Csa O86-14 Commentary.pdf 5z4d44

Commentary on CSA O86, “Engineering Design in Wood” 2014 Edition

Cl. 1-3

Scope, Reference Publications, Definitions, Symbols, Dimensions

Cl. 4

Objectives and Design Requirements

Cl. 5

General Design

Cl. 6

Sawn Lumber

Cl. 7

Glued-laminated Timber (Glulam)

Cl. 9

Structural s

Cl. 10

Composite Building Components

Cl. 11

Lateral-load-resisting Systems

Cl. 12

Connections

Cl. 15

Proprietary Structural Wood Products—Design

Cl. 16

Proprietary Structural Wood Products—Materials and Evaluation Appendix

The first CSA O86 Commentary was written by past Chair of CSA O86, Robert F. DeGrace, based on the 1984 limit states design edition and published by CSA in 1986. Since 1989, the CSA O86 Commentary has been written by of the CSA O86 Technical Committee and published in the Canadian Wood Council’s Wood Design Manual. This is the 7th edition of the CSA O86 Commentary and includes references to the 2014 edition of CSA O86. The contribution of the following to the development of this and earlier editions of the CSA O86 Commentary is gratefully acknowledged: J.D. Barrett, S. Boyd, Y.H. Chui, R.F. DeGrace, G.A. Dring, R.O. Foschi, D. Himmelfarb, D. Janssens, E. Jones, E. Karacabeyli, P. Lepper, C. Lum, G. Newfield, C. Ni, M. Popovoski, P. Quenneville, D. Rice, I. Smith, C.K.A. Stieda, J. E. Turnbull, J.B. Wang and G. Williams

Clauses 1 to 3 Scope, Reference Publications, Definitions, Symbols, Dimensions The scope of CSA Standard O86, “Engineering Design in Wood,” covers structural design and appraisal of structures and structural elements made from wood or wood products. The Standard includes design provisions for graded lumber, glued-laminated timber, plywood, oriented strandboard (OSB), composite building components, shearwalls and diaphragms, timber piling, pole-type construction, prefabricated wood I-joists, structural composite lumber products, preserved wood foundations and their structural fastenings. For structures or components of a specialized nature where the provisions in CSA O86 may not be directly applicable, designs still need to demonstrate an equivalent level of safety and performance, in accordance with the objectives of Part 4 of the National Building Code of Canada (NBC). Clause 3.3 notes the equivalence of “soft-” and “hard-” converted metric dimensions that relate to imperial unit values. For example, a maximum stud spacing of 600 mm actually connotes a maximum spacing of 610 mm (2 feet).

Clause 4

Objective and Design Requirements

Clause 4.2 – Limit states design Starting in the 1989 edition, design provisions were derived partly from principles of reliability-based (or probabilistic) design. The general objective of such an approach was to develop safe design guidelines by considering, in a consistent manner, all of the uncertainties involved in the design process. The following text describes the general approach to reliability analysis undertaken by Foschi et al (1989) at the University of British Columbia in the late 1980’s for the CSA O86 Technical Committee. Any design problem includes the interaction of several variables. For example, the design of a beam under combined bending and axial loads involves the beam’s axial and bending strengths, the modulus of elasticity, the beam dimensions and the loads themselves. Some of these variables may be known with more certainty than others. Test results provide information on strength and stiffness, surveys permit the development of statistics on maximum expected loads and duration of load pulses, and fabrication tolerances can be used to quantify the uncertainty in dimensions. Most design variables are random in nature, and reliability-based design procedures take these uncertainties into by using principles of probability theory. In general the design problem can be written in of a “performance function” G(X1, X2,…, XN), which is a function of the N design variables X. Some of these variables influence the “demand” D on the system, and some others characterize the “capacity” C to withstand the demand. If the function G is written, by convention, as G

= C–D

(Eq. 1)

non-performance of the system is indicated by negative values of G. That is, the combination of variables leads to demands greater than the capacity and G < 0. Conversely, G > 0 implies performance as intended, and G = 0 corresponds to the “limit state” between performance and non-performance. The probability of non-performance can be calculated as the probability of the event G < 0. This probability calculation can be performed using a well-established computer algorithm (Madsen, 1986). This algorithm computes the reliability index β.

