| Author |
Message |
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
 15 Jun 2004 22:53 REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi friends,
some days ago I placed a request in the “EDA E-Book, Articles & Specification Requests” forum and there wasn’t any reply. Now I think this forum is a better place for my request.
I'm looking for information on the Orchard technique for pattern antenna synthesis. The method was introduced in this paper form Proceedings of the IEE (not IEEE):
H. J. Orchard, R. S. Elliott, and G. J. Stern, “Optimising the synthesis of shaped beam antenna patterns,” Proc. Inst. Elec. Eng., vol. 132, part H, pp . 63-68, Feb. 1985.
Any information is welcome (documents that describe the method, information about programs that implement it, etc.).
Thanks
Z
|
|
| Back to top |
|
 |
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 05 Jul 2004 23:34 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi Zorro,
some ears ago I envolved in Orcherd-Elliott-Stern array synthesis optimization  . I found that technique very powerful for the following reasons:
- fast
- good pattern shape prediction and control
with the following drawbacks
- it is a "shape synthesis" (no Directivity specification are allowed)
- only equispaced array (and essentially only linea array)
- not element beam (it assumes omnidirectional radiator, but can be easy implemented a version with a common radiator shape)
- fixed ripple in shaped region and lobe level, you don't have any control in transition zone.
essentually this kind of synthesis comes from theory of "digital filter Design" Field (see "Digital Signal Processing" - "Oppenheim-Shafer" pages regarding Parks and McClellan).
Please feel free to ask more about that.
Bye!
|
|
| Back to top |
|
|
satellite
Joined: 29 Mar 2004
Posts: 321
Helped: 7
|
| 06 Jul 2004 11:52 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
| Could you upload the reference paper or some illustrative material?
|
|
| Back to top |
|
|
mamali
Joined: 29 Apr 2004
Posts: 381
Helped: 1
Location: between hell and heaven
|
| 08 Jul 2004 1:40 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
| satellite wrote: |
| Could you upload the reference paper or some illustrative material? |
some interesting papers, tell me if you cant find it:
1 New surface expansion for fast PO synthesis of shaped reflector antennas
Pinsard, B.; Renaud, D.; Diez, H.;
Antennas and Propagation, Tenth International Conference on (Conf. Publ. No. 436) , Volume: 1 , 14-17 April 1997
Pages:25 - 29 vol.1
--------------------------------------------------------------------------------
2 Shaped reflectors for multiple beam applications
Philippou, G.Y.; Adatia, N.;
Multiple Beam Antennas and Beamformers, IEE Colloquium on , 21 Nov 1989
Pages:3/1 - 3/3
--------------------------------------------------------------------------------
3 Sidelobe suppression in shaped reflectors for contour beams
Ramanujam, P.; Tun, S.M.; Adatia, N.A.;
Antennas and Propagation, 1989. ICAP 89., Sixth International Conference on (Conf. Publ. No.301) , 4-7 Apr 1989
Pages:117 - 121 vol.1
--------------------------------------------------------------------------------
4 Optimised synthesis of shaped line-source antenna beams
Ares, F.; Elliott, R.S.; Moreno, E.;
Electronics Letters , Volume: 29 , Issue: 12 , 10 June 1993
Pages:1136 - 1137
--------------------------------------------------------------------------------
5 A novel phase-only method for shaped beam synthesis and adaptive ing
Gatti, R.V.; Marcaccioli, L.; Sorrentino, R.;
Microwave Conference, 2003. 33rd European , Volume: 2 , 7-9 Oct. 2003
Pages:739 - 742 vol.2
--------------------------------------------------------------------------------
6 Smooth aperture distribution synthesis for shaped beam reflector antennas
Westcott, B.S.; Zaporozhets, A.A.; Searle, A.D.;
Electronics Letters , Volume: 29 , Issue: 14 , 8 July 1993
Pages:1275 - 1276
--------------------------------------------------------------------------------
7 Power synthesis of shaped beam antenna patterns
Mangenot, C.; Judasz, T.; Combes, P.F.;
Antennas and Propagation Society International Symposium, 1989. AP-S. Digest , 26-30 June 1989
Pages:420 - 423 vol.1
--------------------------------------------------------------------------------
8 Beam shaping techniques based on analytical gradient iteration procedures
Westcott, B.S.; Zaporozhets, A.A.;
Novel Techniques for Antenna Beam Control, IEE Colloquium on , 16 Jan 1995
Pages:1/1 - 1/6
--------------------------------------------------------------------------------
9 Optimising pattern synthesis using classical variational method
Zhang, X.-z.;
Antennas and Propagation Society International Symposium, 1993. AP-S. Digest , 28 June-2 July 1993
Pages:1582 - 1585 vol.3
--------------------------------------------------------------------------------
10 Phase-only shaped beam synthesis via technique of approximated beam addition
Kautz, G.M.;
Antennas and Propagation, IEEE Transactions on , Volume: 47 , Issue: 5 , May 1999
Pages:887 - 894
--------------------------------------------------------------------------------
11 A method for the optimal pattern synthesis of linear arrays with prescribed nulls
Shpak, D.J.;
Antennas and Propagation, IEEE Transactions on , Volume: 44 , Issue: 3 , March 1996
Pages:286 - 294
--------------------------------------------------------------------------------
12 Optimisation of synthesised array excitations using array polynome complex root swapping and genetic algorithms
