The design strength of nails is greater when a nail is driven into the side rather than the end grain of a member. Withdrawal information is available for nails driven into the side grain; however, the withdrawal capacity of a nail driven into the end grain is assumed to be zero because of its unreliability. Furthermore, the NDS does not provide a method for determining withdrawal values for deformed shank nails. These nails significantly enhance withdrawal capacity and are frequently used to attach roof sheathing in high-wind areas. They are also used to attach floor sheathing and some siding materials to prevent nail back-out. The use of deformed shank nails is usually based on experience or preference.
The design shear value Z
for a nail is typically determined by using the following tables from NDS•12:
- Tables 12.3A and B. Nailed wood-to-wood, single-shear (two-member) connections with the same species of lumber using box or common nails, respectively.
- Tables 12.3E and F. Nailed metal plate-to-wood connections using box or common nails, respectively.
The yield equations in NDS•12.3 may be used for conditions not represented in the design value tables for Z
. Regardless of the method used to determine the Z
value for a single nail, the value must be adjusted, as described in Section 7.3.2. As noted in the NDS, the single nail value is used to determine the design value.
It is also worth mentioning that the NDS provides an equation for determining allowable design value for shear when a nailed connection is loaded in combined withdrawal and shear. The equation appears to be most applicable to a gable-end truss connection to the roof sheathing under conditions of roof sheathing uplift and wall lateral load owing to wind. The designer might contemplate other applications but should take care in considering the combination of loads that would be necessary to create simultaneous uplift and shear worthy of a special calculation.
Bolts may be designed in accordance with NDS•8 to resist shear loads in wood-to-wood, wood-to-metal, and wood-to-concrete connections. As mentioned, many specialty bolt-type fasteners can be used to connect wood to other materials, particularly concrete and masonry. One common example is an epoxy-set anchor. Manufacturer data should be consulted for connection designs that use proprietary fastening systems.
The design shear value Z
for a bolted connection is typically determined by using the following tables from NDS•8:
- Table 8.2A. Bolted wood-to-wood, single-shear (two-member) connections with the same species of lumber.
- Table 8.2B. Bolted metal plate-to-wood, single-shear (two-member) connections; metal plate thickness of 1/4-inch minimum.
- Table 8.2D. Bolted single-shear wood-to-concrete connections; based on minimum 6-inch bolt embedment in minimum fc = 2,000 psi concrete.
It should be noted that the NDS does not provide W
values for bolts. The tension value of a bolt connection in wood framing is usually limited by the bearing capacity of the wood, as determined by the surface area of a washer used underneath the bolt head or nut. The bending capacity of the washer should be considered. For example, a wide but thin washer will not evenly distribute the bearing force to the surrounding wood.
The arrangement of bolts and drilling of holes are extremely important to the performance of a bolted connection. The designer should carefully follow the minimum edge, end, and spacing requirements of NDS•8.5.
Any possible torsional load on a bolted connection (or any connection, for that matter) should also be considered in accordance with the NDS. In such conditions, the pattern of the fasteners in the connection can become critical to performance in resisting both a direct shear load and the loads created by a torsional moment on the connection. Fortunately, this condition is not often applicable to typical light-frame construction. However, cantilevered members that rely on connections to anchor the cantilevered member to other members will experience this effect, and the fasteners closest to the cantilever span will experience greater shear load. One example of this condition sometimes occurs with balcony construction in residential buildings; failure to consider the effect discussed above has been associated with some notable balcony collapses.
For wood members bolted to concrete, the design lateral values are provided in NDS•Table 8.2 E. The yield equations (or general dowel equations) may also be used to conservatively determine the joint capacity.
Lag screws (or lag bolts) may be designed to resist shear and withdrawal loads in wood-to-wood and metal-to-wood connections, in accordance with NDS•9. As mentioned, many specialty screw-type fasteners can be installed in wood. Some tap their own holes and do not require pre-drilling. Manufacturer data should be consulted for connection designs that use proprietary fastening systems.
The withdrawal strength of a lag screw (inserted into the side grain of lumber) is determined in accordance with either the empirical design equation below or NDS•Table 9.2A. It should be noted that the equation below is based on single lag screw connection tests and is associated with a reduction factor of 0.2 applied to average ultimate withdrawal capacity to adjust for load duration and safety. Also, the penetration length of the lag screw Lp
into the main member does not include the tapered portion at the point.
The allowable withdrawal design strength of a lag screw is greater when the screw is installed in the side rather than the end grain of a member. However, unlike the treatment of nails, the withdrawal strength of lag screws installed in the end grain may be calculated by using the Ceg
adjustment factor with the equation above.
