Many of us use F connectors and 75 ohm CATV RG-6 coax for GPS antennas when
both antenna and receiver are specified for 50 ohms.

Don't sweat the difference between 75 and 50 ohms.

Tim N3QE

On Fri, Jun 5, 2020 at 7:22 PM Robert DiRosario ka3zyx@comcast.net
wrote:

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Dana...

I think that you are neglecting two important mitigating factors.

1. the cable loss at 1575MHz, even for a 25' run of RG-6, reduces those
   reflections quite a lot from one end to the other.    It amounts to 2 - 3
   dB in 25', depending on cable quality.

2. a 1.5:1 SWR is not a very big reflection to begin with, on the order of
   20% of the incident power, about  7 dB. I am rounding a lot here just to
   keep the math easy...for me.

By the time a reflection has made the round trip from the receiver back to
the antenna and them back to the receiver, which is how the delay would
have
to manifest itself, it will be down at least 15 dB from its original self,
and probably more. Given the coding of GPS signals which allows several
satellites to share a common frequency band, that is not going to be much
of
a problem. And if only one end of the path actually is 75 ohms, then there
won't be a delayed signal.

Tom Holmes, N8ZM

-----Original Message-----
From: time-nuts time-nuts-bounces@lists.febo.com On Behalf Of Dana
Whitlow
Sent: Friday, June 05, 2020 9:01 PM
To: Taka Kamiya tkamiya9@yahoo.com; Discussion of precise time and
frequency measurement time-nuts@lists.febo.com
Subject: Re: [time-nuts] ThunderBolt question

I'd like to point out that mismatches at the ends of an antenna cable can
cause trouble.  When both ends are mismatched, each bit of detail in the
signal
gets partially reflected back and forth, each time delayed by the round
trip propagation
delay in the cable, and so you have something like multipath going on.
Fortunately the
successive reflections get weaker with time, generally quite rapidly.
Since many
GPS users seem very concerned about multipath resulting from poor antenna
placement,
I think this factor should be considered as well and not just get swept
under the rug.

The amplitude of the "multipath" resulting from cable mismatches depends on
the product
of the voltage reflection coefficients at the two ends of the cable.  If
either end is perfectly
matched, then the quality of the match at the other end is not significant
vis-a-vis apparent
multipath problems and only affects transmission loss.

But when there is a mismatch on both ends, then the length of the cable
comes into play
as well.  A longer cable means more delay between successive reflections,
which is just
like multipath involving longer delays between the direct and the reflected
signals.

Cheers,

Dana      (K8YUM)

On Fri, Jun 5, 2020 at 7:13 PM Taka Kamiya via time-nuts <
time-nuts@lists.febo.com> wrote:

have

makes

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Dana...

I think that you are neglecting two important mitigating factors.

1. the cable loss at 1575MHz, even for a 25' run of RG-6, reduces those
   reflections quite a lot from one end to the other.    It amounts to 2 - 3
   dB in 25', depending on cable quality.

2. a 1.5:1 SWR is not a very big reflection to begin with, on the order of
   20% of the incident power, about  7 dB. I am rounding a lot here just to
   keep the math easy...for me.

By the time a reflection has made the round trip from the receiver back to
the antenna and them back to the receiver, which is how the delay would
have
to manifest itself, it will be down at least 15 dB from its original self,
and probably more. Given the coding of GPS signals which allows several
satellites to share a common frequency band, that is not going to be much
of
a problem. And if only one end of the path actually is 75 ohms, then there
won't be a delayed signal.

Tom Holmes, N8ZM

-----Original Message-----
From: time-nuts time-nuts-bounces@lists.febo.com On Behalf Of Dana
Whitlow
Sent: Friday, June 05, 2020 9:01 PM
To: Taka Kamiya tkamiya9@yahoo.com; Discussion of precise time and
frequency measurement time-nuts@lists.febo.com
Subject: Re: [time-nuts] ThunderBolt question

I'd like to point out that mismatches at the ends of an antenna cable can
cause trouble.  When both ends are mismatched, each bit of detail in the
signal
gets partially reflected back and forth, each time delayed by the round
trip propagation
delay in the cable, and so you have something like multipath going on.
Fortunately the
successive reflections get weaker with time, generally quite rapidly.
Since many
GPS users seem very concerned about multipath resulting from poor antenna
placement,
I think this factor should be considered as well and not just get swept
under the rug.

The amplitude of the "multipath" resulting from cable mismatches depends on
the product
of the voltage reflection coefficients at the two ends of the cable.  If
either end is perfectly
matched, then the quality of the match at the other end is not significant
vis-a-vis apparent
multipath problems and only affects transmission loss.

But when there is a mismatch on both ends, then the length of the cable
comes into play
as well.  A longer cable means more delay between successive reflections,
which is just
like multipath involving longer delays between the direct and the reflected
signals.