The computer algorithm can be adjusted for the distribution of the random variables. For example, maximum annual snow loads are better represented by extreme-type probability distributions. The algorithm can also be adjusted for correlated variables. For example, tests show that the bending strength and the modulus of elasticity of wood are positively correlated (it is likely that a stronger piece is also stiffer). The algorithm used for the calculation of β was implemented in the computer program RELAN, which was developed at the University of British Columbia. This is a general program for First Order Reliability Analysis, implementing the Rackwitz-Fiessler algorithm for non-normal variables (Madsen, 1986) and the work of Der Kiureghian (1986) for handling the correlation between variables. Since the algorithm produces the reliability index β, a graph can be constructed relating β to the resistance factor φ. The resistance factor is assigned to a specified material property, and could be set to for variability of dimensions and properties, type of failure and uncertainty in the prediction of resistance. A typical graph is shown in Figure 1.

Figure 1 Typical relationship of reliability index β to the resistance factor φ

A target value for β for an adequate level of safety can be used as shown in Figure 1 to obtain the required value for φ in design. The curve in Figure 1 will not be the same if the statistical distribution of the component strength and/or loading is changed. Since only one parameter is being calibrated (φ), it is impossible to achieve the same level of reliability for all cases. The study of the performance function G shows how reliability can depend on the random variables. In a sense, G represents the actual structural behaviour, while the design equation is a conventional means of obtaining structural dimensions using nominal loads and strengths, with adjusting factors calibrated to ensure an adequate reliability. The calibration of these factors is done by a simultaneous study of the function G and the probability that G < 0. The different adjusting factors have been computed to ensure a reliability level comparable to that of steel structures and to provide for a nearly uniform reliability across different applications. For lumber, as an example, the design guidelines result in nearly uniform reliability across species, grades and sizes.

To illustrate the calibration procedure, consider the case of a simple beam under short term uniform loading. Let: R

= bending strength, MPa, a random variable for given beam size, species and grade

D

= long term distributed load, including dead and permanent loads, N/m, a random variable because loads are estimates

Q

= distributed transient load maximum over the service life of the beam, N/m, a random variable because of uncertainty in the estimation of this extreme load

L

= beam span, m

b, d

= beam dimensions, m

Dn

= nominal or “design” dead load, N/m, usually obtained by using average weights for materials

Qn

= nominal or “design” transient loads, N/m, (for example, the 30 year return load for snow at a given location)

Ro

= “characteristic” or “specified” bending strength for the beam, a lower percentile (the fifth) of the distribution of bending strength R

Assuming linear elastic behaviour to failure for the limit state of bending strength, the performance function G can be written as:

G = R − (D + Q)x

6L2 8bd2

(Eq. 2)

while the format for the design equation in the Code is:

(1.25Dn + 1.50Q n )x

6L2 = φRo 8bd2

(Eq. 3)

The factors 1.25 and 1.50 in this design equation are load factors common to all materials, but the factor φ is calibrated so that use of Eq. 3 ensures a proper reliability level. Eq. 3 permits writing:

Qnx

φRo 6L2 = 2 1.25γ + 1.50 8bd

(Eq. 4)

where the constant γ is the ratio between the nominal dead load and the nominal transient load:

γ=

Dn Qn

(Eq. 5)

The failure function G of Eq. 2 can also be written as

G = R − Q nx

Q  6L2  D +  2  8bd  Q n Q n 

(Eq. 6)

Eq. 4 can now be used to modify the performance function G in of the resistance factor φ. Introducing Eq. 4 into Eq. 6, G=R−

