Back to basics: Switchgear, transformers and UPSs

Switchgear, transformers and uninterruptible power supplies are among the largest, costliest and most critical devices in a facility

By Tom Divine, PE, LEED AP July 29, 2020


Learning Objectives

  • Learn about basic construction and operation of switchgear, transformers and uninterruptible power supplies.
  • Understand the fundamental applications of this equipment.
  • Know the most significant codes, standards and ratings applicable to each.

An understanding of the operation, construction and application of operation of switchgear, transformers and uninterruptible power supplies is important for designers, specifiers, facility owners and construction managers who may be called on to render decisions about design, project budgets and available space.


Switchgear is electrical distribution equipment: it accepts power from a source, routes it to a number of outputs and provides overcurrent protection and control functions. Of the types of distribution equipment described in the NFPA 70: National Electrical Code Article 408: Switchboards, Switchgear and Panelboards, switchgear is generally the most robustly constructed, the largest and the most expensive. It’s typically applied in high-reliability facilities, like hospitals or data centers, where continuity of power is critical to effective operation.

Switchgear is available in a wide range of voltage ratings, from below 1,000 volts to more than 200 kilovolts. Medium-voltage switchgear, rated above 1,000 volts, is manufactured in a variety of configurations. Assemblies are available for exterior padmount installation, vault installation or installed in dedicated freestanding metal buildings, with air, gas, vacuum or oil as insulating media. This discussion will focus on interior low-voltage switchgear.

The alternative to switchgear is switchboard construction. Switchboards generally require less space and are less expensive. Both are typically constructed of a number of vertical sections. Each section is enclosed in sheet metal, with openings in front for overcurrent protection devices, monitoring equipment and control devices. A section may contain a main overcurrent protection device, metering devices, automatic control and monitoring systems, overcurrent protection devices for distribution feeders or a combination of these or other equipment specific to the installation. Overcurrent protection is typically accomplished with circuit breakers, with fused switches are less frequently.

LV switchgear is constructed to UL 1558: Standard for Metal-Enclosed Low-Voltage Power Circuit Breaker Switchgear. Switchboards are constructed under UL 891: Switchboards. UL 1558 incorporates a number of requirements that enhance the reliability, durability and maintainability over UL 891.

Switchgear breakers are typically installed four high in a vertical section, individually mounted. Each circuit breaker is separated by solid barriers from other breakers and from the rest of the assembly. In a typical switchgear assembly, the horizontal and vertical buses are enclosed in a bus compartment to the rear of the breaker compartments and this bus compartment is isolated from the rest of the assembly using insulating barriers.

Finally, the cable connections are in the rear compartment, which is isolated from the bus compartment with an insulating barrier. These separations and barriers, prescribed by UL 1558, are intended to increase the reliability and maintainability of switchgear by limiting the possibility of contact between conductors attached to adjacent breakers during installations or maintenance and to minimize any damage to adjacent components in the event that an arcing fault should develop. Switchboards, under UL 891, are not required to provide the same level of isolation between components.

Circuit breakers installed in LV switchgear are required to meet UL 1066: Standard for Low-Voltage AC and DC Power Circuit Breakers Used in Enclosures. This standard requires that circuit breakers have a 30-cycle withstand rating, describing the level of fault current that they can tolerate for 0.5 seconds without damage. The instantaneous trip function can thus be delayed, to allow downstream breakers to clear a fault without tripping the switchgear breaker, facilitating selective coordination.

The switchboard standard allows breakers built to UL 489: Molded-Case Circuit Breakers, Molded-Case Switches and Circuit-Breaker Enclosures. Breakers built to this standard are required only to have a 3-cycle withstand, 0.05 seconds. For these breakers, the instantaneous trip function cannot be delayed to facilitate selective coordination. Use of fused switches is also allowed. The applicable standard for enclosed switches is NEMA KS1: Heavy Duty Enclosed and Dead-Front Switches.

Switchgear ratings include:

  • Insulation level.
  • Maximum continuous current.
  • Maximum voltage.
  • Power frequency.
  • Short-circuit withstand current.
  • Short time withstand current.