Cheers,

Dana      (K8YUM)

On Fri, Jun 5, 2020 at 7:13 PM Taka Kamiya via time-nuts <
time-nuts@lists.febo.com> wrote:

have

makes

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

On Jun 6, 2020, at 7:14 AM, Dana Whitlow k8yumdoober@gmail.com wrote:

It's one thing to maintain lock in a multipath environment, quite another
thing
to get "full" accuracy of GPS measurements of PVT.

An interesting difference between my scenario of poorly matched impedances
and "ordinary" multipath is this:  In the poor matching scenario, all the
received
signals will be impaired identically, while in the ordinary multipath
scenario,
signals from different satellites will suffer different (and time-varying)
multipath
impairments.  I'm not at all sure what effect this difference will have on
final
outcome, but my gut feel is that the case where all signals are impaired
identically
could lead to worse effects.

Dana

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Dana...
The question that comes to mind is just how much effect a weak a long
delayed reflection will have on overall system performance since it will
only matter to SV’s with poor S/N. The modulation scene which allows all
the SV's to transmit on the same frequency has to be pretty robust in the
face of both widely varying signal strengths and multiple signals arriving
at different times. It’s a similar scheme to CDMA cell phones, which
operate in a much more difficult environment with regard to signal
strengths, multi-path, and number of on channel signals. And those work
amazingly well.

Further, I am led to believe that once you have enough SV’s in view to
get a good set of ‘readings’, ionospheric effects are the limiting factor
until you go to a multi-band receiver.

Yes, to wring the last ounce of performance out of GPS takes attention to
the details, but don’t lose sight of how it was designed to work for users
in less than optimum (military field operations) in the first place.
From Tom Holmes, N8ZM

another

impedances

time-varying)

on

• 3

of

to

self,

much

there

can

antenna

depends on

significant

reflections,

reflected

system.  I

50

bought

for

for

less

10

50

http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

On Jun 6, 2020, at 8:51 AM, Dana Whitlow k8yumdoober@gmail.com wrote:

Basically true, Tom, but with some caveats:

The GPS system was originally designed, as you say, to provide
positions and velocity
information to military users in the field.  Errors of a few meters were of
little consequence.

But now, we have the time-nuts, surveyors, etc, who are decidedly pushing
the envelope
of GPS accuracy.  Some users want sub-nanosecond timing accuracy, surveyors
want
mm-level position accuracy, etc, and doing these things indeed requires
extreme attention
to detail.

As I understand it, multipath components that are delayed by about a few
chips or more
basically appear as a wee bit of added noise and have negligible
consequence unless
they are comparable to the direct signal in power.  However, MP components
whose
delay is less than about a chip time have the effect of tugging the
apparent temporal
location of the direct signal to some new value.  Note that this is in the
range of practical
round trip cable delays in at least some installations.  At Arecibo, for
example, the one-way
cable lengths from GPS antennas on the roof to their respective receivers
was about
200 ft.  Round trip was thus 400 ft physical, hence up to about 600 ft
depending on
what the cables were using for the dielectric.  That's about 0.6 chip
length for the C/A
code GPS signal.

Another factor which will influence the error introduced will be the RF
phase relationship
between the direct and delayed signal component.

Gee, this stuff gets complicated ...

Dana

On Sat, Jun 6, 2020 at 7:30 AM Tom Holmes tholmes@woh.rr.com wrote:

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Dana...
The question that comes to mind is just how much effect a weak a long
delayed reflection will have on overall system performance since it will
only matter to SV’s with poor S/N. The modulation scene which allows all
the SV's to transmit on the same frequency has to be pretty robust in the
face of both widely varying signal strengths and multiple signals arriving
at different times. It’s a similar scheme to CDMA cell phones, which
operate in a much more difficult environment with regard to signal
strengths, multi-path, and number of on channel signals. And those work
amazingly well.

Further, I am led to believe that once you have enough SV’s in view to
get a good set of ‘readings’, ionospheric effects are the limiting factor
until you go to a multi-band receiver.

Yes, to wring the last ounce of performance out of GPS takes attention to
the details, but don’t lose sight of how it was designed to work for users
in less than optimum (military field operations) in the first place.
From Tom Holmes, N8ZM

another

impedances

time-varying)

on

• 3

of

to

self,

much

there

can

antenna

depends on

significant

reflections,

reflected

system.  I

50

bought

for

for

less

10

50

http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

No caveats required.

Give some numbers on how loud that MP signal has to be to cause a problem.
My original example of line losses for a relatively short cable still
suggests to me that it is a minimal problem. If the S/N of a satellite is
poor, the receiver algorithm is likely to not use it if there are better
choices, and there often will be.

