Indirect GPS time source options

[Olivier Drouin]

Hello,
I'm located in Canada and I'm trying to find indirect GPS time sources. I like the idea of using CDMA as a time source but it seems that this technology will not be around for a lot longer.
For exemple, Endrun Technologies sells the Sonoma N12 which seems to allow time synchronization against CDMA signal without the need for a carrier subscriptions.


Is the CDMA method still available with these other cellulare network technologies like HSPA or LTE? If not, do I have other options beside a GPS antenna with a sky view?



Thank you and best regards,
Olivier Drouin.

[William Unruh]

On 2014-03-12, Olivier Drouin <ol.d...@gmail.com> wrote:
> Hello,
> I'm located in Canada and I'm trying to find indirect GPS time sources. I like the idea of using CDMA as a time source but it seems that this technology will not be around for a lot longer.

What is an "indirect gps timing source"? How does that differ from using
ntp to go to a level 1 server which gets its time from gps. While cdma
uses accurate time, it is not really designed to deliver accurate time.
GPS is.

> For exemple, Endrun Technologies sells the Sonoma N12 which seems to allow time synchronization against CDMA signal without the need for a carrier subscriptions.
>

And a gps receiver with PPS output is probably 1/100 the price.

>
> Is the CDMA method still available with these other cellulare network technologies like HSPA or LTE? If not, do I have other options beside a GPS antenna with a sky view?

CDMA also needs an antenna with cell tower view. What's the difference?
(Note my gps on an east facing window through a large cedar tree works
just fine.)

[Terje Mathisen]

The Endrun unit worked very nicely inside the server room of our Tampa
facility when I installed it 10+ years ago.

There was no possible outside antenna location and at that time
absolutely no gps units that would work from inside the metal glazed
windows of the high rise building.

Terje

--
- <Terje.Mathisen at tmsw.no>
"almost all programming can be viewed as an exercise in caching"

[Olivier Drouin]

On Wed, Mar 12, 2014 at 3:59 PM, William Unruh <un...@invalid.ca> wrote:

> On 2014-03-12, Olivier Drouin <ol.d...@gmail.com> wrote:
> > Hello,
> > I'm located in Canada and I'm trying to find indirect GPS time
> sources. I like the idea of using CDMA as a time source but it seems that
> this technology will not be around for a lot longer.
>
> What is an "indirect gps timing source"? How does that differ from using
> ntp to go to a level 1 server which gets its time from gps. While cdma
> uses accurate time, it is not really designed to deliver accurate time.
> GPS is.
>

Well, from what I understand the cell networks are closely synchronized
against the GPS constellation. So, synchronizing against the cellular
network is called "indirect gps", at least according to the endrun website
(I haven't seen this nomenclature anywhere else).

It is different of GPS because if your WAN links are down, which can happen
once every few years, then, hopefully, you can still reach the cellular
network in your neighborhood or else its the end of the world or
something...

Also, from what I understand and please correct me if I'm wrong but
cellular networks are indeed designed to be accurately timed because it's
needed for the normal operations of the cell network.


Thank you for your reply, best regards,
Olivier Drouin

>
> > For exemple, Endrun Technologies sells the Sonoma N12 which seems
> to allow time synchronization against CDMA signal without the need for a
> carrier subscriptions.
> >
>
> And a gps receiver with PPS output is probably 1/100 the price.
>
> >
> > Is the CDMA method still available with these other cellulare
> network technologies like HSPA or LTE? If not, do I have other options
> beside a GPS antenna with a sky view?
>
> CDMA also needs an antenna with cell tower view. What's the difference?
> (Note my gps on an east facing window through a large cedar tree works
> just fine.)
>
>

As Terje is mentioning, the cell antenna doesnt need a line of sight to a
cell antenna, so It can be installed inside the server room and you dont
have to deal with the maintenance problems (lots of snow around here) of a
GPS antenna on the roof...

> >
> >
> >
> > Thank you and best regards,
> > Olivier Drouin.
>

> _______________________________________________
> questions mailing list
> ques...@lists.ntp.org
> http://lists.ntp.org/listinfo/questions

[Olivier Drouin]

Oops, I said

It is different of GPS because if your WAN links are down, which can happen
once every few years, then, hopefully, you can still reach the cellular
network in your neighborhood or else its the end of the world or
something...

but it should read

It is different from using ntp to go to a level 1 server because if your

WAN links are down, which can happen once every few years, then,
hopefully, you can still reach the cellular network in your neighborhood or
else its the end of the world or something...

