Observations on Supplemental Grounding and Bonding: Part 2

In Part 1 of this article (July 2007), we described (per the Emerald Book and ANSI J-STD-07-A) supplemental grounding and bonding from the grounding electrode system (GES) to the telecommunications main grounding bus bar [T(M)GB]. Note: The notation “T(M)GB” indicates that the bar may be either the TMGB or TGB, depending on the location within the building. Looking forward from the T(M)GB, there's still a need for supplemental grounding and bonding between the T(M)GB and the telecommunications and information technology equipment systems, which we'll refer to in this article as ITE.

Just as identifiable supplemental grounding and bonding entities exist for the power distribution system, they also exist for ITE, which inherently incorporates supplemental grounding and bonding as provided in the design of the equipment and typical installation recommendations from the manufacturer. This ensures the ITE operates properly. Furthermore, the manufacturer may assemble ITE into functional blocks by including cabinets, racks, or units into a specified grouping. However, supplemental bonding and grounding contained in such a block is outside the scope of this article.

Multipoint or common-type bonding networks

These bonding networks (BNs) for ITE are specific arrangements that are intended to reflect the grounding and bonding philosophy chosen by the manufacturer or end-user for that specific application or location. Industry-recognized bonding networks are generally equally applicable to AC- or DC-powered ITE.

IEEE Std 1100 (2005), the “Emerald Book,” in Sec. 9.9.17.1, informs (by reference): “For equipment bonding networks, lightning and both AC and DC power faults are the energy sources that cause the greatest concern. Of less concern are quasi-steady-state sources, such as AC power harmonics and function sources, such as clock signals from digital ITE. The energy sources that cause concern are referred to as emitters. The ITE that can suffer adversely from these emitters are referred to as susceptors. The coupling between the emitter and a susceptor is characterized as a transfer function. Therefore, the purpose of a BN is to reduce the magnitude of the transfer function by controlling the design of how the BN is attached to the CBN.”

Essentially, two attachment methods are employed: diverting or shunting (common galvanic connections) and blocking (isolated to one galvanic connection).

Note that these connections are at the “systems” level and primarily only address the galvanic connections. Parasitic capacitance and inductance are not specifically addressed. These are more important in the sub-system and board levels of the ITE and would be addressed by the manufacturer. (See Bonding Network Interference below.)

The Emerald Book describes both multipoint bonding networks and single-point bonding networks. It's important to recognize that these bonding networks are harmonized with international and national standards. The Table (click here to see Table) provides identification and a brief description of four variations of common-type BNs. The distinction of these four variations allows all interested parties to readily identify the variation(s) of BNs addressed at a given location.

Note that bandwidth describes a range of frequencies over which the structure is said to operate at nearly constant low impedance. The declared bandwidth of a signal reference grid (SRG) is often not readily declared or even known. Verbiage such as “handles high frequencies” is more often offered in writings. However, it seems somewhat reasonable to assume that the intent is to “handle” conducted power line emissions (up to 30 MHz). Let's take a look at these ITE supplemental grounding and bonding entities in more detail.

Common bonding network

The infrastructure, discussed in Part 1, is part of the common bonding network (CBN) as described in the Emerald Book. Historically, the CBN was developed by the telecommunications industry, supporting two equipment bonding networks (EBNs) deemed acceptable for high-availability installations, such as in a public switched telephone network (PSTN).

Due to the rapid convergence of technology and markets for telecommunications and computing, commercial and regulated installations are no longer so easily distinguishable. For example, regulated installations now often deploy soft switches that use Internet Protocol (IP) and are planning the next-generation network (NGN), which is also based on IP. Of course, IP is dominant in data center installations. Furthermore, telecommunications network (regulated) equipment (sometimes referred to as telecommunications load equipment or TLE) can be installed within the data center. Thus, it's important to become knowledgeable of CBN and EBNs — even for commercial locations. In many instances, the ITE/TLE may have similar characteristics.

ANSI T1.333-2001, “Grounding and Bonding of Telecommunications Equipment,” describes a CBN as “the principal means for effecting bonding and grounding inside a telecommunication building. It is the set of metallic components that are intentionally or incidentally interconnected to form the principal bonding network (BN) in a building. These components include: structural steel or reinforcing rods, metallic plumbing, AC power conduit, AC equipment grounding conductors, cable racks, and bonding conductors. The CBN always has a mesh topology and is connected to the grounding electrode system.”

The CBN is three dimensional. The key point here is that there are many conducting “loops” intended within each of the three dimensions that collectively promote electromagnetic compatibility (EMC) via shielding. The prime example of a CBN component is the multi-grounding and bonding that normally occurs when the AC power system is installed into the commercial building in accordance with the NEC and industry-recommended practices. The grounding electrode system, although a separate entity, becomes a part of the CBN because the CBN must always be grounded. As contributors to the electromagnetic shielding capability of the CBN, multiple interconnections of metallic structures and objects are desirable and increase the utility (density) of the CBN.

