Microwave Circuit Fabrication

The new £1M microwave circuit fabrication laboratory is targeted at providing fabrication facilities for system-in-package solutions (inc LTCC) for communications and sensing applications. The major items of equipment are a laser prototyping system, automatic screen printer and isostatic laminator. They are supported by a full range of standard fabrication equipment such as a spinner, etching and developing tanks, UV exposure system, digital measuring microscope, laminar flow cabinet, etc.

Thin-film inductors

Thin-film inductors offer many practical advantages over air wound cores (though of course they can not achieve the same Q factors). Thin-film inductors are easier to pick and place in a SMT process than air core inductors. They can easily be processed in the IR, vapor phase and wave processes typically used in today’s assembly. Further, they will maintain value through those processes, as well as through handling and through high-vibration environments. Though they can not be tuned in circuit like air cores can, thin-film inductors can be used to replace air cores once an exact inductance value is determined for proper circuit function (assuming the Q factor is acceptable).

As in the case of thin-film capacitors, ESR and loss are dramatically reduced due to line width control and dielectric laydown quality/accuracy. This results in an end product which can be as small as an 0402 package with virtually any inductance value imaginable plus tolerance accuracy as close as 0.05 nH. Further, consistent metalization allows relatively high-current-carrying capability in thin-film inductors – up to 1,000 mA depending on device selected.

Other thin-film structures:

A variety of other structures have been created with the knowledge and process capability gained in manufacturing thin-film capacitors and inductors. Among these are couplers and harmonic low pass filters.

Miniature SMT couplers are a welcome addition to the designer’s toolbox. These devices provide high directivity, repeatable coupling with low insertion loss. They handle large amounts of power in a PCB footprint as small as 0402 with low profiles. As in the case with other thin-film components their electrical response and consistency on a lot-to-lot and inter lot basis is unmatched.

An example of thin-film inductor implementation might be in frequency compensation on broadband amplifiers. Previously, a resistor/inductor combination was used. As in the case of thin-film capacitors, the use of a thin-film inductor can reduce the number of components used in the circuit thereby saving size, weight, assembly and cost as well as improving reliability.

Just as thin-film capacitors, thin-film inductors are limited in maximum value.

In particular, a thin-film inductor provides designers with a good solution at extremely high frequencies. A common example is in multi-gigahertz oscillators. At high frequencies, wire-wound inductors may simply not be available, due to the absence of cost effective manufacturing techniques to build such low value wire-wounds.

At this point the designer is left with the choice of creating a low value inductor with serpentine PC board trace designs or choosing a miniature SMT thin-film inductor.

Though a PCB-based solution can be considered low cost, it uses valuable board space, and can vary based upon the PCB supplier. The thin-film inductor will have the same extremely repeatable and consistent frequency response on a lot-to-lot basis and on an inter lot basis as that of thin-film capacitors.

Band-reject filters

A real-world example of this is band-reject filters. A band-reject filter is a circuit designed to block the passage of signals in a specific spectrum of RF frequency while allowing other signals to pass un-attenuated. Other common names are notch filters, bandstop filters or band suppression filters. A common implementation of a band-reject filter is between a power amplifier and matching circuit prior to an antenna.

For example, in a typical application, the narrow notch filter is used to attenuate noise from heterodynes and harmonics unintentionally generated by complex, multiband, wide coverage receivers. The use of a single high-quality thin-film capacitor can essentially replace the use of six components in a twin T design due to thin film’s near-ideal characteristics.

Thin-film capacitors (see Fig. 1) have an additional performance advantage not discussed earlier: a single resonant point response due to the fact that the devices use a single-layer dielectric design packaged as a multilayer ceramic capacitor (MLCC). A few of the thin-film capacitor’s S21 forward transmission loss characteristic curves are shown in Fig. 2

Fig. 2. S21 forward transmission loss characteristics curve.When using a thin-film capacitor, manufacturers can reap the electrical benefits of a single-layer capacitor while being rewarded with the processing ease of an MLCC-type component. A thin-film capacitor’s consistent performance impact on electrodes and oxide thickness and quality impact on dielectric K is shown in Fig. 3.

Fig. 3. A thin-film capacitor has an extremely repeatable frequency response compared to MLCCs.

It is also important to realize the limitations of thin-film capacitors used as band-reject filters. Since thin-film capacitors are typically only available in low capacitance values they are limited to relatively high-frequency band-reject filter designs. If dealing with low frequency designs, other filter methods must be utilized typically using high-Q multilayer RF capacitors.