On Wed, Mar 12, 2014 at 3:59 PM, William Unruh <un...@invalid.ca> wrote:

[Jochen Bern]

On 13.03.2014 13:00, question...@lists.ntp.org digested:
> From: Olivier Drouin <ol.d...@gmail.com>

>
> For exemple, Endrun Technologies sells the Sonoma N12 which seems to
> allow time synchronization against CDMA signal without the need for
> a carrier subscriptions.

(This thing supposedly answers more requests per second than I've set up
NTP clients in my entire life:
http://www.endruntechnologies.com/ntp-server.htm
- do you have thousands of *local* clients to serve and/or a Gigabit
uplink to bring that throughput to *remote* ones?)

> If not, do I have other options beside a GPS antenna with a sky view?

Horror story: I once set up a customer's housing rack, right next to a
window, with a DCF77 receiver cable-tied onto its top. (That was less
than 50km from the DCF77 antenna array supposedly serving the whole of
Central Europe, mind.) Half a year into operation, reception plain
stopped and never came back, with no change in nearby racks apparent to
the naked eye. If quietly watching your N12 turn into a $5,000 rack
weight is not an option, then IMHO neither is an indoor antenna that the
building administration doesn't officially know about and support.

(I've also seen server rooms where you have to actually stand next to a
specific window a dozen meters from the nearest rack to get any GSM
reception, so potential reception problems are definitely not limited to
DCF77 and/or GPS.)

You specifically asked about syncs where the ultimate source is GPS.
There are lots of forms of time syncing built into building automation
systems (ISDN, single-wire, fire alarms, ...), as well as into WAN
connections you might have in your very rack, but there's likely some
DIY involved and you very likely will not get the same precision ...

(Personal opinion: I just don't get why there apparently are so many
hosting/housing/colo providers where you *cannot* get access to a local
NTP server or other central sync source even if you threaten to throw
money at them. Tamperproofness-wise, two miniature PCs with GPS
receivers on a roof your provider owns and controls run circles around
everybody and his VMs talking to public no-crypto NTP servers through a
plain Internet uplink ...)

Regards,
J. Bern
--
*NEU* - NEC IT-Infrastruktur-Produkte im <http://www.linworks-shop.de/>:
Server--Storage--Virtualisierung--Management SW--Passion for Performance
Jochen Bern, Systemingenieur --- LINworks GmbH <http://www.LINworks.de/>
Postfach 100121, 64201 Darmstadt | Robert-Koch-Str. 9, 64331 Weiterstadt
PGP (1024D/4096g) FP = D18B 41B1 16C0 11BA 7F8C DCF7 E1D5 FAF4 444E 1C27
Tel. +49 6151 9067-231, Zentr. -0, Fax -299 - Amtsg. Darmstadt HRB 85202
Unternehmenssitz Weiterstadt, Geschäftsführer Metin Dogan, Oliver Michel

[Jochen Bern]

> I work for a college and a lot of my students are complaining that
> their cell phones are off by 4 min. My servers poll us.pool.ntp.gov.

(I trust that that's ....org .)

> The cell phones poll from time.gov.
> If you look at the website for time.gov it does display the time
> being off by 4 min.

It's correct within wristwatch-and-eyeball accuracy for me right now.

> Would be nice if pool.ntp.org would display it's time on the site
> so I can see if it's actually an internal issue.

*.pool.ntp.org entries are *pools* of NTP servers run by whoever owns a
particular one, so technically, the pool doesn't *have* a single
inherent time (available from all members equally) that the website
could display. Identify the (single ... ?) *IP* your servers currently
poll so that *that one* public server can be scrutinized.

And prepare to use *several* pool servers instead of one, recommended
precisely to avoid such a scenario.

[William Unruh]

On 2014-03-13, Olivier Drouin <ol.d...@gmail.com> wrote:
> On Wed, Mar 12, 2014 at 3:59 PM, William Unruh <un...@invalid.ca> wrote:
>
>> On 2014-03-12, Olivier Drouin <ol.d...@gmail.com> wrote:
>> > Hello,
>> > I'm located in Canada and I'm trying to find indirect GPS time
>> sources. I like the idea of using CDMA as a time source but it seems that
>> this technology will not be around for a lot longer.
>>
>> What is an "indirect gps timing source"? How does that differ from using
>> ntp to go to a level 1 server which gets its time from gps. While cdma
>> uses accurate time, it is not really designed to deliver accurate time.
>> GPS is.
>>
>
> Well, from what I understand the cell networks are closely synchronized
> against the GPS constellation. So, synchronizing against the cellular
> network is called "indirect gps", at least according to the endrun website
> (I haven't seen this nomenclature anywhere else).