φR o x(γd + q) 1.25γ + 1.50

(Eq.7)

where dn and qn represent, respectively, the dead and transient loads normalized with respect to their corresponding design values: D Q dn = qn = Dn Qn ; (Eq. 8) Given statistical information on the random variables R, dn and qn, Eq. 7 can be used to study the probability of G < 0 for different values of the resistance factor φ. For the example, Figure 2 shows the β - φ relationships for lumber, averaged over three species, two grades and two sizes, when different cases for snow and occupancy loads are considered. The factor φ specified in the Standard has been chosen to provide an average reliability comparable to steel beams, over the range of species, sizes and grades, for five snow locations and two occupancy loads (offices and residential), for a nominal ratio of dead to transient loads of 0.25 (representative of light frame buildings). The resistance factor φ = 0.90 provides an average reliability index β = 2.6 for a design loads based on 1in-30 year return period with an overall range 2.4 < β < 2.9. (Note: this range is slightly larger than shown in Figure 2, due to the effect of material variability.) It should be noted that a calculated reliability index can be influenced by the methods chosen for assessment and characterization of material properties, data variables and distributions, as well as load assumptions. Dead-to-transient load ratios are another source of variation in calculating β values. Reliability decisions are based on review and committee judgment on these and other parameters, taking statistical assumptions and the historical record into consideration. Some relevant aspects of the method used for CSA O86 include the following: •

snow and occupancy load distributions were used independently to characterize load data;

•

a snow-to-dead load ratio of 4:1 was used as the reference point for comparisons with working stress design;

•

a 2-parameter Weibull distribution was used to characterize material property data; and

•

the lower tail of the material property distributions were used in reliability analysis to ensure a closer fit to the data and lessen the impact of distribution assumptions on the analysis.

The resistance assumptions were intended to accommodate many different input sources ranging from very large to very small data sets. Specified snow and wind loads in the 2005 National Building Code of Canada (NBC) were revised to a 1in-50 year return period. A 1-in-50 year return period means that on average the nominal value will be exceeded once in 50 years, or in other words, the probability of exceedance of the nominal load in any given year is 1/50. A review of reliability calculations on a 1-in-50 year basis indicated that this did not significantly alter the reliability basis for wood design, since load effects were offset by span or capacity effects. As a result of this review, no changes were made to the basic reliability method. A reliability analysis similar to the procedure described above for bending strength was followed for all other limit states whether related to strength or serviceability. The reliability calculations used the Canadian data available for strengths and loads. For lumber strength and stiffness, for example, the data were obtained from the lumber properties testing program conducted by the Canadian Wood Council (see Commentary on Clause 5). The data included short-duration testing results in bending, tension and compression. Models for loads included snow loading distributions provided by the National Research Council of Canada.

Figure 2 Summary of β - φ relation for five snow loads and two occupancy loads

The formulation of the performance functions G made use of research on the behaviour of wood structures. For example, in the case of columns axially loaded: G = Pmax – P

(Eq. 9)

where P is the applied end load (demand) and Pmax is the column resistance (capacity). Pmax depends on the slenderness of the column, the modulus of elasticity, the compression and tensile strength, and the amount of end eccentricity for the load P. Furthermore, the behaviour of wood in compression is nonlinear. The calculation of Pmax was carried out by a finite element analysis which closely approximated the column’s behaviour at all slenderness ratios. The complete set of research results used in the background calculations for the Standard have been compiled in Foschi et al, 1989.

References: Der Kiureghian, A. and P.L. Liu. 1986. Structural Reliability Under Incomplete Probability Information. J. Engineering Mechanics Division, ASCE, EM1. Foschi, R.O., Folz,B., Yao, F. 1989. Reliability-Based Design of Wood Structures. Structural Research Series, Report No. 34, University of British Columbia, Vancouver, BC. Madsen, H.O. et al. 1986. Methods of Structural Safety. Prentice-Hall, Inc. Englewood Cliffs, N.J.

Clause 5

General Design

Clause 5 of the Standard provides a concise summary of the design and load requirements in Part 4 of the National Building Code of Canada (NBC), as well as general considerations and factors affecting wood member resistance. The load and design provisions in the clause are updated periodically in accordance with changes to the NBC. s of the CSA O86 Standard should refer to the NBC and the ’s Guide – NBC Structural Commentaries (Part 4 of Division B) for complete information on design and load requirements.

