The European Southern Observatory (ESO) is an intergovernmental European organization for astronomical research, coordinating many projects. The leading one in radio astronomy is the construction of the Atacama Large Millimeter Array (ALMA),1,2 a project for an interferometer aimed at the study of microwave and sub-millimeter signals coming from space.

ALMA is expected to operate in the 30 to 950 GHz band, subdivided into ten sub-bands. It will comprise sixty-four dual reflector antennas that will be located in the Atacama Desert in northern Chile. This millimeter and sub-millimeter instrument will be very helpful in the study of some major challenges in modern astrophysics: the formation and evolution of galaxies and quasars in the early universe and the formation of stars and planets.

Due to the excellent atmospheric transparency at the site, the efficiency of the receiving system is of utmost importance; hence, the performance of the front-end is critical. As a first step, a cryogenic cooler will be used to reduce as much as possible the noise temperature. The most efficient design tools and fabrication techniques must then be used to guarantee optimal design, high dimensional accuracy and low surface roughness.

In this article, the design and fabrication aspects related to the front-end, horn3 and backing ortho-mode transducer (OMT)4 are presented. The device operates between 84 and 116 GHz, corresponding to sub-band number 3 of the ten sub-bands used to cover the overall ALMA spectrum. While the horn itself has been specifically designed for ALMA, and has already been tested by the Institut de Radioastronomie Millimetrique (IRAM),5 the OMT has not been developed within the ALMA project but the ALMA specifications have been taken as guidelines. The OMT for ALMA has been produced with conventional machining, as reported in an ALMA memo by Wollak, et al.6 The primary aim of this research was indeed to develop a new fabrication technique for the manufacturing of entire high frequency front-ends.

It is worth noticing that even if the described application is very specific, the design tools and technological solutions presented can have a mass-market follow-up in the near future. As a matter of fact, there are already European projects for wireless communication in the 90 GHz range, like the European Space Agency (ESA) Data and Video Interactive Distribution (DAVID) project, which may benefit from these results.

Horn Design

The horn must illuminate the parabolic reflector according to the specification provided by the astro-physicists community. These specifications are reported in Table 1 and are given in terms of beam waist. Typical requirements, related to horn antennas, are usually given in terms of gain, or of an equivalent parameter such as the edge taper, that is the direction from the horn axis at which the radiated power reaches a given level below the maximum. Alternatively, since the horn is a directive source, its main lobe can be modeled as a Gaussian beam. Since Gaussian beams are completely characterized in terms of their waist,7 the latter is the specification given. The other design constraints are given in terms of return loss and cross-polarity level, and attaining such tight specifications over a 30 percent band is far from being trivial. These two latter specifications can be met only by resorting to a circular corrugated horn.

The horn design is then carried out by taking into account the fact that the main beam radiation is controlled by the aperture size and the horn’s overall length. The first part of the corrugated region, the horn throat (see Figure 1), principally determines the input matching (or return loss) and the proper excitation of the fundamental hybrid mode HE.11 It is very important to take care in designing this part and very useful hints are given in the literature.8 The corrugation geometry also governs the cross-polarity level,9 an important parameter if the horn, as in this case, will be used in dual-polarization.

Fig. 1 Circular corrugated horn geometry.

The first approach to the horn design is essentially similar to the method presented in detail by Clarricoats,10 based on the use of formulas and graph charts, given a set of specifications. This procedure almost never leads to a structure whose behavior is rather close to the requirements. To improve the performance, numerical techniques are then necessary. An accurate full-wave solver, based on the Mode Matching (MM) and the Method of Moments (MoM) hybrid technique,11 has been used to refine the first rough design.

Such a refinement can be carried out by a trial and error approach or by applying some sort of optimization technique. Automatic optimization allows for the unsupervised search of the horn that best matches the requirements in the geometrical parameters space. The algorithm can be based on both the Quasi-Newton method12 and Genetic Algorithms,13 suitably combined to prevent local minima problems and speed up the execution time. This approach has been already successfully applied to horn design.14

This procedure has lead to a horn design, which has been presented previously.5 Prototypes were subsequently built by an electroforming process, which is basically the same as the one described below for the OMT. Figure 2 shows the simulated and measured results taken on such prototypes.

Fig. 2 Simulated and measured radiation patterns of two 100 GHz horns; (a) E plane and (b) H plane.

OMT Design

The OMT is based on a dual junction in rectangular waveguide, introduced and accurately described by Bøifot.15 This design requires a thin septum and two matching pins, all of which would be too small at these high frequencies to be fabricated and to be mechanically stable. To make the OMT fabrication possible in the original Bøifot configuration, the design has been revised with the aim to avoid such critical mechanical parts. The approach is quite similar to the one proposed by Narayanan.16 Basically, a thick septum is adopted and the two side arm matching pins are eliminated. Matching is then obtained by using a pair of symmetrical E-plane steps (see Figure 3).

