Mitigating EMI in High-Speed Digital Transmission Networks, Part II

P. S. Neelakanta and Jesada Sivaraks

With the rapidly growing demand for digital communication via cellular networks, EMI solutions must be found for cellular base-station interference.

The seamless integration of high-speed, broadband digital and analog transmissions is the new frontier for today's cellular communication systems. As technology continues to improve, it is ever more likely that typical mobile users of the future will be able to use a single cellular device to transmit and receive several types of information, such as voice, text, and video. For example, two technologies are now available to realize the goal of integrated wireless communication: cellular digital packet data (CDPD) and wireless asynchronous transfer mode (ATM) transmission.

As communication systems supporting heterogeneous traffic become more commonplace, controlling EMI inherent in mixed-signal wireless conditions is increasingly critical. (Mitigating EMI and the role of cabling products and shielding accessories are discussed in Part I of this article in the November/December 1999 issue.) Although there are some methods available for EMI control, additional research is required to fully address interference in the cellular networks of the future. A very promising area for creating EMI solutions is the development of novel materials for equipment housings at cellular base stations. New composite materials could be fabricated to exhibit a variety of useful properties.

To see how new materials can offer solutions to EMI in integrated cellular networks, it is necessary to understand how interference affects such networks.

ATM, CDPD, and EMI

At least two approaches are available for progressive broadband, high-speed networking via a wireless medium: wireless ATM and CDPD.^1,2

ATM technology is a fixed-size packet or cell mode of transmission for statistically multiplexed information from heterogeneous sources including voice, video, and data over broadband integrated services digital networks (B-ISDN). Though initially conceived for wireline transport, this broadband strategy has recently begun to be used in wireless systems as well.

Designed for seamless integration of wide- and local-area networks, ATM used in wireline systems provides access among varied transport technologies, such as SONET/SDH and Frame Relay. Wireless ATM, shown in Figure 1, offers a backbone for supporting multimedia traffic.

 

Figure 1. A wireless ATM architecture.

CDPD networks have been developed as overlays for conventional 800-MHz analog cellular systems, such as the Advanced Mobile Phone System (AMPS). These networks are intended to seamlessly support applications that are based on Internet Protocol (IP) via existing wireless cellular phone systems. In a CDPD network, like the one shown in Figure 2, voice and data traffic share a set of wireless channels. Network planners configure the radio frequency (RF) channels as either dedicated or capable of supporting channel hopping by more than one data type. In such networks, voice takes priority over digital data in accessing the AMPS channels that support data hopping.

 

Figure 2. CDPD transmission system.

Any cellular site is vulnerable to RFI from neighboring sources, such as AM or FM broadcast and TV stations. The proximity of such sources plagues cellular communication sites with high ambient RF levels. Resulting EMI on the cellular equipment could even crash the whole system. ATM and CDPD networks are both particularly susceptible to interference and require adequate shielding at the base site to prevent noise and crosstalk in cellular communication channels, as well as erratic operation of the associated digital control systems.

For example, CDPD transmission involves careful timing, as the network must wait for free AMPS channels and then send bursts of data packets until the channels are taken over by voice transmissions. This channel hopping requires very fast digital control that is free of systemic inference. This synchronous effort, illustrated in Figure 3, is extensively prone to logic upsets, which can upset the local computer control and cause devastating interference problems at unprotected cellular sites in high-RF environments.

 

Figure 3. Digital control of CDPD implementation.

Modern digital control circuits use a variety of logic families, such as emitter-coupled, complementary metal oxide semiconductor (CMOS), or transistor-transistor logic (TTL), which have specified noise margins or noise immunity levels. Typically, a high-speed pulse has a narrow pulsewidth () with a corresponding bandwidth approximately equal to 1/. The invasion of RFI could alter this spectral width significantly, changing the shape of the pulse in the time domain in amplitude and width. The changed pulse may represent an erroneous logic state at a digital gate, causing undesirable effects in the gate's output.

