Properly applying automatic transfer switches

Circuit protection analyses have to look at the routine, standby, and emergency power systems. This shows how selective coordination needs to include generators and transfer switches—not just circuit breakers and fuses in healthcare facilities.

By Dean J. Eriksen, PE, Affiliated Engineers Inc., Seattle July 12, 2012

The proper design of power distribution systems entails numerous considerations: overcurrent and short-circuit protection, voltage drop, harmonics, grounding, equipment loads, short-circuit rating, and selective coordination, to name a few. These issues relate not only to the devices used for overcurrent and short-circuit protection (circuit breakers and fuses), but to the transfer switches and alternate power sources used in emergency distribution systems.

This article focuses on the transfer switch and the alternate power source or generator that is commonly found in commercial building applications to create optional standby, emergency, legally required, and essential branches of power. To illustrate the impact of selective coordination on the application and specification of transfer switches and generators, this article will consider the implications of the design of an essential electrical system for a healthcare facility.

Healthcare facilities demand very reliable power to support critical and life safety system equipment. This imperative affects the design of both the normal power supply from an electric utility and the alternate power supply, such as that from an emergency diesel engine generator plant. The National Electrical Code (NEC) has been modified over the past several editions to address and prescribe requirements that are intended to increase the reliability of healthcare facility power distribution systems. The combination of high demand for reliability and the recent code requirements for emergency power systems to be selectively coordinated creates challenges for the design and application of power distribution equipment.

Specifically, the recent requirement for selective coordination imposes exceptional performance characteristics on automatic transfer switches (ATSs) the devices used to control the source of power serving essential electrical systems and the optional standby in a healthcare facility. Considering that the transfer switch is commonly sourced from both the electric utility and a generator, it is important to evaluate both circuits ahead of the transfer switch to understand the effects of selective coordination.

A power distribution system is considered to be selectively coordinated between two points in the system when an overcurrent (fault) condition is cleared by the operation of the device closest to the point in the system where the overcurrent (fault) occurred. No other devices operate, and continuity of power is maintained in all other parts of the system. There are numerous articles in the NEC that lead the design engineer toward the interpretation that within the healthcare facility the essential power distribution system’s overcurrent devices must be selectively coordinated with all supply-side overcurrent devices—implying power sources from both the normal source of power (i.e., utility) and from the alternate source of power (i.e., diesel engine generator).

The following articles can be referenced in the NEC to help the reader follow the logic behind these requirements. Fundamentally, Article 700.27 prescribes that overcurrent devices in emergency power systems must be selectively coordinated with all supply-side overcurrent protective devices. Meanwhile, Article 517.26 requires that the essential electrical system shall meet Article 700, which requires that overcurrent devices in essential electrical systems shall be selectively coordinated as well.

  • Article 100 – Definitions, Coordination (Selective)
  • Article 517.2 – Definitions, Emergency System
  • Article 517.2 – Definitions, Equipment System
  • Article 517.2 – Definitions, Essential Electrical System
  • Article 517.26 – Application of Other Articles
  • Article 700.27 – Coordination

Note that while the definition of selective coordination as described in Article 100 excludes any reference to a period of time, the commentary found in the 2011 NEC Handbook prescribes that the overcurrent device closest to the overcurrent condition must operate quickly to isolate the condition.

Transfer switches

ATSs (specifically, rated up to 600 V, 3-phase, and with a rating up to 4,000 amps) are listed per the results of testing prescribed in UL 1008. The withstand and close-on rating (WCR) is one of the most important considerations when applying ATSs. This is a root mean squared (RMS) symmetrical ampere rating of the transfer switch’s ability to withstand and close into a circuit under a fault condition. The WCRs are published by the transfer switch manufacturer and represented as shown in Table 1. ATSs must have a WCR at least greater than the available fault current at the point where the transfer switch is in the circuit. In addition to the transfer switch WCR, it is important to understand the duration for which the transfer switch must withstand the fault current. The transfer switch WCR must be based on the maximum duration of time the transfer switch will be exposed to the fault condition. Therefore, both the fault magnitude and fault duration at the point where a transfer switch will be applied in the circuit must be known in order to properly apply a transfer switch.

Many parameters affect the transfer switch WCR. Some of the more important parameters that must be considered are:

  • Voltage: The WCR varies depending on the application voltage (240, 480, 600).
  • Switch frame size: Different manufacturers have varying frame sizes that result in different WCRs.
  • Bypass/non-bypass: A switch may be specified as a bypass/isolation type transfer switch, which can alter its WCR rating.
  • Number of poles (2, 3, 4): WCR rating may be limited if the switch is specified as 4-pole with overlapping neutral contact.
  • Transfer switch overcurrent protection:  The WCR of the transfer switch is dependent upon the type of overcurrent protective device applied ahead of the transfer switch.
  • A transfer switch protected with “any circuit breaker” may have a different WCR than a transfer switch used with a specific manufacturer’s circuit breaker.
  • A current-limiting fuse, located ahead of the transfer switch, will result in the highest WCR for a given transfer switch. The current-limiting feature of the fuse significantly reduces the magnitude of the fault current seen by the transfer switch, and the fault is cleared in less than one electrical cycle.