Clause 5.2 – Specified loads, load effects and load combinations This section was revised in the 2005 Supplement to the 2001 edition of the Standard due to several major changes to Part 4 of the 2005 NBC, including the following: •

Specified snow and wind loads were converted from 1-in-30 year basis to 1-in-50 year basis,

•

Seismic loads and design procedures were revised significantly to take into structural rigidity and energy dissipation,

•

Buildings were divided into low, medium, high and post-disaster importance classes, with corresponding design factors,

•

Load factors were divided into principal and companion factors (“companion load” approach) and as a result combined loads were reduced in some design cases, and

•

Serviceability load factors were delegated to the material design standard committees.

Most of these revisions are not material-specific but they raised some issues for material design standard committees. For example, the CSA O86 reliability-based design method was based on a 30-year design life, as discussed in the Commentary on Clause 3 (Foschi et al, 1989). It should be noted that reliability analysis depends on several assumptions, including parametric distributions for loading and material property data. Although some assumptions have changed since this work was carried out in the late 1980’s, the basic conclusions have remained the same. After review, the CSA O86 Technical Committee considered it unnecessary to revise the approach to reliability calculations at the time. Serviceability considerations are important for wood elements, since wood member design is often governed by serviceability criteria. The “companion load” approach to load combinations increased the potential for deflection, although this was partially offset in some cases by increases in member size as a result of higher snow loads based on the 1-in-50 year return period, and other design changes. Load factors for serviceability in Tables 5.2.3.2 and 5.2.4.2 were approved after a review of bending member impact studies. See also Commentary on Clause 5.5 for additional information on serviceability. Further revisions were made in the 2014 edition of the Standard to reflect major changes to Part 4 of the 2015 NBC, including the following: •

For Ultimate Limit States design with Snow (S) and Live Load (L) combinations, the companion load factor for Snow or Live Loads was changed from 0.5 to 1.0. The companion load factor for L shall be further increased by 0.5 for storage areas, equipment areas and service rooms.

•

Seismic hazard data were updated based on the probabilistic model incorporating Cascadia subduction zone and updated shaking relations. Also seismic loads and design procedures were revised significantly.

•

Wood-frame construction of up to six storeys is allowed in NBC 2015.

Clause 5.2.1 – Buildings A note was added to the 2009 edition of CSA O86 advising designers that Clause 11 contained seismic provisions additional to those in Division B of the NBC. This guidance was consistent with a provision in the NBC directing designers to the material design standards for specific seismic design provisions.

Clause 5.2.3 – Specified loads Consistent with the 2005 NBC, the 2005 Supplement to the 2001 edition of CSA O86 introduced companion load combinations for dead, snow, live (occupancy), wind and earthquake loads. Specified loads were redefined including the importance factors in Table 4.2.3.2. Note that loads P (prestress effects) and T (temperature change effects) do not typically figure in calculations for wood structures, although designers should be aware that Clause 4.3.4 identifies upper bounds for wood temperature in the Standard, and moisture content effects are considered in factored resistance of each member and its fastenings (see Clause 5.3). In the 2015 edition of NBC, the sections on Snow Load and Wind Load were significantly rewritten incorporating most of the Commentary material in previous editions of the NBC with some revisions. Readers should refer to Part 4 of NBC 2015 for updated provisions on the calculation of specified wind and snow loads.

Clause 5.2.4 – Load combinations The principal and companion load factors for ultimate limit states are in accordance with the NBC. The factors for serviceability limit states were adopted by CSA O86 specifically for wood design.

Clause 5.3 – Conditions and factors affecting resistance Specified strengths are located in appropriate clauses of the Standard. The CWC publication on “Standard Practice Relating Specified Strengths of Structural to Characteristic Structural Properties” explains the basis for deriving specified strengths for sawn lumber. These same principles have been used to guide derivation for other wood products in CSA O86, with the intent of using the same target reliability index and statistical analysis methods.

