Microwave Journal
www.microwavejournal.com/articles/32622-fixed-frequency-analog-filters

Fixed Frequency Analog Filters

From “Handbook of RF, Microwave, and Millimeter-Wave Components” by Leonid A. Belov, Sergey M. Smolskiy and Victor N. Kochemasov by Artech House

July 30, 2018

Methods of radio signal filtering, summing and separation in the frequency domain are widely used during signal generation and processing. The implementation of filters with high quality factors (Q) up to frequencies in the mmWave range is challenging. Frequency filtering can be achieved through analog means (in hardware) or digitally (in software). This discussion considers only analog frequency filters in the RF, microwave and mmWave frequency ranges.

A frequency filter is a linear electrical circuit with lumped or distributed parameters. A frequency filter is generally characterized by the complex transfer coefficient Ẇ(f), which can be described by amplitude-frequency  | Ẇ(f )|  and phase-frequency  φ(f ) = arg[Ẇ(f)]  characteristics  (AFC  and  PFC,  respectively).  The function  Ẇ(f )  describes the pulse response h(t) at the filter’s output after applying a short pulse (impulse) to its input. The slope of the filter PFC defines the group delay (tgroup) of the response, which depends upon frequency

tgroup = - (1/2p)(df/df)     (1)

The ideal nondistorted filter has one or several bandwidths, for which the ideal transfer coefficient modulus |S21(f)| = 1 and one or several suppressed bands (rejection), for which the ideal transfer coefficient is equal to zero. Ideally, a signal whose spectrum width is narrower than the filter bandwidth is reproduced without loss or distortion at the output, but with a time delay.

Physically realizable frequency filters have loss at their bandwidth boundaries (cutoff frequencies) and AFC ripple within their bandwidths; this causes frequency distortion in the output signal due to PFC nonlinearities. They also have finite gain slopes at their bandwidth boundaries, and accordingly, finite group delays.

Filters are classified according to the location of passbands and attenuation bands, the character of amplitude-frequency and phase-frequency characteristics, selectivity, topology of construction and manufacturing technology. The main classes are lowpass filters (LPF), highpass filters (HPF), bandpass filters (BPF) and band-reject filters (BRF).

The band-reject or band-stop filter has the combined characteristics of a LPF and a HPF; a notch filter is a BRF with a narrow stopband (i.e. a high Q). More complicated/combined filtering products include multiplexers (diplexers and triplexers), duplexers, dispersive filters, preselectors, tunable and switchable filters and filter banks, filter-amplifiers, filter-couplers, and duplexer-filter/low-noise amplifier assemblies.

The complex filter transfer coefficient  Ḣ can be represented by the ratio of polynomials of the Laplace operator p = σ + j2πf

Ḣ = (a0 + a1p + … + aMpM)/(bo + b1p + … + bNpN);  N>M     (2)

Here a0, a1,…, aM, b0, b1,…, bN are constant coefficients, M and N are integers indicating the number of zeros and poles in Equation 2. Their mutual locations on the complex plane and quantitative values determine the main parameters of  the AFC and the PFC (e.g. shape-factor, in-band ripples and pulse response). N is the filter order.

FILTER RESPONSE CHARACTERISTICS

The following amplitude characteristic responses are typical: Chebyshev-1 or Symmetrical Equiripple Lumped Filter (SELF). The Chebyshev filter is arguably the most popular filter response type. It provides the greatest stopband attenuation but also the greatest overshoot. It has the worst group delay flatness.

Bessel (maximally flat group delay). This filter has the best in-band group delay flatness, with no overshoot, but the lowest stopband attenuation for a given order, and percentage bandwidth. It is ideal for receiver applications such as image-rejection filters.

Butterworth (maximally flat amplitude-frequency characteristic). This filter has the best in-band amplitude flatness, but has lower stopband attenuation than a Chebyshev filter; it is better than Chebyshev for group delay flatness and overshoot and is usually used as a compromise. It is common to use a Butterworth lowpass filter, which gives a continuous and monotonic response without discontinuity and without ripples in the frequency domain. The equation for a Butterworth lowpass filter is given by:

| Ẇ(f )|  = 1/sqrt (1 + (f/fcutoff)2N)     (3)

where  fcutoff  is the cutoff frequency and N is the filter order (the order is determined by the number of capacitors and inductors in the circuit). The Butterworth highpass filter is similar; the only difference is that the fraction in the denominator is inverted:

| Ẇ(f )|  = 1/sqrt (1 + (fcutoff/f)2N)     (4)

All of the above filters are realizable in parallel-coupled, direct-coupled and interdigital filter topologies.

