GPS antenna length correction
I am taking this out of the "New timing receivers" thread as it is only
peripherally related.
As I think about the geometry of satellite position and path length, it
seems to me that, since the geometry is determined by the antenna position
and not the receiver position, additional antenna cable introduces a fixed
delay value and hence a fixed constant that gets added to each path
regardless of direction. It seems to me that this would produce a much
"fuzzier" solution to position and/or variation in timing. Knowing cable
length and propagation velocity, would allow the software to subtract that
constant from all ranges and thus provide a more correct position and time
solution. Is this not the case? Does it do something simpler but "good
enough"?
--
Brian Lloyd, WB6RQN/J79BPL
706 Flightline Drive
Spring Branch, TX 78070
brian@lloyd.com
+1.916.877.5067
On Mon, Apr 28, 2014 at 6:52 AM, Brian Lloyd brian@lloyd.com wrote:
..., additional antenna cable introduces a fixed
delay value and hence a fixed constant that gets added to each path
regardless of direction. It seems to me that this would produce a much
"fuzzier" solution to position and/or variation in timing. Knowing cable
length and propagation velocity, would allow the software to subtract that
constant from all ranges and thus provide a more correct position and time
solution. Is this not the case?
I think you have it exactly correct. The antenna location determines the
geometry and the cable length adds only a fixed offset. Many GPS receivers
have a command where you can tell them a "cable delay". But also you
have to think about the other end of this. How long is the cable that is
used for the output FROM the GPS receiver. This cable introduces a delay
also. You might even have a distribution amplifier of at least a TTL chip
acting as a driver or maybe a MAX232 chip doing level conversion.
There is a total system delay. But really this only matters if you are
keeping absolute time of day. Delay does not mater for frequency
measurement or for time interval measurement.
--
Chris Albertson
Redondo Beach, California
On Mon, 28 Apr 2014 13:10:05 -0700, Chris Albertson wrote:
On Mon, Apr 28, 2014 at 6:52 AM, Brian Lloyd brian@lloyd.com wrote:
..., additional antenna cable introduces a fixed
delay value and hence a fixed constant that gets added to each path
regardless of direction. It seems to me that this would produce a much
"fuzzier" solution to position and/or variation in timing. Knowing cable
length and propagation velocity, would allow the software to subtract that
constant from all ranges and thus provide a more correct position and time
solution. Is this not the case?
I think you have it exactly correct. The antenna location determines the
geometry and the cable length adds only a fixed offset. Many GPS receivers
have a command where you can tell them a "cable delay". But also you
have to think about the other end of this. How long is the cable that is
used for the output FROM the GPS receiver. This cable introduces a delay
also. You might even have a distribution amplifier of at least a TTL chip
acting as a driver or maybe a MAX232 chip doing level conversion.
There is a total system delay. But really this only matters if you are
keeping absolute time of day. Delay does not mater for frequency
measurement or for time interval measurement.
As mentioned earlier the position of the receiver is indeterminate unless
the vector displacement from the antenna to the receiver is accounted for.
Accounting for the cable delay will only correct the absolute time.
Imagine a 100m antenna feed line; the receiver could be anywhere within
100m of the antenna (even above it or at it). The algorithm that computes position
needs to know this. It is possible the programmer assumes that the receiver
is directely below the antenna; but I don't know what programmers assume.
It would be safest to display the position of the antenna. This discussion
suggests that manufacturers should make it clear that the position is of the
antenna or otherwise....
--
Chris Albertson
Redondo Beach, California
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Bill Beam
NL7F
As mentioned earlier the position of the receiver is indeterminate unless
the vector displacement from the antenna to the receiver is accounted for.
Accounting for the cable delay will only correct the absolute time.
Imagine a 100m antenna feed line; the receiver could be anywhere within
100m of the antenna (even above it or at it). The algorithm that computes
position
needs to know this. It is possible the programmer assumes that the
receiver
is directely below the antenna; but I don't know what programmers assume.
It would be safest to display the position of the antenna. This
discussion
suggests that manufacturers should make it clear that the position is of
the
antenna or otherwise....
The receiver computes the position of the antenna phase center. The
position of the receiver is not of interest and also cannot be computed.
People who do not accept this, should keep their antenna cables really
short... ;-)
/Björn
On Mon, Apr 28, 2014 at 3:53 PM, Bill Beam wbeam@gci.net wrote:
As mentioned earlier the position of the receiver is indeterminate unless
the vector displacement from the antenna to the receiver is accounted for.
Accounting for the cable delay will only correct the absolute time.
Imagine a 100m antenna feed line; the receiver could be anywhere within
100m of the antenna (even above it or at it). The algorithm that computes
position
needs to know this.
Yes, of course. The GPS receiver reports the location of the antenna. If
you need to know some other location you have to apply some (x,y,z) offset
from the antenna. Actually I doubt anyone would care about the location
of the receiver. They might want to know the location of the vehicle's
center of mass or the location of some display screen. The cable delay
accounts only for the offset in time.
Time and location really are different things, especially because a cable
can take a non-straightline path or even be coiled up.
