Design and Optimization of Butterworth and Elliptic Band Pass Filters in 5G Application

Filters are two-port networks that may pass or attenuate frequencies within defined ranges and can alter the frequency response of any system. In the present research study, the goals and optimization controller embedded with (ADS2019) are used to design Butterworth and Elliptic bandpass filters with frequency ranges of (18GHz – 38GHz), a bandwidth of (7GHz), stopband attenuation of (S21=-60dB), and passband attenuation of (S21=-1dB). Three types of each filter (Hp-Lp 6th order – 3rd order – 6th order) are simulated and optimized to choose the best (C, L) values. The selected filters are redesigned using the Design Filter Guide, and the simulation during this phase yields different values for (C, L). The designed circuit is then transformed into a microstrip model using transmission lines for open and short circuits. The study investigates the differences between each filter in BW-f center- attenuation at the stopband. In the last phase of the study, the circuit of each filter is transformed using a microstrip transmission line to obtain the (W, L) for each component of each filter. Finally, the study compares past studies and research projects in this field.


INTRODUCTION
Filters are selective networks that separate the required signals from a mixture of signals [1]. They are electric circuits designed to process (attenuate, amplify, or reshape) signals. In any communication system, filters are used for noise extinction to isolate communication signal bands from many [2]. In communication/radar systems, filters have essential roles in rejecting unwanted frequency ranges and canceling noise and interference [3].

PREVIOUS WORK
In [2006], a 2.4GHz bandpass filter was designed, fabricated, and tested using microstrip technology. The parallel coupling lines filter topology simulated, analyzed, and verified the obtained results. A comparative analysis was later used to optimize the acquired experimental and analytical data. Other software tools such as MatLab, CorelDraw 12, and Microwave Office were later used to enhance those results [4]. In [2010], the C++ program was used to design a Butterworth passive low-pass filter with n th order using modulating techniques. A passive synthesis network was implemented to develop such a filter of any order [1]. In [2011], the T-shaped patches-based filter was proposed. The T-shaped patches and folded open stub. The designed filter has a cut off frequency at (2.37GHz) (-3dB) and a (2.44GHz) (-40dB) rejection. The return loss was more significant than (-14.5 dB), and the insertion loss was less than ( 0.164dB) [5].
In [2012], the anti-aliasing filter was designed at (2GHz -4GHz) S-Band with a center frequency of (2.491GHz). The lambed component could not be used to achieve the Al-Rafidain Engineering Journal (AREJ) Vol.27, No.2, September 2022,pp.68-81 minimum aliasing interface at the desired enter frequency due to its distributed effects at this band. Alternatively, the optimization process was used [6].
In [2012], the low-pass filter was designed, simulated, and manufactured using a microstrip line in different software with a cut-off frequency at (1.8GHz) and attenuation of (-26dB) at stopband (4.6GHz).
The features of the S-parameter of the lowpass filter were also studied and compared [7]. In [2013], a bandpass filter of (4.6GHz -4.8GHz) bandwidth was designed using the ADS software tool and cascading a low pass filter for (4.8GHz) and a high pass filter for (4.6GHz) [8]. In [2014], the structure and design of a microstrip lowpass filter were analyzed through circuit simulation and layout simulation using Richard's transformation and Kuroda rule description. The filter design and simulation were carried out based on the ADS 2011 software platform [9]. In [2014], a stepped impedance filter was designed and optimized at a center frequency of 5 GHz to operate between 4 GHz and 6 GHz frequencies. The designed microstrip low pass filter was used to attenuate microwave frequency signals beyond the cutoff frequency of 5GHz and stopband attenuation of (32dB) [10]. In [2014], a low pass filter was proposed using the maximally flat Butterworth Technique, and the filter's seventh order was realized on a microstrip transmission line using MatLab and (AWR) software [3]. In [2015], the microstrip transmission line topology was used to design and implement a J-band symmetric coupled line pass band Butterworth filter. The structure of the designed filter was developed using the ADS. The physical features of the designed filter's resonators were optimized to have the desired response. The Network Analyzer was used to measure the insertion loss of the optimized filter [11]. In [2016], a Butterworth filter up to 8 th and 9 th order was designed using the combination of 2 nd and 3 rd order [12]. In [2017], a dualband passband filter was designed for the GSM application. Two frequency bands were used to pass the GSM1800 and GSM 900 signals. The first frequency band started from 1812.5 MHz to -1887.5 MHz with a bandwidth of 75MHz and a center frequency of 1850 MHz. The second frequency band ranged from 962.5 MHz to 937.5 MHz with a 25MHz bandwidth and a center frequency of 950 MHz. The designed dual bandpass filter began with a center frequency of 950 MHz -1850 MHz, respectively. The return loss of the combining filter was S11<-10dB), and the insertion loss was S21>-3dB [13]. In [2018 ], a complete procedure was adopted to develop, design, and simulate a microstrip bandpass filter at a center frequency of 5.25 GHz with lower and upper cut-off frequencies as 5 GHz and 5.5 GHz, respectively. In this work, the design of the low pass filter prototype was explained. The impedance and frequency scaling was then performed to achieve a bandpass filter. The bandpass filter was further designed using the lumped components (L & C), ideal microstrip lines, and practical microstrip lines. Finally, the microstrip layout version was also presented [14]. In [2019], a 5G band pass filter was designed and simulated to support a broadband application in telecommunications at a center frequency of 28GHz using ADS 2011. The designed filter was compared with an equivalent lumped circuit. The insertion loss was -0.12dB and the return loss was -10dB [15]. In [2019], the ADS software was used to design a microstrip low pass filter with a cutoff frequency of 2.4GHz and with more than 60dB attenuation at 4GHz. The band stop filter was designed with a reject characteristic at 2.4GHz using a step impedance resonator (SIR) and couple line structure techniques. At the notch, the attenuation of a 2.4GHz frequency was more than 117dB in (SIR). This filter was easy to implement and more compact compared with its coupled line structure equivalent. The designed filter can be used in wireless communication systems [16]. At [2019], a Butterworth 5 th order bandpass filter was designed based on a quarter-wave resonator in ISM (2.4GHz -2.48GHz) band with a fractional bandwidth of 50% for a center 2.1GHz frequency. The insertion loss was 21dB at a cut-off frequency of 2.1GHz and a passband of 1.8GHz -2.7GHz used for GSM and Wi-Fi applications [17].

