[Antenna Beam Forming] Antenna Pattern & Element Pattern

[Jefferson86]

Hey guys,

currently I am working on my diploma thesis in the field of beam design for mobile communication.

I created a linear array with N elements and computed the corresponding Array Factor with the help of the chebyshev distribution. I assume my Element Factor to have its maximum at 90° broadside of the array, let`s say between 75° and 115°.
So, my Total Gain will be the addition of the Element Factor and the Array Factor .
This is clear to me, as long as the array is steered towards the region of 75°-115°.

My question is:
When the array is steered more degrees than the above region, the Element Factor will not have its maximum anymore, resulting in the suppression of the Array Factor.
--> What happens to the suppressed energy? Where does it go? It must be somewhere, right ?

Ideas highly appreciated,
best regards.
Jeff

 

[Baluncore]

The beam will be broadside, i.e. perpendicular to the line of the array if all elements are driven with the same phase.
The resulting beam pattern is not the addition, but is the product of the element factor by the array factor.

You will get a greater “gain” or EIRP in the main beam. That extra energy in the main beam is not being radiated elsewhere.
An energy null in some direction will be balanced by an energy gain in some other direction.

The position of your side lobes and nulls will be determined by the element phase and positioning in the array.

 

[sophiecentaur]

My question is:
When the array is steered more degrees than the above region, the Element Factor will not have its maximum anymore, resulting in the suppression of the Array Factor.
--> What happens to the suppressed energy? Where does it go? It must be somewhere, right ?

Ideas highly appreciated,
best regards.
Jeff

The two factors (array and element) are just multiplied together to give the pattern. (variable separable situation, in an ideal case) BUT that only tells you the normalised shape of the pattern. The side lobe levels will increase to bring the total Power radiated to the same - because, as you say, the Energy has to go somewhere and that consideration is the pivotal one.
In practical terms, of course, there are limits to how much you can control the element currents (contributions) because of mutual impedance between the elements will come into play.

 

[Jefferson86]

The two factors (array and element) are just multiplied together to give the pattern. (variable separable situation, in an ideal case) BUT that only tells you the normalised shape of the pattern. The side lobe levels will increase to bring the total Power radiated to the same - because, as you say, the Energy has to go somewhere and that consideration is the pivotal one.
In practical terms, of course, there are limits to how much you can control the element currents (contributions) because of mutual impedance between the elements will come into play.

So, if I neglect all effects due to mutual coupling, I could argue that the suppressed energy goes fully in the increased side lobe level?

 

[davenn]

which of course isn't what you want ... for any multi-element beam ( Yagi) the idea is to get the biggest lobe off the front smaller off the rear and minimal off the sides

There will always be rear and side lobes, they just need to be minimised as much as possible by the placement of the elements and if its a multi-Yagi array the spacing of the 2 arrays. Power radiated anywhere else other than off the front lobe is just wasted

( you may already know all this)

 

[Baluncore]

Constructive and destructive interference of fields cannot disobey the principle of conservation of energy.

If you ignore internal power losses, then the total 3D integral of radiated energy will be the same no matter what the element pattern or array distribution is.
The only difference between arrays is the directional distribution of that energy flow.

 

[sophiecentaur]

Something the initial theory doesn't tell you is that skewing the beam of an array (and weighting the element contributions of an unslewed array, too) is not a trivial task. It is not easy to calculate the real gain that an array can achieve, even though the 'directivity gain' can be predicted to some extent from the pattern.

 

[sophiecentaur]

So, if I neglect all effects due to mutual coupling, I could argue that the suppressed energy goes fully in the increased side lobe level?