Typically, RFI assaults logic circuits via capacitive coupling. A design-level approach to avoid RFI-based logic upsets is to select and use components that are rated with large noise margin tolerances. Still, as a result of statistical unpredictability of the extent of RFI strengths, logic upsets are inevitable and they can enhance the probability of blocking data calls, leading to inefficient CDPD operation.

In addition to logic upsets, receiver overload, desensitization, harmonic reception, and intermoderation effects can also occur in the cellular site equipment that is exposed to ambient RFI.

Because ATM and CDPD site equipment installations are often retrofits and may even be physically located outside existing cellular boards, these networks may be significantly prone to RFI. ATM and CDPD installations outside equipment racks with enormous wires and cables penetrating into the racks may also compromise whatever shielding is inherent in the existing equipment. Control units or computers that are part of the installations can also be susceptible to RFI ingress.

The interference that threatens CDPD and wireless ATM technologies can lead to excessive bit-error corrections and even data packet dropping. These possible effects could place a colossal burden on the system infrastructure and lead to low throughput conditions, which could cause a variety of symptoms, including signal outages, unexplainable handoffs, unwanted digital audio and video superimposed on a desired signal, unusual power-regulation problems, and unreliable coverage radius.

Some Methods for Reducing EMI

Management of EMI at base-station sites can be challenging; shielding RFI-menaced equipment at the site is not easy.³ Placing the equipment within a classically conceived "sandwich box" to block out the RFI is often impractical and expensive, and may not be effective. The passage of undesirable RFI in radiated mode can be stopped by the walls of the shielding box, but in conducted mode, the RFI can penetrate via other routes, such as antenna transmission lines, main power hookups, telephone lines, doors, grounding cables, and plumbing.

Base-site equipment shielding is in essence an architectural problem. It can be accomplished conveniently while the base site is constructed. If the ATM or CDPD systems are retrofits, however, several other solutions are indicated, such as using fiber attenuation composites, coating suspensions, filters as needed, and gasketing at clearly designated locales where EMI tests indicate inevitable "treatments" are imminent.

Filter component integration can offer one solution to the conducted-mode EMI menace. At the board level, the use of high-current filters or filtering arrays, filtered connectors, and discrete EMI filters are possible quick-fix strategies. Deployment of D-sub and customized connectors, filter plates, and filtered terminal blocks may be required at the rack level to ward off the invading RFI. The equipment's power lines can also be fitted with power line filters, high-current filters, and filter terminal blocks. In effect, an EMI "fence" can be established around the base-site equipment.

In addition, back plane, large-cabinet, handheld enclosure, large-surface, environmental enclosure, computer, and server shielding may be required to protect the paraphernalia of the base-station electrical system from EMI or RFI. These options may not provide a complete solution, but their use may reduce the extent of EMI or RFI significantly.

Although the shielding techniques proposed above may mitigate some EMI effects at base stations, the challenges of RFI on cellular-site equipment and the anticipated system malfunctions when high-speed digital service is introduced on a massive scale will require more effort to overcome.

Addressing EMI with Novel Materials

The answer to the problem of EMI at the base site is ensuring that the housings of equipment subsystems are not prone to RFI invasion. These housings include enclosures, panels, and racks, as well as the electronics and cables they contain.⁴

A variety of metallic and nonmetallic materials can be used in making these housings, as discussed in Part I of this article. However, more research is needed to develop additional metal-in-insulator or insulator-in-metal composites, which could be judiciously used in the manufacturing of EMI-retarding housings. Composite EM materials are needed because monolithic materials, either conductors or insulators, will not match the constrained performance considerations and compliance requirements at base stations. The main considerations for synthesizing such materials are as follows:

• Shielding effectiveness versus frequency.
• UL-rated thermal durability.
• Mechanical properties.
• Conformability for fitting into complex structured profiles.
• Ease of installation.
• Cost-effectiveness.

Making an EM composite is both a science and an art.⁵ It requires a clear understanding of the EM (dielectric, magnetic, and conductive) properties of the constituents and the artful blending of them to form an end product that has the required shape, mechanical characteristics, thermal properties, and above all, shielding effectiveness.