Tom Holmes, N8ZM

-----Original Message-----
From: time-nuts time-nuts-bounces@lists.febo.com On Behalf Of Dana
Whitlow
Sent: Saturday, June 06, 2020 8:52 AM
To: Discussion of precise time and frequency measurement <
time-nuts@lists.febo.com>
Subject: Re: [time-nuts] ThunderBolt question

Basically true, Tom, but with some caveats:

The GPS system was originally designed, as you say, to provide
positions and velocity
information to military users in the field.  Errors of a few meters were of
little consequence.

But now, we have the time-nuts, surveyors, etc, who are decidedly pushing
the envelope
of GPS accuracy.  Some users want sub-nanosecond timing accuracy, surveyors
want
mm-level position accuracy, etc, and doing these things indeed requires
extreme attention
to detail.

As I understand it, multipath components that are delayed by about a few
chips or more
basically appear as a wee bit of added noise and have negligible
consequence unless
they are comparable to the direct signal in power.  However, MP components
whose
delay is less than about a chip time have the effect of tugging the
apparent temporal
location of the direct signal to some new value.  Note that this is in the
range of practical
round trip cable delays in at least some installations.  At Arecibo, for
example, the one-way
cable lengths from GPS antennas on the roof to their respective receivers
was about
200 ft.  Round trip was thus 400 ft physical, hence up to about 600 ft
depending on
what the cables were using for the dielectric.  That's about 0.6 chip
length for the C/A
code GPS signal.

Another factor which will influence the error introduced will be the RF
phase relationship
between the direct and delayed signal component.

Gee, this stuff gets complicated ...

Dana

On Sat, Jun 6, 2020 at 7:30 AM Tom Holmes tholmes@woh.rr.com wrote:

arriving

users

wrote:

the

impaired

those

order

to

back

would

several

the

round

swept

If

cable

have

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

On Jun 6, 2020, at 9:54 AM, Dana Whitlow k8yumdoober@gmail.com wrote:

I wish I could give some numbers.  Sadly, I don't know how to calculate
them nor do I
have the equipment to do empirical measurements of that level of accuracy.

But I could do an arm-waving estimate.  Suppose that the delayed signal
component
is in phase with the direct signal, but 20 dB down, and is delayed by 100
nsec.  Then
I could see how the apparent centroid of a given chip could be displaced by
about
10 nsec, which would be enough to cause complaint from many time-nuts.

Personally I'm not into time keeping of that order of magnitude, but I'd
like my GPS-
disciplined oscillator to be phase stable to within 10's of ps over about a
1 minute
time frame.  At present I cannot seem to do this well unless I let my Rb
free-run.

Dana

On Sat, Jun 6, 2020 at 8:43 AM Tom Holmes tholmes@woh.rr.com wrote:

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

No caveats required.

Give some numbers on how loud that MP signal has to be to cause a problem.
My original example of line losses for a relatively short cable still
suggests to me that it is a minimal problem. If the S/N of a satellite is
poor, the receiver algorithm is likely to not use it if there are better
choices, and there often will be.

Tom Holmes, N8ZM

-----Original Message-----
From: time-nuts time-nuts-bounces@lists.febo.com On Behalf Of Dana
Whitlow
Sent: Saturday, June 06, 2020 8:52 AM
To: Discussion of precise time and frequency measurement <
time-nuts@lists.febo.com>
Subject: Re: [time-nuts] ThunderBolt question

Basically true, Tom, but with some caveats:

The GPS system was originally designed, as you say, to provide
positions and velocity
information to military users in the field.  Errors of a few meters were of
little consequence.

But now, we have the time-nuts, surveyors, etc, who are decidedly pushing
the envelope
of GPS accuracy.  Some users want sub-nanosecond timing accuracy, surveyors
want
mm-level position accuracy, etc, and doing these things indeed requires
extreme attention
to detail.

As I understand it, multipath components that are delayed by about a few
chips or more
basically appear as a wee bit of added noise and have negligible
consequence unless
they are comparable to the direct signal in power.  However, MP components
whose
delay is less than about a chip time have the effect of tugging the
apparent temporal
location of the direct signal to some new value.  Note that this is in the
range of practical
round trip cable delays in at least some installations.  At Arecibo, for
example, the one-way
cable lengths from GPS antennas on the roof to their respective receivers
was about
200 ft.  Round trip was thus 400 ft physical, hence up to about 600 ft
depending on
what the cables were using for the dielectric.  That's about 0.6 chip
length for the C/A
code GPS signal.

Another factor which will influence the error introduced will be the RF
phase relationship
between the direct and delayed signal component.

Gee, this stuff gets complicated ...

Dana

On Sat, Jun 6, 2020 at 7:30 AM Tom Holmes tholmes@woh.rr.com wrote:

arriving

users

wrote:

the

impaired

those

order

to

back

would

several

the

round

swept

If

cable

have

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.