Markus, K.; Vaskelainen, L.;
Microwaves, Antennas and Propagation, IEE Proceedings - , Volume: 145 , Issue: 6 , Dec. 1998
Pages:460 - 464
--------------------------------------------------------------------------------
13 Comparison between real and power optimisation methods for arrays synthesis of antennas
Eclercy, D.; Rammal, M.; Reineix, A.; Jecko, B.;
Electronics Letters , Volume: 32 , Issue: 2 , 18 Jan. 1996
Pages:84 - 85
--------------------------------------------------------------------------------
14 Designs of Chebyshev-type complex FIR filters and digital beamformers with linear-phase characteristics
Zhang, X.; Dai, S.;
Vision, Image and Signal Processing, IEE Proceedings- , Volume: 141 , Issue: 1 , Feb. 1994
Pages:2 - 8
--------------------------------------------------------------------------------
15 Array synthesis with excitation constraints
Franceschetti, G.; Mazzarella, G.; Panariello, G.;
Microwaves, Antennas and Propagation [see also IEE Proceedings-Microwaves, Antennas and Propagation], IEE Proceedings H , Volume: l35 , Issue: 6 , Dec. 1988
Pages:400 - 407
--------------------------------------------------------------------------------
16 Synthesis of shaped-beam reflector antenna patterns
Bergmann, J.; Brown, R.C.; Clarricoats, P.J.B.; Zhou, H.;
Microwaves, Antennas and Propagation [see also IEE Proceedings-Microwaves, Antennas and Propagation], IEE Proceedings H , Volume: 135 , Issue: 1 , Feb. 1988
Pages:48 - 53
--------------------------------------------------------------------------------
17 Phased array shaped multi-beam optimization for LEO satellite communications using a genetic algorithm
Sherman, K.N.;
Phased Array Systems and Technology, 2000. Proceedings. 2000 IEEE International Conference on , 21-25 May 2000
Pages:501 - 504
--------------------------------------------------------------------------------
18 Use of modified Gaussian beams to optimize shaped reflectors
Schott, P.; Pascal, O.; Lemaitre, F.;
Antennas and Wireless Propagation Letters , Volume: 2 , Issue: 1 , 2003
Pages:14 - 17
--------------------------------------------------------------------------------
19 Radiation pattern synthesis for arrays of conformal antennas mounted on arbitrarily-shaped three-dimensional platforms using genetic algorithms
Allard, R.J.; Werner, D.H.; Werner, P.L.;
Antennas and Propagation, IEEE Transactions on , Volume: 51 , Issue: 5 , May 2003
Pages:1054 - 1062
--------------------------------------------------------------------------------
20 Linear program synthesis for direct broadcast satellite phased arrays
Richie, J.E.; Kritikos, H.N.;
Antennas and Propagation, IEEE Transactions on , Volume: 36 , Issue: 3 , March 1988
Pages:345 - 348
--------------------------------------------------------------------------------
21 Synthesis of shaped line-source antenna beams using pure real distributions
Ares, F.; Elliott, R.S.; Moreno, E.;
Electronics Letters , Volume: 30 , Issue: 4 , 17 Feb. 1994
Pages:280 - 281
--------------------------------------------------------------------------------
22 Synthesis and analysis of the long range radar transmitting antenna feeding network
Kreczkowski, A.; Rutkowski, T.;
Microwaves, Radar and Wireless Communications. 2000. MIKON-2000. 13th International Conference on , Volume: 2 , 22-24 May 2000
Pages:592 - 596 vol.2
--------------------------------------------------------------------------------
23 An elliptical aperture dual shaped reflector antenna for SNG applications
Skyttemyr, S.A.;
Antennas and Propagation Society International Symposium, 2000. IEEE , Volume: 2 , 16-21 July 2000
Pages:558 - 561 vol.2
--------------------------------------------------------------------------------
24 Acceleration on the synthesis of shaped reflector antennas for contoured beam applications via Gaussian beam approach
Hsi-Tseng Chou; Theunissen, W.; Pathak, P.H.;
Antennas and Propagation Society, 1999. IEEE International Symposium 1999 , Volume: 4 , 11-16 July 1999
Pages:2336 - 2339 vol.4
--------------------------------------------------------------------------------
25 Application of the fractional Fourier transform to the synthesis of low sidelobe pencil shaped beams for nonseparable planar arrays
Buckley, M.J.; Nelson, W.S.;
Antennas and Propagation Society International Symposium, 1998. IEEE , Volume: 2 , 21-26 June 1998
Pages:744 - 747 vol.2
--------------------------------------------------------------------------------
26 An improved method for the design of linear arrays with prescribed nulls
Shpak, D.J.;
OCEANS '95. MTS/IEEE. 'Challenges of Our Changing Global Environment'. Conference Proceedings. , Volume: 2 , 9-12 Oct. 1995
Pages:1303 - 1310 vol.2
--------------------------------------------------------------------------------
27 Synthesis of SAR radiation patterns incorporating mutual coupling by using genetic methods
James, P.;
Antennas and Propagation, 1995. ICAP '95. Ninth International Conference on (Conf. Publ. No. 407) , 4-7 Apr 1995
Pages:383 - 386 vol.1
--------------------------------------------------------------------------------
28 Fast synthesis of shaped reflector antennas for contoured beams
Stirland, S.J.;
Antennas and Propagation, 1993., Eighth International Conference on , 1993
Pages:18 - 21 vol.1
marti
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
| 09 Jul 2004 19:19 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi juppydu,
Thank you for your answer.