And in npt terms it would be called a stratum 1 server.

>
> It is different of GPS because if your WAN links are down, which can happen
> once every few years, then, hopefully, you can still reach the cellular
> network in your neighborhood or else its the end of the world or
> something...

Or a power blackout, while the sattelites keep beeping away even if the
whole eastern seaboard is in the dark.


And except a gps receiver is far cheaper and more accurate than a cell
timing receiver.

>
> Also, from what I understand and please correct me if I'm wrong but
> cellular networks are indeed designed to be accurately timed because it's
> needed for the normal operations of the cell network.

Yes, but they are not designed to deliver accurate time to the rest of
the world AFAIK.

[David Taylor]

On 13/03/2014 15:24, Jochen Bern wrote:
[]

> (Personal opinion: I just don't get why there apparently are so many
> hosting/housing/colo providers where you *cannot* get access to a local
> NTP server or other central sync source even if you threaten to throw
> money at them. Tamperproofness-wise, two miniature PCs with GPS
> receivers on a roof your provider owns and controls run circles around
> everybody and his VMs talking to public no-crypto NTP servers through a
> plain Internet uplink ...)
>
> Regards,
> J. Bern

Example of low-cost, low-power miniature NTP server as mentioned by Jochen:

http://www.satsignal.eu/ntp/Raspberry-Pi-NTP.html

--
Cheers,
David
Web: http://www.satsignal.eu

[Hal Murray]

In article <lfqeb1$m6$5...@dont-email.me>,

William Unruh <un...@invalid.ca> writes:

>CDMA also needs an antenna with cell tower view. What's the difference?

1/r-squared.

The GPS satellites are way up in the air. A cell tower is a lot closer.
You can get a signal from a cell tower into a lot of places where GPS
doesn't work.

--
These are my opinions. I hate spam.

[Steve Kostecke]

On 2014-03-13, William Unruh <un...@invalid.ca> wrote:

> On 2014-03-13, Olivier Drouin <ol.d...@gmail.com> wrote:
>
>> Also, from what I understand and please correct me if I'm wrong but
>> cellular networks are indeed designed to be accurately timed because
>> it's needed for the normal operations of the cell network.
>
> Yes, but they are not designed to deliver accurate time to the rest of
> the world AFAIK.

http://www.cdg.org/technology/cdma_technology/a_ross/systemtime.asp

CDMA requires accurate time synchronization among all base stations and
mobile stations. The accuracy must be within a few microseconds among
base stations because the pilot code phase is used to distinguish them.
When a mobile station is communicating with a base station they must be
synchronized to within a fraction of a chip (814 ns). And the "clocks"
(the PN generators) that must be synchronized have a period of 37
centuries.

http://www.endruntechnologies.com/gps-cdma3.htm

GPS Absolute timing accuracy of unit is under 30 nanoseconds.

CDMA Absolute timing accuracy of unit is typically under 10
microseconds.

Network timing accuracy 1/2 - 2 milliseconds, typical for both GPS and
CDMA NTP products.

Also see: http://www.endruntechnologies.com/cdma.htm

--
Steve Kostecke <kost...@ntp.org>
NTP Public Services Project - http://support.ntp.org/

[Olivier Drouin]

Thank you Greg, Jochen, William,

Great answers.

Diversity is really what I'm looking for and
I dont really need microsecond accuracy.
CDMA seemed to be the most 'convenient' way
to add diversity without a need for a working WAN link and/or line of sight
view to the sky.
Also, I already know that I get good cell
signal from inside the server room.
The thing with CDMA is that it looks like
it'll not be around for many years and I haven't seen any equipement for
4g, hspa, LTE, etc...
For your information, I think I can get away
with a 10K budget but I need to be absolutely sure It'll work beforehand.

I'll talk with the facility owner and do
some tests with handheld GPS so I can verify what kind of signal I can get
(and where). I'm a bit worried because I dont have roof access whenver I
want and I dont know how the antenna will behave with the snow we get
around here (quebec city).