The constituents that are useful in making EM composites can be classified as nonmagnetic or magnetic conductors and nonmagnetic or magnetic dielectrics. These materials are available as metals, semiconductors, insulators (organic and synthetic), and ceramics. Typical RF-attenuation characteristics versus frequency of some materials are given in Figure 4 and Figure 5.

 

Figure 4. RF attenuation of typical materials where (a) is copper in epoxy (volume fraction of copper 33%); (b) is silver in acrylic plastic (volume fraction of silver 30%); (c) is nickel whiskers in acrylic plastic (volume fraction of nickel 15%); and (d) is graphite in polystyrene (volume fraction of graphite 33%). Thickness of all materials equals 1/8 in.

Figure 5. Relative RF attenuation provided by identical enclosures with different surface finishes. Conductive-paint-coated enclosure (a), and conductive-fabric (with 9% copper per sq ft)­covered enclosure (b).

To meet both the mechanical and thermal requirements for high-frequency shielding, a judicious choice is a diphasic mixture of a dielectric and a conductor or semiconductor, with a binder if necessary. Such a composite should meet shape, size, and resilience requirements and also be machinable.

In their diphasic form, the EM composites are composed of simple, discrete, conducting inclusions dispersed in a polymeric or ceramic host. Different types of EM composites are shown in Figure 6. Apart from conventional metals like copper or aluminum, carbon is also widely used in graphite form for this purpose because of its high electrical conductivity, low cost, and low density. With the exception of aluminum, metals have higher density than carbon. Semiconductors, solid-state electrolytes, and salts are other possible conducting inclusions for synthesizing EM composites.

 

Figure 6. Different types of EM composites, including (a) particles in a polymer, (b) disk-loaded composite, (c) spheres in a polymer, (d) diced composite, (e) rods in a polymer, (f) sandwich composite, (g) glass-ceramic composite, (h) transverse reinforced composite, (i) vertical honeycomb composite, (j) horizontal honeycomb composite, (k) single-side-perforated composite, (l) two-side-perforated composite, (m) replamine composite, (n) burps composite, (o) crisscross sandwich composite, and (p) ladder-structured composite.

Three classes of conductors are used to synthesize conductor-loaded composites: metallic inclusions (aluminum, copper, iron, stainless steel, and silver), nonmetallic inclusions (carbon black, graphite, ferrous oxide, salts of copper and aluminum with and without binders, solid electrolytes, and semiconductors), and metal-coated dielectrics (metal-coated glass, nickel-coated fiberglass). There is a handbook that details the general characteristics and the availability of these conducting inclusions.⁵

The host materials used in the fabrication of conductor-loaded composites are selected on the basis of minimum degradation of host medium due to the catalyzation and oxidation of conducting inclusions, wetting of metallic inclusions under heavy-current operations, morphology of the polymer or ceramic material (amorphous, semicrystalline, or polycrystalline), and compatibility with the processing techniques involved. The choice of polymers or ceramics that meet the above requirements is rather limited. A few examples that are in use are polycarbonate, graphite, cadmium oxide, and polymer concrete (see Table I).

 

Material	Application
Polycarbonate	Moldable cable products. With copper inclusions, however, this host material may suffer degradation due to catalyzation of copper.
Graphite	This is a popular matrix material in constituting cermet composites. It is conducive for baking at high temperatures (1000°­1400°C). Copper- graphite cermet is a good candidate as an EM composite for shielding the cabled network under RFI environment.
Cadmium oxide	Oxide-based cermets can be used with silver, nickel, or tungsten as the conducting inclusion components in making shielding parts.
Polymer concrete	Conductor-based polymer concrete can be tried in architectural applications for large units of shielding structures.

Table I. Polymers and ceramic host media for EM composites.

