On 13.06.2016 17:31, adel.mamin.private@gmail.com wrote:
> In GPS L1C/A signal tracking, there is a condition of a false phase lock, which is possible in PLL. This source: http://tinyurl.com/hhc3v8b refers to multiples of 25 Hz being the typical error value. The context, in which the typical error value is mentioned, is, I assume, the FLL assisted PLL.
>
> From another source ("Global Positioning System: Theory and applications Volume I"): "Unfortunately, if transitioning to a PLL, the PLL would could also false lock at this (25Hz) frequency offset".
>
> I hope, I understand why there could be the false lock condition in FLL, if, for example, atan2 type of discriminator is used and the signal correlation time is 20ms. Then there could be more than one stable tracking point in FLL due to the fact that atan2 is a periodic function. The error could be, indeed, multiple of 25Hz. The periodicity of the atan2 discriminator is also described here http://tinyurl.com/zkj5wen.
>
> My question is: is there a genuine property of PLL that could result in multiple 25Hz frequency error, when tracking GPS L1C/A signal?
>
A bit more general point... Kaplan is a good text (the best GPS text in
my opinion). But it's not a communications textbook. It's a GPS
textbook. You will only learn the stuff from it that is relevant to
building a working GPS receiver.
There are other books with more in-depth coverage of communications,
which not restricted to a single specific protocol.
Gene
Reply by Evgeny Filatov●June 13, 20162016-06-13
On 13.06.2016 17:31, adel.mamin.private@gmail.com wrote:
(snip)
> My question is: is there a genuine property of PLL that could result in multiple 25Hz frequency error, when tracking GPS L1C/A signal?
Basically, yes. Because "With alternating data bits and even with random
data, the Costas loop can false lock to a data sideband such that the
frequency remains offset by half the data rate." See
http://sci-hub.cc/10.1109/TCOM.1977.1093790
for more discussion of a false lock in BPSK phase tracking.
Gene
> On Mon, 13 Jun 2016 07:31:31 -0700 (PDT), adel.mamin.private@gmail.com
> wrote:
>=20
> >In GPS L1C/A signal tracking, there is a condition of a false phase lock=
> >multiples of 25 Hz being the typical error value. The context, in which =
the=3D
> > typical error value is mentioned, is, I assume, the FLL assisted PLL.=
=3D20
> >
> >From another source ("Global Positioning System: Theory and applications=
Vo=3D
> >lume I"): "Unfortunately, if transitioning to a PLL, the PLL would could=
al=3D
> >so false lock at this (25Hz) frequency offset".
> >
> >I hope, I understand why there could be the false lock condition in FLL,=
if=3D
> >, for example, atan2 type of discriminator is used and the signal correl=
ati=3D
> >on time is 20ms. Then there could be more than one stable tracking point=
in=3D
> > FLL due to the fact that atan2 is a periodic function. The error could =
be,=3D
> > indeed, multiple of 25Hz. The periodicity of the atan2 discriminator is=
al=3D
> >so described here http://tinyurl.com/zkj5wen.
> >
> >My question is: is there a genuine property of PLL that could result in =
mul=3D
> >tiple 25Hz frequency error, when tracking GPS L1C/A signal?
>=20
> False lock is typically due to properties of the phase detector rather
> than the loop. You've described the basics of how this can happen,
> so I think you get it. In a sampled system it's always possible to
> lock on aliases, too, if the system permits it to happen.
>=20
> The link you provided went to a page that is unviewable by me, so I'm
> not sure of the details of the system you're looking at.
Your point of phase detector being a typical culprit of the false lock is h=
elpful.
Reply by Tauno Voipio●June 13, 20162016-06-13
On 13.6.16 22:24, adel.mamin.private@gmail.com wrote:
> Indeed, there is the 50bps ephemeris data which is modulated onto the ranging code. The ranging code is sent with 1.023MHz frequency. The ranging code is BPSK modulated onto 1575.42MHz. However, assuming the ephemeris bit stream is variable, the 50bps data should not impose a stable 25Hz spectral component. Or am I wrong?
>
It is enough that the transitions are at fixed 20 ms intervals,
though part of them is missing.
--
-TV
Reply by ●June 13, 20162016-06-13
Indeed, there is the 50bps ephemeris data which is modulated onto the ranging code. The ranging code is sent with 1.023MHz frequency. The ranging code is BPSK modulated onto 1575.42MHz. However, assuming the ephemeris bit stream is variable, the 50bps data should not impose a stable 25Hz spectral component. Or am I wrong?
Reply by Tauno Voipio●June 13, 20162016-06-13
On 13.6.16 21:05, Tim Wescott wrote:
> On Mon, 13 Jun 2016 07:31:31 -0700, adel.mamin.private wrote:
>
>> In GPS L1C/A signal tracking, there is a condition of a false phase
>> lock, which is possible in PLL. This source: http://tinyurl.com/hhc3v8b
>> refers to multiples of 25 Hz being the typical error value. The context,
>> in which the typical error value is mentioned, is, I assume, the FLL
>> assisted PLL.
>>
>> From another source ("Global Positioning System: Theory and applications
>> Volume I"): "Unfortunately, if transitioning to a PLL, the PLL would
>> could also false lock at this (25Hz) frequency offset".
