A small multifamily dwelling service change reveals interesting Code applications.

One of the best features of the electrical trade is that every day, problems need to be solved that aren't quite the same as the ones we solved yesterday. Changing the service at a small multifamily dwelling proved this point dramatically.

When we think of service design, we often think in terms of designing the size of the service equipment so it can safely carry the projected load. In this case, that was secondary because the service was designed with a great deal of headroom in the first place. The actual load at the time of installation was under 200A, but the owner wanted 400A in total capacity to accommodate possible commercial loading in the future. In this case, the majority of the design issues related to projected occupancies, tenant access, and the probable likelihood and effect of future utility changes to the supply system.

Tenant access. The existing service was on the opposite side of the basement from the hatchway. Since there was no continuous management supervision, this meant that the entire basement had to be accessible to the tenants so they could reach their overcurrent devices. A key objective was to relocate the service and tenant disconnects so the remainder of the basement could be locked and kept for the exclusive use of the owner. In this way, the tenants would still have the access required by Sec. 230-72(c) for the service equipment and Sec. 240-24 (b) for the overcurrent devices applicable for each such tenancy.

Future occupancies. At the present time, the building has only two apartments, one very large and one at the far end of the building with a remote 125A panel in that apartment. However, the building was extensively reconfigured in the early 1960s, including separate hot water plumbing risers, such that the one large apartment could be divided into three apartments. One of these "apartments" even had its own panel, served by a 100A feeder. A key objective was to preserve this flexibility.

In addition, the owner didn't want to be subject to the six-disconnect constraint in Sec. 230-71(a) and was willing to pay for a single main disconnect as part of this work. Similarly, the owner wanted the full 400A capacity of the service switch, even though, as mentioned previously, the load calculation came out much lower.

Commercial use. With the increasing prevalence of people working at home, it seemed likely that a tenant or possible future owner might want to use one of the spaces for a home occupation, such as a psychologist's office. That had been done in the past and was allowable under the prevailing zoning. Therefore, another objective was to provide for future commercial metering in at least one occupancy. The serving utility requires demand metering for such occupancies, and a full lever bypass for those meters. The normal residential meters cannot be used, so the design had to accommodate some commercial metering.

The present fire alarm system serves the large apartment; however, if that apartment is subdivided, a complete system for the building would be required, if not at the time of construction then certainly at the time of resale, in accordance with local law. In addition, there would be some public area lighting. At that time, these common systems would require a separate meter loop in accordance with Sec. 210-25. The serving utility requires commercial-grade metering with demand meters and a full lever bypass for these circuits as well. Therefore, the design had to provide space for at least two commercial meters in the future.

Limited space. The logical place for the new service location was at the bottom of the hatchway. Many years ago, this area had been partitioned with a brick wall to allow for coal storage. Now, oil tanks occupied that storage area, with the brick wall remaining. The other side of that wall, adjacent to the hatchway, made a logical space for the electrical equipment, if it would fit. See plan view diagram of the electrical room on page 64. In addition to horizontal space limitations, after the ceiling was insulated and sheet-rocked, the ceiling height was only 6 ft 7 in.

Since there must be clear workspace in front of the panels extending to a height of 6 ft 6 in., this meant that any ceiling fixtures (deeper than 1 in.) had to be at least 3 ft back from the panels. Since the doors were 6 ft 6 in. as well, this made for an interesting geometry problem on the ceiling.

Energy efficiency. The space was perfectly positioned to provide a thermal air-lock from the rest of the basement. The owner was willing to bring in a carpenter to frame and insulate the walls. As noted, the insulated steel doors were 6 ft 6 in. high, so anything on the ceiling had to be out of the way so they could swing.

Available fault current. At present, the available fault current is below 10kA, and this seems likely to continue due to the type of distribution on the street. In the unforeseen event that the utility system is upgraded to the point that this exceeds 10kA, the tenant mains for each feeder would need to be upgraded to breakers of the same physical dimensions that would have an acceptable series-connected rating with the 10kA branch breakers. The owner was prepared to cross that bridge when he came to it.

He didn't, however, want to put all his protection eggs in the same type of basket. The electrical room area is fairly humid, and he wanted a fuse as the main protection. He said, "I never saw a fuse corrode so it wouldn't blow." Therefore, the contractor used a fused switch for the service disconnect, with circuit breakers for the other protective devices for convenience.

The owner was well aware that the fuse did not, in and of itself, create a series-rated system, even though it could clear 200kA safely. There are two basic procedures for handling high available fault current. One is to fully-rate the system, and the other is to series-rate the system. A fully-rated system utilizes overcurrent protective devices at all levels in the distribution that are independently capable of clearing the fault current available at their supply terminals. A series-rated system relies on two or more overcurrent devices in series to clear a heavy fault. For example, a 22kA-rated feeder breaker ahead of a 10kA-rated branch-circuit breaker may allow the 10kA breaker to safely clear 22kA faults by distributing some of the energy on clearing over its own contacts. These ratings must be verified by test and cannot be calculated in the field.

