My inclination is that one would need three things to participate in such a system---

  1. A GPS receiver. 
  2. A local reference oscillator. 
  3. A way to measure their relative phase.

I understand that the Thunderbolt can perform all of these functions, so expect that this would be the favoured option for those without an Rb oscillator.   I do not have one, so will move on to the next setup.  For development the ResT could work the same way, but it's oscillator isn't as good. So, we consider the next option.  

For those without a receiver that has itself a high-quality oscillator, the next option would be to use a local reference and build a separate phase-measurement circuit. I have already written a time-tagging system with an FPGA, which could be reworked into what we need. This is probably overboard, though, and might be achieved with a micro, possibly paired with a PICTIC for the final interpolation.

The next step would be to either transfer directly or to upload the phase data to a central location where it could be processed together to estimate  everyone's offset to the average at once.

A DMTD will be needed somewhere to evaluate performance, but this needs only to be done somewhere rather than everywhere. 

Since there has been some interest I'll start whacking some simulations together in the near future.

Thanks, 
Lachlan Gunn

-------- Original message --------
From: Mark Spencer mspencer12345@yahoo.ca
Date:
To: time-nuts@febo.com
Subject: Re: [time-nuts] Common-View GPS Network

I'd be interested in understanding what it would take to participate in this type of experiment.   I'm unlikely to have any free time for this over the next several months but if for example there was a particular GPS receiver that was well suited for this type of work that would be useful info for me.  
(Recently at work I've started to become involved with precision GPS location systems and the claimed 3D accuracies that can be achieved using post processing seem very impressive.)
Thanks in advance for any comments.
 
 
Regards
Mark Spencer

Message: 6
Date: Mon, 15 Apr 2013 22:49:38 +0200
From: Magnus Danielson magnus@rubidium.dyndns.org
To: time-nuts@febo.com
Subject: Re: [time-nuts] Common-View GPS Network
Message-ID: 516C67E2.80501@rubidium.dyndns.org
Content-Type: text/plain; charset=ISO-8859-1; format=flowed

Lachlan,

This would be fun to try with a few sites involved. It requires a fairly
decent GPS and antenna at each location, and loging and then
post-processing. Would be fun to see how well it can be made to perform.

Cheers,
Magnus

On 04/15/2013 02:55 PM, Lachlan Gunn wrote:

time-nuts mailing list
time-nuts@febo.com
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts

End of time-nuts Digest, Vol 105, Issue 46

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.

Hi

Unfortunately there's more to it than that. Unless both TBolts are looking at the same sats, they will get very different answers. Ideally you want to compare single sats, and have very precise station locations. That lets you eliminate a lot of geometric stuff and individual sat errors. Fancy sat orbit data also helps this.

The measurement stuff is pretty easy. Anything that will measure 100 to 200 ps is likely to be plenty good enough. There are a number of counters and boards that will do that.

Then you get down to ionosphere stuff, and fiddly crustal tide shifts. There are lots of layers to this onion.

Bob

On Apr 15, 2013, at 8:27 PM, Lachlan Gunn lachlan@twopif.net wrote:

I too would be interested. The topic of common view GPS comes up every now and then. The incentive to use common view time transfer (CVTT) is reduced when plain old time transfer (TT) is so good already. Still, it's worth a try.

I'd be curious what level of improvement is possible. It will depend on the receiver and the antenna. I believe the NIST project uses fancy antennas but normal M12 receivers. So there's hope for the amateur. I don't think the TIC is that important: the sawtooth granularity is 1 ns so precision greater than that is wasted. One averages over many minutes anyway.

Simulations would be a great idea. If you need raw data let me know.

Google for CGGTTS and RINEX. Not all GPS receivers provide the level of per-satellite information required for common view. I know the M12 does, as well as high-end specialty timing receivers. I don't know if the Trimble TBolt or Res-t give out enough information to generate RINEX. But you can get the individual SV contribution to the solution using TBolt packet 0x8F-A7.

/tvb

----- Original Message -----
From: "Lachlan Gunn" lachlan@twopif.net
To: time-nuts@febo.com
Sent: Monday, April 15, 2013 5:27 PM
Subject: Re: [time-nuts] Common-View GPS Network

On Mon, Apr 15, 2013 at 8:23 PM, Tom Van Baak tvb@leapsecond.com wrote:

I'd be curious what level of improvement is possible. It will depend on the

The M12 is certainly affordable, something like $60. but what is a "fancy
antenna"?  How are they different from a normal timing antena?

--

Chris Albertson
Redondo Beach, California

On 4/15/13 8:36 PM, Chris Albertson wrote:

Probably a choke ring? With small variation in phase center over viewing
angles.  The "secret sauce" is in the design of the crossed dipole
elements which are not flat, nor are they constant width.

http://facility.unavco.org/kb/questions/311/Choke+Ring+Antenna+Calibrations
has dimensions

fins are 50-63 mm tall, 25 mm apart, etc.