It is tempting to think of this rather like an airbed that you are trying to deflate by sitting on part of it - other bits pop up. Whilst this is indeed what it seems like, it's only a crude metaphor.
All that is at work here is the geometry of the situation, when you make assumptions about the actual currents flowing in the elements. You will never generate a maximum in the direction of a minimum in the basic element pattern, for instance. As I remarked before, in the ideal situation you have separable variables.
In a real array (particularly one with low gain (not very directive) elements, you are dealing much more with the 'airbed' situation because the elements will interact with each other very significantly and it becomes harder and harder to control the basic element currents as you try to force a pattern that's very different from the co-phase excitation situation.
I remember looking into the problem of making a 'super-gain' array, based on the Chebychev distribution. To get a sharper beam than will occur 'naturally', you needed huge values of drive currents, in inconvenient phases, which would require lunatic matching networks with individually fed elements. Nothing you could hope to get with a simple passive splitter / phase shift network.
Phase slewing, to get beam directivity will always (? I think) give you a slightly wider main beam shape than you start with and asymmetrical sidelobes.

 

[Baluncore]

To get a sharper beam than will occur 'naturally', you needed huge values of drive currents, in inconvenient phases, which would require lunatic matching networks with individually fed elements.

It is not really that difficult. By inserting series capacitance in a wire antenna, the phase velocity can be increased to become superluminal. The mainlobe then becomes extremely narrow with very high directional gain. The practical limit is reached when the beamwidth or bandwidth of the antenna become too narrow due to frequency sensitivity of the velocity.

 

[sophiecentaur]

N

It is not really that difficult. By inserting series capacitance in a wire antenna, the phase velocity can be increased to become superluminal. The mainlobe then becomes extremely narrow with very high directional gain. The practical limit is reached when the beamwidth or bandwidth of the antenna become too narrow due to frequency sensitivity of the velocity.

Not quite as easy as you imply, if you want a 'good' radiation pattern; I am not sure of the application you are referring to but there are others - a small loop antenna being an example of superdirectivity, by virtue of the null.
The Chebychev style of element weight gives, iirc, side lobes of equal amplitude and of minimal height and these side lobes become very dependent on accurate feed currents. There are a lot of erudite links about this, to be seen on Google but the only ones I followed are not free (From Springer etc.). I got my information, long ago, from IEEE papers, I seem to remember, and possibly from a selection of textbooks in the departmental library. One serious problem with super directive rays is the low radiation resistance, which makes matching difficult and efficiency low. (This applies to a small loop antenna, too.)

 

[Baluncore]

- a small loop antenna being an example of superdirectivity, by virtue of the null.

No, I am referring to the main beam, not the null.
The vf on a bare copper wire can approach 99.8%. If the wire has many series capacitors inserted that cancel part of the self inductance, then the vf can exceed 100% luminal. That is a game changer. For superluminal velocities the radiation pattern can become a very narrow high gain beam.

 

[sophiecentaur]

A loop has a narrower beam than its physical size would suggest - because of two currents in anti phase. That's the only 'super gain' aspect of it that I meant. Of course, if it has a null then the 'main beam' will be higher, giving gain in that direction, wrt an omni dipole.
This is well off topic but I am assuming, as you haven't described the details, that you are referring to an end fed (omni in the horizontal plane) monopole, with Capacitors along its length. That could well increase the vertical directivity. It would be interesting to know how the Radiation Resistance is affected or what the side lobes are like.
The Array that the OP is discussing is a broadside array of omni or directive elements, individually fed. It gives uniform (optimal) side lobes and it fulfils a different purpose. Slewing or attempting super gain involves more difficulty in the feeding arrangement.

 

[Baluncore]

It would be interesting to know how the Radiation Resistance is affected or what the side lobes are like.

Yes, it attracted my attention in the late 1980s which is why I modeled it and built prototypes using varactors in a long wire to sweep the beam. It was way more sensitive than the arrays of small loops used for DF at the time. Small loops do have a simple dipole pattern, but they are relatively deaf in dBi when compared to real half wave dipoles. I was designing for receive only, hence the delicate varactors. I never had any problem efficiently coupling the antenna to the line. Unfortunately, with superluminal wires it was a challenge to find intermittent signals, mainly because the primary lobe, (conical about the wire), was too narrow and side lobes were too small. It was certainly interesting.