Fig. 3 Dual junction diagram.

Four physical ports are present, corresponding to five electrical ports: a common port CP carrying both the vertical (V, electrical port no. 1) and the horizontal (H, electrical port no. 2) polarization, two side ports (SPL and SPR, ports no. 3 and 4, respectively) that will be recombined in the whole OMT, and a through port (TP, port no. 5). The following scattering matrix gives the ideal electrical behavior of this junction

where ?, ? and ? are phase terms whose value depends on the locations of the electrical port reference planes.

A 0.1 mm thick septum has been used, constituting a finite thickness septum from an electromagnetic point of view. This has a significant influence on the V-polarization channel, mainly in terms of return loss, and a symmetrical double E-plane step is used to control the input matching over approximately a 32 percent bandwidth. The septum shape is designed in order to have a good H-polarization return loss.

The dual junction has been designed using Ansoft HFSS, a finite element method (FEM)-based program able to accurately characterize the electromagnetic performances of such devices. In the design phase, a return loss better than 20 dB was achieved for both polarizations, as shown in the plots of Figure 4, where the operating bandwidth is also highlighted.

Fig. 4 Dual junction design results.

To achieve the complete OMT configuration, the side arm ports need to be recombined by means of H-plane bends and a combiner. Furthermore, output transitions are used, both for the V and H channel, to allow standard WR10 flange connections to the device. The entire OMT is sketched in Figure 5. To design the bends, the combiner and the transitions, different analysis techniques have been used, based on a hybrid mode-matching (MM)-FEM approach.

Fig. 5 OMT sketch.

OMT Fabrication

Electroforming consists in growing the item in a galvanic bath (usually nickel or copper) over a ‘negative’ of the desired shape, usually made of aluminum. The negative is dissolved later with a specific chemical solution and the final device is obtained.

Thanks to its intrinsic ability to reproduce any given shape with extremely high accuracy, it is a very interesting technique for the fabrication of metallic passive microwave components used at frequencies ranging from tens to a few hundred gigahertz (upper microwaves and far infrared).

At such high frequencies, the precision of common tool machining becomes comparable with the wavelength. The advantage of electroforming over direct machining is due to the fact that machining of the “negative” — or mandrel — of small waveguide objects is usually easier (and hence the accuracy is greater) than machining the object itself directly. This is particularly true for the OMT under study, where bends in the waveguides would require, for direct machining, to manufacture several separate parts to be assembled later on. Moreover, as opposed to direct machining, electroforming is a very slow process, giving no stress on the produced device and hence an overall better behavior when the device undergoes thermal and pressure variations, as is the case for cryogenically cooled devices.
Electroforming has already given very good results in the fabrication of microwave devices such as ALMA reflector antennas, designed to work up to 900 GHz.17,18 Although the OMT electroforming procedure requires two steps, a negative machining phase and the electroforming itself, the total procedure is simpler than direct machining and hence attractive from a cost-benefit point of view.

As a first step, a mandrel that is shaped as the inner empty space of the OMT is produced by machining an aluminum piece (see Figure 6). Particular care has been taken to locate accurately the dual junction septum inside the mandrel, since its correct placement is very important for the electromagnetic performance. The gold-plated nickel septum exhibited two very little wings protruding from the mandrel so that it could be attached to the electro-deposited material during the electroforming process.

Fig. 6 OMT aluminum mandrel.

The electroforming process is subdivided into two parts, corresponding to different galvanic baths. First, a thin gold layer is electro-deposited on the aluminum mandrel to ensure good electrical performance; second, a thick nickel layer is electro-deposited to give a strong mechanical support to the device. In the electroforming phase, particular care has been devoted to preserve intact the end-points of the mandrel (two rectangular and one circular), since they will serve as very precise references for later flange alignment. The electroformed OMT is shown in Figure 7, at the end of the nickel-deposition phase, together with the flanges and the related alignment tools. Two flanges are in standard WR10 rectangular waveguide while one is in circular waveguide for a through connection to the corrugated feed horn.

Fig. 7 Electroformed OMT with the mandrel still in place and with flanges and alignment tools.

Once the flanges have been aligned and soldered, the inner aluminum part is dissolved with caustic soda. The final OMT appearance is shown in Figure 8. An OMT prototype, cut to inspect the dual junction septum area, is also shown.

Fig. 8 OMT final devices; (a) sliced part for dual junction inspection and (b) ready for test.


