Comparative Study of the Power Handling Capability ...

11 downloads 0 Views 939KB Size Report
5. Ridged-waveguide filter [10]. 6. Waffle-iron filter [13],[14]. Additionally, the most usual structures found in the literature have been used for each filter [8]-[14].
Comparative Study of the Power Handling Capability of Space Broadband Antenna Filters in Ku-band 5th International Workshop on Multipactor, Corona and Passive Intermodulation in Space RF Hardware 12-14 September 2005, ESTEC, Noordwijk, The Netherlands Pablo Sarasa (1), Álvaro González (2), Héctor Esteban (2), Philippe Mader (1), Kokou Tossou (1), Philippe Lepeltier (1) (1)

Space Antenna Department, Alcatel Space 26, Avenue J.-F. Champollion, 31037, Toulouse Cedex 1, France E-mail: [email protected], [email protected], Kokou.Tossou @space.alcatel.fr, Philippe.Lepeltier @space.alcatel.fr (2)

Departamento de Comunicaciones, Universidad Politécnica de Valencia Camino de Vera s/n, 46022, Valencia, Spain E-mail: [email protected], [email protected]

INTRODUCTION In the field of telecommunication satellites, the trend is to increase more and more the EIRP and the number of channels. With the development of high-power spacecraft in the 12 to 20 kW range, the RF power at the antenna input can now reach several kWs. The antenna feed can be realised using different components depending on its system architecture. Usual elements are horns, OMTs (if there are two polarizations), polarizers or hybrid couplers (if circular polarization is required), diplexers (if two or more bands are used), magic-Tees, waveguide bends… Among these components, transmission filters in diplexers or in orthomode junctions have turned out to be one of the most limiting elements in power handling capability. Generally, this is not too critical in case two different antennas are used, one for transmission and the other for reception. As a matter of fact, no diplexer is required in this case and more RF power can go through the feed. Nevertheless, using two different antennas implies a significant cost increase and some potential spacecraft-implementation difficulties. Therefore, a single transmit/receive antenna, using a diplexer to separate the two different frequency bands, is often preferred. This involves the use of multipactor-free and high-power-handling broadband antenna diplexers and filters, topic of this article. These days, Ku is one of the most commercially used frequency bands. In fact, both the FSS (Fixed Satellite Service) and the BSS (Broadcast Satellite Service) use this band. It is also exploited for telecommunication broadband services such as DBS (Direct Broadcast System) or VSAT (Very Small Aperture Terminals) networks. For this reason, service providers demand more transponders in Ku-band, which involves handling a greater amount of RF power in spacecraft. Therefore, multipactor analysis becomes an important issue in the hardware design process in this band. A comparative study of the power handling capability of broadband antenna filters has been carried out in order to know which kind of filter can handle the greatest amount of RF power. The study consists of S-parameter modelisation and full-wave E-field computation in order to perform an accurate multipactor analysis and margin evaluation. For this, the same S-parameter specifications and interface ports have been chosen for all the filters. The frequency band under study is the Ku-band, focusing on the Fixed Satellite Service (FSS) and the Broadcast Satellite Service (BSS) frequencies (from 10.95 GHz to 12.75 GHz for the pass band and from 13.75 GHz to 14.50 GHz for the rejected band). The feed is located out of the repeater thermal protected zone. Hence, the filters under study will go through extreme temperature variations. This will make the scattering-parameter response suffer a significant frequency shift. In this paper, the considered temperature range will be ±150 °C, which will shift the S-parameter response about ±60 MHz. Therefore, a slightly wider band (±100 MHz) has been considered to take into account the thermal effects on the filter response due to the particular operating environment.

The aim of this paper is to analyse some well-known filters and to establish an accurate comparison to determine the best candidate to be implemented inside the antenna feed. The obtained results provide a complete study in antennafilter performance, considering not only power handling capability but also some other relevant features such as insertion loss and size.