 

[sophiecentaur]

Yes, it attracted my attention in the late 1980s which is why I modeled it and built prototypes using varactors in a long wire to sweep the beam. It was way more sensitive than the arrays of small loops used for DF at the time. Small loops do have a simple dipole pattern, but they are relatively deaf in dBi when compared to real half wave dipoles. I was designing for receive only, hence the delicate varactors. I never had any problem efficiently coupling the antenna to the line. Unfortunately, with superluminal wires it was a challenge to find intermittent signals, mainly because the primary lobe, (conical about the wire), was too narrow and side lobes were too small. It was certainly interesting.

Your antenna was a bit 'Rhombic-like' then? I haven't yet visualised the orientation of the wire and here the main beat was pointing.

 

[Baluncore]

Your antenna was a bit 'Rhombic-like' then?

Yes, think of it as a single traveling wave wire above ground. For a TW antenna the length of the wire in wavelengths along with the vf on that wire determines the angle between the wire and the main lobe, the pattern is conical about the wire axis.

The rhombic is an array of four traveling wave wires. A fixed rhombic can really only be optimised for a fixed frequency, azimuth and elevation. That is because rhombic gain is a function of the superposition of four conical main lobes, along with their ground images.

 

[sophiecentaur]

Yours would be more like a short Beverage Antenna then. (If you confirm it's a horizontal wire; we never got that sorted yet.

 

[Baluncore]

Yours would be more like a short Beverage Antenna then.

No, not short, but multiple wavelengths long.

(If you confirm it's a horizontal wire; we never got that sorted yet.

Yes, think of it as a single traveling wave wire above ground.

Since it is a traveling wave antenna, over ground, it should be horizontal so that both the driven-end and the termination-end can be referenced to the ground.

If a TWA has a slope relative to the ground then it needs to be constructed as a tapered cage in order to maintain a constant impedance along the radiator.

 

[sophiecentaur]

Unfortunately, you have not quoted any dimensions or the wavelengths involved (it can't be a secret? ;))
I know from experience that measuring the actual gain of an antenna, particularly one that can't be mounted on a turntable / goniometer, can be very problematical. I have been involved in helicopter based measurements of the radiation patterns of HF Curtain arrays - and they turned out to be, only roughy, what the design predicted and also the performance of a directional MF transmitting array, which was only assessed by monitoring (quantitatively) the signal levels in Central Europe. I therefore have to question about the actual performance you quote for your TW antenna. However, the use of varactor diodes is a clever idea because you could maximise the performance for any station that you want to receive.
I was pleased to find that my design of MF array gave the expected received field strength (I forget the results from 30 years ago). But the transmitting site was near- ideal, being on a salt marsh and with a massive earthing net.

It isn't absolutely necessary to have a constant impedance along a TW structure, although it makes the design easier. But I was only speculating because you had not been explicit about your system.

 

[Baluncore]

... it can't be a secret?

I only talk in generalities about this field, not because of ignorance, but because I prefer to sleep at home in my own bed.

I therefore have to question about the actual performance you quote for your TW antenna.

Go ahead and cast your doubts. I originally looked into the field because others doubted it. Super-luminal long wires with matched termination are not normally experienced by RF engineers. Others came up with all sorts of excuses to avoid the subject. Fundamentally, they avoided investigation because their promotion was based on seniority and quite independent of performance. They could only lose.

I know from experience that measuring the actual gain of an antenna, particularly one that can't be mounted on a turntable / goniometer, can be very problematical.

By comparing the receive performance of the antenna against a reference standard antenna you can have a host of natural and man-made signal sources from many different directions and elevations. That can be at a very low cost because it can be done without transmitting or travelling. Another big advantage is that the opposition do not know you are active in their field, indeed they deserve a medal for their helpful assistance. The only sensible rule to follow then is; “Never ever transmit”. As you know, the commissioning of a transmit antenna is quite a different problem.

But the transmitting site was near- ideal, being on a salt marsh and with a massive earthing net.

Likewise, when the ground-image is part of the pattern, a TWA needs a well defined Earth network. As beam-width becomes narrower it becomes more important to maintain velocity and impedance, or the sharp pattern will rapidly deteriorate.