The OMT prototype has been measured. The test setup was based on the Agilent 8757D scalar network analyzer, giving good accuracy measurements up to 110 GHz. As shown in the photograph in Figure 9, the two OMT WR10 rectangular ports were connected to the ports of the instrument while the common port is connected to the feed horn. Since the feed horn is quite a good termination, with a return loss better than 30 dB over the whole band, the scattering parameter measurements directly read on the instrument are a quite accurate estimate of the return loss at the OMT rectangular ports (corresponding to the through port and the recombined side ports of the dual junction) and of the isolation between them. The reflection coefficient curves of the OMT are shown in Figure 10 for the through port and in Figure 11 for the side port, showing performance better than 15 dB over almost the whole band. The isolation between the two input ports is better than 30 dB, as shown in Figure 12.

Fig. 9 OMT test setup.

Fig. 10 OMT through port (V) reflection coefficient.

Fig. 11 OMT side port (H) reflection coefficient.

Fig. 12 OMT through port (V) to side port (H) isolation.


In this article, the design of a front-end receiver at 100 GHz, comprising a horn and an OMT, has been presented and a new design solution for a high frequency OMT has been discussed. Particular emphasis was given to the fabrication technique, which is of prime importance due to the small dimensions of the device parts. The electroforming process adopted to fabricate the prototypes has been discussed. Finally, some measurements on the device have been presented, showing very good performance and agreement with the simulated data.


This activity has been partially supported by the Italian Space Agency (ASI) within contract ASI ARS/99/85. The authors wish to thank Dr. Sergio Mariotti of the IRA-INAF and Dr. Luca Valenziano of the IASF-INAF for their useful help during the OMT testing phase.



2. G. Tofani, “ALMA: The Project Becomes Real,” 47th Annual Meeting of the Italian Astronomical Society, Nuove Frontiere dell’ Astronomia Italiana, Memorie della Società Astronomica Italiana, Trieste, L. Girardi and S. Zaggia, Eds., 14–17 April 2003, Vol. 3, 2003, p. 292.

3. L. Lucci, R. Nesti, G. Pelosi and S. Selleri, “Corrugated Horns,” The Wiley Encyclopedia of RF and Microwave Engineering, March 2005.

4. G.G. Gentili, R. Nesti, G. Pelosi and S. Selleri, “Orthomode Transducers,” The Wiley Encyclopedia of RF and Microwave Engineering, March 2005.

5. P. Bolli, V. Natale, R. Nesti, G. Pelosi, G. Tofani, G. Valsecchi and R. Banham, “100 GHz Corrugated Horn: Design and Fabrication,” 25th ESA Antenna Workshop on Satellite Antenna Technology, ESTEC, Noordwijk, The Netherlands, September 18–20, 2002.


7. P.F. Goldsmith, “Quasi-optical System: Gaussian Beam Quasi-optical Propagation and Applications,” IEEE Press, Piscataway, NJ, 1997.

8. X. Zhang, “Design of Conical Corrugated Feed Horns for Wideband High Frequency Applications,” IEEE Transactions on Microwave Theory and Techniques, Vol. 41, No. 7, August 1993, pp. 1263–1274.

9. A.C. Ludwig, “The Definition of Cross-polarization,” IEEE Transactions on Antennas and Propagation, Vol. 21, 1973, pp. 116–119.

10. P.J.B. Clarricoats and A.D. Olver, Corrugated Horns for Microwave Antennas, Peter Peregrinus Ltd., London, UK, 1984.

11. R. Coccioli, G. Pelosi and R. Ravanelli, “Combined Mode-matching-integral Equation Technique for Feeders Optimization,” Software for Electrical Engineering Analysis and Design, P.P. Silvester, Ed., Computational Mechanics Publications, Southampton, UK, 1996.
12. R. Fletcher, Practical Methods of Optimization, John Wiley & Sons Inc., New York, NY, 1981.

13. D.E. Goldberg, Genetic Algorithms in Search, Optimization and Machine Learning, Addison Wesley Inc., Reading, MA, 1989.

14. L. Lucci, R. Nesti, G. Pelosi and S. Selleri, “Design of an Improved Profiled Corrugated Circular Horn at 320 GHz,” Journal of Electromagnetic Waves and Applications, Vol. 18, No. 3, 2004, pp. 387–396.

15. A.M. Bøifot, E. Lier and T. Shaug-Pettersen, “Simple and Broadband Orthomode Transducer [Antenna Feed],” IEE Proceedings H, Vol. 137-6, December 1990, pp. 396–400.

16. G. Narayanan and N.R. Erickson, “Design and Performance of a Novel Full-waveguide Band Orthomode Transducer,” 13th Symposium on Space THz Technology Digest, Cambridge, MA, 2002.

