Frequency management is an essential goal of modern high-frequency design, whether for avoiding interference in cellular communications systems or ensuring security in military electronic systems. But generating and maintaining stable signals, especially at RF and microwave frequencies, is not a trivial task.

Time alone can cause the designed frequencies of some circuit materials to drift, not to mention the deleterious effects of temperature, vibration, environmental contaminants, and humidity. Fortunately, with long experience in materials science, Dielectric Laboratories, Inc. (Cazenovia, NY) has developed proprietary ceramic materials that are well suited to the modern challenges of precise frequency management.

With the XTREME-Q™ resonators, filters, and custom ceramic components, coverage extends well beyond 67 GHz. DLI refers to its innovative frequency-management design and manufacturing approaches as a "disruptive technology."

DLI's skill in designing and manufacturing ceramic circuit materials has been honed over the years in the development of extensive lines of high-frequency, high-quality-factor (high-Q) single-layer SLCs and multilayer capacitors (MLCs). Manufactured to precise tolerances, these extremely stable capacitors maintain their values over time, temperature, and vibration in the most demanding commercial, industrial, and military and space applications. DLI's capacitor expertise is also a building block for higher-level design and manufacturing services offered as part of DLI's disruptive ceramic technology. Manufacturing capability utilizing thin-film technology facilitates the integration of capacitors, inductors, resistors, and other high-frequency microwave structures, forming high-performance components on ceramic substrates.

DLI offers a variety of different proprietary ceramic formulations in support of the technology, including CG, CF, and FS materials for frequencies through 67 GHz. These low-loss materials feature excellent stability over temperature (Fig. 1), as judged by their temperature coefficients of frequency (a measure of the variations in frequency over a certain temperature range). For temperatures from –54 to +125°C, the temperature coefficients of frequency for the CG, CF, and FS ceramic materials range from 2.3 to 8.8 PPM/°C. In terms of practical resonators and filters, these numbers translate into frequency shifts of less than one percent at wide temperature extremes. Compare this to the temperature stability of alumina (Al2O3) substrates commonly used in microwave circuits, at about 120 PPM/°C. These novel ceramic materials offer greater miniaturization capabilities compared to alumina and other substrates. And they are stable over time, and offer a space heritage of no out-gassing and no aging having been proven for decades in the company's lines of SLCs and MLCs.

Ceramic technology can be applied to any number of custom frequency-management components, including duplexers, diplexers, filters, and resonators. It is the resonator that plays a key role as a frequency-determining device that can be used in oscillators and frequency synthesizers to create the stable signals needed by cellular communications, radios, radar systems, and electronic warfare receivers. Resonators come in many shapes and sizes at high frequencies, including transmission-line resonators, coaxial-line resonators, and dielectric puck resonators. Figure 2 shows where different resonator families fit in terms of Q and spectrum management while Table 1 provides a comparison of some of the traditional resonator approaches to the patent-pending XTREME-Q™ single-frequency cavity resonators (SFCRs) produced with DLI's ceramic technology. Table 2 provides a quick summary of the performance levels for several of these SFCRs at different frequencies. Typical SFCR designs are nominally rectangular, with length-to-width aspect ratios of less than 1.2:1. Narrowband (3%) and wideband (10%) tunable variants are also available.

These temperature-stable XTREME-Q™ SFCRs compare in size with the highest-frequency resonator technologies (microstrip) and are chip-and-wire compatible while also well suited for designs employing surface-mount-technology (SMT) components. The actual size of a SFCR will depend on the frequency (wavelength) of the resonator (higher frequencies result in smaller resonators) and the dielectric constant of the ceramic material chosen for the resonator.

How is the performance of a resonator evaluated? One of the main performance parameters for resonators is quality factor or Q. Q, for a resonant structure, relates the energy stored in the system to the energy lost per cycle. For an electrical system, the Q is the ratio of the energy stored in the reactive elements to the power loss in the resistive elements. Unloaded Q or the Q of the resonator alone is limited by the loss mechanisms internal to the resonator. Loaded Q, a somewhat lower number, takes into account the energy lost to the external network. The loaded Q of an XTREME-Q™ SFCR can be improved by increasing the height of the resonator. In general, for a given frequency resonator, selecting material that leads to a larger resonator size will also yield higher Q performance. When a resonator is to be used in an oscillator, for example, high loaded Q is critical for achieving low phase noise.

A 36-page brochure, "Resonators, Filters, and Custom Ceramic Components: Disruptive Technologies For Spectrum Management," contains several charts for estimating the size of an XTREME-Q™ SFCR and how to select the right ceramic material for a given resonator application.