 

[sophiecentaur]

That all sounds far to conspiratorial for me, I'm afraid. I am only doubting your results on the grounds that you do not quote how you got them (or the results themselves). The term 'super luminal' is just a term and doesn't imply anything special about a loaded transmission line, does it? Series Inductors will produce the reverse effect.
You will have to accept / forgive my skepticism on the grounds that you are not supplying enough information for me to make a serious appraisal of what you are describing. Without a very special environment for AB testing of antennae (large, uncluttered and a known Earth situation etc) then any results are not going to be super-reliable. Could you at least give a clue as to how 'super' the supergain turned out to be? Was it impressive enough to be outside your margins of error? (Numbers might help) - I don't think the opposition are listening to our conversation. My room is not bugged afaik.

 

[Baluncore]

That all sounds far to conspiratorial for me, I'm afraid.

Then you should avoid signals intelligence and Government employment.

Without a very special environment for AB testing of antennae (large, uncluttered and a known Earth situation etc) then any results are not going to be super-reliable.

I used a nationl site developed and maintained for the purpose of radio monitoring, it had an EM exclusion zone and closed airspace. I audited the site and it met all the environmental EM requirements for accurate DF at the time. I then included all other antennas at the site in a numerical model to check that they had minimal effect on the systems I had under test.

The term 'super luminal' is just a term and doesn't imply anything special about a loaded transmission line, does it? Series Inductors will produce the reverse effect.

If to you a term does not imply anything special, then how can it have an opposite? IMHO only a barbarian would add series inductance to a TWA. The fact that series inductance destroys the pattern implies that series capacitance may improve it. My aim was to improve the pattern and gain, not to destroy it.

I don't think the opposition are listening to our conversation. My room is not bugged afaik.

This is a public forum.

 

[sophiecentaur]

There are examples of a traveling wave structure that operates 'sub luminary' - it just depends upon what you want to achieve. Do you imply there's something magic about a phase velocity greater than c?

I really don't see the point in keeping up this particular conversation. Your hints about dark forces and implications of conspiracy are much too far from the business of PF to keep my interest. If you didn't want to discuss the Engineering aspects of this project of yours then you really shouldn't have introduced it. You haven't supported your remarks with any substantial evidence or references and that goes against PF principles, imo.
BTW, what was the increase in gain that you reckon you achieved? Was your antenna slewable? If not, how did you choose a suitable transmission to do your calibration against?
What you have written about this seems to imply that broaching the subject is somehow violating security in some way - or were you just following some pet project of your own that the management hadn't approved? That happens frequently in some establishments; the boss can be very unreasonable at times.

But how does any of this have anything to do with OP?

 

[Baluncore]

To get a sharper beam than will occur 'naturally', you needed huge values of drive currents, in inconvenient phases, which would require lunatic matching networks with individually fed elements.

Your statement in post #9 may be true for transmitting antennas, but I know it is not true of all receive antennas.

But how does any of this have anything to do with OP?

Since post #9 you have pursued this matter by demanding I answer your questions about a subject unrelated to the OP. I do not know why you are pursuing the matter since it is not related to the OP and you are clearly not working in the field. I have answered your extended questions as best I can within the bounds available to me at this time. Let that be the end of it.

 

[sophiecentaur]

Fair enough, but you, yourself did introduce the subject of your TW antenna (quite unrelated to my mention of Chebychev feed problems - including the super gain arrangement). It would be an idea not to introduce claims about stuff you have used if you are not prepared to substantiate them publicly. I should be grateful if you would send me a PM with a few details. I could guarantee it would go no further. Having been involved with assessing and optimising the performance of Beverage antennae, I would find it interesting to see how much better they might perform. (Albeit there are some significant differences in the effect of the ground in the two cases - and the polarisation of the signals involved could be different.)

 

[Baluncore]

Blocked conversations so:
My results are not comparable to others previous experience.