Elliptic or Cayer is a filter whose transfer function is derived through the use of Jacobean elliptic functions. The amplitude is equal ripple in both the passband and stopband. This type of filter offers the most selectivity per pole but has the worst delay and phase linear performance. These filters can sometimes be designed wider than Butterworth or Chebychev filters thereby offering reduced insertion loss.

Gaussian. This filter provides a Gaussian response in both frequency and time domains and creates minimal group delay time. Its transfer characteristic g(t) has no overshoot:

g(t) = (BW-3dBT)(2p/sqrt(ln 2))exp[-(2p2(BW-3dBT)/T2ln2)t2]     (5)

where BW-3dB is the bandwidth at half-power and T is the time characteristic.

Linear phase. This filter has a maximally linear PFC. For example, with N = 8 at a center frequency of 100 MHz and with a relative bandwidth of 5 percent in the frequency band 75 to 125 MHz one may obtain over 540 degrees of phase, a deviation from linear of not more than 3 degrees (see Figure 1).

FiltersFig1

Fig. 1 Phase-frequency characteristic of a highly linear BPF.



FILTER TOPOLOGIES

The different configurations include:

  • Lumped LC (on discrete elements)
  • Cavity (combline, interdigital)
  • Ceramic (monoblock, resonator)
  • Stripline, microstrip, suspended substrate stripline (SSS)
  • Coaxial
  • Tubular
  • Waveguide
  • SAW and BAW
  • Monolithic crystal
  • Thin-film
  • YIG (yttrium iron garnet, ferrite)
  • MEMS (micro electro-mechanical systems)

Lumped LC filters cover the frequency range from DC to hundreds of MHz with relative bandwidths up to 200 percent. Cavity filters offer high values of Q (to 1000) with a number of sections up to 15. Strip and stripline filters are suitable for surface mount components up to frequencies of about 40 GHz. Tubular filters have increased power handling. Crystal, SAW and BAW filters have rather high Q values. Garnet (YIG) filters enable multi-octave electrical tuning.

Many quantitative characteristics are common among them; for example, increasing the number of filter sections can be used to decrease the relative bandwidth and increase selectivity. However, losses within the passband increase as well as the duration of transients. Size and cost increase as well.

The operating frequency and relative bandwidth of microwave resonators differ depending upon the manufacturing technology (see Figure 2). Thermal properties, environmental factors and the application (transmitter, receiver, GPS/GLONASS, VSB, CDMA, GSM, wireless) depend upon the manufacturing technology as well.

Typical filter performance specifications include:

Insertion Loss (IL) within the bandwidth is proportional to filter Q and inversely proportional to the relative bandwidth at half-power BW-3dB:

IL = 20log10 Q/(Q-1.41/BW-3dB)     (6)

Loaded Q (QL) is the ratio of the center frequency to the 3 dB bandwidth of a bandpass filter:

FiltersFig2

Fig. 2 Operating frequencies and a passbands for various filter technologies

QL = CF/BW-3dB      (7)

where CF is the center frequency.

The Frequency Bandwidth of the passband for a given loss is typically defined by the frequencies at the 3 dB points (the loss at band center minus 3 dB); and the stopband, for example, may be defined from the frequency for which the loss is 30 dB greater than the passband peak. The relative bandwidth (percent bandwidth) for BPFs and BRFs is normalized as a percent of the center frequency value. Moreover, values of the minimum and maximum frequencies are specified.

Shape Factor is the ratio of the AFC bandwidth at a given reject level (for example, BW-30dB) and the half power bandwidth (BW-3dB), i.e.

SF = f/fco for  LPF;   fco/f for  HPF;                                  (8)

BW-30dB/BW-3dB for  BPF; BW-3dB/ BW-30dB for BRF

Rise Time of the amplitude transient process at the output after an applied impulse is the time that the output signal amplitude changes from 10 to 90 percent of its maximum.