Clause 5.3.2 – Load duration factor, KD One of the distinctive characteristics of wood is that its strength is influenced by the intensity and duration of the applied load. In some sense, this phenomenon is similar to that of fatigue in metals. In wood, however, strength loss is observed even under static loading of a given intensity. Should the applied stress be sustained, the strength loss phenomenon may cause failure after a given time. It is important to quantify this “load duration” effect to ensure reliability throughout the structure’s service life when subjected to various loading sequences. The load duration effect is treated in design as an adjustment to the design stress. These adjustments were based on experimental results collected at the US Forest Products Laboratory, Madison, WI, during the 1940s. These experiments were conducted with small Douglas fir specimens, free of defects, tested in bending. Using these results, a curve (called the “Madison curve”) was developed in earlier standards to explain load duration effects on strength (Wood, 1960). Based on the curve, short term strengths from ramp-load tests (lasting from two to ten minutes) were divided by 1.62 to estimate long term strengths corresponding

to a ten year load duration. This ten year reference point was termed “normal” duration, corresponding to loads normally encountered in floors of residential, business, commercial and assembly occupancies. For shorter cumulative durations, the design stresses were adjusted upward. For example, the effect of snow loads over the service life of the structure was made equivalent to a two-month continuous application, for a factor of 1.15. A downward adjustment of 0.90 was applied when a longer term load lasted for the entire life of the structure (longer than ten years). Research conducted at the University of British Columbia during the early 1970s (Madsen, 1971) using long term bending tests of Douglas fir lumber in structural sizes, showed that the load duration effect for material with defects was substantially different from that of small, clear specimens. The UBC research showed that the strength degradation was, overall, less severe than predicted from the earlier tests. These results were later replicated by Forintek Canada Corp. (now FP Innovations) using 38x 140 mm No.2 and Better hemlock lumber in bending and two qualities and two sizes (38 x 89 mm and 38 x 184 mm) of spruce lumber, also in bending. Long term tests of spruce lumber in tension and laterally restrained compression have replicated the findings in bending. All these data were used in the development of revised load duration provisions (Barrett, 1996; Karacabeyli and Soltis, 1991). Other experimental results have been obtained from long term shear testing using small torque tubes and from bending tests of plywood and waferboard. All of the experiments were conducted using an applied constant stress, at different fractions of the average strength as predicted by matched short term tests of one minute average duration. The measured variable was the time to failure. Because service loads are not generally constant during the life of the structure, studies were conducted to develop the means to predict times to failure for load sequences of varying intensity. Following the procedures used in metal fatigue to predict number of cycles to failure, the process leading to failure in wood was modeled as an accumulation of damage. The damage variable was defined as being zero at the beginning of load application and reaching one at failure. A process model was postulated to represent the speed with which damage would accumulate and this rate was assumed dependent both on the intensity of the applied stress and the magnitude of the damage already accumulated. As a consequence, damage under a constant load grows in a nonlinear fashion; slow at the beginning and accelerating as failure nears. Parameters in the damage accumulation model were assumed to be random variables to represent the variability in long term performance across specimens. Research at the University of British Columbia (Foschi et al, 1989) calibrated the model parameters using the particular experimental load history and showed that the proposed damage accumulation model fit the experimental trends.

Figure 3 Effect of load duration on β - φ relationship

The advantage of an accumulation model is that it permits the prediction of damage produced by an arbitrary random load sequence. The model was thus used in computer simulations to estimate the damage that would be produced in a beam of given dimensions by a combination of long term and snow loads at the end of a service life of 30 years. This damage α30 can be used to estimate the reliability of the design at the end of the service life. Defining the performance function G as: G = 1.0 – α30

(Eq. 10)

the possibility of non-performance, or probability of G < 0, corresponds to the probability that the damage α30 exceeds 1.0 and the beam fails at or before the end of the service life. The dimensions of the beam can then be adjusted to reduce such probability to acceptable levels. In the Standard, the duration of load effect is taken into by modifying the short term characteristic or specified strength. The following example shows the manner in which the adjustment factor KD was obtained for the case of lumber in bending: 1.