DESIGN PROCEDURE AND MULTIPACTION MODELLING Firstly, the design has been carried out using the package WaspNet [1], a hybrid electromagnetic simulator based on several analysis and optimization methods [2],[3]: Mode-Matching (MM), Finite Elements (FE), Method of Moments (MoM) and Finite Differences (FD). In most cases, WaspNet uses the Mode-Matching Method and combines the results with the other methods if necessary. The RF specifications have been met with the minimum total length. Secondly, the performance has been checked out using HFSS [4], a different electromagnetic simulator based on the Finite Elements Method [5]. Therefore, the design process reliability has been proved, since similar performance has been obtained using different methods. In addition, the effect of using real conductors, instead of PEC, has been studied with HFSS for the insertion-loss calculation. Finally, the study of the multipactor effect has been performed using HFSS. The Hatch and Williams model has been used for the multipactor breakdown voltage calculation [6]. The electric field in the structures has been worked out at different frequencies in the chosen pass band. The rejected band is not considered as voltages are much fainter. The voltages in the most critical distances have been calculated at each frequency. After that, the Voltage Magnification Factor (VMF) has been calculated. The VMF is defined as the ratio between voltage along the critical distances (Vd) and the input voltage (VIN). The VMFmax is defined as the highest VMF value at a fixed frequency. The most critical distance can be different at different frequencies.

VMF =

 Vd VMFmax = max  i ∀d i  V IN

Vd VIN

   

(1)

The input voltage can be theoretically calculated as follows:

VIN = 2 ⋅ Z 0 ⋅ PIN

(2)

where PIN is the input RF power and Z0 is the input impedance defined by:

Z0 =

2b 2b λ g ⋅ Z TE = η 0 ⋅ ⋅ a a λ

(3)

where a is the width and b is the height of the waveguide cross-section, ZTE is the fundamental-mode impedance, λg is the wavelength in the input guide, λ is the wavelength in open space and η 0 = µ 0 ε 0 is the vacuum intrinsic impedance. The multipactor breakdown voltage (Vmulti) along distance d has been computed as proposed in [7]. The model has been particularised for the case that silver is used as conductor.

Vmulti [V ] = 63 ⋅ f [GHz ] ⋅ d [mm]

(4)

The threshold input power at each frequency (Pthr) is finally obtained from the highest voltage magnification factor in the filter, the multipactor breakdown voltage and the input impedance:

Pthr ( f ) =

1

[VMFmax ( f )]2

2

V (f) ⋅ multi 2 ⋅ Z0

(5)

The worst case of the threshold input power in the chosen transmission band will be finally the measure which will be used to carry out the comparison of the different antenna filters.

FILTER COMPARISON The band under study is from 10.95 GHz to 12.75 GHz for transmission and from 13.75 GHz to 14.5 GHz for reception. This covers the FSS, both transmission and reception, and the BSS transmission band. The BSS reception band has not been considered for this study. The S-parameter specifications are typical and not very demanding ones, in order not to focus excessively on this feature: Return loss (Tx) > 25 dB and Rejection (Rx) > 70 dB. The chosen access waveguides are the standard WR-75, this is 19.05 mm x 9.525 mm. The insertion losses in these filters are really a key factor in the design due to the conditions they are going to work under. In fact, they will be inside the satellite feed and there will not be any further amplification. Therefore, all the insertion losses the filter adds are going to be a loss to the final system EIRP. Moreover, losses in high-power filters (like the ones under study) increase dramatically the heat to be evacuated from the feed. This could be critical taking into account the antenna environment is out of the thermal control provided by the satellite bus. Hence, the insertion losses must be minimal. The study has been carried out using HFSS. The conductivity σAg and relative permeability µAg considered for silver are σAg = 6.1x107 S/m and µAg = 0.99998 The studied antenna filters are: 1. H-plane filter using irises [8],[9] 2. E-plane filter using stubs [10] 3. H-plane filter using metallic posts [11], [12] 4. E-plane corrugated filter [10] 5. Ridged-waveguide filter [10] 6. Waffle-iron filter [13],[14] Additionally, the most usual structures found in the literature have been used for each filter [8]-[14]. Their geometrical structures are shown in Fig.1. From the point of view of size, there are some remarkable differences between these filters.