Maximum Input Level (continuous and/or pulse), is specified in accordance with requirements for heat dissipation and/or distortion. For example, in order to decrease geometric size, some manufacturers use ferrite inductive elements, for which loss and magnetic permeability are affected by signal amplitude. Limitations on power handling are inherent in SAW and BAW filters as well.

Return Loss (RL) is the ratio of reflected to incident power or, equivalently, the voltage standing wave ratio (VSWR):

RL = 20 log [(VSWR – 1)/(VSWR + 1]     (9)

Ripple is a wave-like variation of the amplitude response within a filter’s passband due to impedance mismatch.

Flatness is the absolute limit of amplitude variation in the passband. It includes ripple due to impedance mismatch as well as the monotonic roll-off at the band edges due to finite Q.

Ringing is a decaying oscillation at the output of a filter due to a transient signal at the input. Overshoot is when a signal or function exceeds its target due to ringing.

Environmental Requirements such as temperature, shock, vibration, moisture and radiation are defined by industry standards. As a rule, components should meet their performance specifications at an input power of no less than 1 W, shock up to 30 g, sinusoidal vibration up to 10 g at frequencies from 5 to 1 kHz, relative moisture of not less than 95 percent, and temperatures between of 40°C and 85°C.

BPF specifications typically include the center frequency, 3 dB bandwidth, lower stopband frequency at an attenuation -30 dB and upper stopband frequency at an attenuation of -30 dB. For higher order filters, ripple is specified both in the passband and in the rejection band.  Requirements on input impedance and environmental conditions are included as well.

BRF, or notch filter, specifications typically include the center frequency of the rejection band and attenuation at this frequency, lower and upper frequencies of the stopband for attenuation of -30 dB, and the lower and upper bandpass frequencies. As with the BPF, ripple, impedance and environmental requirements are included as well.

Lumped LC Filters

FilterFig3

Fig. 3 Typical frequency response of a BPF. (Courtesy of K&L Microwave).

Filters with lumped LC elements using various technologies for frequencies in the RF and microwave ranges are offered by several manufacturers. LPF and HPF examples are listed in Table 1, while several BPF and BRF type filters are presented in Table 2. Figure 3 shows the frequency response of a typical BPF where insertion loss is low and return loss is high within the passband, but rapidly transitions to high insertion loss and low return loss (i.e. rejection) in the regions below and above.

FiltersTable1

In Table 2, there are several examples of a notch filter. This is a narrowband rejection filter designed to suppress an unwanted concentration of spectral components within a desired band of interest.

FiltersTable2



Notch filters may be connected in cascade to improve suppression, having matched impedances of 50 or 75 W at the input and output. Some models integrate several stages in a single package.

Some of the features that may be incorporated in individual products may include anti-aliasing, absorptive input, constant impedance, flat time delay, high rejection/isolation, high linearity, low inter-modulation distortion, high-power rating, high mean time between failures, low insertion loss, high temperature stability and high shock resistance.

Cavity Filters

In cavity filters, Q-factors are high and can be traded against smaller package sizes. Cavity filters are available from 30 MHz to 40 GHz with bandwidth options from less than 0.5 percent to over 66 percent. They offer lower insertion loss, steeper skirt selectivity and narrower bandwidths than discrete component filters. At lower frequencies a helical coil is used to excite the electromagnetic field, while a 1/8 to 1/4 wave capacitively loaded stub is used at higher frequencies. Waveguide designs are used for narrow bandwidths and high-power.

Ceramic Resonator Filters

FiltersFig4

Fig. 4 Coaxial ceramic resonator filters (Courtesy of Skyworks Trans-Tech Inc.).

A high-Q dielectric or ceramic resonator is made of a ceramic material with metallized surfaces in the form of a disk, cylinder, or parallelepiped, frequently with an internal hole. Resonators with the internal holes are called coaxial resonators (see Figure 4). Ceramic resonator filters provide cost-effective solutions enabled by their compact low profiles and high Q values. They are typically used at frequencies from 0.4 to 6 GHz, have from 2 to 10 sections and handle RF power up to 10 W. Relative bandwidths are 1 to 10 percent with temperature stabilities of  2 to 11 ppm/°C operating from -40°C to +85°C. Losses are typically low and mechanical durability (resistance to shock and vibration) is high. Table 3 lists some examples.