Ill also verify if I can synthonize the CHU
radio frequency inside the server room if I can get my hand on a scanner.

Olivier Drouin.




On Thu, Mar 13, 2014 at 1:26 PM, Dowd, Greg <Greg...@microsemi.com> wrote:

> You are not incorrect but I think that "indirect GPS time" may be more of
> a marketing term. Basically, GPS is the predominant method of time
> dissemination. When you think of a commercial ntp server, really you just
> moved the GPS antenna from your PC (getting rid of the receiver and coax
> cable) over to the ntp server, and then distributed the timing information
> over a different cable (your ethernet cable). Since NTP is layer 3, and
> the bus isn't dedicated, you introduce error but it's still GPS time.
>
> My term for the concept I understand you to be discussing is timing
> diversity. I prefer to have a number of different methods of accessing
> time data. GPS is the baseline but access to the L1 GPS signal is limited
> in some locations, particularly indoors. Keep in mind the analogy that the
> GPS broadcast is like a light bulb shining from 22000 km away. CDMA is
> nice as it's a powerful signal (compared to GPS) but it is limited in
> geographic deployment. Most 3G/LTE cellular systems these days are tightly
> syntonized (frequency) but not synchronized (time). So, no, you won't get
> microsecond level timing from GSM. Next gen LTE has a profile for LTE-TDD
> which will require about the same level of synchronization as CDMA.
>
> If you are looking to add diversity to an installation, your first best
> bet is to move to a GNSS receiver. Like your smartphone, a GNSS receiver
> typically works with both GPS and the Russian Glonass satellite systems.
> There's some diversity.
>
> For Canada, you could also look at eLoran. That operates on a completely
> different frequency plan to satellite timing. I think Canada has CHU radio
> timing as well (3330kHz). There are a number of signals of opportunity
> that require custom hardware to access either time or frequency. You
> mentioned CDMA. Digital television is another example.
>
> As someone mentioned earlier, to get a more precise answer you have to
> provide a more detailed analysis of your requirements and budget. Anything
> is possible. There are neutron stars out there dieing to provide you with
> a frequency reference :-)

[William Unruh]

Thanks. By the way, network timing I have found is much better than
1/2ms. If I get the time from a gps and a stratum 1 source 2000 miles
away, the timing accuracy of that source is in the 50usec range. (the
delay is clearly much larger than that ). From a local stratum 1 source
(ie in the same building) it is more like 15usec.

[William Unruh]

On 2014-03-13, Hal Murray <hal-u...@ip-64-139-1-69.sjc.megapath.net> wrote:
> In article <lfqeb1$m6$5...@dont-email.me>,
> William Unruh <un...@invalid.ca> writes:
>
>>CDMA also needs an antenna with cell tower view. What's the difference?
>
> 1/r-squared.
>
> The GPS satellites are way up in the air. A cell tower is a lot closer.
> You can get a signal from a cell tower into a lot of places where GPS
> doesn't work.

There are loads of places here where the cell phone signal disappears.
The gps signal is there everywhere outside. Agreed that in favourable
circumstances, the cell signal may penetrate into the building much
deeper, but it can also be non-existant even outdoors.


>

[William Unruh]

On 2014-03-13, Olivier Drouin <ol.d...@gmail.com> wrote:

> Thank you Greg, Jochen, William,
>
> Great answers.
>
> Diversity is really what I'm looking for and
> I dont really need microsecond accuracy.
> CDMA seemed to be the most 'convenient' way
> to add diversity without a need for a working WAN link and/or line of sight
> view to the sky.
> Also, I already know that I get good cell
> signal from inside the server room.
> The thing with CDMA is that it looks like
> it'll not be around for many years and I haven't seen any equipement for
> 4g, hspa, LTE, etc...
> For your information, I think I can get away
> with a 10K budget but I need to be absolutely sure It'll work beforehand.

A gps will cost you $50 assuming you have a computer already (if not
another $50 for a Raspberry Pi +power source and ethernet cable. )
It will also cost a bit of wiring up time.

I do agree that gps will almost certainly require an unalumized window, and in
Quebec, not a north facing window. (gps avoids the north).

Ie, you can try out the gps option with spare change. If that does not
work, then you can go to the more expensive CDMA option. I have no idea
how long Bell and Telus will continue their cdma offering. Rogers never
used it, and all the newer players do not use G2 type offerings.