I used the Parks-McClellan method for the synthesis of digital filters. I thought that this method could be appropriate for the synthesis of shaped antenna patterns; in the case of nonsymmetrical patterns the complex version should be used. But I encountered some diffuculties that I will try to expose.
The Parks-McClellan method takes a specification of the desired transfer function (both amplitude and phase) in one or more bands. In the case of antenna array synthesis, the desired pattern is usually specified only in amplitude. One can specify an arbitrary phase characteristic, (e.g. linear phase), but this constraint strongly reduces the degrees of freedom for the optimization of the amplitude shape.
I planned to synthetize |H(z)|^2 instead of the desired amplitude |H(z)|, because |H(z)|^2 is linear-phase and for this reason the zeroes are symmetrically placed with respect to the unit circle (or the Shelkunoff circle, using the antenna jargon). This means that if Z0 is a zero, 1/Z0* is a zero as well. Once |H(z)|^2 is synthetized the zeroes could be split between H(z) and H*(z). This could be done with the freedom to choose the more convenient zero (inside or outside the unit circle) of each pair to be assigned to H(z). But the zeroes that lie the unit circle (they are produced in the low-sidelobe regions) cannot be split because they are simple (not double).
So I have some questions:
Is the Orchard-Elliot-Stern procedure related with that I described above?
Is the Orchard-Elliot-Stern procedure based in the Remez exchange algorithm (as Parks-McClellan method)?
How to apply the Parks-McClellan procedure for the beam synthesis taking only amplitude specification?
Best regards
Z
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 10 Jul 2004 15:28 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
very good list of documents, mamali.
anyway Zorro probably is looking for documents on Orchard-Elliott-Stern, I can suggest the following:
"Array Antenna Synthesis", IEEE Antennas and Propagation Society Newsletter, Oct 1985 (part-I)
"Array Antenna Synthesis", IEEE Antennas and Propagation Society Newsletter, Apr 1986 (part-II)
"On Optimising the synhtesis of shaped Beam Antenna Patterns", Advanced Antenna Technology, Edt. Peter Clarricoats
"Extension of the Orchard-Elliott Synthesis Method to Pure-Real Nonsymmetrical Patterns, J.A.Rodriguez, E.Botha, F.Ares, IEEE Trans.,Vol. 45, No 8, Aug 1997.
"A note on the Limitations of Orcherd's Method", F.Ares-Pena, IEEE Antenna's and Propagation Magazine, Vol. 44, No 1, Feb 2002.
also in
"pahsed array" antenna handbook, R.J.Mailloux, Artech House you can find at page 152 the synthesis of Orc. et all.
To Zorro I try to do some answers (I am going to the Beach ! later I hope to reply to the other Zorro topics)
Zorro:
"In the case of antenna array synthesis, the desired pattern is usually specified only in amplitude".
- This is not always true I just envolved in Antenna design with phase pattern constraints.
Zorro:
"One can specify an arbitrary phase characteristic..".
- Obviously if you want optimise a modulus the extra phase constrint is a complication.
Zorro:
"I planned to synthetize |H(z)|^2..".
- It's ok also the Orc. Synthesis is regardin |F|^2, it is apower synthesis.
Zorro:
"This means that if Z0 is a zero, 1/Z0*.."
- I have to check, but in Orch. we have: if Z0 = A x exp(fi) is a root in shaped region then also Z1 = (1/A) x exp(fi), that's is not the same. But peraphs in Parks-McClellan they have differnt meaning (!?).
For the other question I hope to answer later ! Anyway thake a look at the two following reference:
T.W.Parks and J.H.McClellan, "Chebysche Approximation for Nonrecursive Digital Filters with Linear phase", IEEE Trans. Circuit Theory, Vol. CT-19, Mar. 1972, pp 189-194
T.W.Parks and J.H.McClellan, "A program for the design of Linear Phase Finite Impulse Response Filters", IEEE Trans. Audio Electroacoust. , Vol. AU-20, No. 3, Aug. 1972, pp 195-199.
Bye!