--
Chris Albertson
Redondo Beach, California
As one responsible for the architecture and design of many military vehicle
GPS based navigation systems I will "un-simplify" my previous email. The
GPS solution (position or Doppler) is always made at the center of the GPS
antenna; that's where the lines of intersection (3D) from the individual
satellites occur (tri-lateration). You can account for the position
difference between the antenna and the receiver, vehicle center of mass or
another sensor (such as an inertial measurement system, velocimeter,
gravitometer etc.) by application of the lever arm vector. This co-locates
the GPS position with that device. This has the function of making the
sensor position and the GPS position both act as if they were at the IMU
location.
An example of this is if you have the GPS antenna located on a ships mast,
and the IMU (whose position defines ships position) located on the bridge,
in order for the navigation Kalman Filter to properly compute the ships
location the lever arm correction must be applied. In this case the antenna
can be several tens, or even hundreds, of meters from the IMU. This needs
to be a vector as can been seen when the ship roles to one side or another.
A second, uncorrelated, correction is for the time (delay) that the RF
signal travels along the transmission line (C x Vf x L). If that's needed
you can compute the delay, hence phase shifts.
Hopefully this helps,
Michael / K7HIL
On Mon, Apr 28, 2014 at 5:00 PM, Chris Albertson
albertson.chris@gmail.comwrote:
On Mon, Apr 28, 2014 at 3:53 PM, Bill Beam wbeam@gci.net wrote:
As mentioned earlier the position of the receiver is indeterminate unless
the vector displacement from the antenna to the receiver is accounted
for.
Accounting for the cable delay will only correct the absolute time.
Imagine a 100m antenna feed line; the receiver could be anywhere within
100m of the antenna (even above it or at it). The algorithm that
computes
position
needs to know this.
Yes, of course. The GPS receiver reports the location of the antenna. If
you need to know some other location you have to apply some (x,y,z) offset
from the antenna. Actually I doubt anyone would care about the location
of the receiver. They might want to know the location of the vehicle's
center of mass or the location of some display screen. The cable delay
accounts only for the offset in time.
Time and location really are different things, especially because a cable
can take a non-straightline path or even be coiled up.
--
Chris Albertson
Redondo Beach, California
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Accounting for the cable delay will only correct the absolute time.
Imagine a 100m antenna feed line; the receiver could be anywhere within
100m of the antenna (even above it or at it). The algorithm that computes position
needs to know this.
No.
What the algorithm computes is position -and- time at the antenna.
It doesn't care where the receiver is, or how long the cable is.
The GPS solution gives the place and time of the antenna only.
And then,
- If you want to translate the antenna solution to some other physical place, then apply a dX,dY,dZ correction.
- If you want to translate the antenna solution to some other physical time, then apply a dT correction.
For example,
- If the antenna is on top of a pole but you actually want the true location of the base of the pole, you apply a spatial correction.
- If the antenna is on top of your roof but you want the true time at a BNC jack in your lab, you apply a temporal correction.
The confusion,
The GPS timing receivers most of us use make it easy to apply a temporal correction, but not a spatial correction.
Consequently, your lab receiver can be configured to pulse the true time at center of the front panel BNC.
But, your lab receiver cannot be configured to show the true location of the center of the front panel LCD.
Instead, the LCD is stuck showing the location of your distant antenna.
/tvb
On Mon, 28 Apr 2014 21:24:56 -0700, Tom Van Baak wrote:
Accounting for the cable delay will only correct the absolute time.
Imagine a 100m antenna feed line; the receiver could be anywhere within
100m of the antenna (even above it or at it). The algorithm that computes position
needs to know this.
No.
Well... Yes!
Please read all of what I stated.
What the algorithm computes is position -and- time at the antenna.
It doesn't care where the receiver is, or how long the cable is.
The GPS solution gives the place and time of the antenna only.
And then,
- If you want to translate the antenna solution to some other physical place, then apply a dX,dY,dZ correction.
- If you want to translate the antenna solution to some other physical time, then apply a dT correction.
For example,
- If the antenna is on top of a pole but you actually want the true location of the base of the pole, you apply a spatial correction.
- If the antenna is on top of your roof but you want the true time at a BNC jack in your lab, you apply a temporal correction.
The confusion,
The GPS timing receivers most of us use make it easy to apply a temporal correction, but not a spatial correction.
Consequently, your lab receiver can be configured to pulse the true time at center of the front panel BNC.
But, your lab receiver cannot be configured to show the true location of the center of the front panel LCD.
Instead, the LCD is stuck showing the location of your distant antenna.
/tvb
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Bill Beam
NL7F
I want to thank everyone for their answers. I was sure I had the geometry
right and that the additional latency would make the position "fuzzier"
because I kept thinking that the clock was absolutely correct.
Of course, the clock in the GPS receiver converges on a time that is offset
from absolute time by the delay in the cable. This time offset is equal and
opposite to the extra path length for each satellite's range, thus ensuring
that the position calculation is independent of connecting cable length. So
this corrects for the (apparent) path length error but still leaves a clock
offset error which is then corrected by the cable length input value.
I get it now. Thank you all very much. This is something that has bothered
me for some time and finally seeing the solution makes me feel a lot better
about GPS installations in aircraft.
--
Brian Lloyd, WB6RQN/J79BPL
706 Flightline Drive
Spring Branch, TX 78070
brian@lloyd.com
+1.916.877.5067