The design engineer is faced with a significant challenge in designing around the “specific breaker” WCR if the scope of the project will be competitively bid, as in a public sector project. This is because the design engineer has limited control over the manufacturers that may supply both the transfer switch and the overcurrent device ahead of the switch in a combination, considering that in each case the low-bid vendor is usually awarded the work. On the other hand, if the design engineer is able to control which manufacturer supplies the overcurrent device and the transfer switch together, then the engineer can design around the specific breaker WCR. For example, if the design engineer is working on a project for a company that standardizes on a particular manufacturer for the transfer switch and overcurrent device, then the challenge is resolved. The design engineer can use the specific breaker WCR to properly apply the transfer switch.

Fault-current analysis

Considering the impact of available fault current on the application of transfer switches, the transfer switch (as noted earlier) must be able to withstand and close-on the magnitude of fault current available at the location in the power distribution system where the transfer switch is applied. The utility available fault duty, kVA rating, and percent impedance of the selected power service transformers will determine the fault current criteria for the transfer switch. The emergency generator system will likely not have the magnitude of available fault current that would typically be available from an electric utility. But this may not always be the case, and the design engineer should always consider the fault current available from both sources. For example, a simple radial system with a single 2500 kVA transformer with 5.75% impedance will produce about 53,000 RMS symmetrical amps at the secondary of the transformer (see Figure 2). Induction motors served from magnetic starters will provide additional contribution to this fault current. Large healthcare facilities, or those with exceptional reliability criteria, may be served with multiple transformers. For example, consider three 2500 kVA transformers arranged in a secondary spot network arrangement. If each transformer has 5.75% impedance and assuming an infinite primary source, the available fault current could exceed 156,000 amps (three times the single transformer scenario), excluding the motor contributions (see Figure 3).

Now consider a 480 V, 3-phase, 4-wire, 750 kW diesel engine driven synchronous generator serving as an alternate power source for the optional standby and essential equipment branches of power. If the generator has a subtransient reactance of 0.10, the available fault current at the terminals of the generator will be approximately 11,250 amps.

However, consider a much larger alternate power system made up of three 480 V, 3-phase, 4-wire, 2000 kW diesel engine driven synchronous generators, each with subtransient reactance of 0.10. The available fault current at the output of this system could be as high as 90,000 amps.

Due to the many possible variations in normal power service and alternate power source configurations, it is always important to consider both the normal source of power and the alternate source of power when performing the short-circuit study. This will define the lower limit of the transfer switch WCR.

Coordination study

To understand the effect of selective coordination on the appropriate selection of a transfer switch, consider the system represented in Figure 4. This is a simple radial normal power service configuration sourced with a 3-phase, 2000 kVA transformer with a 480 V, 3-phase, 4-wire secondary and an ATS served from a feeder directly from the service switchgear. The transfer switch serves a single 480 V distribution panel that serves multiple loads, including a distribution transformer. The transformer in turn feeds a 208 V, 3-phase distribution panel that serves multiple branch panel boards. This is a common configuration used to create an emergency branch of power in a healthcare facility. The overcurrent devices in this system include a combination of power circuit breakers with adjustable trip settings, as well as molded case circuit breakers with fixed trip settings (nonadjustable).

Figure 5 is the single line diagram representing the emergency power distribution system components, including transformers, feeders, overcurrent devices, and the transfer switch as described above. This diagram provides the basis for the coordination study of this circuit. Figure 6 represents a series of graphs that define the trip characteristics of the overcurrent devices in the circuit under study for the components of Figure 5. Selective coordination is realized graphically when none of the graphs overlap, that is, when there is sufficient space between adjacent graphs (approximately 0.1 sec) to avoid an overlap anywhere between adjacent graphs. In practical terms, this is when there is sufficient time for the downstream device, nearest the fault, to clear the fault before the upstream device detects (picks up) the fault.

In this diagram, note the lack of selective coordination between the circuit breaker ahead of the transfer switch (CB-LS-ATS) and the feeder circuit breaker (CB XFMR) in Distribution Panel LS after the transfer switch. The curves overlap between the time of 0.02 and 0.06 sec. Also notice the overlap in the curves of the circuit breaker feeding Transformer A (FDR CB XFMR) and the main circuit breaker in Distribution Panel A (MB-DIST PNLA). These two circuit breakers are in series; circuit breaker CB XFMR serves only the transformer and circuit breaker MB-DIST PNL A, no other loads. A lack of selective coordination between these two devices will not affect continuity of service to other loads so they do not need to be selectively coordinated. (Refer to exception [2] noted in NEC Article 700.27.)

It is not unusual to have a lack of selective coordination when an initial study is performed; it’s an iterative process, one that may require numerous “runs” to arrive at a selectively coordinated circuit. In our example, the next step is to select the correct devices and adjust the trip settings of the circuit breakers in the model to get to a point where there is selective coordination between all devices in the circuit.