2

1

4

3 6 5 Fig. 1. Comparison of the size of the filters under study

H-Plane Filter Using Irises This filter is shown in Fig.1.¬. In order to minimize the geometrical optimization parameters, all the irises have the same thickness and every cavity has constant width and height (standard WR-75). The required S-parameter specifications have been achieved using 15 irises and 14 coupled cavities. This leads to a total length of 199 mm. The insertion loss is rather high in the pass band: 0.29 dB (worst case) and 0.16 dB (upper bound in 80% of the pass band). The worst-case threshold input power is 4490 W and it takes place in the central cavity at 10.95 GHz. The mean threshold input power (over the pass band) is 10200 W. The electric-field cartography at 10.95 GHz is shown in Fig. 2.

(a)

(b)

Fig. 2. (a) Electric fields at 10.95 GHz and (b) worst-case of the VMF and Pthr. In Fig. 2.(a), the electric field is concentrated in the middle of each cavity and the worst case is found in the middle of the filter, this is, in the seventh and in the eighth cavity. The electric field goes through the irises in an evanescent way and it resounds in the cavities. Therefore, each cavity resounds with the coupled field coming from the former cavity.

E-Plane Filter Using Stubs This filter is shown in Fig.1.-. In order to minimize the geometrical optimization parameters, all the waveguides between stubs have the same height, and the width is constant all along the filter. The required S-parameter specifications have been achieved using 6 stubs. A taper is necessary to have WR-75 access waveguides. This leads to a total length of 75 mm. The insertion loss is quite low in the pass band: 0.08 dB (worst case) and 0.05 dB (upper bound in 80% of the pass band). The worst-case threshold input power is 5015 W and it takes place in the central stub at 12.75 GHz. The mean threshold input power (over the pass band) is 13640 W. The electric-field cartography at 12.75 GHz is shown in Fig. 3.

(a)

(b)

Fig. 3. (a) Electric fields at 12.75 GHz and (b) worst-case of the VMF and Pthr. In Fig. 3.(a), the most important feature is the resonance that takes place in every stub.

H-Plane Filter Using Metallic Posts This filter is shown in Fig.1.®. In order to minimize the geometrical optimization parameters, all the posts have the same diameter and the waveguide they are in has constant width and height (standard WR-75). The required Sparameter specifications have been achieved using 15 pairs of metallic posts. This leads to a total length of 203 mm. The insertion loss is rather high in the pass band: 0.32 dB (worst case) and 0.17 dB (upper bound in 80% of the pass band). The worst-case threshold input power is 4620 W and it takes place in the cavity between the two central pair of posts at 10.95 GHz. The mean threshold input power (over the pass band) is 10370 W. The electric-field cartography at 10.95 GHz is shown in Fig. 4.

(a)

(b)

Fig. 4. (a) Electric fields at 10.95 GHz and (b) worst-case of the VMF and Pthr. The field distribution is similar to the one obtained in the filter using irises and so are the rest of results. This is due to the fact that both are based on resonant cavities and the physics beneath this are the same for both of them.

E-Plane Corrugated Filter This filter is shown in Fig.1.¯. In order to minimize the geometrical optimization parameters, width is constant all along the filter. The required S-parameter specifications have been achieved using 7 corrugations. This leads to a total length of 30 mm. The insertion loss is low in the pass band: 0.14 dB (worst case) and 0.07 dB (upper bound in 80% of the pass band). The worst-case threshold input power is 2285 W and it takes place in the central stub at 12.75 GHz. The mean threshold input (over the pass band) is 9210 W. The electric-field cartography at 12.75 GHz is shown in Fig. 5.