A curve relating reliability index β and the resistance factor φ was developed as previously discussed, using the results of short term strength tests. This curve is shown as (I) in Figure 3;

2.

A curve relating reliability index β and the resistance factor φ was developed by computer simulations, using the performance function G from Eq. 10, and different sequences of loads over a period of 30 years. This curve (Curve (II) in Figure 3) reflects the effect of load duration. The resistance factor φ must be reduced from that of curve (I) to maintain consistent reliability under long term loading.

3.

The adjustment factor KD to be applied to the short term resistance factor is determined from the ratio of the φ values from curves (I) and (II) at the same reliability level.

The adjustment factor is always less than 1.0, since it is applied to the short term strength.

The adjustment factor was obtained for snow loads recorded in different locations across Canada, using statistics on snow magnitude and duration provided by the National Research Council of Canada. Also, the adjustment factor was obtained for two types of occupancy loads, corresponding to office and residential uses. Finally, the factor was obtained for constant loads lasting 30 years, 7 days, 1 day and 1 hour. All research results are reported in Foschi et al, 1989. The effect of load duration is most severe for long term loads. For such loads, the adjustment factor is 0.50 (the characteristic short term strength must be reduced by 50%). As the duration of the constant load is shortened, the adjustment is less severe. Combinations of long term and transient loads like snow or occupancy are less severe than long term loads only. The results of KD studies for these cases showed a dependency on the magnitude of the ratio between design long term load and design transient load. Regardless of the type of the transient load (snow or occupancy) a factor of 0.80 can be used with sufficient accuracy where the design long term load is less than or equal to the design transient load. The factor KD, which is to be used when considering the superposition of long term and several distinct transient loads (snow + wind or snow + earthquake), is controlled by the load of the shortest duration. The KD factors described above form the basis for those actually shown in CSA O86. This describes three classes of load duration: Short Term, Standard Term and Long Term. Standard Term load duration includes all snow and occupancy load cases, to which a KD = 0.80 should apply. For simplicity, this factor has already been included in the specified strengths, resulting in a specified KD = 1.00 for Standard Term loadings as shown in Table 1.

Table 1 Duration of load factors for wood

Calculated KD

KD in CSA O86

Short Term

0.9 to 1.0

1.15

Standard Term

0.8

1.0

Long Term

0.5

0.65

Accordingly, the factor in CSA O86 for Long Term loads is 0.65, and for Short Term (wind, earthquake, formwork, impact) is 1.15. These factors for Short and Standard Term apply as long as the Long Term load is not greater than the design Standard Term load. When the design Long Term load is more than five times the design Standard Term load, the factor KD for Long Term load must be applied. CSA O86 provides an equation for a linear interpolation of the factor KD in intermediate cases in Clause 5.3.2.3. The equation was modified slightly in 2005 to recognize that the Standard Term loads could be acting alone or in combination with other Standard Term loads, and to clarify that the load importance factor included in the specified loads was to be set at 1.0. Clause 5.3.2 was further modified in 2009 to clarify two other aspects of load duration: •

The term “Permanent” load duration (as well as “Dead” load duration) was changed to “Long Term” duration to recognize that reduction in strength due to long term load effects can also be caused by variable loads including lateral earth pressure, controlled fluid and sustained live loading; and

•

Clause 5.3.2.4 was changed to identify that the KD factor for the duration of the shortest load in a combination is permitted to be used to determine the resistance to that combination, except where specified otherwise in the Standard (e.g., wood foundation studs in accordance with A.6.5.12).

The adjustment factor KD, which was revised in the 1989 edition, can be considered to be a reliable estimate of the load duration effect for lumber, given that the bulk of the experimental data was collected for that particular product. Without the benefit of a similar database, KD for other applications (glulam beams, connectors, products) must be considered to be approximations. However, the practice is to use the same adjustment factors for these other products, which was the case prior to the 1989 update.