FiltersTable3

Microstrip and Stripline Filters

Microstrip consists of a conducting strip separated from a ground plane by a dielectric, or substrate. Stripline is a conductor embedded in dielectric between a pair of ground planes. Suspended substrate stripline consists of a transmission line structure printed on a substrate suspended between two conducting planes (see Figure 5). The conducting planes provide shielding and prevent the excitation of higher order modes.

FilterFig5

Fig. 5 Cross-section of suspended substrate stripline.

Suspended substrate stripline filters are often used due to their higher Q-factors than printed stripline and microstripline. A suspended substrate filter has a higher unloaded Q, which results in lower passband loss than a conventional stripline filter. In addition, the lack of dielectric surrounding the circuit makes the filter less sensitive to ambient temperature variations. Combining lumped and distributed elements onto one suspended substrate board provides enhanced unloaded Q, lower insertion loss and smaller package sizes, than the use of a single topology.

Tubular Filters

Tubular filters have many advantages over other topologies for bandpass and lowpass filters. Because of their mechanical configuration, they offer very broad stopbands and very high rejection. Using capacitively coupled dielectric spacers, tubular filters are ideal for handling power levels up to 5000 W. Tubular filters are implemented for frequencies from 100 MHz to 20 GHz with bandwidths from 4 to 40 percent and have two to eight sections. They are compact, vibration resistant up to 20 g and operate from -55°C to +85°C.

Waveguide Filters

Waveguide filters operate at frequencies from 2.5 to over 300 GHz in the rectangular waveguide TE101 mode. They may have one to 20 sections and passbands from 0.1 to 10 percent. Circular waveguide filters using the TE111 mode can have 2 to 6 sections and relative bandwidths from 0.1 to 1.8 pecent. Compared to coaxial, tubular and microstrip implementations, BPFs and BRFs in waveguide have higher levels of rejection, higher power handling capabilities, the ability to operate in mm and sub-mm wavelength ranges and greater connector choices. They have also been used to realize mechanically tuned filters and diplexers with high transmit and receive channel isolation of not less than 85 dB.  In many systems, waveguide filters are more easily integrated with other components in the antenna-feed path. All filter types can be implemented in waveguide. The parameters of some models of waveguide filters are shown in Table 4.

FiltersTable4



Yttrium-Iron Garnet (YIG) Filters

Many companies offer YIG filters for civil and military applications in spectrum analyzers, sweep generators, ECM receivers, broadband test equipment and frequency synthesizers (see Table 5). These filters use single crystal synthetic YIG spheres as microwave resonators and are magnetically tunable over broad frequency ranges. YIG filters are used for their high Q factors, typically between 100 and 200, and their electronic tunability.

FiltersTable5

YIG filters use a permanent magnet and a small coil for frequency correction and thermal compensation. For the most models, the operating temperature is from -54°C to +85°C. The slope Scontrol  of the modulation characteristic is 10 to 30 MHz/mA depending on the model, and its deviation  from linear over the  tuning range is typically not more than 0.5 percent. The saturating power level of the input signal Pin sat is only about 10 dBm.

Tuning hysteresis depends on the tuning step and is proportional to the operating frequency (it does not exceed 20 MHz at a frequency of 20 GHz). The minimal width of the passband BW3dB is determined by the resonator’s unloaded Q and spurious parameters associated with filter construction. It may be constructed as a separately packaged module with coaxial connectors (see Figure 6), or in a PCB surface mount package.

FiltersFig6

Fig. 6 YIG-tuned bandpass filter (Courtesy of Micro Lambda Wireless).

Some filters are distinguished by relatively small tuning ranges with narrow rejection bandwidths, high linearity and little hysteresis, while others have wide tuning ranges (2 to 18 GHz), but with relatively broader passbands, poorer linearity and greater hysteresis.

Thin and Thick Film Filters

Most BPFs are based either on discrete components or on an integrated ceramic structure. In the ceramic structure, layers of ceramic dielectric material and metal alloy electrodes are interleaved and then sintered at high temperature. This process introduces component variability in dielectric properties (loss, dielectric constant and insulation resistance) as well as variability in electrode conductivity and physical size.

Thin-film BPFs eliminate these uncertainties. For frequencies from 0.75 to 67 GHz, filters thin film technology may be used, the same thin film technology that is commonly used for producing semiconductor devices. Applying this to the manufacture of BPFs has enabled the development of components where both electrical and physical properties are tightly controlled. For example, line width variations are less than 1 μm and layer thickness can be controlled to 100 Angstrom (see Figure 7). These manufacturing features create a thin film capacitive-based filter that has a repeatable frequency response.