>>
>> I hope, I understand why there could be the false lock condition in FLL,
>> if, for example, atan2 type of discriminator is used and the signal
>> correlation time is 20ms. Then there could be more than one stable
>> tracking point in FLL due to the fact that atan2 is a periodic function.
>> The error could be, indeed, multiple of 25Hz. The periodicity of the
>> atan2 discriminator is also described here http://tinyurl.com/zkj5wen.
>>
>> My question is: is there a genuine property of PLL that could result in
>> multiple 25Hz frequency error, when tracking GPS L1C/A signal?
>
> Yup, it's behind a paywall.
>
> I'd suggest looking for aliases -- is there some 25 or 50Hz sampling
> process in there, or some pair of frequencies with a difference of 25Hz?
> Surely the author explains _why_ the problem.
>
> If you have enough money to throw at the problem, lock at the apparent
> good frequency and at 25Hz off, and see which is better.
The ephemeris data with supplements id modulated at 50 bit/s BPSK.
--
-TV
Reply by Tim Wescott●June 13, 20162016-06-13
On Mon, 13 Jun 2016 07:31:31 -0700, adel.mamin.private wrote:
> In GPS L1C/A signal tracking, there is a condition of a false phase
> lock, which is possible in PLL. This source: http://tinyurl.com/hhc3v8b
> refers to multiples of 25 Hz being the typical error value. The context,
> in which the typical error value is mentioned, is, I assume, the FLL
> assisted PLL.
>
> From another source ("Global Positioning System: Theory and applications
> Volume I"): "Unfortunately, if transitioning to a PLL, the PLL would
> could also false lock at this (25Hz) frequency offset".
>
> I hope, I understand why there could be the false lock condition in FLL,
> if, for example, atan2 type of discriminator is used and the signal
> correlation time is 20ms. Then there could be more than one stable
> tracking point in FLL due to the fact that atan2 is a periodic function.
> The error could be, indeed, multiple of 25Hz. The periodicity of the
> atan2 discriminator is also described here http://tinyurl.com/zkj5wen.
>
> My question is: is there a genuine property of PLL that could result in
> multiple 25Hz frequency error, when tracking GPS L1C/A signal?
Yup, it's behind a paywall.
I'd suggest looking for aliases -- is there some 25 or 50Hz sampling
process in there, or some pair of frequencies with a difference of 25Hz?
Surely the author explains _why_ the problem.
If you have enough money to throw at the problem, lock at the apparent
good frequency and at 25Hz off, and see which is better.
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
I'm looking for work -- see my website!
Reply by Eric Jacobsen●June 13, 20162016-06-13
On Mon, 13 Jun 2016 07:31:31 -0700 (PDT), adel.mamin.private@gmail.com
wrote:
>In GPS L1C/A signal tracking, there is a condition of a false phase lock, w=
>hich is possible in PLL. This source: http://tinyurl.com/hhc3v8b refers to =
>multiples of 25 Hz being the typical error value. The context, in which the=
> typical error value is mentioned, is, I assume, the FLL assisted PLL.=20
>
>From another source ("Global Positioning System: Theory and applications Vo=
>lume I"): "Unfortunately, if transitioning to a PLL, the PLL would could al=
>so false lock at this (25Hz) frequency offset".
>
>I hope, I understand why there could be the false lock condition in FLL, if=
>, for example, atan2 type of discriminator is used and the signal correlati=
>on time is 20ms. Then there could be more than one stable tracking point in=
> FLL due to the fact that atan2 is a periodic function. The error could be,=
> indeed, multiple of 25Hz. The periodicity of the atan2 discriminator is al=
>so described here http://tinyurl.com/zkj5wen.
>
>My question is: is there a genuine property of PLL that could result in mul=
>tiple 25Hz frequency error, when tracking GPS L1C/A signal?
False lock is typically due to properties of the phase detector rather
than the loop. You've described the basics of how this can happen,
so I think you get it. In a sampled system it's always possible to
lock on aliases, too, if the system permits it to happen.
The link you provided went to a page that is unviewable by me, so I'm
not sure of the details of the system you're looking at.
Reply by ●June 13, 20162016-06-13
In GPS L1C/A signal tracking, there is a condition of a false phase lock, w=
hich is possible in PLL. This source: http://tinyurl.com/hhc3v8b refers to =
multiples of 25 Hz being the typical error value. The context, in which the=
typical error value is mentioned, is, I assume, the FLL assisted PLL.=20
From another source ("Global Positioning System: Theory and applications Vo=
lume I"): "Unfortunately, if transitioning to a PLL, the PLL would could al=
so false lock at this (25Hz) frequency offset".
I hope, I understand why there could be the false lock condition in FLL, if=
, for example, atan2 type of discriminator is used and the signal correlati=
on time is 20ms. Then there could be more than one stable tracking point in=
FLL due to the fact that atan2 is a periodic function. The error could be,=
indeed, multiple of 25Hz. The periodicity of the atan2 discriminator is al=
so described here http://tinyurl.com/zkj5wen.
My question is: is there a genuine property of PLL that could result in mul=
tiple 25Hz frequency error, when tracking GPS L1C/A signal?