On 21 abr, 23:30, "Dave (from the UK)" see-my-signat...@southminster-
branch-line.org.uk wrote:
Wimpie wrote:
Hello,

Your formula for far field distance (Fraunhofer region) assumes a path
difference between the inner and outer antenna with respect to an
observation point of 1/16 lambda.

Someone has said that the formula I gave is not valid for a phased array.

His comment (about the 2 D^2/lambda) is below:

-----
That estimation does not apply in this case. It can be considered to be
valid for aperture antennas which is not the case here. It would only
require to have the transmitting antenna illuminating the pleased array
within its 3dB mainlobe which of course is by far the case at a distance
of 3000m or even more.
------

I'm not to bothered about the odd factor of two. I have seen a
derivation of the formula, but it was based on a rectangular aperture,
not an array of them. It don't know if that may mean the equation is
just not appropriate at all.

Using that forumal puts the far-field distance at about 10 km in my
case. Using someone elses idea, puts it at only a few hundred meters.
There is at least a factor of 10 difference.
--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form:
Hitting reply will work for a few months only - later set it manually.

http://chessdb.sourceforge.net/- a Free open-source Chess Database

Hi Dave,

Whether or not the formula is applicable, depends on many factors as
mentioned in my previous posting.

For a broadside array, the formula holds with same accuracy as for
continuous aperture antennas. In my antenna courses I use the
broadside array approach to derive the 2B^2/lambda formula.

For an end-fire case, the situation is different. When you are
interested in main lobe gain only (so not the complete radiation
pattern), you can reduce the distance significantly. The reason for
that is that when you come closer to the antenna, the path difference
doesn't change; the amplitude contribution of each array element is of
importance now. However when you need to know the complete pattern
(including broadside directions), you need the large distance.

It is just a matter of change in path length difference amplitude
unbalance when you come too close to the antenna. If you keep this in
mind, you can figure out the measuring distance for you application
(for example with a spread sheet).

I would reserve the term "far field distance" for that distance where
the complete radiation pattern does not change with measuring
distance. In that case, the 2B^2/lambda formula is a good rule of
thumb.


Best regards,

Wim
PA3DJS

Wimpie wrote:

Hi Dave,

Hi Wim

Whether or not the formula is applicable, depends on many factors as
mentioned in my previous posting.

For a broadside array, the formula holds with same accuracy as for
continuous aperture antennas. In my antenna courses I use the
broadside array approach to derive the 2B^2/lambda formula.

My situation is very odd. As I said at the start, this is not an amateur
antenna.

The array of "antennas" are not designed to work as one nice antennas,
but are an essentially random(ish) collection of radiating centres.
(However, they are all energised from the same signal source). So they
can be considered like a phased array, as they are regularly spaced all
in one long line.

Hence my original diagram

A---A---A---A---A---A---A---A---A---A

accurately describes the situation. Each "A" is an antenna. The
amplitude and phase can be arbitrary.

I do *not* want them to behave as a nice phased array with decent gain
and low side-lobes! Each antennas is radiating an *unwanted* signal. But
the fact remains that the gain could conceivably be high under some
circumstances, which would create interference.

Hence I need to test this.

I would reserve the term "far field distance" for that distance where
the complete radiation pattern does not change with measuring
distance. In that case, the 2B^2/lambda formula is a good rule of
thumb.

In this case, I am interested in any direction. The direction of the
main lobe will be essentially random.

--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form:
Hitting reply will work for a few months only - later set it manually.

http://chessdb.sourceforge.net/ - a Free open-source Chess Database

"Dave (from the UK)"
wrote in message ...
Wimpie wrote:

Hi Dave,

Hi Wim

Whether or not the formula is applicable, depends on many factors as
mentioned in my previous posting.

For a broadside array, the formula holds with same accuracy as for
continuous aperture antennas. In my antenna courses I use the
broadside array approach to derive the 2B^2/lambda formula.

My situation is very odd. As I said at the start, this is not an amateur
antenna.

The array of "antennas" are not designed to work as one nice antennas, but
are an essentially random(ish) collection of radiating centres. (However,
they are all energised from the same signal source). So they can be
considered like a phased array, as they are regularly spaced all in one
long line.

