диафрагмированные волноводные фильтры / 2c40bfdd-e6bc-463f-ad4e-10719fc415cf

Figure 3 BER vs. the number of simultaneous users (N 29, L 841,29)

termines the necessary bandwidth, is analyzed under the amplifiergain condition G 30 dB.

It can be seen that a BER lower than 10 4 is achieved for standard access, low-power level, and low-cost filtering conditions with Rb 155 Mbits/s, P1 54 dBm (before amplification), and Bo/Be 100). This value can easily improved by using low-cost forward correcting codes (FECs), such as the Reed– Solomon FEC [11].

In Figure 3, the effect of multiaccess users on the BER is depicted. The graphs establish the relation between the MAI and optical noise. The MAI effects are limited by the noise introduced during the transmission via the optical channel, and the upper bound of the associated error probability is limited by the error probability of the optical channel. Indeed, the BER that results from 29 simultaneous users is identical to the BER obtained for an equiprobable optical sequence, in the same physical condition, without using CDMA. Thus, in our optical model, it is demonstrated that the effects of the MAI on the BER are conditioned and limited by the signal distortions generated by optical noise.

5. CONCLUSION

The performances of an optical CDMA system associated with a Gaussian model of an optical amplified channel have been presented. The results demonstrate that appropriate time-domain optical codes (PSs) can be used to reduce and control the effect of multiuser interference.

An analytic expression of the error probability has been developed. This formulation includes the most important parameters of the optical channel, and arises the effect of filtering conditions by pointing out the benefit of a low optical-to-electrical bandwidth ratio on the resultant BER. The simulations carried out thereafter confirm the theoretical expression of the error probability.

On the other hand, our model simultaneously takes into account the MAI and optical noise generated by the optical channel, and demonstrates that the effect of the MAI and optical codes on error probability is directly related to the optical noise introduced by the components of the optical channel.

REFERENCES

1.H.M.H. Shalaby, A performance analysis of optical overlapping PPMOCDMA communication systems, J Lightwave Technol 17 (1999), 426 – 432.

2.L.B. Nelson and H. Vincent Poor, Performance of multiuser detection for optical CDMA part I: Error probabilities, IEEE Trans Commun 43 (1995), 2803–2811.

3.L.B. Nelson and H. Vincent Poor, Performance of multiuser detection for optical CDMA part II: Asymptotic analysis, IEEE Trans Commun 43 (1995), 3015–3024.

4.M. Brandt-Pearce and B. Aazhang, Performance analysis of single user and multiuser detectors for optical code division multiple access com-

munication systems, IEEE Trans Commun 43 (1995), 435– 444.

5.J.A. Salehi, Code division multiple access: Technique in optical fiber network part I: Fundamental principles, IEEE Trans Commun 37 (1989), 824 – 833.

6.A. Keshavarzian and J.A. Salehi, Optical orthogonal code: Acquisition in fiber optic CDMA system via the simple serial search method, IEEE Trans Commun 50 (2002), 473– 483.

7.J.G. Zhang and G. Picchi, Tunable prime code encoder/decoder for all optical CDMA applications, Electron Lett 29 (1993), 1211–1212.

8.C. Hong and G. Yang, Concatenated prime codes, IEEE Commun Lett 3 (1999), 260 –262.

9.M. Lourdiane, R. Vallet, P. Gallion, and E. Bridoux, Perspectives of optical CDMA, Europnet Proc, Paris, France, 2002, pp 200 –208.

10.P. Gallion, Basics of digital optical communication, Undersea fiber systems, J. Chesnoy (Editor), Academic Press, New York, 2002, pp 51–93.

11.J.G. Proakis, Digital communications, 3rd ed., McGraw-Hill, New York, 1995, pp 464 – 466.

©2004 Wiley Periodicals, Inc.