Public information includes ham experiments such as in:
The RSGB “Amateur Radio Techniques” 6th edn;
It has a section on capacitively “stretched” dipoles, pages 279 to 281.
See Fig. 97, for amateur experiments with capacitance loaded TW wires.

For the numerical evaluation for different values of vf.
Get a copy of “Antennas” by Kraus. Original edn, 1950, is viewable at,
http://ok1ike.c-a-v.com/literatura/Kraus_antennas.pdf
Look at section 5-8. Pages 148 to 153. (It is on pages 228 to 234 in the 2'nd edition).
Use equation on page 151 to generate similar plots to fig 5-17. Verify against fig 5-16 patterns.
Notice that fig 5-17 is for p = 1. There is no consideration of p > 1.
Evaluate pattern for various p, including for p > 1.

 

[sophiecentaur]

Kraus is a well respected book, which I frequently referred to in the 70s and 80s but it only shows the 'established' traveling wave antenna ideas. It includes the helical beam antenna where the element is loaded to slow the wave down (barbarian?).
I am not really motivated to do your suggested sums as I don't have the book to hand and the PDF version is not convenient to use. This is strange, though. You leaped into say that the supergain is "easy" yet you won't even quote a ballpark figure for the improvement. Why did you post in the first place? You know the way PF works; people need the sustenance of performance figures and facts.

 

[Baluncore]

I am not really motivated to do your suggested sums as I don't have the book to hand and the PDF version is not convenient to use

Some elderly people have trouble with their eyesight and today's computers. You may be interested to know that PDF readers have an option to print a specified range of pages from the reference. If you print just the few pages I referenced, then you may find it more convenient.

You leaped into say that the supergain is "easy" yet you won't even quote a ballpark figure for the improvement.

Performance is limited by the length of the wire in wavelengths, by the ability to maintain a constant vf along a long wire and by the ability to acquire the signal. Only you can evaluate your own ability to improve on the historical TWAs such as a bare wire Beverage or Rhombic. You will see that it is steerable when you model it for various vf and wire length.

Why did you post in the first place?

Because you made a false statement …

To get a sharper beam than will occur 'naturally', you needed huge values of drive currents, in inconvenient phases, which would require lunatic matching networks with individually fed elements. Nothing you could hope to get with a simple passive splitter / phase shift network.

A statement that I know is not true of all antennas.

You have failed to justify your “impossible to do without lunatic matching networks” claim, indeed you cannot. I have presented you with a solution that demonstrates the failure of your claim. Your unwillingness to evaluate the traditional TWA equation for some configuration that you have encountered in the past demonstrates that you are not really interested in admitting that your unconditional statement was in fact conditionally based on your previous experience.
Those who claim that something is impossible should not interrupt those who are doing it.

 

[sophiecentaur]

Some elderly people have trouble with their eyesight and today's computers. You may be interested to know that PDF readers have an option to print a specified range of pages from the reference. If you print just the few pages I referenced, then you may find it more convenient.

Matron doesn't allow us to use the printer, in my care home for the partially sighted and weak minded, I'm afraid, and I definitely couldn't hope to do calculations with all those complicated formulae you quoted. That's why all I asked for was a ball park figure for the extra gain that your TWA achieves. You must know that information would not possibly be particularly confidential so why did you not just supply it, if it exists? Are we talking 3dB, 10dB, 20dB (is all I wanted to know)?
However, the TWA is not an array of individually driven elements and is not as flexible for steering and is probably not even suited to the same frequencies that the OP is discussing. I don't need to tell you how to google for information on Chebychev arrays; you can also attach the supergain requirement to the search. The efficiency problem is mentioned in several of the links I scanned through. One link says it's not as bad as people think but if you look at the sort of current amplitudes and phases required, you can see the potential problems in feeding the array. The reason that the Chebychev array is studied is the fact that it has optimal side lobe levels, as opposed to a TWA, from which you more or less have to put up with what you get. Horses for courses, I think.

I notice that you appear to be getting a bit rattled, when challenged about the actual meat of your claim. Perhaps we should stop before that goes too far. It only gets threads closed - as we have seen recently.