Hence my original diagram

A---A---A---A---A---A---A---A---A---A

accurately describes the situation. Each "A" is an antenna. The amplitude
and phase can be arbitrary.

I do *not* want them to behave as a nice phased array with decent gain and
low side-lobes! Each antennas is radiating an *unwanted* signal. But the
fact remains that the gain could conceivably be high under some
circumstances, which would create interference.

Hence I need to test this.

I would reserve the term "far field distance" for that distance where
the complete radiation pattern does not change with measuring
distance. In that case, the 2B^2/lambda formula is a good rule of
thumb.

In this case, I am interested in any direction. The direction of the main
lobe will be essentially random.

--
Dave (from the UK)


Hi Dave

I'm curious about two things.

1 - Do you intend to actually make and record measurements of the
radiated field, or do you want to determine the minimum distance at which
the measurements can be made?
2 - What prevents the use of a computer modeling program to predict the
pattern?

Jerry

Jerry Martes wrote:

Hi Dave

I'm curious about two things.

1 - Do you intend to actually make and record measurements of the
radiated field, or do you want to determine the minimum distance at which
the measurements can be made?

yes

2 - What prevents the use of a computer modeling program to predict the
pattern?

nothing. I think that will be done. But a theoetical analysis would be
nice if possible.


Jerry




--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form:
Hitting reply will work for a few months only - later set it manually.

http://chessdb.sourceforge.net/ - a Free open-source Chess Database

"Dave (from the UK)"
wrote in message ...
Jerry Martes wrote:

Hi Dave

I'm curious about two things.

1 - Do you intend to actually make and record measurements of the
radiated field, or do you want to determine the minimum distance at which
the measurements can be made?

yes

2 - What prevents the use of a computer modeling program to predict
the pattern?

nothing. I think that will be done. But a theoetical analysis would be
nice if possible.


Jerry


--
Dave (from the UK)

Hi Dave

Would you consider building a scale model of this array to allow a polar
orbiting satellite to be the illuminator?
I get some pretty good radiation pattern plots at VHF, using Patrik Tast's
SignalPlotter program and NOAA satellites.

Jerry

Jerry Martes wrote:
Hi Dave

Would you consider building a scale model of this array to allow a polar
orbiting satellite to be the illuminator?

It's not really practical to do that for various reasons. The structure
is not a nice conventional antenna that can be scaled up/down size.

I get some pretty good radiation pattern plots at VHF, using Patrik Tast's
SignalPlotter program and NOAA satellites.

Jerry

That's a concept I have not heard of before. But for me to work at VHF,
I'd need to make the structure larger by a factor of 10, so it would be
several hundred metres long.


--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form:
Hitting reply will work for a few months only - later set it manually.

http://chessdb.sourceforge.net/ - a Free open-source Chess Database

"Dave (from the UK)"
wrote in message ...
Jerry Martes wrote:
Hi Dave

Would you consider building a scale model of this array to allow a
polar orbiting satellite to be the illuminator?

It's not really practical to do that for various reasons. The structure is
not a nice conventional antenna that can be scaled up/down size.

I get some pretty good radiation pattern plots at VHF, using Patrik
Tast's SignalPlotter program and NOAA satellites.

Jerry

That's a concept I have not heard of before. But for me to work at VHF,
I'd need to make the structure larger by a factor of 10, so it would be
several hundred metres long.


--
Dave (from the UK)

Hi Dave

You are not restricted to VHF. It is possible that there is a polar
orbiting satellite transmitting continuously on the frequency you are
interested in. NOAA satellites transmit continuously at about half the
frequency range you refer to.

Jerry

Dave (from the UK) wrote:
Jerry Martes wrote:

Hi Dave

I'm curious about two things.

1 - Do you intend to actually make and record measurements of the
radiated field, or do you want to determine the minimum distance at
which the measurements can be made?


yes


An interesting problem. What you're presumably trying to do is
determine how far do I need to be to bound the uncertainty on a
measurement in an arbitrary direction. Or, another way, at what distance
is the collective effects of the phase error for each of the signals
(due to path length differences) smaller than your measurement
uncertainty (so you don't care anymore).

This can be quite challenging if you want to worry about -40dB nulls,
for instance, because a very small phase error can result in a -40dB
null becoming a -30 dB null.