In a typical installation, LV switchgear is connected to the secondary of a power transformer — either the utility’s service transformer or a facility transformer. Where service is at medium voltage, a power transformer may be close-coupled to the switchgear, with the two assemblies bolted together to form a single unit. The resulting assembly is called a “unit substation.” The distribution breakers of the switchgear will typically serve feeders to large facility loads, such as chillers, large transformers or large UPSs — or other distribution equipment, such as switchboards, motor control centers, panelboards or, rarely, other switchgear assemblies.

Switchgear has definite advantages over switchboard construction in terms of reliability and maintainability. The decision of which system to use on a particular project will depend on a variety of factors. Switchboard construction requires a considerably smaller footprint to provide the same distribution and protection functions, so available space will have an impact on the selection. Switchgear is considerably more expensive, with a cost penalty on the order of 60% to 100%, so a tight project budget will bias the decision toward switchboard constructions. And, in projects where selective coordination is challenging, particularly on an emergency system where strict coordination is required by NEC Article 700.28, switchgear may be the necessary solution.


A transformer is an alternating-current electromagnetic device that magnetically moves power from one or more primary circuits to one or more secondary circuits. The primary and secondary circuits secondary circuits typically operate different voltages and currents, with the ratio between them determined by the transformer’s characteristics. Requirements for transformers are described in NEC Article 450.

Transformers are ubiquitous in modern life, with a variety of characteristics, ratings and uses. On the high-power end of the scale, electric utilities use large power transformers to connect transmission systems operating at different voltages. On the small end, tiny signal transformers are used to connect communication equipment to Ethernet systems and microscopic transformers have even been printed in integrated circuits. Transformers used in facility distribution systems fall between those extremes.

A transformer operates on the principle of magnetic induction, an electromagnetic principle that states that a voltage will develop across a conductor in the presence of a changing magnetic field. Magnetic induction was discovered and quantified in the 19th century by scientists whose contributions were so significant that their names have been attached to electrical units of measure and laws of physics. A thorough treatment of magnetic induction would require many times the space available here, so it will be treated qualitatively in this discussion of transformer operation.

In an elementary implementation, a simple transformer might consist of an iron ring, called the “core,” with one primary and one secondary each making multiple loops around the ring, called “coils,” as shown in Figure 1. When the primary is energized with alternating current, the primary coil generates a magnetic field that varies in magnitude and direction with the input power.

In theory, that magnetic field exists throughout all space, but the magnetic characteristics of the iron core concentrate nearly all of the magnetic field within the body of the ring, where it passes through both the primary and secondary coils. The time-varying magnetic field running through the secondary coil induces a voltage across those coils by magnetic induction. The quotient of the number of primary loops and the number of secondary loops is called the “turns ratio,” where turns refers to turns of wire around the core. In the end, the secondary voltage is equal to the primary voltage divided by the turns ratio.

Real-world transformers are much more complex than the naïve implementation described here. For example, most transformers installed in facilities are three-phase units, whose core geometry must accommodate three primary and three secondary coils. Transformers are often provided with taps on the secondary coil — additional connection points whose output voltage is slightly higher or lower than the nominal voltage, for use in applications where voltages lower or higher than normal chronically occur due to system loading, utility voltage levels or for other reasons. Transformer cores are typically made from sheets of specialty steel, bonded together with an insulating adhesive, rather than solid iron or steel, to reduced magnetically induced currents that circulate in the core during operation. A typical facility transformer is mounted inside a metal enclosure, usually with openings for ventilation.

No conductive connection exists between the primary and secondary coils of a transformer. The magnetic interaction between the coils forces the voltage between the secondary conductors to a specific value, but the voltage between either conductor and its surroundings is, in theory, undefined. For most systems, one of the secondary conductors must be intentionally connected to ground, to ensure that the voltage on the secondary doesn’t stray too far from earth potential. Exceptions to that rule are systems that must be tolerant of a single ground fault, such as isolated power systems in medical facilities.

Transformer ratings include:

  • Capacity, typically expressed in kilovolt amperes, the maximum apparent power that the transformer can supply to its loads.
  • Primary voltage or line voltage — the operating voltage of the primary coil.
  • Secondary voltage or load voltage — the operating voltage of the secondary coil.
  • Temperature rise, typically expressed in degrees Celsius — the difference between the temperature of the transformer coils and ambient temperature when the transformer operates at full load.

Other features of transformers that normally appear in specifications are number of phases, number and spacing of transformer taps, enclosure characteristics, insulation medium, impedance and efficiency.