The design shear value Z
for a lag screw is typically determined by using the following tables from NDS•9:
- Table 9.3A. Lag screw, single-shear (two-member) connections with the same species of lumber for both members.
- Table 9.3B. Lag screw and metal plate-to-wood connections.
The yield equations in NDS•9.3 may be used for conditions not represented in the design value tables for Z
. Regardless of the method used to determine the Z
value for a single lag screw, the value must be adjusted.
System Design Considerations
As with any building code or design specification, the NDS provisions may or may not address various conditions encountered in the field. There may be alternative or improved design approaches. Similarly, some considerations regarding wood connection design are appropriate to address here.
First, as a general design consideration, crowded connections should be avoided. If too many fasteners are used (particularly nails), they may cause splitting during installation. When connections become crowded, an alternative fastener or connection detail should be considered. Basically, the connection detail should be practical and efficient.
Second, while the NDS addresses system effects within a particular joint (i.e., element) that uses multiple bolts or lag screws (i.e. the group action factor Cg
), it does not include provisions regarding the system effects of multiple joints in an assembly or system of components. Therefore, some consideration of system effects is given below based on several relevant studies related to key connections in a home that allow the dwelling to perform effectively as a structural unit.
Sheathing Withdrawal Connections
Several past studies have focused on roof sheathing attachment and nail withdrawal, primarily as a result of Hurricane Andrew (HUD, 1999a; McClain, 1997; Cunningham, 1993; Mizzell and Schiff, 1994; and Murphy, Pye, and Rosowsky, 1995). The studies identify problems related to predicting the pull-off capacity of sheathing based on single-nail withdrawal values and determining the tributary withdrawal load (i.e., wind suction pressure) on a particular sheathing fastener. One clear finding, however, is that the nails on the interior of the roof sheathing panels are the critical fasteners (i.e., initiate panel withdrawal failure) because of the generally larger tributary area served by these fasteners. The studies also identified benefits of the use of screws and deformed shank nails. However, the use of a standard geometric tributary area of the sheathing fastener and the wind loads, along with the NDS withdrawal values, will generally result in a reasonable design using nails. The wind-load duration factor should also be applied to adjust the withdrawal values, since a commensurate reduction is implicit in the design withdrawal values relative to the short-term, tested and ultimate withdrawal capacities.
It is interesting to note, however, that one study found that the lower-bound (i.e., 5th percentile) sheathing pull-off resistance was considerably higher than that predicted by the use of single-nail test values (Murphy, Pye and Rosowsky, 1995). The difference was as large as a factor of 1.39 greater than the single-nail values. While this would suggest a withdrawal system factor of at least 1.3 for sheathing nails, it should be subject to additional considerations. For example, sheathing nails are placed by people using tools in somewhat adverse conditions (i.e., on a roof), and not in a laboratory. Therefore, this system effect may be best considered as a reasonable construction tolerance on actual nail-spacing variation relative to that intended by design. Thus, an 8- to 9-inch nail spacing on roof sheathing nails in the panel’s field could be tolerated when a 6-inch spacing is targeted by design.
A couple of studies have investigated the capacity of roof-to-wall (i.e., sloped rafter-to-top plate) connections using conventional toe-nailing and other enhancements (i.e., strapping, brackets, gluing, etc.). Again, the primary concern is related to high wind conditions, such as those experienced during Hurricane Andrew and other extreme wind events.
First, as a matter of clarification, the toenail reduction factor Ctn
does not apply to slant-nailing, such as those used for rafter-to-wall connections and floor-to-wall connections in conventional residential construction. Toe-nailing occurs when a nail is driven at an angle in a direction parallel to the grain at the end of a member (i.e., a wall stud toenail connection to the top or bottom plate that may be used instead of end nailing). Slant nailing occurs when a nail is driven at an angle, but in a direction perpendicular to the grain through the side of the member and into the face grain of the other (i.e., from a roof rafter or floor band joist to a wall top plate). Though this is a generally reliable connection in most homes and similar structures built in the United States, even a well-designed slant-nail connection used to attach roofs to walls is impractical in hurricane-prone regions or similar high-wind areas. In these conditions, a metal strap or bracket is preferable.
Based on the studies of roof-to-wall connections, five key findings are summarized as follows (Reed et al., 1996; Conner et al., 1987):
Heel Joint in Rafter-to-Ceiling Joist Connections
- In general, it was found that slant-nails (not to be confused with toenails) in combination with metal straps or brackets do not provide directly additive uplift resistance.
