How to design an electrical system in which the upstream protective device (fuse or circuit breaker) nearest to the system fault clears the fault without affecting the protective devices that are upstream from it
Selective coordination generally describes the design of an electrical system in which the upstream protective device (fuse or circuit breaker) nearest to the system fault clears the fault without affecting the protective devices that are upstream from it. Figure 1 (click here to see Fig. 1) illustrates this concept, where a fault is seen on a feeder circuit breaker in a panel. In this example, the fault on the feeder serving panel (M) is protected by circuit breaker (L). With proper action, this will not trip protective devices (G), (E), or (A) upstream — and the fault is contained so that buses (B) and (H) are unaffected, as are the panels marked (N).
Reviewing a typical protective device trip curve (click here to see Fig. 2), it can be divided into three regions. First is the “instantaneous region,” which is the region intended to interrupt high-level faults. On a UL breaker (typical molded-case unit), this area operates with no intentional delay. For ANSI-rated breakers, this region may operate with a delay, or may be defeated entirely, within the withstand capability of the breaker.
Next is the “short-time region,” which operates in a general range of between three and 30 cycles. The short-time region will cause the breaker to trip at a lower current than the instantaneous region, but acts after a short delay (by definition). This is a feature of many electronic-trip circuit breakers, and is not intentionally provided on standard thermal-magnetic, molded-case units. The final region is the “long-time region,” operating in the range of 30 cycles to beyond 8 min. (Fig. 2). This region causes a trip at the trip setting of the circuit breaker (i.e., a 1,600A circuit breaker will trip when the nominal current exceeds 1,600A for significantly more than 8 sec — approximately 60 sec for the breaker shown).
By their nature, breakers are more difficult to selectively coordinate than fuses. Their time current curves are not as smooth, and the tolerances of thermal-magnetic molded-case breakers (as indicated by their trip curves) are not as narrow as those of fuses. (Note that digital-trip units have tighter manufacturing tolerances, indicated by “thinner” curves.) However, they are reusable, reliable, and inherently prevent “single-phasing.”
Consider a typical electrical system, again referencing Fig. 1. A 75kVA step-down transformer (F) is serving a 208V emergency receptacle panel (H). Based upon a typical impedance of 2.6%, the available fault current at the secondary terminals of the transformer is more than 7,900A. The typical 250A molded-case secondary circuit breaker (G) will have an instantaneous trip of between five and 10 times the trip rating (depending on the manufacturer, model, and frame), or 1,250A to 2,500A. UL 489 standards allow for a manufacturing tolerance of -20% to +30%, which effectively means that a unit with a 2,500A rated instantaneous trip setting will actually trip instantaneously on a fault somewhere between 2,000A and 3,250A, which accounts for the width of the trip band on the manufacturer's coordination curves. Clearly, this will not easily coordinate in the instantaneous region.
However, selective coordination of circuit breakers is a possibility. In fact, in many cases, breakers that do not appear to selectively coordinate (based solely upon their individual curves) do selectively coordinate in specific pairings. This is possible using the same principle that allows series rating of breakers — dynamic impedance.
Circuit breakers, by their nature, do have current-limiting properties, which are inherently tied to the action of the breaker. When interrupting a fault, the breaker will “blow open,” separating its contacts and opening the circuit. This action causes a variable (or dynamic) impedance, which changes the short circuit current of the system, changing the action of the other circuit breakers. This impedance can allow certain pairs of breakers to coordinate to a much higher short circuit level than would be expected if the only available information were the individual protective device curves. The ability of these pairs to coordinate is determined by the manufacturer; therefore, only breakers from the same vendor may be paired, based upon the manufacturer's recommendation.
As the requirements for selective coordination have grown, breaker manufacturers are making information more readily available. This information is being provided in “coordination tables,” which are somewhat similar to (although more cumbersome than) the fuse manufacturer coordination ratios. A typical coordination table may indicate that a type QO breaker rated 20A (L) would coordinate with a type LA 200A circuit breaker (G), set to its maximum instantaneous trip of approximately 1,600A, indicating coordination to this level. However, a type LA-MC (mission-critical) 200A main breaker (G) coordinates with a 20A, single-pole, QO-type breaker (L), up to a fault current of 18kA, even though their individual trip curves would indicate coordination only to approximately 2,700A. This is an important and valuable tool in designing a selectively coordinated system.