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 10 Jul 2004 15:53 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Sorry for my mistake Zorro your sentence:
"This means that if Z0 is a zero, 1/Z0*.." is correct. (* = complex an coniugated !)
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 12 Jul 2004 23:33 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi Zorro.
I try to give you as much help as possible regarding your doubts about Orchard-Elliot-Stern.
"Is the Orchard-Elliot-Stern procedure related with that I described above? ".
I think yes. The Orchard-Elliot-Stern procedure is based on the "equiripple approximation" and I see in the Oppheneim-Shafer that Parks-McClellan worked on this kind of Filter approximation.
"Is the Orchard-Elliot-Stern procedure based in the Remez exchange algorithm (as Parks-McClellan method)? "
Really I don't know the Remez algorithm but in the Orchard paper "Optiomising .." is clearly said "..by using a variant of Remez algorithm."
"How to apply the Parks-McClellan procedure for the beam synthesis taking only amplitude specification? "
I don't know Parks-McClellan algorithm, but in "Oppheniem-Shafer" the algorithm is applied to FIR filters with null phase. The synthesis procedure is described and no phase specification is shown !
If you want I can provide to you some optimized array antenna coefficients by Orch-Ell-St procedure for some shaped beam like flat-top beam and you can verify your results obtained with Parks-McClellan procedure.
PS: do you have the Orchard-Elliott-Stern paper ? all synthesis procedure details are described in that paper.
Bye.
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
| 13 Jul 2004 18:23 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi mamali,
| mamali wrote: |
some interesting papers, tell me if you cant find it:
................
|
thank you for your suggestion and your offer. Please, could you share the papers 4, 13, 15 and 21 of your list?
Thanks
Hi juppydu,
thank you again. I hope you had enjoyed the beach.
Now, I am working with the Mailloux book.
| juppydu wrote: |
If you want I can provide to you some optimized array antenna coefficients by Orch-Ell-St procedure for some shaped beam like flat-top beam and you can verify your results obtained with Parks-McClellan procedure.
|
Thank you juppydu, I will take into account your offer.
| juppydu wrote: |
PS: do you have the Orchard-Elliott-Stern paper ? all synthesis procedure details are described in that paper.
|
I don’t have the Orchard-Elliott-Stern paper. If you have it in electronic version and you could kindly send me the file, it wold be very helpful.
| juppydu wrote: |
"How to apply the Parks-McClellan procedure for the beam synthesis taking only amplitude specification? "
I don't know Parks-McClellan algorithm, but in "Oppheniem-Shafer" the algorithm is applied to FIR filters with null phase. The synthesis procedure is described and no phase specification is shown !
|
N.B. the standard Parks-McClellan method does have a phase specification. In the case of lowpass, hipass, multiband, etc. this specification is linear phase (or null phase w.r.t. the center of the time response). In the case of Hilbert filters it is +/- pi/2, and so on.
By the way, there was an interesting discussion in the Antennas and Propagation Society Newsletter (jun 1988 p.43, oct 1988 p.48, dec 1988 p.28-29, feb 1989 p.35-36, apr 1989 p.55-56, jun 1989 p.49-50) between Elliot and Steyskal regarding Woodward-Lawson vs. Orchard-Elliott-Stern synthesis methods.
Best regards
Z
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 13 Jul 2004 22:36 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi Zorro,
What I think is that a phase zero filter is nothing else (in antenna terms) a linear array with the phase centre at the center of the arrray whereas if the phase centre is at one edge of the array you have a linear phase, it is only a change (time traslation of sample in filtering language) of the reference point. (a linear array is a "spatial filter").
About the Parks-McClellan procedure I have good news. I have just look at my library and I found this:
the book is: "Optimum Array Processing", Part IV, H.L. Van Trees, Wiley-interscience.
At chapter 3.6 "MiniMax Design" the Parks-McClellean Works are recalled, in particular in 3.6.1 the "Alternation Theorem" the minimax criteria is explained (it seems very like to Orch-Ell-St) and then paragraph 3.6.2 is "Parks-McClellan-Rabiner Algorithm" (BINGO!!) in this paragraph step by step the synthesis procedure is described !!
it is described as on Oppenheim-Shafer, anyway in this book is esplicitilly described in trems of Antenna Synthesis.
Your question about the link between Parks-McClellan is not still definitively solved, but for the possibiliy to use the Parks-McClellan procedure for array synthesis the answer is YES.
Let me know any progress in your investigation about the synthesis procedure you are developing !
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
| 16 Jul 2004 22:21 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi juppydu,
| juppydu wrote: |
| What I think is that a phase zero filter is nothing else (in antenna terms) a linear array with the phase centre at the center of the arrray whereas if the phase centre is at one edge of the array you have a linear phase, it is only a change (time traslation of sample in filtering language) of the reference point. (a linear array is a "spatial filter"). |
I agree. In addition, I think that nonlinear relationship between phase and frequency (in filter terms) means that the phase center of the antenna (in antenna terms) changes with the angle.
| juppydu wrote: |
| ... the book is: "Optimum Array Processing", Part IV, H.L. Van Trees, Wiley-interscience. |
Unfortunately I have not part 4 of Van Trees' work. I'm trying to get it.