THEORITICAL BASIS OF PROPOSED
FILTERS Filters, in general, can be classified into two categories; active and passive. This classification is based on the essential components used in the design [2]. Active filters are constructed from resistors, capacitors, and op-ampsno inductors are needed [18]. In contrast, passive filter circuits containing passive element capacitors, inductors, . The insertion loss (S21 ) and return loss (S11 ) are used to determine the characteristic of the filter [19].

Butterworth Filter
Butterworth filters are maximally flat in the passband, but their out-of-band attenuation slopes are unsuitable [13]. These filters are designed to process signals with flat frequency response in the passband (no ripple) and zero roll-off in the stopband [12]. A steep attenuation transmission from passband to stopband requires Butterworth filters to have more components and provide monotonic attenuation for the low pass filters, as shown in Fig 1. [18].

Elliptic Filter
Elliptic filters are particular types of analog and digital filters characterized by passband ripples of equal amplitude. Their passband is associated with a maximal nonlinearity regarding their phase response. Elliptic filters give the smallest filter order compared to the other filter types for the same values of the filter design parameters [20]. Elliptic filters have the sharpest out-of-band attenuation, but they have undesired ripples in and out of the passband [3]. Elliptic filters have a steeper transition from passband to stopband as shown in Fig 2. [18].

OPTIMIZATION PROCESS
The proposed Butterworth bandpass filter comprises 3rd order high pass and 3rd order low pass filter, which introduces the bandpass filter of 6 th order as shown in Fig 3. This filter is designed using ADS simulation; hence the specified GOALs are evaluated on the cockpit OPTIM controller [6]. The filter offers the desired performance with frequency ranges illustrated in Table (     The result can be rearranged from the frequency response of all filters, as shown in Table (2). The optimization error, ripple, and transition region are the best for the bandpass filter (HPLP) 6 th order. The Butterworth bandpass 6 th order characteristic is the second, and bandpass 3 rd order with constant bandwidth (7GHz) comes last.  . shows the Elliptic filter design with the same frequency ranges and goals of the Butterworth filter with a variance in circuit diagrams. Fig 12. Explain frequency response of insertion loss (S21) and return loss (S11) of Elliptic filter of 6 th order consisting of 3 rd order high pass and 3 rd order low pass filter.    Table (3) shows the results of the designed Elliptic filter and indicates that the optimization error of the Elliptic bandpass 6 th order filter is the best with a greater bandwidth (7.76 GHz). There is a ripple in the passband region. The second filter in terms of quality characteristics is the Elliptic bandpass filter (HPLP) 6 th order. The Elliptic bandpass 3 rd order comes last in this category.

MICROSTRIP TRANSMISSION LINE
Based on the simulation results and for the Butterworth bandpass filter in Table (2), the optimization error was 2.5 minimum for the HP-LP 6 th order with a flat passband region and minimum value of insertion loss (S 21 ) in the low pass and high pass transition region. Thus, to obtain practical HP-LP 6 th order filters, the lumped component filters must be converted into distribution element realizations using ADS simulation software tool [10] and a microstrip transmission line according to the following steps: Step 1: Open a new Schismatic and add filter parameters (S-parametersoptimization -Goal1 for S21 -Goal 2 for S11) as shown in Fig 17. Step 2: Add filter ( LC-Band Pass) block and enter the suggested specification ( fs1 -fp1 -fp2 -fs2 -Ap -As ), as shown in Fig 18.        Table (3) shows the simulation result of the Elliptic filter. It can be noted that this filter has a tiny ripple for all types of the proposed filters. For this reason, the bandwidth is investigated, and it is pointed out that the 3 rd order has less bandwidth than the others. Fig. 27 shows the circuit of the schematic form used to design the filter Guide. The frequency response of insertion loss (S21) and return loss (S11 ) of Open Circuit Elliptic BPF 3 rd order is shown in figure (29). After optimizing all the proposed filters and comparing them in terms of optimization error and bandwidth and based on tables (2) and (3), it can be noted that the Butterworth bandpass 6 th order (HP-LP) has a good optimization error with constant bandwidth. However, the Elliptic bandpass 3 rd order has less bandwidth. This filter is selected to be redesigned using a design filter guide then converted into a transmission line model to transform it into a microstrip using the Line Calc option.      To match the MLIN (Microstrip Line) components with the 50-ohm circuit, they are added to both sides of the filter, whose characteristic impedance is 50 ohms. The length and width of the transmission line sections can be found using the Line Calc tool [10]. Finally, the transmission line model is converted into a microstrip with each filters component (W, L). Fig. 36 shows how to start line calculation.     This comparison gives a positive point for the Butterworth BPF (HPLP) 6th order short circuit. In the Elliptic BPF 3 rd order, the bandwidth is constant (7 GHz) for the open and short circuit form, and the values of (W, L) in the open circuit are less than the short circuit. Each proposed filter specification has its application in the (5G) communication system.