Complicating it a bit is that what you're probably really concerned with
is a statistical problem.. you've got multiple sources, a random
direction of observation, (and practically speaking, some propagation
uncertainties between antenna and observation point).

You might want to look for a paper by Dybdal and Ott: "Coherent RF
Error Statistics", IEEE Trans MTT, v34,n12, Dec 86, pp1413-1420 which
discusses this in some detail, and, as well, provides some nice
approximations that are useful in practical systems.



2 - What prevents the use of a computer modeling program to
predict the pattern?


nothing. I think that will be done. But a theoetical analysis would be
nice if possible.

One can come up with a "bound" for the performance from analytical
means, and a Monte Carlo analysis can give you some statistics.

Jim

Jim Lux wrote:

1 - Do you intend to actually make and record measurements of
the radiated field, or do you want to determine the minimum distance
at which the measurements can be made?



yes



An interesting problem. What you're presumably trying to do is
determine how far do I need to be to bound the uncertainty on a
measurement in an arbitrary direction. Or, another way, at what distance
is the collective effects of the phase error for each of the signals
(due to path length differences) smaller than your measurement
uncertainty (so you don't care anymore).

This can be quite challenging if you want to worry about -40dB nulls,
for instance, because a very small phase error can result in a -40dB
null becoming a -30 dB null.

I'm not really that bothered about the depth of nulls to any great
extent. Since this is radiating an unwanted signal, the concern is
finding where the gain is highest and how high it is.

Complicating it a bit is that what you're probably really concerned with
is a statistical problem.. you've got multiple sources, a random
direction of observation, (and practically speaking, some propagation
uncertainties between antenna and observation point).

You might want to look for a paper by Dybdal and Ott: "Coherent RF
Error Statistics", IEEE Trans MTT, v34,n12, Dec 86, pp1413-1420 which
discusses this in some detail, and, as well, provides some nice
approximations that are useful in practical systems.

In this case, it is certainly a statistical thing. As I said before, the
amplitudes and phases of the radiators are random(ish) and will be
changing all the time.

2 - What prevents the use of a computer modeling program to
predict the pattern?



nothing. I think that will be done. But a theoetical analysis would be
nice if possible.


One can come up with a "bound" for the performance from analytical
means, and a Monte Carlo analysis can give you some statistics.

Jim

I will look at doing some MC analysis.
--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form:
Hitting reply will work for a few months only - later set it manually.

http://chessdb.sourceforge.net/ - a Free open-source Chess Database

On 22 abr, 12:25, "Dave (from the UK)" see-my-signat...@southminster-
branch-line.org.uk wrote:
Wimpie wrote:
Hi Dave,

Hi Wim

Whether or not the formula is applicable, depends on many factors as
mentioned in my previous posting.
For a broadside array, the formula holds with same accuracy as for
continuous aperture antennas. In my antenna courses I use the
broadside array approach to derive the 2B^2/lambda formula.

My situation is very odd. As I said at the start, this is not an amateur
antenna.

The array of "antennas" are not designed to work as one nice antennas,
but are an essentially random(ish) collection of radiating centres.
(However, they are all energised from the same signal source). So they
can be considered like a phased array, as they are regularly spaced all
in one long line.

Hence my original diagram

A---A---A---A---A---A---A---A---A---A

accurately describes the situation. Each "A" is an antenna. The
amplitude and phase can be arbitrary.

I do *not* want them to behave as a nice phased array with decent gain
and low side-lobes! Each antennas is radiating an *unwanted* signal. But
the fact remains that the gain could conceivably be high under some
circumstances, which would create interference.

Hence I need to test this.

I would reserve the term "far field distance" for that distance where
the complete radiation pattern does not change with measuring
[all text deleted]

Hi Dave,



I don't know what you are going to do with the array. As long as you
understand how a radiation pattern (whether within or outside the far
field distance) can be calculated based on the array elements, you
should be able to find a comfortable distance. I think references to
scientific documents will not help you any further, maybe a physics
book on electromagnetism or a specialized book on beam forming
antennas may help you.

If your organization is not able to do this in-house, you might hire
an expert.

Best regards,

Wim
PA3DJS