Transformers are not 100% efficient. While most of the input power is delivered to the secondary terminals, some is lost as heat. These losses can be characterized as load losses, primarily due to resistance of the coil conductors and no-load losses, primarily due to magnetic effects inside and outside of the core. These two types of losses are interdependent, in that designing to reduce one type of losses can raise the other.

For example, load losses can be reduced by constructing the coils from larger wire, reducing their series resistance. However, larger conductors will place outside layers further from the core, reducing the effectiveness of magnetic coupling between the coil and core and raising no-load losses. For most transformers, Department of Energy rules describe required efficiency levels and specify that transformer efficiency will be optimized at a load level at or near 35%. Those regulations generally dictate what trade-offs are between load losses and no-load losses are permissible.

Uninterruptible power supplies

A UPS is an electrical assembly that is designed to provide nearly perfect alternating current power continuously, with nearly 100% reliability. A UPS is typically deployed to support electrical loads that are critical to the business conducted in a facility. UPSs are available in very small desktop units to power loads in hundreds of volt-amperes, to very large enterprise systems rated in thousands of kilowatts.

The function of a UPS is to provide high-quality power to its load when the primary power source, usually an electric utility, fails or becomes unacceptable. A UPS maintains power to its load during blackouts, brownouts, voltage sags and swells, loss of a single phase and other system disturbances, protecting from both loss of power and from damage.

All UPSs contain an energy storage system, most often in the form of chemical batteries (lead-acid, nickel-cadmium, lithium-ion). When the input power fails, a UPS draws energy from its batteries, converts it to AC and delivers it to the load. A number of schemes for providing replacement power, called “topologies,” are in common use.

A “double-conversion” UPS, also called an online UPS, continuously converts incoming AC to direct current using an internal rectifier. The resulting DC power is used to generate AC power for the load, using an internal inverter and to maintain the charge on the system batteries. Should the AC input be disrupted, the batteries provide power to the DC bus and the conversion to AC and delivery to the load continue without interruption.

The term “double-conversion” refers to the fact that the UPS continuously converts AC to DC and then converts that DC back to AC. With this scheme, the quality of the output AC does not depend on the quality of the input power, since the output is independently generated from the DC bus. Because conversion is continuous, there is no requirement for detection of input power disturbances in order to protect the load. This topology is seen as highly reliable. It is also generally more expensive and less efficient, than the alternatives.

Because a double-conversion UPS continuously generates the output AC, a failure inside the UPS can put the continuity of power to the critical load at risk. To address this vulnerability, these units typically include a static switch — a high-speed electronic switch connected between input and output — that will connect the input power directly to the load. The UPS monitors its own output and, should the output fall out of acceptable limits, the UPS closes the static switch and disconnects itself from the load.

A “single-conversion” or “standby” UPS continuously passes its input power directly to the load, while the input is acceptable. The UPS monitors the input power for disturbances and, should any appear, it disconnects input power and begins to serve the load from its batteries through its inverter. This process necessitates a delay between the input disturbance and the commencement of replacement power, for detection, realignment of the system and starting the inverter. A standby UPS is therefore applicable to loads with a higher tolerance for system disturbances. This topology is seen as less reliable than double conversion. It is, however, more efficient, since it doesn’t incur losses in its rectifier or its inverter in normal operation.

Ratings for uninterruptible power systems include:

  • Full load runtime — a function of the battery capacity.
  • Input voltage.
  • Maximum output apparent power, expressed in volt-amperes.
  • Maximum output wattage, expressed in watts.
  • Output voltages.

A UPS is typically sized at about 125% of its expected maximum load, estimated for its entire life cycle. Data center applications call for estimates of aggressive load growth that sometimes don’t materialize, stranding excess capacity. To address this issue, some systems are available with modular, hot-swappable power supply and battery modules to allow incremental capacity and runtime upgrades as load increases.

UPSs require routine maintenance and, like everything else, they sometimes fail in service. For some systems, a wrap-around maintenance bypass, connecting the load directly to utility power, is an adequate provision for maintenance and repair activities. More sensitive systems will require a level of redundancy. Units can be connected in parallel or in series to provide redundant capacity, with communication and monitoring among redundant units.

Author Bio: Tom Divine is a senior electrical engineer at Johnston, LLC. He is a member of the Consulting-Specifying Engineer editorial advisory board.