[time-nuts] Positional accuracy of the M12+T

[Magnus Danielson]

From: "Randy Warner" <Randy at synergy-gps.com>
Subject: Re: [time-nuts] Positional accuracy of the M12+T
Date: Thu, 4 Jan 2007 09:59:48 -0800
Message-ID: <D6D8D27B5219D649A005529978F791371CCE62 at SBSERVER.Synergy-GPS.sbs>

Randy,

> Good points. I think that a lot of people are unaware of the diurnal
> shifts that occur due to atmospherics. These can be many 10's of
> nanoseconds compared to UTC. This is true for every receiver I have ever
> worked with. The ionospheric correction algorithms are good, but they
> are not perfect.

This confirms my suspicion that nobody is doing the single frequency
ionospheric measurement that I have been trying to hint about. The research was
done by Robin Giffard at Agilent, published at the 31st PTTI meeting in 1999
(page 405-418): http://tycho.usno.navy.mil/ptti/ptti99/PTTI_1999_405.PDF
The work then was evolved into the patent US 6.407.700.

Do read. While the PTTI paper is a research paper not giving very straight-
forward methods for inclusion into a practical receiver (but certainly all the
hints for "somebody so skilled in the art") the patent is more practically
oriented. The PTTI paper is also interesting since he used a Motorola Oncore VP
which was modified so that it's 19.09575 MHz TCXO was PLL locked to the 1 MHz
output of a HP 5061B in a 3590/188 relation in order to improve the phase
stability of the oscillator and in particular the ADEV(1s). This to be able to
ensure the measurement of phase slips. In there he extracted the raw code and
carrier phase measures every second and logged it for later analysis.

> Can't wait for a civilian L2..................... 

There is three operational sats up there now. You only need to invest in a dual
freq receiver. In future we might actually see L2/L5 receivers with no L1 since
that might actually be more economic. L2 and L5 is only 51.15 MHz apart and
their full footprint is covered in 61.38 MHz, which means a single common
frequency antenna and RF path. Using a single A/D at sufficient rate will allow
for the sampling to be say 66 MHz or so. In the meanwhile some of the costs for
a L1 C/A and L2C reciever is really common to that of a full L1/L2 receiver
even if the chipping rate (and thus necessary bandwidth) is lower. However,
with tighter E/L detector distance and wider bandwidth in the receiver, the
multipath suppression for a normal C/A receiver is greatly improved, for a
certain range of multipath even slightly better than the P(y) code. Also, if
you care that much to go L2C you also need a choke ring, and you then need a
dual frequency choke ring. The benefit of L2C/L5 is that you could have a
single frequency choke ring, with about 5% bandwidth rather than about 1%
bandwidth. Should be an interesting but rewarding challenge. The ICD is already
prepared for such a solution if you look carefully. You could also envision a
three-frequency system of L1/L2C/L5 (which would effectively be a 2 frequency
RF solution) for the high-end market. On top of that, you would still be able
to do the P(Y) tricks with Z-tracking or whatever algorithm you choose. You
could also do similar tricks to the M-code if you dare, but the benefits would
be less than the P(Y) locking with all the spreaded civial signals around.

Cheers,
Magnus