17. K. van’t Klooster, G. Valsecchi and J. Eder, “A Spin-off Space Technology: Highly Accurate Reflector Panels for a Prototype ALMA Radio Telescope,” International Conference on Antenna Theory and Techniques Digest, Sevastopol, Ukraine, 9–12 September 2003, pp. 611–616.

18. G. Valsecchi, J. Eden, G. Grisoni, C.G.M. van’t Klooster and L. Fanchi, “High Precision Electroformed Nickel Panel Technology for Sub-millimeter Radio Telescope Antennas,” IEEE AP-S/URSI International Symposium Digest, Columbus, OH, June 22–27, 2003, pp. 124–127.

Robert Banham obtained his BSc degree in metallurgy. He worked for British Aerospace in the Commercial Aircraft Division, Advanced Material Development, as a production planner for aircraft assembly from 1989 to 1995. Since then, he has been with Media Lario, where he currently works on electroformed reflectors and mirror plates research and development activities, as responsible for the R&D activities relevant to the production of optical terminals for communications.

Guiseppe Valsecchi obtained his Laurea degree in chemical engineering at the Polytechnic of Milan. From 1990 to 1993, he was production manager at MACDIT for X-ray electroformed mirrors for the Beppo SAX X-ray astronomy satellite. He joined Media Lario in 1993, where he is currently responsible for R&D activities for future applications of electroformed mirror technology and has been project manager for the design, manufacturing and verification of high accuracy panels for the prototype for the Atacama Large Millimeter Array (ALMA).

Leonardo Lucci received his Laurea degree (cum laude) in electronics engineering from the University of Florence in 2001, where he worked until June 2002 under a one-year research grant. From June to September 2002, he was a visiting scientist at the European Space Research and Technology Center (ESTEC) of the European Space Agency (ESA), Noordwijk, The Netherlands. He is currently a PhD student in informatics and telecommunications at the University of Florence.

Guiseppe Pelosi received his Laurea degree in physics (summa cum laude) from the University of Florence in 1976. He has been with the department of electronics and telecommunications at the same university since 1979, where he is currently a full professor. He was a visiting scientist at McGill University, Montreal, Canada, in 1994 and 1995. His current research activity is mainly devoted to the development of numerical procedures in h context of the finite element method, with particular emphasis on microwave and millimeter-wave engineering (antennas, circuits, devices and scattering problems). He is co-author of Finite Elements for Wave Electromagnetics (IEEE Press, 1994), Finite Element Software for Microwave Engineering (Wiley, 1996) and Quick Finite Elements for Electromagnetic Fields (Artech House, 1998).

Stephano Selleri obtained his Laurea degree (cum laude) in electronics engineering and his PhD degree in computer science and telecommunications from the University of Florence in 1992 and 1997, respectively. In 1992, he was a visiting scholar at the University of Michigan, Ann Arbor, MI, in 1992; a visiting scholar at McGill University, Montreal, Canada, in 1994; and a visiting scholar at the Laboratoire d’ Electronique of the University of Nice, Sophia Antipolis, France, in 1997. From February to July 1998, he was a research engineer at the Centre National d’ Etudes des Telecommunications (CNET), France Telecom. He is currently an assistant professor at the University of Florence, where he conducts research on numerical modeling of microwave devices and circuits with particular attention to numerical optimization. He is co-author of Quick Finite Elements for Electromagnetic Fields (Artech House, 1998).

Enzo Natale received his Laurea degree in physics from the University of Torino in 1967. From 1969 to 1988, he worked as a scientist at the Istituto di Ricerca sulle Onde Elletromagnetiche of CNR in Florence. Since 1988, he has been a senior scientist at the Istituto di Radioastronomia Sezione di Firenze. His research has dealt with the observation of diffuse galactic emission in millimetric and submillimetric regions and cosmic background radiation temperature and anisotropy. His current interests include the development of radio receivers and large bandwidth acousto-optical back-ends for millimeter and submillimeter astronomy.

Renzo Nesti received his Laurea degree in electronics engineering from the University of Florence in 1996 and his PhD degree in computer science and telecommunications in 2000. He joined the National Institute for Astrophysics at the Arcetri Astrophysical Observatory in Florence in 1999, where his main activity is in the area of passive microwave devices for radio astronomy receivers. His research interests include numerical methods for the analysis and design of millimeter-wave components.

Gianni Tofani received his degree in electronics engineering in 1964 and his PhD degree in radio astronomy in 1970. From 1964 to 1982, he worked at the Arcetri Astrophysical Observatory as a CNR research scientist, and served as an associate professor of radio astronomy at the University of Florence from 1982 to 1986. He is now senior astronomer and director of the Institute of Radio Astronomy-INAF in Bologna. His interests include radio telescope technologies and astrophysics of the stars formation