I suspect that this is not a high precision sort of thing.. (that is, if
your fin spacing isn't exactly 25.1 mm it doesn't make much difference)

L1 is 1575 MHz (approx) or a wavelength of 19 cm.  So the fins are 0.315
to 0.26 lambda high and 0.13 lambda apart.  If you were making a fancy
corrugated horn, you make the fins range from 1/2 lambda to 1/4 lambda
so you get a nice transition from shorted to open, but that's not what
they're doing here.

Google for: NIST SIM GPS common view pinwheel

/tvb

On 4/15/13 9:27 PM, Tom Van Baak wrote:

described in one of the NIST reports as an aperture coupled slot fed
array that is better than a patch, but not as large and heavy as a choke
ring.

W. Kunysz, 2000, “High Performance GPS Pinwheel Antenna,” in
Proceedings of the 2000 International Technical Meeting of the Satellite
Division of the Institute of Navigation (ION GPS 2000), 19-22 September
2000, Salt Lake City, Utah, USA (ION, Alexandria, Virginia), pp. 2506-251

http://www.novatel.com/assets/Documents/Papers/GPS-704xWhitePaper.pdf
Patented by Novatel & Pinwheel is a trademark

Performance is almost as good as a choke ring but a heck of a lot
smaller and lighter.

On 4/15/13 10:22 PM, Jim Lux wrote:

of course, cake pans with 2-2.5 inch high walls are readily available.

There's a Wilton cakepan set with 6",8" and 10" diameter pans with 3"
walls.. hmm, an inch between fins..

Oddly, the package shipping size is 12x12x2"... I wonder how they fit a
3" high pan in a 2" thick box..

a real restaurant/pastry supply has a mindboggling variety of pans

http://www.fantes.com/cake-pans-round.html

every integer inch diameter from 4" to 18" and ditto for heights from 2"
to 4"...

People like those machined or cast choke rings because they're easier to
fabricate: Slap a block of aluminum in the lathe or milling machine,
push GO on the CNC, and stand back.

Or for those with a taste for hot metal.. you could cast it with scrap
aluminum you've melted in the forge in your time nuts lab..  Turn those
empty beer cans into something useful.

If you have a fancy multiaxis mill, you could probably do one of those
porcupine looking things.

Or, if you have a swimming pool or pond, and some sheet aluminum, and
some suitable high explosives.. hydroforming is your friend.

If you want true timenuts.. do the explosive hydroforming without a
mold/buck, and instead use precision timing of shaped charges.  Finally,
a use for those krytron switches you found at the surplus place.

On 04/16/2013 05:23 AM, Tom Van Baak wrote:

They use the Novatel 700 pin-wheel antenna.

Setting up my M12M-T for this would be fun. That should be attainable
for most folks around here.

I wonder if one should only measure the PPS thought. Looking directly at
the clock could help to separate the clock drift and the time, even if
you get sufficient clues from the PPS and sawtooth correction.

Cheers,
Magnus

I've never tried that. Start with checking the ADEV of the LO. I always assumed the 1PPS was pre-compensated for instantaneous LO offset and rate, as calculated during the previous second(s). If so and ADEV(1 s) is less than 1e-9 then the 1PPS + sawtooth will have sufficient accuracy for your needs. Message @@Ha contains the oscillator and clock parameters.

As you know professional receivers bypass this issue because they take an external LO (e.g., Rb, Cs, HM).

When you look at the actual clock solutions (which are in the @@Hn message) you will be surprised at the variance. At the per-SV per-second level it is not uncommon to see errors of many ns, even tens of ns. But when you average up to 12 SV together every second for 600 seconds then you start to see something trustable. You can see first hand why GPSDO need long time-constants.

Here's a one second sample of a single Hn message:

@@Hn: sigma 36 sawtooth 10 0 0 00000000 mean frac 861564.000 (0.000)
ch  0 sv  6 frac 0.000861576 s = mean +  12 ns
ch  1 sv  5 frac 0.000861586 s = mean +  22 ns
ch  2 sv 10 frac 0.000861581 s = mean +  17 ns
ch  3 sv 19 frac 0.000861573 s = mean +    9 ns
ch  5 sv 13 frac 0.000861571 s = mean +    7 ns
ch  6 sv 26 frac 0.000019831
ch  7 sv 28 frac 0.000861555 s = mean +  -8 ns
ch  8 sv  7 frac 0.000861548 s = mean +  -15 ns
ch  9 sv  3 frac 0.000861571 s = mean +    7 ns
ch 10 sv  8 frac 0.000861546 s = mean +  -17 ns

/tvb

In message A687BF7F4A1642E8BA7EDDA7432A3969@pc52, "Tom Van Baak" writes:

A lot of that variance is because the position-hold coords are wrong.