Conductor-loaded polymers or ceramics should be selected according to a variety of electrical and nonelectrical properties. Electrical properties include

• Bulk or volume resistivity (effective dc conductivity).
• Surface resistivity.
• Effective permittivity (at static and high-frequency limits).
• Effective permeability (if the conductive filler is a magnetic material).
• Effective loss tangent at a given frequency of operation.
• Effective dielectric response of the material under time-varying conditions as specified by the effective complex permittivity (dielectric relaxation).
• Effective permittivity of the material versus the volume fraction of the conducting fillers.
• Degradation of electrical parameters with shelf life and aging.
• Variation of electrical characteristics with ambient conditions such as temperature, humidity, or corrosive agents.
• Effects of the shape (in bulk or in thick- or thin-film forms) of the composite in dictating its effective electrical characteristics.
• Effects of the resiliency of the end product (its softness, flexibility, or toughness) as decided by the type of polymeric or ceramic chosen in characterizing the effective electrical properties of the composite.

Nonelectrical properties include

• Cost-effectiveness.
• Density.
• Thermal endurance.
• Corrosion resistance.
• Mechanical strength, hardness, or durability.
• Brittleness and flexibility.
• Strength-to-weight ratio.
• Aesthetic appearance in terms of color and surface finish.
• Porosity and defect sites.
• Compatibility for use with other conventional monolithic or composite materials with adhesive or fastening feasibilities.

Processing the Composites

The processing procedures for the EMI shielding composite materials are essentially decided by the constituent phases of the composite. The type of conducting inclusion, in terms of shape, size, and electrical conductivity, and the properties of the host medium, such as its complex permittivity, determine not only the electrical and nonelectrical characteristics of the composite, but also the processing methods that are required.

Spherical or irregular-shaped inclusions permit conventional fabrication methods, in which the inclusions are blended into polymers using a Banbury mixer, and the mixtures are processed with extrusion-compounding techniques.

To predict the effective conductivity of mixtures that are processed as described above, one can assume that the particulates are randomly dispersed. Further, the critical volume fraction, which is the amount of conducting inclusions added that brings the composite nearly from the insulation to the conduction phase, is not only controlled by the concentration of the inclusions, but also by their shape and particle size distribution. Some evidence of this observation can be seen in experiments with aluminum or iron particles plus styrene-acrylonitrile copolymer composites.

The processing of a composite with the electrical characteristics required for shielding becomes even more complex with shaped particles, such as flakes or fibers. When flakes are used, the processing could significantly affect the effective conductivity of the composite. Injection molding may align the flakes in the flow direction, thereby creating a directional anisotropy of conductivity. Synthesizing with conducting flakes as inclusions, in general, results in more-stable conducting composites.

EM Composites for Cabling and Equipment

A number of application-specific conductor-loaded composites that exhibit characteristics helpful to EMI shielding of telecommunication equipment have already been developed. Such materials have great potential for present and future developments in material technology vis-à-vis EM applications for shielding home-based cabling and base-station equipment.

Metal-included plastics. These are substitutes for metals in certain electrical engineering applications.^5,6 They are composed of a metal in particulate, fibrous, or flaky form added to a polymeric, or plastic, host material.

These materials exhibit considerably larger electrical and thermal conductivities in comparison with pure plastics, but are also lightweight and have the aesthetic appearance of plastics. Because of their high electric conductivity, they provide adequate EM shielding effectiveness and prevent electrostatic buildup. They are superior to plastics with metallic surface coatings in terms of cost and damage resistance. They are also better than carbon-filled plastics considering the plastic strength, impact resistance, and color options.

The metal filling in plastics refers to the use of chunky fragments, near-spherical particulates, or fibrous whiskers. Of these, fibers are most effective in controlling electrical resistivity and providing thermal conductivity.

The electrical conductivity of metal-included plastics depends on the volume fraction of the inclusions such that the average number of contacts with neighboring particles (coordination number) reaches a minimum value. The critical volume fraction at which the conductivity of the composite begins to increase depends on the particulate shape: chunky, fibrous, spheroidal, or near-spherical. Invariably, a low volume fraction is required with fibers to achieve a specified conductivity because the geometry of the fibers provides greater points of contact with neighbors even at lower concentrations of the fibers. In general, controlling the conductivity of metal-included plastics is rather a stringent design consideration, because resistivity of such materials alters drastically even with small changes in the volume fraction of metallic inclusions.