A CBN is typically more explicit and visible in a restricted access area, such as a dedicated equipment room. Compare this to a typical office area in a commercial building where the CBN components may be sparse (limited number of grounding and bonding conductors).

This distinction regarding the density of the CBN is important. To accomplish the EMC objectives cited in the originating documents for a CBN, you must consider the density and placement of the CBN components. If not, you may unwittingly build the CBN to a layout and density below original objectives.

The usefulness of the CBN/EBN concept was adopted in the 1999 edition of the Emerald Book, in Chapter 9, and recognized via reference in ANSI J-STD-607-A in 2002 for commercial buildings. This harmonization of standards means the concept is now the international basis for describing grounding and bonding networks in both regulated and commercial ITE environments.

Although the CBN was developed under considerations for EMC, the CBN by itself does not ensure ITE will meet EMC requirements or objectives. ITE having a regulatory mark (such as CE) does not ensure its electromagnetic immunity when placed into the CBN. The CBN and the chosen EBN perform as a system, and you must coordinate the desired immunity margin to accomplish EMC. Supplemental grounding and bonding provided by the CBN and EBN are key factors in achieving acceptable EMC.

Sparse common bonding network

A situation where ITE in a CBN becomes effectively single-point grounded (IBN equivalent) is known as a worst-case or sparse CBN (SCBN). In this arrangement, only single-point grounding is afforded the ITE from the serving power distribution circuit. This situation is likely more prevalent in small commercial office spaces, smaller wooden framed buildings, or residential home offices.

Note that the SCBN is not intended, it just happens as a result of the installation characteristics. Therefore, a CBN cannot be readily assumed for all situations. For example, consider the multitude of personal computers in a commercial office environment. Many may be single-point grounded via:

• An isolated grounding receptacle (IGR) circuit,

• A power outlet unit or a multiservice-multiport surge protection unit, which, in turn, is connected by its “single” power cord, or

• Branch circuits made up with non-conductive (plastic) conduit or raceway.

Mesh common bonding network

For increased density within the ITE area, a nested CBN can be intentionally “meshed” by the manufacturer(s) or end-user onto the ITE at the deployment area. This variation is known as a mesh common bonding network (MCBN or MESH-BN) and is typically designed into the ITE complex by means of metal racks, cable tray, raceway, etc. The MCBN can be arranged for installation under the raised floor, at the floor level (metal structure of the raised floor), or above the cabinet or rack (i.e., superstructure). You can also describe such structures as mats.

Figure 1 (click here to see Fig. 1) illustrates an MCBN superstructure arranged for both under-floor and at-floor level installation. This superstructure is effectively a “dual entity,” as it depicts inter-bonding at several stanchions. Essentially, they can be considered single superstructure entities. The under-floor superstructure is a grid network of copper conductors that may be within a few inches of the at-floor level or placed onto the under floor, which is usually concrete. The at-floor superstructure is a grid network comprised of the raised floor stanchions, stringers, and, to some extent, the panels. Usually, the at-floor superstructure is supplemented by the under-floor superstructure due to construction issues in maintaining proper grounding and bonding or for improved access for connecting to the ITE.

The figure also illustrates some important points and raises some concerns. First, the raised floor stanchion-and-stringer system is usually not considered robust enough (mechanically and electrically) to stand alone as an MCBN superstructure. Second, the under-floor copper grid is intended to improve the electrical capability of the at-floor metal grid. Third, both the raised floor stanchion and stringer system, and the under-floor copper grid must be inter-bonded at least once (for safety). Typically, they are inter-bonded periodically (for performance) within the grid pattern. Finally, the perimeter of the MCBN superstructure grid can be distinguished as a ring-bonding conductor (RBC) to provide placement for easy attachment to the CBN [such as at the T(M)GB].

The following statements regarding MCBNs are generally supported in the Emerald Book:

• Cabling routes should avoid significant electromagnetic interference sources, such as motor drives.

• ITE should be located a distance from the perimeter due to probable lightning and power fault currents occurring at higher amperages in those areas.

• The MCBN should always be grounded/bonded at the T(M)GB and any accessible building steel and metal piping, etc., near the T(M)GB. For the other three sides of the MCBN, bonding to the building steel and metal piping, etc., is recommended unless the MCBN is separated by around 6 feet to prevent a lightning side flash or shock hazard.

An MCBN provides these benefits:

Signal reference structure

You can further arrange the MCBN superstructure into a design that more intentionally accounts for the effects of different frequencies along the grounding and bonding conductors. The typical form of the signal reference structure (SRS) is an SRG, such as shown in Fig. 2.