After numerous iterations, Figure 8 shows that all non-series circuit breakers selectively coordinate, from the downstream 20-amp branch circuit breaker back to the main circuit breaker (MB-SWGR) in the main switchgear. This was accomplished by eliminating or “turning off” the instantaneous trip function in circuit breaker CB-LS-ATS. While this seems to be a minor change to achieve selective coordination, let’s now consider the impact this has on the required performance of the transfer switch.

The transfer switch will need to be able to withstand a fault magnitude of at least 34,000 RMS symmetrical amps for duration of at least 0.11 secs or 12 electrical cycles. The fault magnitude is a result of the short-circuit study performed on the normal power distribution system, and the withstand duration is a result of the coordination study and corresponding circuit breaker settings to achieve selective coordination.

But we must now consider the same system served by the alternate source of power. For example, let’s assume the alternate source is a 480 V, 3-phase, 4-wire, 750 kW diesel engine-driven synchronous generator. The coordination study must also look at the system while served by the generator source. Figure 8 represents the system single-line diagram when powered by the generator; the overcurrent devices downstream of the transfer switch are identical to those used in the normal power coordination study. Figure 9 shows the output of the alternate source coordination study; note that the circuit breakers after the transfer switch must have the same trip setting values as those resulting from the normal power coordination study. The overcurrent devices have to coordinate in both cases, so the settings have to be the same in both cases. This system is selectively coordinated from the branch circuit overcurrent device back to the generator circuit breaker (GEN CB).

The software tool used in our examples also computes short-circuit current values, represented in both Figure 5 (33,786 amps) and Figure 8 (7,256 amps), for the case when the utility is the source and when the generator is the source, respectively. In our example, the case when the transfer switch is served from the utility source defines the WCR as well as the duration for which it must withstand the fault (as a result of the coordination study under this condition). However, this may not always be the case., It is important to analyze the circuit under both utility power and generator power to determine which of the two conditions is the stricter. For example, the decrement curve of a 750 kW alternator with subtransient reactance of 0.14 can produce a fault current if driven by an engine of sufficient capacity.

Other criteria

There are a few additional points to understand in addition to the analysis described above. A final coordination study should be performed using the data from the approved overcurrent device manufacturer that will supply equipment for the designed facility. Variations in manufacturers’ circuit breaker trip functions (and corresponding time-current graph) will affect potential solutions to achieve selective coordination and subsequently the transfer switch withstand period of time. This creates a dilemma for the design engineer working on projects that are design/bid/build. The final coordination study, completed after the bids and contracts for equipment have been awarded, may impact the WCR of the transfer switch.

Molded case circuit breakers have minimal adjustments. A system with only molded case circuit breakers may be difficult to coordinate without oversizing the circuit breakers and may force the design engineer to oversize the transfer switch.

Application of current limiting fuses ahead of transfer switches in circuits that otherwise use circuit breakers downstream is very difficult to selectively coordinate. The fast-acting nature of the current limiting fuse will clear the fault well before a downstream circuit breaker has time to clear the fault. However, a circuit with all fuses, from source to final branch overcurrent device, is the easiest to coordinate. The design engineer will have a variety of manufacturers from which to choose to specify for the transfer switch in such applications. 

The design engineer must assure that the transfer switch manufacturers identified in the project specifications can provide a UL-listed and labeled transfer switch with the required WCR for the required time period. These requirements must be met with the other specific features necessitated by the application: voltage, bypass/non-bypass, number of poles, connection type, type of overcurrent protection device, etc. Not all manufacturers have their switches tested and listed for the ratings published in their product literature. It is necessary to pay close attention to footnotes in their product literature that qualify the published WCR.

Not all jurisdictions enforce the NEC as written, and some issue amendments that modify particular articles. For example, the city of Seattle issued amendments to NEC 2011, in particular amending article 700.27, effectively stating that selective coordination is required starting 0.1 sec or six cycles after an overcurrent condition occurs. Seattle has also prescribed that selective coordination need not be considered on the normal power supply side of the emergency transfer switch, effectively defining that selective coordination need only occur from the alternate source of power. Be wary of amendments issued by the jurisdiction of your project region and understand how these amendments might affect the coordination study and the transfer switch application. The design engineer faces a significant challenge in designing selectively coordinated emergency systems—in particular, the essential electrical branches in healthcare facilities. Proper specification and application of transfer switches is influenced by both the short-circuit study (WCR) and the results of the coordination study (withstand duration). Normal power sources and alternate power sources must by modeled in the studies to completely understand which source will drive the WCR of the required transfer switches.


Eriksen is principal at Affiliated Engineers’ Seattle office. His expertise is designing electrical power distribution systems for the built environment. With more than 29 years of experience in the field, Eriksen has designed electrical power systems for facilities serving numerous market segments including healthcare, research and development, museum, and performing arts.