I'm working on an implemantation of Orch-Ell-St procedure. I hope it will work.
Regards
Z
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 16 Jul 2004 22:48 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi Zorro,
unfortunately this week-end I will be at work  (We are at manufacturing phase of an array entenna !  ) and I cannot be able to send you the article of Orch. (I have only the paper and I have to scan it in electronic format, moreover I am new in the "site" so I have to look if I can and how to upload a file !). I'll try at the begin of next week !
For the phase centre, I agree, a phase centre must be defined "on a specific coverage area" (i.e. within a specific anglular sector) changing the area the same patter can have a different phase centre ! On other words an antenna can have several phase centre !!
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
| 20 Jul 2004 1:11 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi juppydu
I hope this working week-end was succesful for you (did the past week-end at beach compensate this one?)
Thank you for your proposal of scan the paper and send it. It is very appreciated.
Analysing both methods, I noticed a main difference between Parks-McClellan (P-MC) and Orchard-Elliott-Stern (O-E-S) procedures, in addition to the difference that P-MC is a function synthesis and O-E-S is a power synthesis method. That difference is related to the way the specifications are posed and how they are treated by the method.
In the P-MC, there are “specified bands” and “transition bands”. The objective is to reach the best equirriple approximation to the desired response in the specified bands, in the minimax sense.
The specifications are:
- order of the filter (total number of zeroes)
- edges between specified bands and transition bands
- desired response in the specified bands
- relative weights in the specified bands
The resulting ripples are as low as possible in the specified bands.
In the O-E-S, there are only “specified regions”, that are “shaped regions” and “low sidelobe regions”. Note that there is no specification for transition regions. The objective is that the maxima and minima in shaped regions reach a specified level, as well as maxima in the sidelobe regions.
The specifications are:
- edges between regions
- number of zeroes in each region
- desired ripple (in shaped regions) and sidelobe level (in sidelobe regions)
The resulting beamwidths depend on the specified ripples (Cf. p. 157 and fig. 3.14 of Mailloux book.)
In O-E-S, you have to choose how many zeros have to be placed in each region.
P-MC manages the zeros automatically.
In P-MC, for a certain order, you have guarantee of constant ripple in the specified regions, but you don’t know a priori their exact value.
In O-E-S, for a certain order, you have guarentee of the value of the ripple, but you don’t know a priori the beamwidth in which the resulting function stays inside the desired ripple band. And you don’t know the “transitions bands”, defining them as the bands in which the resulting function exits the ripple bands.
(Sure, the above is true if the algorithm converges....)
Regards
Z
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 24 Jul 2004 17:01 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi Zorro,
The past week-end I worked hard but the first test results on the array are quite good  !
Regarding your last post, I agree with you.
In my post of 10-July I said that Orh-El-St is a power synthesis and in the post of 5-July that the transition zone cannot be specified. Ripple and transition zone are not independent !
Anyway Orh-El-St is realted, in my mind, to the theorem of "Alterantion Theorem" as Parks-McCl, both assume to know the number of min e max in shaped and null region. I have to study in deep the Parks-MCl to better put in evidence the common aspects (I hope next-week).
For what concern the "no control" over the transition zone, I think that Orch-Ell-St procedure gives the "Best" transition zone (i.e. the minimum) for fixed ripple in both shaped and lobes region. I think also to classic "Chebyschev synthesis" where "the best" directivity (i.e. narrow beam width) is obtained for fixed level of side lobes.
I can send you an example of Orch-Ell-St, and you, by your code can try to obtain a "Filter" with less transition zone than that obtained by me.
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
| 02 Aug 2004 18:34 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi juppydu
Using the O-E-S procedure, I tried to reproduce the results shown in fig. 3.14 of Mailloux book and in fig. 5 of the Elliot’s article ["Array Antenna Synthesis", IEEE Antennas and Propagation Society Newsletter, Oct 1985 (part-I)]. I assumed that the element spacing is 0.5 wavelenghts.
I got results that seem to be OK but they are not the same as in the figures. Maybe I’m not using the correct specifications (they are not clearly stated in the references I have), or I made some error in the implementation of the method. Could you send me the specifications from the O-E-S paper or an example with specifications?
Thank you
Z
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 03 Aug 2004 19:39 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hello Zorro,
I think the shape is the following:
(cosec(Th-Pi/2))^2 * cos(Th-Pi/2) with
pi = "P-Greek"
Th = elevation angle, varying from 100° to 140° (at 100* the peek, at 140° the end of shaped region).
Be aware that, as you said, you have not control over the transition zone, this means that for fixed side lobe level, you have to try how many zeros are in the shaped region (too few zeros in the shaped region means "reduced" shaped region and too many zeros means "enlarged" shaped region. This is the main drawback of this kind of synthesis !
Let me know your results !