Unfortunately, these tables do not “reflect” through a transformer. Therefore, although you can see that the 200A mission-critical main circuit breaker coordinates with the downstream QO branch breaker, the QO breaker must also coordinate with the breaker protection on the primary side of the transformer. In the case noted above, the short circuit current seen by the primary of the transformer, based upon a fault on the secondary side of the transformer (commonly known as the “through-fault current”) would be approximately 3,400A. Thus, the primary side breaker must be able to hold 3,400A+. In order to accommodate this requirement, it would be necessary to provide a minimum circuit breaker of 225A (mission-critical), which would have an instantaneous trip rating of approximately 4,000A. This is much larger than the typical 125A to 150A molded-case circuit breaker used for protection of a 75kVA transformer.
Where coordination of molded-case circuit breakers proves to be unwieldy or impossible, there are other options available for selective coordination. These include the use of electronic trip breakers, which can be “trimmed” to provide greater coordination. In certain instances, ANSI-rated breakers will allow the elimination of the instantaneous region of the electronic breakers, which requires the use of 30-cycle-rated (ANSI-rated) switchgear. However, it should be noted that most breakers also have an “instantaneous override,” a factory setting that is not adjustable. This will cause the breaker to trip in order to protect itself, at a specified rating, effectively causing an instantaneous trip — regardless of the lack of an instantaneous trip rating.
These methods are more versatile than coordination of the fixed trip curves available for molded-case circuit breakers. Nevertheless, this flexibility comes at a price. Electronic trip circuit breakers are more costly than the molded-case variety, and must be properly set and field maintained to assure the curves in the field match the intent of the drawings. Furthermore, use of 30-cycle (ANSI-rated) equipment requires that items protected downstream be properly braced to withstand a 30-cycle fault, and will likely expose downstream components to greater arc flash energy in the event of an arcing-type fault.
In addition to the methods described above, there are a few other ways to enhance the ability to selectively coordinate the system. The most important is to minimize the number of “levels” in the path of coordination. Naturally, a coordination path with three protective devices will be easier to design, maintain, and coordinate than a path with six such devices.
One of the easiest ways to reduce these levels is to eliminate main circuit breakers in panel risers. An example of this is a typical emergency lighting riser. In some cases, a 400A riser feeder will be used to feed several emergency lighting panels. Normally, each of these panels would have its own main circuit breaker (typically 100A). By eliminating these breakers, you can eliminate a “level” in the path of coordination. This increases the ability to coordinate the system, but has the drawback of requiring the entire riser to be shut down for service, as the individual panels do not have main protective devices (use of non-automatic main breakers and unfused disconnect switches is limited, as many such devices do not have a sufficient withstand rating for use in these applications). Furthermore, a fault on the panel bus must, by its nature, be protected by the 400A upstream breaker, taking down the entire riser.
In systems with larger fault currents or large transformers, additional levels in the system can quickly force the use of drawout equipment and ANSI breakers. However, the same principles described above still apply. Referring back to Fig. 1, you can assume the fault current at 4,000A main breaker (B) is approximately 45kA, and that circuit breaker (E) is 800A, serving a 500kVA transformer (F). Secondary breaker (G) is 1,600A, and breaker (L) is a 200A feeder serving panel (M). Working backward, you can see how the breaker requirements are affected.
The 500kVA transformer, with an impedance of 5%, has a secondary fault current of 27.8kA and a through-fault current of 12kA. Assuming a sufficient distance to breaker (G) in panel (H) to reduce the available fault current below 22kA — and a further distance to lighting panel (M), which reduces the available fault current at the panel to 5kA — you will need to select a 20A breaker and 200A circuit breaker to be coordinated to a fault level of 5kA. If the lighting panel has standard “QO-VH”-type breakers, breaker (L) will need to be a mission-critical type breaker, which will coordinate to 10kA for 3-pole breakers.
Moving upstream, breaker (G) must also coordinate with breaker (L). This 1,600A breaker needs to remain closed under 5kA of fault current on panel (H). This is relatively easy, given the great difference in sizes. A typical R-frame breaker has an adjustable instantaneous rating of five times the pickup rating of the breaker, or 8,000A, in this case.
However, on the primary of the transformer, breaker (E) does not need to coordinate with breaker (G), as they have the same zone of coverage — but it must coordinate with downstream breaker (L). Because there are no manufacturer tables for testing of breakers through a transformer, you must select a breaker that can carry an instantaneous current of 2.1kA, which is the 5kA fault current at breaker (L) reflected onto the primary of the transformer. Again, the standard 800A breaker (in this case M-frame) is adjustable from five to 10 times the pickup rating of the breaker, for a minimum setting of 4,000A, which will easily clear a 2.1kA fault.