>
> I'll talk with the facility owner and do
> some tests with handheld GPS so I can verify what kind of signal I can get
> (and where). I'm a bit worried because I dont have roof access whenver I
> want and I dont know how the antenna will behave with the snow we get
> around here (quebec city).

If it is on the roof, put the antenna onto
a pole on the roof high enough that it will not get covered. But on a
window, the chances are pretty slim that it will get covered in snow.

[Steve Kostecke]

On 2014-03-13, William Unruh <un...@invalid.ca> wrote:

> Ie, you can try out the gps option with spare change. If that does not
> work, then you can go to the more expensive CDMA option. I have no idea
> how long Bell and Telus will continue their cdma offering. Rogers never
> used it, and all the newer players do not use G2 type offerings.

If you had thoroughly read all of the material at the links I previously
posted you would know that CDMA is predicted to be available through
2020+ to support M2M (machine to machine) communications.

> [---=| TOFU protection by t-prot: 138 lines snipped |=---]

...

[Mauricio Tavares]

On Thu, Mar 13, 2014 at 11:48 AM, Jochen Bern <Joche...@linworks.de> wrote:
>> I work for a college and a lot of my students are complaining that
>> their cell phones are off by 4 min. My servers poll us.pool.ntp.gov.
>
> (I trust that that's ....org .)
>
>> The cell phones poll from time.gov.
>> If you look at the website for time.gov it does display the time
>> being off by 4 min.
>
> It's correct within wristwatch-and-eyeball accuracy for me right now.
>

Call me odd, but being off by 4 minutes does not seem to be bad.
I mean, even a default kerberos install would be cool with that.

>> Would be nice if pool.ntp.org would display it's time on the site
>> so I can see if it's actually an internal issue.
>
> *.pool.ntp.org entries are *pools* of NTP servers run by whoever owns a
> particular one, so technically, the pool doesn't *have* a single
> inherent time (available from all members equally) that the website
> could display. Identify the (single ... ?) *IP* your servers currently
> poll so that *that one* public server can be scrutinized.
>
> And prepare to use *several* pool servers instead of one, recommended
> precisely to avoid such a scenario.
>

I always thought the best thing to do is to have your own ntp
servers which then are used by your servers. And then those ntp
servers can query whatever pool you want to use.

> Regards,
> J. Bern
> --
> *NEU* - NEC IT-Infrastruktur-Produkte im <http://www.linworks-shop.de/>:
> Server--Storage--Virtualisierung--Management SW--Passion for Performance
> Jochen Bern, Systemingenieur --- LINworks GmbH <http://www.LINworks.de/>
> Postfach 100121, 64201 Darmstadt | Robert-Koch-Str. 9, 64331 Weiterstadt
> PGP (1024D/4096g) FP = D18B 41B1 16C0 11BA 7F8C DCF7 E1D5 FAF4 444E 1C27
> Tel. +49 6151 9067-231, Zentr. -0, Fax -299 - Amtsg. Darmstadt HRB 85202
> Unternehmenssitz Weiterstadt, Geschäftsführer Metin Dogan, Oliver Michel

[Brian Inglis]

On 2014-03-13 11:52, Olivier Drouin wrote:
> Thank you Greg, Jochen, William,
>
> Great answers.
>
> Diversity is really what I'm looking for and
> I dont really need microsecond accuracy.
> CDMA seemed to be the most 'convenient' way
> to add diversity without a need for a working WAN link and/or line of sight
> view to the sky.
> Also, I already know that I get good cell
> signal from inside the server room.
> The thing with CDMA is that it looks like
> it'll not be around for many years and I haven't seen any equipement for
> 4g, hspa, LTE, etc...
> For your information, I think I can get away
> with a 10K budget but I need to be absolutely sure It'll work beforehand.
>
> I'll talk with the facility owner and do
> some tests with handheld GPS so I can verify what kind of signal I can get
> (and where). I'm a bit worried because I dont have roof access whenver I
> want and I dont know how the antenna will behave with the snow we get
> around here (quebec city).
>
> Ill also verify if I can synthonize the CHU
> radio frequency inside the server room if I can get my hand on a scanner.