(a)

(b)

Fig. 5. (a) Electric fields at 12.75 GHz and (b) worst-case of the VMF and Pthr. This filter presents a quite interesting field distribution. In opposition to the filter using stubs, the corrugations are quite short and there is no resonance in them. In fact, the field concentrates at the beginning of each corrugation and becomes weaker the more you advance through them.

Ridged-Waveguide Filter This filter is shown in Fig.1.°. In order to minimize the geometrical optimization parameters, all the irises, except the first and the last one, have the same height and width, and teeth in them have the same size. Besides, all the cavities have the same height and width. The first and the last irises have different height from the others in order to achieve the VWSR requirements. The required S-parameter specifications have been achieved using 16 irises. A taper is necessary to have WR-75 access waveguides. This leads to a total length of 160 mm. The insertion loss in the pass band is 0.19 dB (worst case) and 0.13 dB (upper bound in 80% of the pass band). The worst-case threshold input power is 920 W and it takes place in the central irises at 12.75 GHz. The mean threshold input power (over the pass band) is 1595 W. The electric-field cartography at 12.75 GHz is shown in Fig. 6.

(a)

(b)

Fig. 6. (a) Electric fields at 12.75 GHz and (b) worst-case of the VMF and Pthr.

The maximum electric field in this kind of filters is located between the teeth in each iris, whereas there is no high resonance in the cavities between irises. This field distribution is not good in terms of multipactor effect, since the electric field concentrates along the shortest distance between metallic plates in the filter. This fact can be observed in Fig.7.

Fig. 7. Electric field distribution in the seventh iris (12.75 GHz)

Waffle-Iron Filter This filter is shown in Fig.1.±. In order to minimize the geometrical optimization parameters, all the irises have 5 teeth and all of them are similar, except the first and the last ones. The first and the last irises differ from the others in the slot height between teeth in order to achieve the VWSR requirements. Besides, all the cavities have the same height and width. The cavity height is the same as the height in the slot between teeth. The required S-parameter specifications have been achieved using 17 irises. A taper is necessary to have WR-75 access waveguides. This leads to a total length of 134 mm. The insertion loss in the pass band is 0.25 dB (worst case) and 0.16 dB (upper bound in 80% of the pass band). The worst-case threshold input power is 1580 W and it takes place in the central teeth in the central iris at 12.75 GHz. The mean threshold input (over the pass band) is 3185 W. The electric-field cartography at 12.75 GHz is shown in Fig. 8.

(a)

(b)

Fig. 8. (a) Electric fields at 12.75 GHz and (b) worst-case of the VMF and Pthr.

The maximum electric field in this kind of filters is located between the teeth in each iris, whereas there is no high resonance in the cavities between irises or in the slots between teeth. This field distribution is not good in terms of multipactor effect, since the electric field concentrates along one of the shortest distances between metallic plates in the filter. This fact can be observed in Fig.9.

Fig. 9. Electric field distribution in the central iris (12.75 GHz)

CONCLUSIONS A complete study of antenna filter performance has been undertaken in Ku-band. No optimization has been tried to be done for the designed filters in order to increase their multipactor power-handling capability. The dimensions of each filter highly depend on the chosen S-parameter specifications. Therefore, a change in them would lead to different power performance. The results provide an accurate comparison considering power-handling capability and other key features such as size and insertion loss for the chosen S-parameters requirements. The results obtained for all these filters are summarised in Table 1. Pthr (W) H-plane filter using irises E-plane filter using stubs H-plane filter using metallic posts E-plane corrugated filter Ridged-waveguide filter Waffle-iron filter