- A basic metal twist strap placed on the interior side of the walls (i.e., gypsum board side) resulted in top plate tear-out and premature failure. However, a strap placed on the outside of the wall (i.e., structural sheathing side) was able to develop its full capacity without additional enhancement of the conventional stud-to-top plate connection (see Table 1).
- The withdrawal capacity for single joints with slant nails was reasonably predicted by NDS with a safety factor of about 2 to 3.5. However, with multiple joints tested simultaneously, a system factor on withdrawal capacity of greater than 1.3 was found for the slant-nailed rafter-to-wall connection. A similar system effect was not found on strap connections, although the strap capacity was substantially higher. The ultimate capacity of the simple strap connection (using five 8d nails on either side of the strap–five in the spruce rafter and five in the southern yellow pine top plate) was found to be about 1,900 pounds per connection. The capacity of three 8d common slant nails used in the same joint configuration was found to be 420 pounds on average, and with higher variation. When the three 8d common toenail connection was tested in an assembly of eight such joints, the average ultimate withdrawal capacity per joint was found to be 670 pounds, with a somewhat lower variation. Similar system increases were not found for the strap connection. The 670-pound capacity was similar to that realized for a rafter-to-wall joint using three 16d box nails in Douglas fir framing.
- It was found that the strap manufacturer’s published value had an excessive safety margin of greater than 5 relative to average ultimate capacity. Adjusted to an appropriate safety factor in the range of 2 to 3 (as calculated by applying NDS nail shear equations by using a metal side plate), the strap (a simple 18g twist strap) would cover a multitude of high-wind conditions with a simple, economical connection detail.
- The use of deformed shank (i.e., annular ring) nails was found to increase dramatically the uplift capacity of the roof-to-wall connections using the slant nailing method.
The heel joint connection at the intersection of rafters and ceiling joists has long been considered one of the weaker connections in conventional wood roof framing. In fact, this highly stressed joint represents one of the significant reasons for using a wood truss, rather than conventional rafter framing (particularly in high-wind or snow-load conditions). However, the performance of conventional rafter-ceiling joist heel-joint connections should be understood by the designer, since they are frequently encountered in residential construction.
First, conventional rafter and ceiling joist (cross-tie) framing is simply a site-built truss. Therefore, the joint loads can be analyzed by using methods that are applicable to trusses (i.e., pinned joint analysis). However, the performance of the system should be considered. As mentioned earlier for roof trusses, a system factor of 1.1 is applicable to tension members and connections. Therefore, the calculated shear capacity of the nails in the heel joint (and in ceiling joist splices) may be multiplied by a system factor of 1.1, which is considered conservative. Second, it must be remembered that the nail shear values are based on a deformation limit, and generally have a conservative safety factor of 3 to 5, relative to the ultimate capacity. Finally, the nail values should be adjusted for duration of the load (i.e., snow load duration factor of 1.15 to 1.25). With these considerations and with the use of rafter support braces at or near mid-span (as is common), reasonable heel joint designs should be possible for most typical design conditions in residential construction.
When wood sole plates are connected to wood floors, many nails are often used, particularly along the total length of the sole plate or wall bottom plate. When connected to a concrete slab or foundation wall, there are usually several bolts along the length of the bottom plate. This points toward the question of possible system effects in estimating the shear capacity (and uplift capacity) of these connections for design purposes.
In recent shear wall tests, walls connected with pneumatic nails (0.131-inch diameter by 3 inches long) spaced in pairs at 16 inches on center along the bottom plate were found to resist over 600 pounds in shear per nail. The bottom plate was spruce-pine-fir lumber and the base beam was southern yellow pine. This value is about 4.5 times the adjusted allowable design shear capacity predicted by use of the NDS equations. Similarly, connections using 5/8-inch-diameter anchor bolts at 6 feet on center (all other conditions equal) were tested in full shear wall assemblies; the ultimate shear capacity per bolt was found to be 4,400 pounds. This value is about 3.5 times the adjusted allowable design shear capacity, per the NDS equations. These safety margins appear excessive and should be considered by the designer when evaluating similar connections from a practical system standpoint.
Design of Concrete and Masonry Connections
In typical residential construction, the interconnection of concrete and masonry elements or systems is generally related to the foundation and usually handled in accordance with standard or accepted practice. The bolted wood member connections to concrete are suitable for bolted wood connections to properly grouted masonry. Moreover, numerous specialty fasteners or connectors (including power-driven and cast-in-place) can be used to fasten wood materials to masonry or concrete. The designer should consult the manufacturer’s literature for available connectors, fasteners, and design values.