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
| 05 Aug 2004 0:26 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi juppydu,
I corrected a bug and now I got good results, but:
- I don’t get convergence in certain cases (that seem to have a solution after the references I have)
- problem with the extra constant C2: I have to tune a constant “at hand” in order to adjust the sidelobes at the desired level with respect to the peak.
The pattern with cosec()^2*cos() shape (fig. 5.a-c in Elliot’s paper and fig. 3.14 in Mailloux) goes from 90 to 140 degrees; and the flat topped beam (fig. 5.d in Elliot’s paper) from 55 to 125 degrees. In fact, if I’m not confused, the procedure places the fixed zero at the specified bound of the shaped region.
I got convergence for the flat topped beam with 0.5 dB ripple and for the cosec()^2*cos() shaped beam with 1.5 dB and 1dB ripples. But my procedure doesn’t converge for the latter with a ripple of 0.7 dB or less.
Still, it would be very nice to get the O-E-S paper.
Thank you. Regards
Z
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 05 Aug 2004 22:14 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi Zorro,
Zorro:
..In fact, if I’m not confused, the procedure places the fixed zero at the specified bound of the shaped region.
You are wrong in the pattern specification, remember that the shaped zone is "not" the zone between nulls as you suppose! in fig 5 of you have to see the shaped zone (that some times is not reached as expected due to the absence of control in shaped region transition), on other word, with little error, see at the zone between the two "peeks" at the edge of shaped region! The shaped region is defined by the other parameters "ripple", "lobes level", "zeros in shaped region". You can think as follow: synthetise a flat top beam (case I) with only one zero (other parameter as you like) then synthetise the same array with 2 zeros in shaped region (case II). Good! compute the shaped region beamwidth (the zone where the costant reference value is obtained with the requested ripple), now the question: Try to synthetise a flat top beam with the same requirements but beamwidth that is less than case II but more than case I, are you able to do that ? let me know !
Zorro:
flat topped beam (fig. 5.d in Elliot’s paper) from 55 to 125 degrees
For the flat-top beam the elevation angle is ranging within 65° to 115°, and as usual the "true" shaped region could be not as the obtained one!
Zorro:
..problem with the extra constant C2: I have to tune a constant “at hand” in order to adjust the sidelobes at the desired level with respect to the peak..
OK, this point was a truble to me too. At University time I left this point unsolved (at hand as you), but now is time to fix this point (as soon as possible).
Zorro:
..I don’t get convergence in certain cases (that seem to have a solution after the references I have) ..
What do you mean with "after" ?
..I got convergence for the flat topped beam with 0.5 dB ripple and for the cosec()^2*cos() shaped beam with 1.5 dB and 1dB ripples. But my procedure doesn’t converge for the latter with a ripple of 0.7 dB or less..
You have to tune the code since the procedure is converging always (a part numerical errors when you put too stringent constraints). Remember to put more zeros in the shaped region when you ask to the code more severe ripple requirements (below 0.2dB).
Zorro:
Still, it would be very nice to get the O-E-S paper.
Yes, I working very hard these days, no holydays this year , starting from 8.00a.m. to 10.00p.m or more every day. Moreover my "old" paper is in very bad conditions, let me see how convert the paper in --- electronic format.
see you later!
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
| 12 Aug 2004 15:44 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hello juppydu,
Sorry for the delay in answering.
| juppydu wrote: |
Zorro:
In fact, if I’m not confused, the procedure places the fixed zero at the specified bound of the shaped region.
You are wrong in the pattern specification, remember that the shaped zone is "not" the zone between nulls as you suppose! in fig 5 of you have to see the shaped zone (that some times is not reached as expected due to the absence of control in shaped region transition), on other word, with little error, see at the zone between the two "peeks" at the edge of shaped region! The shaped region is defined by the other parameters "ripple", "lobes level", "zeros in shaped region". |
I agree with you in what is the “useful” or “true” shaped region. But what is called “shaped region” and what is called “controlled sidelobes region” by Orchard? I followed what Mailloux states in p. 155:
“For the purpose of ordering the pattern zeros, and with no loss in generality, the shaped beam edge (region I) is arranged to end at psi=pi (as in the figure)”.
And the fixed zero (Nth root) is at w=-1+j*0=exp(0+j*pi).
But the abovementioned figure is fig. 3.13, in which the region marked as “Region I” doesn’t end at pi (!). [And there are some typographical errors].
My implementation places the fixed zero at a given angle, that is the theta specified as the bound between the two regions. The other zeros are placed in such a way that the specified number of ripples are obtained. For example, for the flat top beam of fig. 5d (Elliot’s paper), I placed the fixed zero for a null at 55°. The algorithm puts six zeros (the required N1) outside the unit circle in the places needed for having the peaks and valleys at +/-0.5 dB; four zeros are put on the unit circle for the four sidelobes at -30 dB, and four more zeros for the sidelobes at -20 dB.