4490 5015 4620 2285 920 1580

Mean Pthr Insertion Loss (dB) (W) Worst-Case 10200 13640 10370 9210 1595 3185

0.29 0.08 0.32 0.14 0.19 0.25

Table 1. Comparative Results

Insertion Loss (dB) Upper Bound in 80% of the pass band

0.16 0.05 0.17 0.07 0.13 0.16

Number of Length cavities (mm) 14 6 14 7 15 16

199 75 203 30 160 134

All this filters present a good power handling capability as they are based on waveguide technology. The electric field distribution and the worst case of the voltage magnification factor versus frequency have been shown for all them. In general, a high number of cavities (poles) leads to high insertion loss as it is shown in Table 1. The E-plane filter using stubs is the one that can handle the highest amount of power. The H-plane filters (using irises or metallic posts) present a really good RF-power performance but have some important handicaps such as a huge size and high insertion loss. The other filters handle a high amount of RF-power but not as much as in the three previous cases. The corrugated filter could be interesting for compact feeds because it is the shortest one. The waffle-iron filter will continue to be used since it is the only filter (in this comparison) that is capable of harmonic rejection. In conclusion, the E-plane filter using stubs has proved to be the filter which suits better all the requirements for Kuband space antennas: small size, low insertion loss and remarkable power handling capability.

REFERENCES [1] WaspNET. MIG Microwave Innovation Group. URL: http://www.mig-germany.com [2] WaspNET. Waveguide Synthesis Program for Waveguide Networks. Tutorial Manual and WaspNET Waveguide Component Design Handbook. pp. 3-10 [3] F. Arndt, R. Beyer, J.M. Reiter, T. Sieverding and T. Wolf. "Automated Design of Waveguide Component using hybrid mode-matching/numerical EM building-blocks in optimization-oriented CAD frameworks—state-of-the-art and recent advances", IEEE Trans. Microwave Theory Tech., vol. 45, pp. 747-760, May 1997 [4] HFSS: 3D High-Frequency Electromagnetic Simulation. Ansoft Corporation. URL: http://www.ansoft.com/products/hf/hfss/ [5] J. Uher, J. Bornemann, U. Rosenberg, "Waveguide Components for Antenna Feed Systems: Theory and CAD",pp. 51-71, Artech House, 1993 [6] A.J. Hatch and H.B. Williams, "The secondary electron resonance mechanism of low-pressure high-frequency gas breakdown", Journal of Applied Physics, vol. 25, pp. 417-423, Apr. 1954 [7] A. Woode and J. Petit, "Diagnostic Investigations into the Multipactor Effect, Susceptibility Zone Measurements and Parameters Affecting a Discharge", ESA Working Paper No. 1556, Noordwijk, The Netherlands, Nov. 1989. [8] Kjetil Folgero and Jan Kocbach, "Yield-Driven Design of Direct-Coupled Waveguide Filters with Minimum Use of Full Wave EM Solvers", 33rd European Microwave Conference, Munich 2003 [9] H. Esteban, J.V. Morro, V.E. Boria, C. Bachiller, B. Gimeno and L. Conde, "Hybrid Full-wave Simulator for the Multipaction Modelling of Low-Cost H-plane Filters", 4th International Workshop on Multipactor, Corona and Passive Intermodulation in Space RF Hardware, 8-11 September 2003, ESTEC, Nordwijk, The Netherlands [10] J. Uher, J. Bornemann, U. Rosenberg, "Waveguide Components for Antenna Feed Systems: Theory and CAD", pp. 185-190 for E-plane filters using stubs, pp. 200-207 for E-plane corrugated filters, pp. 207-212 for ridgedwaveguide filters, Artech House, 1993 [11] P. G. Li, A. T. Adams, Y. Leviatan, and J. Perini, "Multiple post inductive obstacle in rectangular waveguide", IEEE Trans. Microwuve Theory Tech., vol. MTT-32, Apr. 1984. [12] R. Gesche and N. Loechel, “Two cylindrical obstacles in rectangular waveguide-resonances and filter applications,” IEEE Trans. Microwave Theory Tech., vol. 37, pp. 962-968, June 1989. [13] Eugene D. Sharp, "A High-Power Wide-Band Waffle-Iron Filter", IEEE Transactions on Microwave Theory and Techniques, March 1963. [14] J. Caputo, F. Bell, "Waffle-Iron Harmonic Suppression Filter", IEEE Transactions on Microwave Theory and Techniques, September 1965.