In general, the thicker, low-dielectric-constant materials have higher Qs. When a resonator has a coupling coefficient of 1, it will exhibit excellent return loss at the resonant frequency and the unloaded Q will be twice the value of the loaded Q. Resonator coupling selection is affected by several variables, including varactor frequency trimming and transistor negative resistance values. For SFCRs produced so far, unloaded Qs have ranged to nearly 2000, an excellent performance standard for a component compatible with automated assembly manufacturing approaches. The SFCRs are also completely self-contained; as a result, their loaded Qs and resonant frequencies can be directly measured using standard RF coplanar probe technology with a vector network analyzer.

DLI has developed accurate computer models for simulating the performance of the XTREME-Q™ SFCRs. The design process is also supported by the use of commercial circuit-simulation software tools. Figure 3 shows an equivalent-circuit model for a 9.95-GHz SFCR based on its lowest resonant frequency mode (typically employed in oscillator and filter designs). The resonant frequency is set by the parallel combination of capacitor Cp and inductor Lp, and the finite unloaded Q by resistor R. Series capacitor Cs connects the resonant LC combination to the input pad, setting the coupling between the external circuitry and the frequency-controlling LC resonator. Shunt capacitance Csh represents the stray capacitance between the input pad and ground. This structure provides integrated DC blocking, eliminating the need for an additional circuit element. S-parameters are available from DLI's website at www.dilabs.com for evaluating this new resonator approach in a commercial circuit simulator.

To show how closely the resonator models agree with measured data, a 9.95-GHz SFCR was fabricated on CF material with typical temperature coefficient of frequency of 2.3 PPM/°C from –54 to +125°C. The measured performance for the patent-pending 9.95-GHz SFCR (Fig. 4) shows typical return loss at resonance of 11 dB. The typical loaded Q for the resonator at 50 ohms is 300. It measures just 5.6 × 4.3 × 0.8 mm.

An 18.65-GHz cavity resonator was fabricated on FS material, a material more commonly employed in the higher regions of the microwave spectrum. The FS material features typical temperature coefficient of frequency of 7.3 PPM/°C at temperatures from –54 to +125°C. This higher-frequency resonator achieves typical return loss at resonance of better than 25 dB (Fig. 5). The 18.65-GHz resonator has a loaded Q of 400 at 50 ohms. It measures 6.1 × 5.6 × 1 mm.

These resonators are well suited for low-noise, narrowband (3%) varactor-tuned oscillators. Their high loaded Qs enable low-phase-noise performance while the inherent stability of the ceramic materials ensures good long-term stability and performance over temperature. Voltage-controlled oscillators (VCOs) represented in the single-loop model (Fig. 6) employ two-port resonators and achieve frequency control or modulation capabilities utilizing a voltage-variable phase shifter in the loop. A 10.5-GHz two-port cavity resonator was fabricated and the measured results for the two-port tunable device show well-behaved phase response and a tunable 3-dB bandwidth that extends from 10.43 to 10.57 GHz (Fig. 7) .

Wideband (~10%) microwave VCOs with low phase noise are difficult to design because of the noise effects of the tuning varactor diode. The relatively poor Q of the varactor degrades the loaded Q of the resonant circuit. The Q of the varactor actually gets worse with increased tuning range. For this reason, the resonator and coupling capacitors in a wideband VCO must exhibit the highest Q performance possible, with tight tolerances and good temperature stability, in order to overcome the negative effects of the tuning varactor.

Using the XTREME-Q™ approach to wideband VCOs, however, the coupling capacitors and resonant circuitry are integrated together on high-stability ceramic substrates. In this way, high-Q performance can be maintained even with the addition of a wide-tuning-range varactor. Figure 8 shows an equivalent circuit based on this approach, with stable, high-Q coupling capacitors denoted as Cs1 and Cs2. In contrast traditional approaches employ resonators and surface-mount MLC coupling capacitors. Variations in their tolerances cause variations in the performance levels of the VCOs, affecting yield. In addition, the high parasitic-element values (additional unwanted inductance) of these discrete coupling MLCs can lead to nonlinear tuning effects and spurious oscillations. DLI has also developed reference designs for complete oscillators, including one-port negative-resistance and two-port feedback oscillators.

Developing a custom resonator with DLI's high-performance ceramic technology is a simple matter of determining a set of performance requirements, including the desired resonant frequency, whether the design is a one-port or two-port resonator, the desired return loss at the resonant frequency (for a one-port resonator), the maximum allowable insertion loss at resonance (for a two-port resonator), and the goal for loaded Q. Mechanical issues include restrictions on case size and whether the resonator is to be surface-mounted or attached by epoxy and wire bonds.

Part 2 of this white paper (in the November issue of Microwaves & RF) on disruptive ceramic technology will focus on filters from 500 MHz to beyond 67 GHz, with typical insertion loss of 0.5 dB and more than 45 dB filter rejection.