FiltersFig7

Fig. 7 Typical topology of a thin film frequency filter (Courtesy of Dielectric Laboratories, Inc.).

Thin film capacitive BPFs save system size, weight and complexity while improving system reliability, manufacturability and performance; and, they have excellent reproducibility. Thin film technology is used for producing MEMS and BAW filters as well.

Monolithic Crystal Filters

FiltersFig8

Fig. 8 Cross section of a monolithic crystal filter (a) and its equivalent circuit (b).

The monolithic crystal filter (MCF) relies upon the piezoelectric effect exhibited in quartz for its operation (see Figure 8a). When signals appear across one pair of electrodes, they set up mechanical vibrations on the crystal. These are affected by the mechanical resonances of the crystal element, and only those signals within the passband of the filter are allowed across the crystal to be picked up by the second pair of electrodes. The equivalent circuit of a double-electrode MCF (see Figure 8b) has the ability to create a narrowband filter since the Q-factor of the quartz crystal is very high (up to several thousand). Coupling between input and output resonance nodes is determined by the electrode arrangement on the crystal surface. Higher-order filters are generally made by connecting several monolithic crystal filters in series. This enables the performance of multisection crystal filters to be replicated in a monolithic format (see Figure 9).

SAW and BAW Filters

FiltersFig9aFiltersFig9b

Fig. 9 Frequency characteristic of a monolithic crystal filter (a) and package (b) (Courtesy of Vectron International).

Electronic devices employing surface acoustic waves (SAW) use one or more interdigital transducers (see Figure 10) to convert SAWs to electrical signals and vice versa by exploiting the piezoelectric effect of certain materials (quartz, lithium niobate, lithium tantalate and lanthanum gallium silicate). These devices are fabricated by photolithography, the process used in the manufacture of silicon integrated circuits. Dimensions of the crystal and transducer arrangement specify the frequency bandpass response between input and output RF ports. To decrease the line length, one sometimes uses a reflective array. Dispersive SAW lines with linear characteristic delay versus frequency (compressed pulse width) have nonequidistant transducer electrodes. SAW filters operate in the RF range from 50 MHz to 2.5 GHz.

FiltersFig10

Fig. 10 Top view of in-line SAW delay line structure.

Bulk acoustic wave (BAW) filters can be used to implement ladder or lattice filters. BAW filters use reflections of the bulk acoustic wave in the piezo-electric crystal and typically operate at frequencies from 300 MHz to around 10 GHz.

A Thin-film bulk acoustic resonator (FBAR) is a device consisting of a piezoelectric material sandwiched between two electrodes and acoustically isolated from the surrounding medium. FBAR devices using piezoelectric films with thicknesses ranging from several micrometers down to a tenth of a micrometer resonate in the frequency range of roughly 100 MHz to 18 GHz. Technology for the production the crystal and SAW units allows for a simpler of implementation with better filtering ability.

Technical characteristics of selected SAW and BAW filters are summarized in Table 6.

FiltersTable6



MEMS Filters

Microelectromechanical systems (MEMS) technology was developed for switching elements in the microwave range and is used for switches, tunable capacitors and inductors with dimensions on the order of of 10s of nanometers. MEMS technology allows implementation of small-size microstrip devices with switching or tunable parameters stable to the influence of the environment. RF MEMS resonators offer the potential of on-chip integration of high-Q resonators and low-loss bandpass filters. The Q-factor of RF MEMS resonators is in the order of 100 to 1000. With the help of MEMS technology chains of coupled resonators are constructed to form bandpass or band-reject filters. The capability of electrical position control of the microelectromechanical cantilever is used for smooth or discrete tuning of the central frequency or bandwidth.

Sandia National Laboratories has developed an aluminum nitride process for fabricating RF MEMS microresonators at frequencies ranging from 1 MHz to 3 GHz. The resonant frequency is determined lithographically. This process uses the same equipment and materials that were developed to fabricate film bulk acoustic resonators that are widely used to implement cellular phone duplexers and filters at 1.9 GHz. The piezoelectric transduction mechanism of these resonators allows the realization of low insertion loss filters. It is this technology that allows the scaling of MEMS resonators without introducing spurious modes or reduction in quality factor, and with acceptable power handling for both the transmit and receive paths in full-duplex radios. This technology is most suited for realizing resonators with the quality factors approaching 5000 and impedances less than 300 Ω.

