Re: [time-nuts] Common sky pps errors for any GPSDOs?
Hi Matt,
having 140ps matching of the 1PPS between units is the equivalent of knowing
your antenna position to within ~0.14 feet total error max.
Thats less than one inch error per antenna!
That would require some serious antenna surveying :) This accuracy is
impossible to achieve with timing GPS receivers without
carrier-phase/post-processing as far as I know.
If you are only looking at second-to-second jitter of less than 140ps then
we can do this without a problem.
Our ADEV at 1s measurement intervalls on the Fury OCXO is about 2E-012, or
2ps. So the pulse to pulse jitter would be much more than an order of
magnitude better than you require.
On the Fury GPSDO one could get better than ~5ns unit to unit variation rms
with proper antenna surveying, and calibration of the 1PPS output in 1ns
steps using the 1PPS shift commands.
bye,
Said
In a message dated 1/5/2009 18:22:01 Pacific Standard Time,
boyscout@gmail.com writes:
But Said just told us that he sees about 25ns difference between
units, which is about 180 times worse than 140 ps.
In VLBI I think that they can deal with an unknown phase error between
the two LOs as long as that phase error remains relatively constant
over the period of integration. That is why they care about ADEV.
They just correlate the two signals, and subtract out that constant
phase error. My problem is that we can't just do that in this
application. We need to know that phase error a priori.
Matt
On 1/5/09 9:27 PM, "SAIDJACK@aol.com" SAIDJACK@aol.com wrote:
Hi Matt,
having 140ps matching of the 1PPS between units is the equivalent of knowing
your antenna position to within ~0.14 feet total error max.
Thats less than one inch error per antenna!
That would require some serious antenna surveying :) This accuracy is
impossible to achieve with timing GPS receivers without
carrier-phase/post-processing as far as I know.
On the other hand, if the antennas aren't moving, then you could spend
arbitrarily long surveying the position.
Or, if you have stable phase relationships (not necessarily known, but
stable), you can do some sort of self survey using am arbitrary receive
signal like Bernie Steinberg, et al.,'s work with the Radio Camera at Valley
Forge Research Center.
There's also been some work with uplink arraying in the Deep Space Network
(but, of course, they have masers and optical fibers, etc.)
On Mon, Jan 5, 2009 at 9:27 PM, SAIDJACK@aol.com wrote:
Hi Matt,
having 140ps matching of the 1PPS between units is the equivalent of knowing
your antenna position to within ~0.14 feet total error max.
Thats less than one inch error per antenna!
That makes it sound a lot more difficult than it really is. The vast
majority of the error in GPS is systematic, such that two GPS systems
with antennas near each other should have highly correlated error.
This is the basis of differential GPS. It doesn't matter if the
absolute error is hundreds of feet, as long as both devices have the
same error.
I spent a couple of years nearly a decade ago doing differential GPS
for steering heavy equipment. You can get sub-centimeter errors over
baselines in the tens of km. Again, this is relative error.
Matt
Hi Matt,
On Mon, 2009-01-05 at 22:03 -0800, Matt Ettus wrote:
On Mon, Jan 5, 2009 at 9:27 PM, SAIDJACK@aol.com wrote:
Hi Matt,
having 140ps matching of the 1PPS between units is the equivalent of knowing
your antenna position to within ~0.14 feet total error max.
Thats less than one inch error per antenna!
That makes it sound a lot more difficult than it really is. The vast
majority of the error in GPS is systematic, such that two GPS systems
with antennas near each other should have highly correlated error.
This is the basis of differential GPS. It doesn't matter if the
absolute error is hundreds of feet, as long as both devices have the
same error.
One idea with a "traditional" GPSDO... I assume the sites have a real
time communication link, Internet or Radio.
Make sure the sites use exactly the same GPS SVs. This can be forced by
telling the GPS to exclude SVs you do not want to use. Either you go
with 1-SV timing mode and manually tell the receivers which SV to use.
Or you juggle 4-6 SVs that are high on the sky on all sites.
This will help the GPSDO to measure against the same satellites which
gives you kind of a differential effect.
I spent a couple of years nearly a decade ago doing differential GPS
for steering heavy equipment. You can get sub-centimeter errors over
baselines in the tens of km. Again, this is relative error.
Sub 10cm, and even sub 5cm over distances up to 40km is possible with
dual freq GPS receivers in real time.
There are some reasonably priced dual freq GPS that can be driven by
your external 5 or 10 MHz OCXO. The Novatel OEMV-2, beeing one
candidate
http://www.novatel.com/Documents/Papers/OEMV2.pdf
Make one site the "base" and the others "rovers". Pass RTK-corrections
from your base to the rovers. You should get a very reasonable position
errors, especially since your sites are stationary and you can let the
RTK algorithms get plenty of time to converge. Then look at the
receivers estimate of local oscillator drift for your frequency
measurement. This is a fairly simple integration of COTS modules. Since
the rovers are not taking advantage of beeing at a surveyed stationary
position this is not the optimal solution.
Someone on the list with the equipment and time to set this system up?
Tvb, have you done this?
Matt
--
Björn