Tom,
On 04/08/2016 02:06 AM, Tom Van Baak wrote:
Hi Bjorn,
For navigation more measurements have always been prefered - that
is use as many GNSS systems as all your receicers support.
I would like to believe this. There's a common myth with clocks that more is
always better. For example, if you have 4 cesium then adding 5 more gives you
an improvement of sqrt(4) to sqrt(9) or 50% better. And similarly, if you go
from receiving 4 SV to 9 SV your position and timing fixes will get 50% better.
But it seems to me that "more is better" only works when each unit has similar
accuracy and stability to begin with. If you have 4 cesium and add 5 quartz you do not
get better performance. Instead the quartz will strongly degrade your net result.
Your analogy is broken in several ways. More clocks helps, but for
optimal performance you weight their contribution. If you add 5 crystal
oscillators against 4 cesiums, a good weigthing mechanism will give them
weight for the ensemble goal (short-term or long-term).
Additional GNSS signals help, but only if they do not compete out
tracking other satellites. If you support say 12 GPS satellites and in
addition can handle 12 GLONASS and 12 Galileo, then with proper weight
you can get a better performance with more. You can get more assistance
with tricky geometry, you can get away with some biases and noise.
Someday when GLONASS and Galileo and BeiDou match GPS in accuracy at the cm and
sub-ns level your claim that more is always better will be true. From what I've
read we're not at that point yet. If you can find some papers to the contrary
please let me know. We're in a very exciting decade or two of GNSS evolution
and coordination.
I think one has to be careful, and that goes for both of you. A
particular platform may perform worse in GNSS mode than pure GPS mode,
but this is not to say that using the other GNSS systems makes GPS
position bad in general, it only says that for this platform it doesn't
give the improvement we thought, in this release.
Meanwhile it would not surprise me if each GNSS system gives a slightly
different position and a slightly different time than GPS does.
They will, but it's for many reasons. It doesn't even say that GPS will
give the *right* position, as we know that that is a problematic issue
on its own. It will be different, and vary somewhat different, but there
is nothing strange about it, it's a different system, with different
signals, orbits, constellation size etc. The FDMA properties of GLONASS
is troublesome.
If someone has a couple of LES-M8T to spare, would you configure each one to a
single GNSS and then post a one day or one month record of 3D position and
timing amongst all 4 systems? That would replace all our mutual conjecture with
actual hard numbers.
Testing is indeed a good idea. Remember that it will not be saying more
than what the current release is able to make out of it, for the platform.
Cheers,
Magnus
/tvb
----- Original Message -----
From: "Björn" <[email protected]>
To: "Discussion of precise time and frequency measurement" <[email protected]>
Sent: Thursday, April 07, 2016 7:43 AM
Subject: Re: [time-nuts] LEA-M8T
For navigation more measurements have always been prefered - that is use as
many GNSS systems as all your receicers support.
That should be true also for common view timing.
--
Björn
<div>-------- Originalmeddelande --------</div><div>Från: Bob Camp <[email protected]> </div><div>Datum:2016-04-07 18:41
(GMT+07:00) </div><div>Till: Bob Stewart <[email protected]>, Discussion of precise time and frequency measurement <[email protected]>
</div><div>Rubrik: Re: [time-nuts] LEA-M8T </div><div>
</div>Hi
Indeed, if you have not turned off the other systems for timing, you will have
issues. Even for
precision navigation, you need to turn them off. Until the European system goes
up, there will
not ba a coordinated approach between any two of the systems. Right now they
each make their
own assumptions and their own definitions. If you are driving a car down the
road, that’s not a
big deal. If you are trying to do TimeNuts stuff … A one meter delta is a big
deal for timing. A “few”
nanoseconds (say >10) is also a big deal.
Bob
On Apr 6, 2016, at 9:39 PM, Bob Stewart <[email protected]> wrote:
Hi Logan,
I seem to remember Bob Camp mentioning that you can't have multiple satellite
sources in the mix, because the other satellites are inferior to the GPS sats
in timing. Maybe Bob or someone could address this. I would love to discover
that I've set something wrong in all the many, many data structures.
Bob
--------------------------------------------
On Wed, 4/6/16, Logan Cummings <[email protected]> wrote:
Subject: Re: [time-nuts] LEA-M8T
To: "Bob Stewart" <[email protected]>, "Discussion of precise time and frequency
measurement" <[email protected]>
Date: Wednesday, April 6, 2016, 8:18 PM
Hi
Bob,
Can't speak to
jitter accuracy but the M8 series is definitely not the same
receiver in the 6 series. As you probably know, M8
introduced multi-GNSS support so in addition to GPS you have
Beidou and Glonass satellites.
At work we've had some gnashing of
teeth about the wider filter passband requirements for
multi-GNSS support since we're operating in a noisy
environment, but I have nothing further on degraded
performance when using only GPS.
Would be interesting to let
it have all the constellations and see what
happens.
-Logan
On Wed, Apr 6, 2016 at
10:04 AM, Bob Stewart <[email protected]>
wrote:
I
recently bought a number of LEA-M8T receivers and I have to
say that I am unimpressed, so far. They don't survey
to the same reported accuracy as the LEA-6T in the same
amount of time. They certainly aren't better in the
jitter after sawtooth correction. So, have I managed to
overlook some new field, or are they just not the same
receiver as the 6T? I did shut all sats off except GPS
sats.
Bob - AE6RV
_______________________________________________
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