I tried using the @@Hn data to "sneak" up on the right coords and got
some pretty good results, but the process too forever (as in: Months)

See:
http://phk.freebsd.dk/raga/sneak/

The improvement in the finished timing solution from the oncore is
quite marginel because on average you have satellites on all
sides of your antenna and the errors mostly cancel out.

The notable exception to that is where I live: at 56N.

56N is at the top of the GPS orbits, so satellites never venture
north of me, and I'm not sufficient north to have any benefits from
the satellites which rise above Canada/Alask on the other side of
the north pole.

--
Poul-Henning Kamp      | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG        | TCP/IP since RFC 956
FreeBSD committer      | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.

Hi Poul:

Once you know the rise and set times (where the signal is stable) in sidereal time for each SV# you can simply
enable/disable them based on time.
I think this may be a better approach than having a mask angle based on each degree of azimuth.
Keeping track of a rise and set time for each SV# is an order of magnitude less data than a 1 deg elevation mask.

Remember that each GPS satellite repeats it's ground track, that's to say that from a fixed antenna each satellite will
always follow the same path in the sky.
This was how/why the GPS orbit altitude was chosen.
That's why I asked for a Sidereal time plot for the Thunderbolt. On a standard time position  plot the satellites appear
to change paths.

Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html

Poul-Henning Kamp wrote:

Hi

Most modern receivers give you corrections in ps rather than ns. That's not to say they are good to 1 ps. It's not uncommon to see adev (after doing the correction) running sub 1 ns at one second when compared to a 5071.

Most people seem to do this in runs many minutes long. That's a bit away from 1 second, so other things do come into the picture.

Bob

On Apr 16, 2013, at 3:42 AM, "Tom Van Baak" tvb@LeapSecond.com wrote:

Hi Tom,

On 04/16/2013 09:42 AM, Tom Van Baak wrote:

Tom, you scare me some times, this is one of them... or the time actually.

Well, it is... but it also solves the position-time each second and that
takes over. If you loose signal, that's when the predicted PPSes will
shine through.

True. That get's you far. Just looking at that and bring it into your
model is also something you can do. It's interesting to notice that the
@@Ha message has a time (to nanosecond), and then clock bias and
oscillator offset and on top of that a delta correction in the @@Hn
message. Would be nice to know exactly how them all fitted together.

BTW. I hooked up my M12M-T to a Novatel 700 pinwheel just to check it
out again.

Oh yes.

The TCXO sitting there looks like a RAKON thing, I haven't looked at the
details, but maybe it can be steered? It's interesting that the TCXO
temperature can be read out.

The urge to get into the dirty details got higher by that. Curious at
seeing a log of data.

Let's see if my M12M-T locks up in the current quick-hack setup.

Cheers,
Magnus

On 04/16/2013 09:55 AM, Poul-Henning Kamp wrote:

Did you look at the pseudo-range correction data as an alternative approach?

Cheers,
Magnus

On 4/16/2013 1:55 AM, Jim Lux wrote:

Thanks for that. The last bit of your post was really cute. I needed a
good laugh :)

I haven't posted much in a while, partly because I've been kinda bummed
out by this list since I got the news about shera... There have been so
many constant reminders of his passing and whatnot (multiple thread
titles mentioning his legacy)

I just have to ask though... cake pans? really? I can't imagine it would
even be possible to modify a cake pan with enough accuracy to get a
usable antenna.

-- Sarah

On 4/16/2013 9:53 AM, Achim Vollhardt wrote:

I'm not sure what all I'll need to participate, but I'd like to
volunteer my thunderbolt to this sort of network as well.

I'm wondering though, is there something inherently different about this
proposed network than a network of continually operating reference stations?

or are CORS stations for something else?

--Sarah

On 4/16/13 5:19 PM, Sarah White wrote:

Sure.. cake pans, like other stamped goods, are actually pretty high
precision, because they're all stamped out of the same die.  As long as
the dimensions are right (and for choke rings that's not real critical),
it works.

Magnus,

I remember why I didn't measure the M12 oscillators directly -- it's a real challenge to get at the signal on the back side of the PCB and to measure it without loading the crystal. It's not like just connecting a wire to a 50R Timepod input. Here are photos of the oscillator on three different versions of the M12 board:

http://leapsecond.com/pages/m12/m12-osc.htm

Perhaps with some 'scope tracing you can find a buffered clock output on one of the ASIC pins. That way the same probing technique could be used on all 3 boards.

/tvb