RoF-Predistortion circuit

Radio-over-Fibre (RoF) networks performances are typically influenced by the non-linearities due to the modulating devices: lasers in direct modulation schemes, electro-optical modulators in external modulation schemes. A completely analog predistortion circuit has been developed to compensate the non-linearities generated by directly modulated semiconductor lasers. This low-cost circuit, of industrial interest, has a multi-decade correction capability, from 300 MHz up to 2 GHz; thus, it can be used effectively to reduce second and third order harmonic (HD2 and HD3) and intermodulation (IM2 and IM3) distortions generated inside the CATV, GSM/DCS, GPRS and UMTS bands. An average broadband compensation of at least 10 dB has been recorded during measurements; the compensation magnitude can be further enhanced by specifically tuning the circuit for CATV or cellular bands, alternatively.

GaAs MMIC prototypes and Radio-over-Fibre modulating devices

GaAs MMIC prototypes:

Simple GaAs MMIC prototypes have been designed and experimented with the purpose to set up and refine both the design tools and the test and measurement facilities. The chips have been manufactured by the Gec-Marconi foundry using the F20 process.

Radio-over-Fibre modulating devices

Radio-over-Fibre (RoF) networks integrate optical fibre backbones and wireless cellular networks to provide broadband services to mobile, nomadic and fixed users. To generate and distribute RFmodulated optical signals, two schemes are used: direct modulation schemes, where a semiconductor laser is directly intensity-modulated by the RF-signal, and external modulation schemes, where the optical carrier generated by the laser is subsequently modulated by the RFinformation through an electro-optical modulator. The main modulating devices of such schemes have been analysed and modelled:

• the optical behaviour of an electro-absorption modulator has been analysed by using the Compact-2D-FDTD method;
• a circuit model has been developed and used to study the non-linear behaviour of directly modulated lasers.

SPDT switch and Quasi-optical frequency multipliers

SPDT switch:

A millimeter-wave “Single Pole, Double Throw” (SPDT) switch, realized with a Si-MMIC microstrip technology has been studied. The analysis has been performed with both commercial, equivalent circuit based simulator and mixed electromagnetic-device simulator. The sinusoidal RF signal (ampl. 0.2 V, freq. 76 GHz) is injected at one end of the structure, and then forwarded along either output branches, depending on the bias supplied by the square-wave generator (ampl. 2.5 V, freq. 214 MHz). Microstrip stubs and matched loads complete the structure. To isolate the inactive

branch, {p-i-n} diodes have been used, due to their high impedance in the off-state, and to the compatibility with the high-resistivity silicon, planar process. Feasibility of such Si-MMICs, suitable for operating frequencies up to 100 GHz, has been demonstrated.

Quasi-optical frequency multipliers:

The basic quasi-optical frequency doubler is shown in Fig.6a. In this device the frequency multiplication is achieved by a diode bridge connected to the center of two /2-dipoles. These dipoles are placed in a cross configuration (referred to as crossed dipole): the longest one receives the incoming power at the fundamental frequency (3.5 GHz in our prototype), while the shortest one transmits the generated power at the doubled frequency (7.0 GHz in our prototype). The advantages of such structure are the following. First, a good isolation between input and output signals is achieved since incoming and outgoing waves are orthogonally polarized. Second, the symmetry of the structure has been exploited to obtain a balanced topology which is characterized by a good conversion efficiency (only even harmonics are generated within the multiplier). However, the conversion efficiency of the above structure is intrinsically limited by the omnidirectional nature of both receiving and transmitting antennas. To overcome the above mentioned problem, the front-to-back ratio (and thus the directivity) of both receiving and transmitting dipoles has been increased by additional layers with parasitic elements. The resulting structure is essentially constituted by two Yagi-Uda antennas working at orthogonal polarizations. The new, multi-layer structure can be better understood with the example of Fig. 6b. In this case only two parasitic elements are adopted, one for each antenna of the frequency doubler. The longest element (left of the photograph) is used as a reflector for the fundamental frequency dipole; the shortest element (right of the photograph) is used as reflector for the doubled frequency dipole.

Coplanar waveguide to slotline transitions

The uniplanar configuration has having a growing popularity in the area of microwave integrated circuits. Fundamentals uniplanar configurations are the coplanar waveguide (CPW) and the slotline(SL). Well known advantages with respect to the microstrip line configuration are: easy shunt and series mounting, elimination of via holes, reduced radiation loss (for CPW), use of thicker substrates overcompensate the necessity for air bridges to suppress the spurious slotline mode in the CPW. The lack of accurate design tools however has prevented an extensive application of uniplanar technology, of which the transition between CPW and SL is a key element. The FDTD method has been employed to analyze several types of CPW-SL transitions including the effects of the coaxial connectors, air bridges, shielding effects as well as interactions between discontinuities. Theoretical simulations show excellent agreement with experimental data appeared in the literature.

