just the clock portion of the budget. I suspect that multiple 5071's and a maser or two at each site will be the ultimate solution if each must stand alone for 6 days and the set of 10 stay within 30 ns p-p. If reliability planning includes 6 days past end of GPS it also likely includes significant redundancy and more than a few 9's in the confidence factor.

I suspect it's cheaper to buy the 5071's than to have a half dozen SR-71 category aircraft ready to climb high enough to act as common view targets. High altitude balloons might work pretty well and they would be cheaper to keep in hot standby mode.

Bob

On Sep 9, 2010, at 8:01 PM, "Rick Karlquist" richard@karlquist.com wrote:

On Sep 9, 2010, at 9:27 PM, Bob Camp wrote:

I expect the more likely outcome to be a reality adjustment and relaxing of the requirements, possibly with some form of satellite common view target.

Ralph

On Sep 9, 2010, at 8:01 PM, Rick Karlquist wrote:

I would imagine the cost of a 5071A per radio station would make the check writers swallow hard, and adding an H Maser into the mix would really get their attention. Especially if you need redundancy.

Thanks for all the input. Being asked to design for Armegeddon begs the question of what will the system do after that.

Ralph

Loran?

Joe

-----Original Message-----
From: time-nuts-bounces@febo.com [mailto:time-nuts-bounces@febo.com] On
Behalf Of Mark J. Blair
Sent: Thursday, September 09, 2010 1:48 PM
To: Discussion of precise time and frequency measurement
Subject: Re: [time-nuts] Timing Distribution in Mountainous Terrain

On Sep 9, 2010, at 10:42, "Ralph Smith" ralph@ralphsmith.org wrote:

Hmm... So the decision makers think that after a solar flare or "some other
reason" (hostile destruction of the birds, perhaps?) takes out the GPS
system, it'll be back up and running within six days? That sounds optimistic
to me.

It sounds to me more like you would need to function indefinitely without
GPS, and just use GPS for initial position survey and as a convenient way to
synchronize to external time reference (with ground-based backup methods in
place and periodically tested).

time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Ralph Smith wrote:

and presumably some form of bistatic radar with an illuminator and
multiple receviers isn't going to cut it?

You're fairly close.. is any RF method out of bounds?  Could you radiate
a low frequency signal that propagates by ground wave?
(e.g. Omega.. but I don't think it was anywhere near nanoseconds..)

Even a really, really good ovenized quartz oscillator isn't going to be
that good.  You're looking for 1E-13 sorts of precision, right (50 ns
out of 500,000 seconds).  I seem to recall numbers like 1 part in 1E10
over 400 days.  You might do better with selected units, but this is
million dollar a copy sort of units.  The Cassini USO drifts between
1E-12 and 1E-11 per day.

You're really looking at Cs standards... (about 2E-14 at 1E6 seconds)

Rick Karlquist wrote:

We also use the maser as a very low phase noise signal in the 10-1000
second tau range..
We multiply it up and send the (very clean) signal out to the
spacecraft, it gets tracked by a loop with a few Hz BW, then sent back
to earth where it's compared to the maser again to measure Doppler.

Basically we measure doppler and doppler rate over a few minutes,
assuming that the transmitted signal is constant during that time (which
the maser is, for all practical purposes).  Over a longer time span
(e.g. the time between transmit and receive, which could be many hours,
implying that we are transmitting from one station and receiving from
another) I assume we use atomic standards and/or astronomical sources
(pulsars).