Some applications of metal-included plastics are

• EMI shields, both small cable products and large-surface architectural units.
• Electrostatic control media.
• Plastic housings for cables to provide good thermal dissipation (plenum-specified), electrical grounding, ESD, and specific shielding against EMI.
• Electrical applications in lieu of metals, which are susceptible to corrosion.  

The electrical resistivity of state-of-the-art metal-included plastics relative to metals and plastics extends from 10^-2 to 10² -cm. Low-cost, high-conductivity materials can be made by appropriate processing, with metal fibers of 20­30 mm length dispersed in a plastic medium that can be fabricated with injection molding and extrusion. The wide availability of plastic host materials and conventional plastic-making technologies allows flexibility in designing end products. Uniformity of fiber dispersion, length, and concentration, and methods of interlinking the fibers, are manufacturing considerations in developing metal-included plastics.

Conducting polymers.⁷ These are polymeric materials doped with substances to increase their electrical conductivity. The doping is accomplished with chemical, electrochemical, or ion implantation methods. Such materials are useful in EMI shielding and RFI absorption and can be fabricated in sheet and film forms.

Potential applications of conducting polymers are high-frequency EMI shielding and RFI absorption. Materials like polyacetylene and PBT can provide shielding effectiveness up to ­40 dB for frequencies extending to GHz levels. They are therefore useful as lightweight shielding for cabinets housing high data-rate electronic equipment as well as for wireless systems operating in the 2.4 GHz range.

Conducting polymers can be used for RFI absorption in configurations such as the Salisbury screen. Owing to the higher conductivity to weight ratio over carbon-impregnated nonmetallic composites, conducting polymers are attractive as RFI-absorbing materials in architectural structures.

Chiralic composites. An interesting set of EM composites are EM chirals, materials in which the electric and magnetic fields are cross coupled.⁵ The potentials of such materials in realizing EM shielding are elaborated further in other sources.⁸

The Future of Base-Site Shielding

New composite EM materials offer an excellent way to mitigate the EMI that will be an inevitable part of cellular base stations that handle mixed data transmissions.

When should new materials be deployed on an existing system? Even though perimeter control of EMI around base-station equipment may have already been put in place, an existing system should be at least partially upgraded and revived periodically. It is appropriate and convenient during such upgrades to consider introducing shielding with new options, such as new EM composites. The use of these new materials offers tremendous potential for dealing with EMI problems that will become ever more critical in an increasingly cellular world.

References

1. D Raychaudhuri and ND Wilson, "ATM-Based Transport Architecture for Multiservices Wireless Personal Communication Networks," IEEE Journal of Selected Areas in Communication 12, no. 8 (1992): 1401­1414.
2. KC Budka et al., "Cellular Digital Packet Data Networks," Bell System Technical Journal (Summer 1997): 164­181.
3. D Björklöf, "EMC Fundamentals: Shielding--Part 7," Compliance Engineering 15, no. 6 (1998): 48­66.
4. J Pryma, "Piecing Together the Residential-Wiring Puzzle," Cabling Installation & Maintenance (May 1999): 39­46.
5. PS Neelakanta, Handbook of Electromagnetic Materials--Monotonic and Composite Versions and Their Applications (Boca Raton, FL: CRC Press, 1995).
6. RB Seymour, ed., Conductive Polymers (New York: Plenum Press, 1981).
7. JM Margolis, ed., Conductive Polymers and Plastics (New York: Chapman and Hall, 1989).
8. N Enghetta, "Chiral Materials and Chiral Electrodynamics: Basic Physical Principles and Background" presented in Workshop of Chiral and Complex Materials, Progress in EM Research Symposium, Cambridge, MA, July 1991).

P. S. Neelakanta, PhD, C.Eng., is professor of electrical engineering at Florida Atlantic University in Boca Raton. Jesada Sivaraks is pursuing a PhD in electrical engineering at Florida Atlantic University.

Back to March/April Table of Contents