FiltersFig11

Fig. 11 Frequency characteristic of a miniature waveguide filter (Courtesy of Memtronics Ltd.).

Memtronics, for example, produces filters based on RF MEMS technology providing the ability to create low-loss, high-linearity tuning while simultaneously reducing volume, weight, and parts count (see Figure 11). A four-pole filter centered at 11.8 GHz with a 5.1 percent bandwidth (11.5 to 12.1 GHz), exhibits a nominal 1.1 dB insertion loss over the passband. Spurious rejection is greater than 35 dB. The unloaded quality factor Q is greater than 450. Performance over temperature is stable, with Δf ~ 3.5 MHz from 25°C to 125°C. The small weight, waveguide filter chip is 12 mm × 35 mm × 0.8 mm and weighs 0.73 g.

RF MEMS switches, switched capacitors, and varactor technology offer the tunable filter designer a compelling trade-off between insertion loss, linearity, power consumption, power handling, size and switching time.

HARMONIC FILTERS

To satisfy RF transmitter electromagnetic compatability requirements in accordance with MIL-STD-469, it is necessary to reduce the spurious radiation of frequencies that are multiples of the fundamental. Harmonics, which are integer multiples of the fundamental (e.g. second and third harmonics) are the most troublesome. To suppress them, we can use band-reject and notch filters (BRF); however, harmonic filters in the form of LPFs, in which the cutoff frequency is 30 to 50 percent higher than the fundamental frequency, are simpler and more compact. One should consider the frequency response of such a filter over a wide frequency band exceeding the operating frequency and to take into account the handling power of the fundamental. Figure 12 shows the attenuation curve of a typical harmonic filter at the fundamental and its harmonics. Some serial harmonic filters produced for specific values of carrier frequency also include notch (rejection) filters tuned to increase attenuation at specific harmonics.

FiltersFig12

Fig. 12 Transfer coefficient of a harmonic filter

Harmonic LPFs are intended for suppressing emission of the higher harmonics of broadcast transmitters, VHF and UHF TV transmitters, and base stations for cellular communication (see Table 7). They are characterized by low passband loss, high suppression at the second and third harmonics, and high power handling.

FiltersTable7

When selecting the type of LPF harmonic filter, it is necessary to check high-frequency interference suppression up to the centimeter wave range. With increasing frequency and at increased rated capacitance of the noise-suppression filter, intrinsic resonance and reduced dielectric permeability become issues. Waveguide implementation (see Figure 13) provides a high level of attenuation at higher harmonics with high power handling capability.

FiltersFig13

Fig. 13 External view of a waveguide harmonic filter (courtesy of K&L Microwave Corp.).

FREQUENCY MULTIPLEXERS AND DUPLEXERS

Multiplexer

The frequency multiplexer is a frequency-separating device with n outputs having different frequency passbands and stopbands. The number of output channels distinguishes it as a diplexer (n = 2), a triplexer (n = 3), a quadriplexer at (n = 4), a quantuplexer (n = 5) and so forth (see Table 8). In most cases, frequency multiplexers possess the property of reciprocity (i.e., they may be used both for signal frequency separation and for combining of two or more signals).

FiltersTable8

To improve isolation of input and output ports and to equalize reflection coefficients over the wide frequency range of input signals, an LPF and HPF with appropriate adjustments are included in the diplexer structure. Bridge-balanced circuits and polarization solutions are used to improve mutual channel isolation.

FiltersFig14

Fig. 14 Typical diplexer structure.

A frequency diplexer (see Figure 14) is the simplest form of a multiplexer, which can split signals from one common port into many different paths. The incoming signals must be offset in frequency by an appreciable percentage, so that filters can do their job sorting them out. A diplexer could be used to route signals to two different receivers, based on frequency, or it could be used to create a matched filter that is nonreflective outside of the intended passband. It could also be used as a bias tee, to feed an active device with DC power.