Hybrid integrated circuit bonding:

The bonding wire interconnection is a key element for the fabrication of hybrid integrated circuits. It is employed to connect solid state devices to passive circuit elements as well as multichip modules. In spite of its small physical length, when millimeter-wave operations are required, the discontinuity introduced by the bonding wire can significantly affect the performance of the whole circuit. At the University of Perugia, the bonding wire interconnection has been studied from the point of view of its modeling and electrical characterization. In particular, two electrical models of the bonding wire have been developed. First, the Finite Difference Time Domain (FDTD) method has been adopted to rigorously analyze several bonding wire configurations (including multi-chip, single- and double-wire structures) and to produce reference results. Then, a quasi-static model of the bonding wire has been derived. This model is based on the representation of the structure with four uniform transmission line sections. Such an approach is suitable for commercial microwave CAD tools since the model parameters can be evaluated analytically from the dimensions of the structure.

Packaged microwave integrated circuits

Because of the increasing operating frequency and complexity of the circuits, the prediction of the package behavior and of the interaction between the package and the enclosed circuit itself is becoming more and more significant. In most of the commercial microwave simulators, the presence of the package is taken into account by simply considering the reactance introduced by the electric walls close to the circuit; unfortunately, when the package supports resonant modes and when electromagnetic couplings between different parts of the circuit are present, this approach is not valid anymore. To overcome this limitation the full-wave FDTD simulator has successfully been used to investigate packaged MMIC circuits. In particular, the behaviour of packaged single and coupled MMIC via-hole grounds have been investigated; the theoretical analysis has been compared with experimental results showing excellent agreement. Moreover, since the package introduces resonances, we have investigated several different possibilities to choke off these resonances. It is shown that the common practice of inserting a damping layer just below the lid is often not effective. In particular, the importance of placing damping layers also on the side walls is demonstrated.

MMIC commercial package:

The behaviour of a commercial MMIC package operating in the range 0-40 GHz has been investigated. It consists of a mechanical support of kovar with a fused quartz substrate above it. The area where the MMIC must be placed is enclosed between glass side walls and a top metallic lid. The latter is connected to a ground plane by via-holes through the glass walls. Bias lines and metal backed CPW terminations are also present. It is interesting to note that this package is substantially an open structure. Within the operating frequency range, the electromagnetic shielding is realized by metallic via-holes connecting the lid with the ground plane. In order to compare theoretical and experimental results, a simple microstrip transmission line, inserted between the input and output ports has been simulated.

What is...

What is the Microwave Circuit Designer’s Duty for Wireless Systems?

• Designing filters, mixers, amplifiers, oscillators, matching networks, packaging, and system level design of the Analog and Digital Systems
• Designing the antennas and matching networks
• Propagation Affects (multipath, signal diversity)

What are the principal future challenges?

• Miniaturized and low cost microwave circuitry (applies to cellular, PCS, GPS, on-board radar).
• Direct hand held unit communication with low earth orbit and mid earth orbit satellites
• Direct high-speed digital communication with low earth orbit and mid earth orbit with portable computers.
• Miniature high-power amplifiers with low signal to noise ratios, miniature high-Q filters, novel printed antennas, new packaging technology for combining RF and digital circuitry, reducing parasitics, improved modeling and analysis capabilities

Microwave Circuit Design.

• Microwave Circuits are composed of distributed elements with dimensions such that the voltage and phase over the length of the device can vary significantly.
• By modifying the lengths and dimensions of the device, the line voltage (and current) amplitude and phase can be effectively controlled in a manner to obtain a specifically desired frequency response of the device.
• Microwave Circuits are used to design microwave amplifiers, oscillators, filters, power dividers/combiners, multiplexers, antennas and mixers.
• The necessary tools for the analysis and design of microwave circuit devices require an understanding of: transmission lines, two-port networks (Z, Y, ABCD Parameters), network theory (S-parameters), impedance matching, and filter design. Much of the course will develop a deeper understanding of these fields while integrating specific applications of microwave circuit design.

Microwave circuits find their use in a plethora of applications, including:

1. Wireless Communications (cellular, PCS)
2. Wireless Networking
3. Digital Communications (Ground-Ground, Satellite- Ground, Satellite-Satellite)
4. Radar Systems (ground based, airborne, personal vehicles) – Target detection and identification, imaging (e.g., SAR)
5. Deep Space Communications
6. Medical Imaging and treatment
7. Radio Spectrometry