Finally, 4,000A breaker (A) must coordinate with breaker (E). This is a somewhat more difficult coordination, as the available fault current at both breakers is known to be 45kA. This will require a power-type circuit breaker (Type NW, for example). This breaker has no instantaneous override, and can withstand a fault current of 65kA, without tripping.
One last concern in coordination of systems is ground fault protection. The NEC requires ground fault protection for systems between 150V and 600V to ground, for all electrical services rated 1,000A and above (per 230.95). For health care systems, at least two levels of ground fault protection are required (per 517.17). It is required that this ground fault protection be coordinated within the system.
Normally, the phase curves are coordinated, and the ground fault curves are coordinated separately. However, in cases where a protective device does not have a ground fault trip sensor, a ground fault on its load side will be treated the same as a phase fault. Therefore, it's imperative the phase trip curve be coordinated with the upstream ground fault protection. As the trip setting for ground fault protection is limited to 1,200A by NEC requirements, coordination with the downstream phase trip curve will require that the delay setting of the ground-fault protective device be adjusted to coordinate with the downstream breaker.
Often, the ground-fault short circuit current available on a 20A feeder circuit is sufficient to trip the ground-fault protection of a building main circuit breaker, which makes it imperative that the ground-fault sensors be properly set up to assure selective coordination.
Selective coordination is an important part of designing emergency electrical systems. Full compliance with the selective coordination requirements of the NEC is mandatory in most jurisdictions. Although there are often tradeoffs with other safety considerations, the latest version of the Code makes it clear that selective coordination is of paramount importance in life safety, critical emergency, and critical operations power systems.
Although this can be achieved in several ways, it's important to remember that changes to systems made to achieve selective coordination will affect how they behave in other ways most notably, the available arc flash energy calculated on the system. The more drastic the method used to accommodate selective coordination requirements, the more likely the solution will have a negative effect on the arc flash energy in the system and drive up the cost of the installation. Therefore, you should consider less drastic methods (such as using coordination tables) to maximize the safety and reliability of the electrical system design.
Finally, it's imperative that ground fault coordination be considered in all cases where selective coordination is a requirement, especially in cases where an upstream breaker has ground fault protection and the downstream breaker does not. Without this key piece, a seemingly minor downstream event could cause the entire electrical system to fail.
Medich, P.E., is a senior electrical project engineer with Ballinger in Philadelphia. He can be reached at bmedich@ballinger-ae.com.
Selective coordination is required by the NEC for the following systems:
Arc flash energy is directly related to the available fault current and directly proportional to the arcing time. Therefore, changing the trip characteristics of a protective device will affect the arc flash energy available on the load side of that device.
As discussed in this article, the goal of selective coordination is to ensure that a fault downstream of a branch breaker does not trip an upstream device, causing a cascading, widespread outage. In order to accomplish this, you usually need to ensure that the upstream breaker will not trip in the instantaneous region for a fault on the load side of a branch breaker (i.e., for a 20A branch with a fault capacity on its load side of 1,500A, you need the upstream “main” to carry 1,500A without tripping in the instantaneous region. This would typically be accomplished by a standard 200A main circuit breaker). However, if the available fault is above 1,500A, you would need a larger “main” breaker. This larger breaker would react more slowly to a fault on the bus ahead of the branch breaker; therefore, it would cause the arc flash energy available from a bus fault to rise.
This increase in arc flash energy could be somewhat mitigated by using the proper breaker pair (where available) from the manufacturers' dynamic impedance tables. This would account for the slower action of the main breaker due to the dynamic impedance of the branch breaker for faults on the load side of that breaker, but still maintain the operation of the main breaker in accordance with its individual rating for a line side fault ahead of the branch breaker (where the dynamic impedance of the branch breaker is not a factor in the operation of the “main”). Naturally, these types of pairs will likely maintain a lower arc flash energy in the system while maintaining the required selective coordination.
Current-limiting fuses are manufactured in such a way that under a high-level fault, the fuse link will melt within the first half-cycle, limiting the amount of fault current in the system to “Ip,” which is known as the “let-through” current. Because of this capability, fuse manufacturers are able to publish a set of “fuse coordination ratios,” which will guarantee selective coordination at any fault level. The fuse ratios are anywhere between 1.5 and 8-to-1, depending upon the type of fuses used, with type RK5 time-delay fuses having the lowest ratio when used upstream, and the highest ratio when used downstream, due to their time delay properties. Conversely, fast-acting fuses (type L, for example) have higher ratios when used upstream and lower ratios when used downstream.