Duplexer

A duplexer is used with wideband antenna for simultaneous transmission and reception in frequency-shifted bands (i.e. duplex communication). As such, it is necessary to separate the transmit signal applied to an antenna from the receive channel, protecting the receiver input circuit (see Figure 15). Along with nonreciprocal or directing devices such as directional couplers, ferrite isolators and circulators, filters are used to prevent the passage of transmitter power to the input of the low-noise amplifier.

FiltersFig15

Fig. 15 Typical duplexer structure directing a transmit signal from the transmitter to the antenna, and directing a received signal from the antenna to the receiver.

Many companies offer duplexers with different frequency ranges (see Table IX) to connect a common antenna with a receiver and a transmitter. To improve isolation between the receiver input signal and the more powerful signal from the  transmitter in a closely located frequency band, duplexers have several sections (as many as eight).

FiltersTable9

FIXED FREQUENCY FILTER ASSEMBLIES

Complex integrated microwave filter assemblies are designed for numerous and varied applications. Among them, for example, are multiband frequency multiplexers, filter banks, filters for automatic recognition of a frequency band, filters for the extraction of a frequency band with narrowband rejection of a frequency zone inside of it. Many, which are classified by the manufacturers as frequency assemblies, may include parameter control (e.g. digitally tuning, auto tuning, switching, tracking agility).


Related Publications:

  1. H. Campanella, Acoustic Wave and Electromechanical Resonators: Concept to Key Applications, Artech House, Norwood, MA, 2010.

  2. K. Hashimoto, RF Bulk Acoustic Wave Filters for Communications, Artech House, Norwood, MA, 2007.

  3. S. I. Baskakov, Radio Engineering Circuits and Signals, 3rd ed., Vysshaya Shcola, Moscow, 2000 (in Russian).

  4. W. Siebert, Circuits, Signals and Systems, McGraw-Hill, 1986.

  5. J. -S. Hong, Microstrip Filters for RF/Microwave Applications, John Wiley & Sons, Hoboken, NJ, 2011.

  6. Pozar, D. M., Microwave Engineering, 3rd ed., John Wiley & Sons, Hoboken, NJ 2005.

  7. G. M. Rebeiz, RF MEMS Theory, Design, and Technology, John Wiley & Sons, Hoboken, NJ 2003.

  8. T. C. N. Clark, “MEMS Technology for Timing and Frequency Control,” IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, Vol. 54, No. 2, February 2007, pp. 251–270.

  9. A. Abramovicz, “Modeling of Wide Band Combline and Interdigital Filters,” Microwave Review, July 2010, pp. 15-22.

  10. A. Tombak, J. P. Maria, F. T. Ayguavives, Z. Jin, G. T. Stauf, A. I. Kingon and A. Mortazawi, “Voltage-Controlled RF Filters Employing Thin-Film Barium-Strontium-Titanate Tunable Capacitors,” IEEE Transactions on Microwave Theory and Techniques, Vol. 51, No. 2, February 2003, pp. 462-467.

  11. J. Brank, J. Yao, M. Eberly, A. Malczewski, K. Varian and C. Goldsmith, “RF MEMS-Based Tunable Filters,” International Journal of RF and Microwave Computer-Aided Engineering, Vol. 11, No. 5, September 2001, pp. 276–284.

  12. K. Entesari and G. M. Rebeiz, “A 12–18 GHz Three-Pole RF MEMS Tunable Filter,” IEEE Transactions on Microwave Theory and Techniques, Vol. 53, No. 8, August 2005, p. 2566-2571.

  13. D. Scarbrough, C. Goldsmith, J. Papapolymerou and Y. Li, “Miniature Microwave RF MEMS Tunable Waveguide Filter,” European Microwave Conference, September 2009, pp. 1860–1863.

  14. M. S. Aftanasar, P. R. Young and I. D. Robertson, “Rectangular Waveguide Filters Using Photoimageable Thick-Film Processing,” 32nd European Microwave Conference, September 2002.

  15. K. M. Lakin, “A Review of Thin-Film Resonator Technology,” IEEE Microwave Magazine, Vol. 4, No. 4, December 2003, pp. 61-67.

  16. J. S. Kim, K. B. Lee, J. Y. Lee and H. Shin, “A Broadband Suspended Substrate Stripline Filter Using Dual-Mode Resonator,” Microwave and Optical Technology Letters, Vol. 53, No. 7, April 22, 2011.