To test CHU reception you need a SW receiver to tune 3330, 7850, and 14670 kHz
- 1kHz tones of different durations are broadcast every second except second 29
of each minute, data from seconds 31 to 39 ending at each half second, and
voice is broadcast from seconds 50 to 59: see time signals at
http://www.nrc-cnrc.gc.ca/eng/services/time/short_wave.html and data format at
http://www.nrc-cnrc.gc.ca/eng/services/time/broadcast_codes.html.

To use CHU, you need a SW receiver, possibly an external antenna, Bell 103 300 bps
8-N-2 compatible modem, an async serial UART port that will read it, and a system
running NTP.

For testing GPS, you could use any smart device with GPS, such as recent iP* or
Android players, phones, or tablets, and a monitor app that shows satellite
counts and visibility.

To use GPS, you need a Sirf receiver module, like a Garmin 18x LVC, or uBlox NEO/
LEA-6T/M8F, ideally mounted inside a south facing roof space without steel,
or window with no buildings in the way, or else an external active GPS antenna
rated for -40C-+40C mounted above snow and obstruction level, hooked to an async
serial UART port, on a system running NTP. See David Taylor's various devices,
systems, and releases at http://www.satsignal.eu/ntp/.

There are also a number of vendors who will sell you commercial GPS gear with
antennas, and consultancies who will buy, install, and set it up for you.

All OEM timing receivers typically perform much better than hand held devices,
and the gold standard seems to be the Trimble Thunderbolt, available locally
from Novotech in Pointe Claire near Montreal.

--
Take care. Thanks, Brian Inglis

[E-Mail Sent to this address will be added to the BlackLists]

Olivier Drouin wrote:
> The thing with CDMA is that it looks like it'll not be
> around for many years and I haven't seen any equipement
> for 4g, hspa, LTE, etc...

IIRC, HSPA, UMB, UMTS, EVDO and EVDV are all CDMA based 3G?
LTE run with GSM or CDMA?

Can you get a modem with LTE, and neither GSM or CDMA?
I'm fairly certain you can get a WiMax Modem without either.


Verizon, Sprint, U.S. Cellular, Cricket, and MetroPCS all use CDMA

MetroPCS (T-Mobil CDMA), is supposed to be toast by 2016.

Cricket is getting borged by ATT
{Who knows what they will do with the CDMA, since ATT is GSM}.


--
E-Mail Sent to this address <Blac...@Anitech-Systems.com>
will be added to the BlackLists.

[Jan Ceuleers]

On 03/14/2014 08:12 AM, E-Mail Sent to this address will be added to the

BlackLists wrote:
> Olivier Drouin wrote:
>> The thing with CDMA is that it looks like it'll not be
>> around for many years and I haven't seen any equipement
>> for 4g, hspa, LTE, etc...
>
> IIRC, HSPA, UMB, UMTS, EVDO and EVDV are all CDMA based 3G?
> LTE run with GSM or CDMA?

Informative white paper here:

https://www.aventasinc.com/whitepapers/WP-Timing-Sync-LTE-SEC.pdf

Among other things, it compares the timing requirements of a number of
cellular technologies. Warning: it also tries to sell you PTP grand
master equipment.

[Paul]

On Thursday, March 13, 2014 1:52:51 PM UTC-4, Olivier Drouin wrote:

>Diversity is really what I'm looking for and
>I dont really need microsecond accuracy.

>I'll talk with the facility owner and do
>some tests with handheld GPS so I can verify what kind of signal I can get
>(and where).

I'm not sure what your goals and constraints are but ...

You should test with a high sensitivity timing GPS receiver (e.g. u-blox or
Resolution SMTx). I get acceptable performance in an inner office off a
machine room on the ground floor of three-story building. In my case
acceptable is 100s of microseconds. If you can sometimes get a GPS signal
in your machine room (interior or exterior antenna) then run that into a
GPS disciplined oscillator which should give 10-100microsec./day drift when
there's no GPS signal -- O(10ms/yr) given stable temperature.

The high sensitivity + timing constraints let you stay locked with just a
single satellite fix.

There are also some time transfer tricks you could do to periodically
discipline an oscillator or you could just switch to a
Rubidium/quasi-Cesium based free-running oscillator that you "manually"
discipline if needed with NTP (e.g. approximate time2 from peer stats).

Some of these things you can buy, some you would have to assemble from
functional modules.

There was a discussion in the last 12 months or so on the Time-Nuts list
about doing time transfer into (really) RF free areas (i.e. secure bunkers).