Really, I don’t use the other bound of the shaped region! As you said, The shaped region is defined by the other parameters "ripple", "lobes level", "zeros in shaped region".
| juppydu wrote: |
| You can think as follow: synthetise a flat top beam (case I) with only one zero (other parameter as you like) then synthetise the same array with 2 zeros in shaped region (case II). Good! compute the shaped region beamwidth (the zone where the costant reference value is obtained with the requested ripple), now the question: Try to synthetise a flat top beam with the same requirements but beamwidth that is less than case II but more than case I, are you able to do that ? let me know ! |
I’m not able because of the abovementioned reasons. The width of the shaped region results as a consequence of the selection of the other parametes. The same happens with the “controlled sidelobe” region. Again we can say that there is no control over the transition regions.
| juppydu wrote: |
Zorro:
flat topped beam (fig. 5.d in Elliot’s paper) from 55 to 125 degrees
For the flat-top beam the elevation angle is ranging within 65° to 115°, and as usual the "true" shaped region could be not as the obtained one! |
Forcing the fixed zero for a null at 55°, I obtain a beamwidth (between the points of -0.5 dB) ranging between 63.7 and 118.5 dB. In order to center it, I have to specify the fixed zero at -53.81 degrees; the beam at -0.5 dB results at (90+/-27.39) degrees.
| juppydu wrote: |
Zorro:
..I don’t get convergence in certain cases (that seem to have a solution after the references I have) ..
What do you mean with "after" ? |
Sorry for my bad English. I meant that the bibliography I had available show solutions for that cases.
| juppydu wrote: |
I got convergence for the flat topped beam with 0.5 dB ripple and for the cosec()^2*cos() shaped beam with 1.5 dB and 1dB ripples. But my procedure doesn’t converge for the latter with a ripple of 0.7 dB or less..
You have to tune the code since the procedure is converging always (a part numerical errors when you put too stringent constraints). Remember to put more zeros in the shaped region when you ask to the code more severe ripple requirements (below 0.2dB). |
OK. I think I have some issue with the convergence of Newton method with the cosec^2()*cos() beam.
| juppydu wrote: |
Zorro:
Still, it would be very nice to get the O-E-S paper.
Yes, I working very hard these days, no holydays this year , starting from 8.00a.m. to 10.00p.m or more every day. Moreover my "old" paper is in very bad conditions, let me see how convert the paper in --- electronic format.
see you later! |
Thank you!
Z
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 14 Sep 2004 22:38 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi Zorro, no news ?
I have finally finished my code for OES synthesis , all works fine now.
My results are better then the result of OES, I thnik because in the 1985 the code of OES was less precise than that of acutal PC!
before the end of this week I upload the executable code for your fun and I will wait your comments.
Stay tuned !
|
|
| Back to top |
|
|
mamali
Joined: 29 Apr 2004
Posts: 381
Helped: 1
Location: between hell and heaven
|
| 15 Sep 2004 11:09 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
| zorro wrote: |
Hi mamali,
| mamali wrote: |
some interesting papers, tell me if you cant find it:
................
|
thank you for your suggestion and your offer. Please, could you share the papers 4, 13, 15 and 21 of your list?
Thanks
Z |
hi friends,
sorry if i am late, i was a little bussy with my thesis. well, here's what you wanted; by the way, i should consider the discussions, it seems interesting  .
marti
|
|
| Back to top |
|
|
mamali
Joined: 29 Apr 2004
Posts: 381
Helped: 1
Location: between hell and heaven
|
| 15 Sep 2004 12:08 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
these i found also practical:
-1 F. Ares-Pena, "A Note on the Limitations of Orchard's Method," IEEE Antennas and Propagation Magazine, Vol. 44, No. 1, February 2002, p.109.
0 J. A. Rodriguez, E. Botha, and F. Ares, "Extension of the Orchard-Elliott Synthesis Method to Pure-Real Nonsymmetrical-Shaped Patterns," IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 45, NO. 8, AUGUST 1997, pp. 1317-1318.
1 M. J. Buckley, "Synthesis of shaped beam antenna patterns using implicitly constrained current elements," IEEE Trans. Antennas Propagat.,
vol. 44, pp. 192-197, Feb. 1996.
2 D. Eclercy, M. Rammal, A. Reinex, and B. Jecko, "Comparison between real and power optimization methods for arrays synthesis antennas," Electron. Lett., vol. 32, no. 2, pp. 84-85, Jan. 1996.
3 Ares, F.; Rengarajan, S.R.; Moreno, E., "Remarks on comparison between real and power optimisation methods for arrays synthesis of antennas," Electron. Lett., vol. 32, no. 15, Page(s): 1338-1339
sorry for wiered numbering!, wish be helpfull,
marti
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
| 30 Sep 2004 21:05 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hi friends,
unfortunately I had other engagements in the last times, but I’m here again.
Juppydu, many thanks for your offer; it is welcome!
Do you get better results than OES? In which sense?
Marti, thank you for the papers. It’s a very good and helpful collection!
By the way, I will pose a question.