Both systems are safe. The series-rated systems tend to be significantly less expensive because they use easier-to-manufacture components that are made in far greater quantities, leading to significant economies of scale. On the other hand, in the event of a major fault, a series rated system is much more likely to produce a wide blackout, since upstream feeder protection normally opens simultaneously with the protective device nearest the fault.

In this case, the owner made the design decision that a major fault, such as would be produced on start-up on a large parallel feeder with mislabeled parallel conductors, was very unlikely. Faults on this type of system are likely to be arcing ground-faults and, as such, they seldom exceed 10kA regardless of what is available due to the impedance of the arc. In addition, on 120-240V systems, arcing tends to extinguish each half-cycle as the voltage goes through its zero crossing. Finally, the consequence for an outage in residential or related occupancies just isn't the disaster it could be for some industrial processes.

The resulting design choice was to use 10kA branch devices, and substitute higher-rated feeder protective devices in the future that had been verified by test as suitable to protect those branch devices. As previously noted, the owner considered the odds of this being a real problem remote, given the nature of the distribution.

In terms of design, this principle of needing to verify a series combination by actual laboratory testing is crucial. You cannot simply calculate the type of fuse that will work in series with a circuit breaker, due to a phenomenon called dynamic impedance. This means that a breaker may begin to open in the first half-cycle of a fault, even though it isn't rated as a current-limiting breaker. If this happens the current flowing into the fault will decrease, perhaps rapidly enough so the fuse never blows. Meanwhile, the breaker has unlatched and continues to open, ending up clearing a fault far larger than its rated ability. This could lead to a disaster.

What fuse type? The owner was concerned about the time-current characteristics of the commonly available fused switches in these metering assemblies (usually using Class T fuses). These fuses, while perfectly safe, are very fast, and switches built exclusively for Class T spacings can't be adapted for conventional fuses. The owner wanted the design flexibility to use RK-1 or even RK-5 fuses that have greater time delay. He foresaw a likelihood that he might have a significant motor load occasionally connected in his part of the building.

A key factor in the brand selection, therefore, was the option, seldom available in today's equipment and oddly unpublicized in this manufacturer's catalog, to order the equipment with conventional spacings. As shipped, the international J rated fuses can be readily installed in this switch as well, using adapters readily available from this manufacturer.

Wiring methods. The owner wanted a clean line visually on the side of the building, and selected aluminum rigid conduit. In addition to mechanical durability, it would be unlikely to have a corrosion problem. This could be carried into the basement through the foundation and into the new electrical room. Note that although aluminum conduit cannot be run in poured concrete without supplementary protection due to listing restrictions, short sections can be embedded in mortar, as in this case where it passed through some brickwork on the fieldstone foundation.

For the communications riser, the contractor selected 1 1/4 IMC. This was the smallest size that would put the fill below 31% based on the actual sizes of the 12-pail telephone cable and the CATV cable that were installed. Note that this configuration of telephone cable is elliptical, so it was taken as a circle equal in diameter to the major axis of the elliptical cross section, in accordance with NEC Chapter 9, Table 1, Note 9.

There was a two-fold reason for using conduit. The outdoor versions of these cables are flammable, which is why Secs. 800-50 and 820-50 don't allow their use within buildings. Although exceptions allow up to 50 ft exposed within buildings, the owner chose to continue the conduit inside. This also allowed the telephone cable protectors and the CATV shield ground to be in the electrical room. Sec. 800-30(b)and Sec. 820-33 require these protective devices to be as near as practicable to the point of building entrance; however, grounded IMC and rigid metal conduit are recognized as extending that point of entrance to their termination.

The overall aesthetic result, as seen in the photo, was that the owner didn't have to look at a large meter grouping or telephone protectors or shield grounds on the outside of the building, just two clean vertical lines. The serving utility agreed to the basement metering location because of the multifamily nature of the building and the fact that it would have ready access.

Grounding electrode conductor. Although skeptical of any real problems from electromagnetic field radiation (EMF), the owner decided to route the grounding electrode conductor as far away as practicable from the living areas upstairs. A grounding electrode conductor is by far the most efficient magnetic field radiator because it only has current flowing in one direction. This type of source decays directly with distance, instead of by the square or even the cube of the distance as in the case of other sources. The contractor managed to find a route along the floor running next to a wall most of the way, and he used rigid metal conduit for physical protection.

Next month

Next month we'll look at many unique construction techniques that were involved in building on this design. We'll show, for example, how you can get 50 or more circuits out of the panels and over to the branch-circuit pull box in a single raceway system without derating.