COMPACT RIDGED-WAVEGUIDE BANDPASS FILTERS AND DIPLEXERS

Aleksandr Shelkovnikov,1 George Goussetis,2 and

Djuradj Budimir1

1 Wireless Communications Research Group

Department of Electronic Systems

University of Westminster

115 New Cavendish Street

London, W1W 6UW, UK

2 Dept. of Electronic and Electrical Engineering

Loughborough University

Loughborough, LE11 3TU, UK

Received 29 November 2003

ABSTRACT: A ridged-waveguide diplexer incorporating integrated compact ridged-waveguide filters and a ridged-waveguide T-junction is presented. The compact size of the filter is achieved by integrating a bandpass periodic filter, and a low-pass structure is used in order to suppress spurious responses. Simulation using the mode-matching method has been carried out for ridged-waveguide filters and T junctions and the results are presented. The measured results are in good agreement with the simulated results. © 2004 Wiley Periodicals, Inc. Microwave Opt Technol Lett 41: 465– 467, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop. 20173

Key words: ridged waveguide; periodic structure; half-wavelength resonators; filters; diplexers

1. INTRODUCTION

Recent developments in digital-communication systems have led to increased utilization of microwave and mm-wave technologies.

A technique to integrate low-pass and bandpass filters in an E-plane waveguide filter was found to improve the stopband performance and significantly suppress the spurious responses [1]. The fabrication process supports the realization of low-cost mass

Figure 1 Configuration of the designed filter

production of these microwave configurations. The high demand for low-cost, low-loss, and compact filters and diplexers has been the reason for further investigations on this subject.

The properties of slow waves in half-wavelength resonators are used in this novel configuration of a bandpass filter. This provides an advantage of the physical dimensions of the filter, and spurious behaviour suppression is achieved by means of an integrated low-pass structure. The filter is further incorporated into the diplexer.

2. CONFIGURATION

The homogeneous section of a rectangular waveguide in the resonators of E-plane filters is replaced with a periodic structure, consisting of a cascade of ridged waveguides with different gaps. This significantly reduces the size of the filter.

The layout of the proposed configuration for a two-resonator filter is shown in Figure 1. It consists of low-pass and bandpass components. The latter is designed so that it satisfies the specification for the desired passband and selectivity, while the former has a cut-off frequency that is close to but does not reach the spurious resonance of the bandpass component, in order to suppress it [2].

The bandpass-component configuration uses a cascade of equal lengths of ridged waveguides with different gaps to form the resonant section, instead of a homogeneous waveguide of length

Figure 2 (a) Cross section of a double ridge waveguide T-junction (3PMM method); (b) configuration of the diplexer

Lr between the two septa of length Ls, as in the standard directcoupled half-wavelength E-plane resonator filter. This establishes the periodic boundary conditions required for slow-wave propagation within the resonators. The less-guided wavelength of a periodic resonator, compared to a homogeneous waveguide, leads to the reduced size of half-wavelength resonators and results in reduction of the filter size. Another reason for this is that the improved selectivity of a periodic filter, due to the dispersion relation of the slow waves, is such that the same out-of-band specifications can be achieved with a lower-order periodic filter instead of a homogeneous filter. The stepped-impedance ridgedwaveguide configuration is used as low-pass structure (see Fig. 1). This is a version of the corrugated waveguide with thin ridges with low-pass characteristics.

The distance between bandpass and low-pass components provided is at least g/4, which is sufficient to avoid higher-order mode coupling. The bandpass component is essentially a directcoupled half-wavelength resonator filter. Hence, in order to apply this design procedure, the propagation characteristics of the slow wave (mainly the guided wavelength) need to be determined numerically. The low-pass component is designed so that its cutoff is prior to the spurious harmonic passband of the bandpass component.

The ridge waveguide T-junction, as shown in Figure 2(a), is obtained by application of the three-plane mode-matching technique (3PMM) [3]. The T-junction has its perpendicular arm shorted at three different planes and the remaining two-port structure, a cascade of five ridge waveguide sections, is simulated using the mode-matching method. The three-port scattering matrix can then be determined by the mathematical manipulation of the three two-port scattering matrices.