It is clear the reason for csc^2(theta) pattern (to obtain a constant height radar coverage). But patterns with shape csc^2(theta)*cos(theta) or even csc^2(theta)*sqrt(cos(theta)) are presented in the papers. Which are their applications? Is it related with the dependence of radar cross section of planes with elevation angle, as seen from the radar?
See you. Best regards
Z
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 06 Oct 2004 0:04 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hello Zorro.
I mean that my soution is little more accurate than the solution of Orchard (as reported in the article). Anyway more probably some differnce are due only to a different level of precision imposed during
the synthesis Loop (so the solution are both quite good).
the test case is
(16 elements, 0.5 spacing, Cosec2*Cos (between 100° and 140°), ripple -3dB, lobe -20dB, four lobe -30dB)
ORCHARD Juppydu
Amp Ph Amp Ph
0.77 177.1 0.77 177.5
0.50 -89.2 0.49 -88.6
0.38 -76.0 0.38 -76.1
0.56 -88.3 0.56 -88.1
0.76 -38.1 0.75 -37.9
0.63 7.7 0.62 7.9
0.56 -5.0 0.56 -4.8
0.99 19.0 0.98 19.1
1.04 66.8 1.04 67.0
0.81 94.1 0.81 94.3
1.03 96.9 1.02 97.0
1.47 132.2 1.46 132.3
1.66 -176.8 1.65 -176.6
1.64 -126.1 1.63 -126.0
1.17 -76.4 1.16 -76.4
1.00 0.0 1.00 0.0
Anyway the discussion in this forum was the cause to finish my Orchard synthesis program developed by me at the university time.
I have uploaded an Executable file for You ! it is very simple to use. It is a Windows GUI application
completely written by me !
For the very first use follow my istruction
1- Run the program (double click on the exe file)
2- Click "OK" button confirming the default synthesis settings.
3- in the new page that you see click "Run" button, You see the synthesis in action !!
4- then go to the 3 edit box at the bottom right and typing
1 in initial box, 4 in end box and -30 in new value box
finally Click on "Set New SPec" button, and then click on "Run" button !!
you see the synthesis as in the reference of Orchard !
5- To be sure that all works as Orchard said, click on "Load Reference" button, you see the Pattern computed by the Orchard coefficients (as in the previous table). Uncheck the CheckBox to hide the marker to better compare the two patterns.
Try all the other pattern type and/or synthesis parameter, the program is able to prevent wrong inputs in many cases (probably except pathological ones !).
You can use the program as you like ! also the other people in Elektroda can do that,
I will wait for any suggestions from you and the others !
good Luck.
PS: I still working very hard and this program has been built during the nights (close to midnight),
so some error in coding are possible ! Sorry for that !
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 06 Oct 2004 0:14 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Probably in the last message I wasn't able to upload the file .
Now I try again !
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 319
Helped: 34
|
| 06 Oct 2004 20:40 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
| juppydu wrote: |
Hello Zorro.
I will wait for any suggestions from you and the others ! |
Thank you!
I ran your program. It seems to work nice. I will try to give you some suggestions.
First of all, Can the coefficients (element excitations) be recovered from the program?
Regards
Z
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 02 Nov 2004 23:41 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Sure ! it can
anyway I have discovered a little bug and I have just corrected it !
Now that I am in deep into Orchard Synthesis I can say that a problem with the code implementation, as suggested by Orchard, is the follow:
The actual desogn fixes the position of the Main Peak at one extrema of the shaped region. This causes two drawbacks:
1. part of the shaped region was not covered as desidered (or not as desidered in real world application)
2. only coverage area with the maximum at the edge of the shaped region can be synthetised
I am actually working on to remove theis limitation
see you the next time !
|
|
| Back to top |
|
|
juppydu
Joined: 05 Jul 2004
Posts: 60
Helped: 11
|
| 19 Feb 2006 22:11 Re: REQ: Antenna pattern synthesis: Orchard method |
|
|
|
Hello,
I have just uploaded a new release of my code !
I would like to known your comments !
PS: there is a "first" help file included, it must be in the same folder of main program. Follow the "eample 1" in the help file before start to use the program, anyway for any questions feel free to contact me (see also my web site)
NB: this program vrsion is limited to beam like "CosecCos" as in the Orchard paper, but is not limite in number of radiating elemets and the optimised excitations are available.
|
|
| Back to top |
|
|
jyfx
Joined: 07 Jun 2004
Posts: 24
|
| 27 Feb 2006 10:09 REQ: Antenna pattern synthesis: Orchard method |
|
|
|
| good
|
|
| Back to top |
|
|
akedar
Joined: 27 Jul 2004
Posts: 251
Helped: 13
|
| 28 Feb 2006 20:08 REQ: Antenna pattern synthesis: Orchard method |
|
|
|
your programme is really good but how to calculate amplitude and phases of elements from your program
reply urgently
Added after 35 seconds:
your programme is really good but how to calculate amplitude and phases of elements from your program
reply urgently on this mail id
ashutosh.kedar(at)gmail.com
|
|
| Back to top |
|
|