The electronic equipment can use the SRG as its return path for the signal to return to its source. This type of circuit is typically designated as “single-ended” because the signal wire can be placed as a single wire (return signal by an available ground system). A prime example of a single-ended circuit is the traditional RS232 circuit and cabling design whereby grounded wiring is used as a common signal reference path for several signals.

Figure 2 illustrates single-ended circuits where the ground path is used for the (logic) signal return path. The ground path represents common impedance whereby noise voltages can be developed when extraneous currents (common-mode currents) flow along the ground path. These noise voltages are a source of noise for the communication path and can influence susceptible equipment such as RS232 links.

However, modern data cabling such as Ethernet (twisted-pair) terminates into a balanced transformer. Common-mode interference is primarily handled by grounding the center tap of the transformer. Common-mode currents can exist between transformers of the Ethernet circuit. However, this situation was recognized during the IEEE 802.3 standards development, and adequate specifications were created to address this issue.

To reduce problems of maintaining a common signal reference for both ends of a single-ended circuit, a (restricted-length) common SRG is admittedly somewhat useful. It helps to reduce the offset of ground voltage between the ends of the wiring (grounded at each end). Notice that these types of circuits are relatively low frequency compared to modern data circuits such as 1,000 Base-T Ethernet over UTP cabling. Although the contribution by the SRG is acknowledged, it's impractical for the SRG to provide a “proper” signal return path. Along with several other issues, the expansive loop area (inductive by nature) between the signal wire and signal return path violates signal-integrity design concepts. This becomes more of a problem as the operating frequency increases to the speeds seen in today's data centers.

Where the connecting cabling is closely coupled to a conductive plane, interference into the cabling is limited. Since the close coupling is dependent upon loop area, routing of the cabling away from an SRG negates the contribution of the SRG to controlling interference coupling. In most data centers, such close-proximity routing isn't routinely performed. Thus, because most data centers are not reporting such problems, the issue of common-mode coupling to modern data cabling must not be notable. This implies there is sufficient common-mode rejection in modern data circuits.

As noted, modern data circuits commonly use highly complex balanced circuits such as Ethernet-over-twisted-pair cabling. The emphasis is on “balanced” as this technology affords much higher data rates due to increased immunity to electromagnetic influences. With the predominance of immune cabling circuits (including fiberoptic cabling) in use and still growing, the usefulness and applicability of the SRG should be questioned and not “routinely” specified over an MCBN.

The Emerald Book provides significant guidance on grounding and bonding electronic equipment and is a recommended practice. ITE should use supplemental grounding and bonding for improving performance and safety, especially for a multipoint grounding arrangement, such as an MCBN. Next month, we'll discuss supplemental grounding and bonding using single-point grounding topology (SPG).

Bush is director of research — power & grounding for Panduit in Tinley Park, Ill. For an extensive treatment on this subject, see white paper titled “ITE Grounding and Bonding - Multipoint” at http://www.panduit.com/resources/whitepapers.asp.

Sidebar: Bonding Network Interference

The quasi-steady-state sources (such as AC power harmonics) and function sources (such as clock signals from digital ITE) can be classified as interference sources. Other sources of electromagnetic interference (EMI) may originate from power electronic devices, motors, etc. These devices can influence the grounding system, primarily by passing through common-mode currents at higher frequencies.

In addition to grounding and bonding, it is vital to recognize other means of improving power quality and attaining electromagnetic compatibility (EMC). These include the following methods or components: 1) separation and shielding of input and output power cables, 2) shielding of control and interface cables, 3) separation of power and signal leads, 4) proper layout of power electronics and controls, 5) high-frequency bypass capacitors, 6) reactors (inductive chokes), 7) EMI filters on power and signal leads, and 8) shielded isolation power transformers.

Immunity of electronic equipment (including ITE) to interference within the bonding network can be traced to two factors:

1. Universal design of products to meet the regulatory environment of locations requiring certain levels of inherent immunity to EMC. Items addressed include electrostatic discharge (ESD), radiated electromagnetic fields, conducted disturbances, electrical fast transient (EFT), surge from lightning, and voltage sag.

2. Predominant use of twisted-pair (TP) data cabling (Ethernet) with significant immunity advantages. Generally:

• At up to at least 30 MHz, impressed common-mode voltages sufficiently cancel out due to “significant balance” of paired conductors.
• Metal shielding (such as aluminum foil) is sometimes added to the construction of the cabling to further control its common-mode impedance and to prevent electromagnetic penetration at around 30 MHz or higher (via foil thickness); the metal shield is grounded/bonded for safety, performance, and continuity across the structured components of the cabling system.

The grounding and bonding network is certainly part of the mitigation mix for interference control within the bonding network. Generally, fewer currents in the bonding network are favored as less current equivocates to lower levels of EMI and less voltage buildup along the network. This brings into consideration important issues addressed in The Emerald Book, such as:

• Extension of the bonding network between rooms, floors, and buildings, and

• Location and connection of surge protective devices (SPDs) within the bonding network.