In order to get the diplexer’s S parameters, the S matrices of the filters were combined with the T-junction’s S matrices. The configuration shown in Figure 2(b) is used, in which the common port of the diplexer is on the perpendicular arm of the T-junction [4, 5]. The diplexer design aims to generate an open circuit at channel 2, a port at the passband of channel 1, and vice versa. The two filters

Figure 3 Simulated response of the low-pass component

are designed separately at the required bands and the distances between the T-junction and the filters are optimized in order to achieve acceptable reflections at all ports.

3. SIMULATED AND EXPERIMENTAL RESULTS

A fifth-order low-pass prototype has been designed and fabricated in order to demonstrate the feasibility of the low-pass component and the accuracy of developed mode-matching simulator. The dimensions are given in Table 1.

Figure 3 shows the measured and simulated results. Mode matching with 20 TE and 12 TM modes has been used. Good low-pass performance is demonstrated. Furthermore, good agreement between the theoretical and experimental results is observed, thus verifying the accuracy of the developed tool.

In order to demonstrate the performance of the proposed integrated filter, a two-resonator X-band filter has been chosen as an example with 0.5-dB ripple, passband between 8.4 GHz and 9.0 GHz, and 15-dB rejection at 9.5 GHz. The dimensions of the designed filter are given in Table 1. A prototype has been fabricated and measured on a network analyzer. The measured response is shown in Figure 4. The improved stopband performance of the proposed periodic filter is evident from Figure 4. Leaving 17.00 mm between the two components brings the total length of the integrated filter to 71.40 mm.

In order to demonstrate the compactness of the proposed structure, a comparison with an equivalent structure containing a standard E-plane bandpass filter is made. In order to satisfy the upper stopband specification (15-dB loss at 9.5 GHz), a standard E-plane filter with the same passband and ripple should be of the third order, with a total length of 67.50 mm. Its simulated response is also shown in Figure 4. Assuming same size for the corresponding

Figure 4 Measured response of the fabricated prototype (dimensions as in Table 1)

Figure 5 Simulated response of a ridged-waveguide diplexer with center frequencies at 8.75 and 10.75 GHz

lowpass component and same distance between the two components, the total length of an integrated filter with suppressed spurious passband would be 116.50 mm. A size reduction of more than 40% is achieved.

To illustrate an example of using the designed filters and E-plane ridge waveguide T-junction in the diplexer configuration, a double ridged-waveguide diplexer was designed. The insertion loss of a double ridged-waveguide diplexer with center frequencies at 8.75 and 10.75 GHz and 500-MHz channel bandwidth is shown in Figure 5.

4. CONCLUSION

Compact ridged-waveguide filters and diplexers have been presented. An integrated compact X-band ridged-waveguide bandpass filter has been designed, fabricated, and tested, and comparable advantages are achieved by using the properties of periodic structures in the resonators of the bandpass filters and integration with a low-pass component in order to suppress resonance. The proposed structure is low cost and mass producible. The filter and T-junction simulation was carried out using the mode-matching method and three plane-mode matching techniques. Numerical and experimental results for the filters have been presented with further incorporation into the diplexer at the X-band.

REFERENCES

1.G. Goussetis and D. Budimir, Integration of low-pass filters in bandpass filters for stopband improvement, Proc Euro Microwave Conf, Milan, Italy, 2002.

2.C. Quendo, E. Rius, C. Person, and M. Ney, Integration of optimised low-pass filters in a bandpass filter for out-of-band improvement, IEEE Trans Microwave Theory Tech 49 (2001), 2376 –2383.

3.G. Goussetis and D. Budimir, Compact ridged waveguide filters with improved stopband performance, IEEE MTT-S Int Microwave Symp Dig, Philadelphia, PA, 2003.

4.J. Uher, J. Bornemann, and U. Rosenberg, Waveguide components for antenna-fed systems: Theory and CAD, ch. 3.4, Artech House, Boston, 1993.

5.A. Morini, T. Rozzi, and M. Morelli, New formulae for the initial design in the optimization of T-junction manifold multiplexers, IEEE MTT-S Int Microwave Symp Dig 2 (1997), 1025–1028.

© 2004 Wiley Periodicals, Inc.