Clause 5.3.5 – System factor, KH Given the variability in properties of wood when taken in isolation, it is important to consider their performance in a system when studying reliability so as to not unduly penalize design. Earlier editions of the Standard used the name “Load Sharing Factor” for this, and assigned an increase factor of 1.15 to bending . However, load sharing is only part of the picture in deg wood systems. Since 1989, the Standard included system modification factors KH that may be used for typical wood frame systems with repetitive all sharing the applied load. An example of such a structure is a floor ed by multiple joists, covered by plywood or OSB s fastened to the joists. These systems are typically designed using a single member equation, considering only one isolated joist. The system modification factor compensates for this simplification and maintains the reliability level that was calibrated for a single member. Justification for the factor KH stems from three considerations: 1.

The reliability analysis of each member requires the calculation of the stresses that the member is actually carrying as part of the structure. These stresses can be estimated very conservatively by considering the member in isolation from the system and taking into only the tributary area for the applied load. A better estimate of the stresses can be obtained by analyzing the system as a redundant structure, taking into the effect that and system components have on each other. In a load-sharing system the effect of other will be related to the relative stiffness of the components.

2.

The composite action of the sheathing and fasteners enhances the performance of each member in the system. This is reflected in the system modification factors.

3.

The factor also s for the reduced probability that weaker be placed in the location of higher stresses in a repetitive system.

Research was carried out for repetitive member systems of the type found in floors or flat roofs using lumber joists or rafters, with nailing or adhesives connecting the lumber to sheathing, under uniformly distributed loads (Vanderbilt et al, 1974). The factor KH depends strongly on several variables: sheathing thickness, ratio of sheathing bending stiffnesses in directions parallel and perpendicular to the joists, fastener stiffness, variability in modulus of elasticity between , variability in bending strength between , magnitude and variability of the applied loads, and the statistical correlation between member stiffness and bending strength. The factor KH is much less dependent on variables that are considered in the design equation, such as joist depth and joist spacing. The factor KH also depends on the definition of performance (in this case, bending failure for the joists or rafters). One definition may consider failure of the most stressed joist, with stresses calculated from a structural analysis of the system. A more stringent definition defines system performance as no failure in any joist.

The first definition follows the philosophy of single member design with due of system-induced stresses. The second is a first level of a more sophisticated system design. One of the major differences between results from the two approaches is that the second is influenced by the number of joists since the greater the number of joists in a floor, the greater the chances of having included a weaker one.

Research conducted at the University of British Columbia (Foschi et al,1989) describes in detail the calculation of KH for this type of system, and Figure 4 shows the procedure. A floor configuration was chosen by selecting the number of joists, their dimensions, stiffness and spacing, and the type and stiffness of fasteners. The reliability index β corresponding to selected floor spans was computed for the complete structure using a computer analysis of the stresses in the floor, as shown by Curve (I) in Figure 4; and for a single joist carrying load on its tributary area, as shown by Curve (II). For a given target reliability, the ratio between the corresponding spans from Curves (I) and (II) was used to obtain the system factor KH. For uniformly distributed loads, this factor is the ratio of the squares of the corresponding spans. Using the first definition of performance (single member approach), the range for KH when considering: three species of lumber (Douglas fir, Hem-fir and S-P-F), two sizes (38 x 184 mm and 38 x 235 mm), three grades (Select Structural, No.2 and No.3), four plywood sheathing thicknesses (9.5 to 18.5 mm) and three loading conditions (two snows and one occupancy) was: 1.41 ≤ KH ≤ 1.77 with an overall average KH = 1.63 and a coefficient of variation (COV) of 4.5%. The highest values corresponded to the thicker sheathing. The results were obtained for a joist spacing of 400 mm, with nominal nailing (stiffness of 2.10 MN/m) with a spacing of 200 mm. Lumber data was obtained from the Canadian Wood Council lumber properties testing program. The calculations were performed for a distributed load with a mean of 2.0 kN/m2 (40 psf), and a (COV) of 30%. The values of KH are smaller if the mean of the load is increased: for 7.0 kN/m2 (140 psf), values are approximately 15% smaller. The average overall for the more stringent performance requirement of no joist failures, was KH = 1.43. CSA O86 specifies a value of KH = 1.40 to be applied to all configurations for this type of floor/roof construction. A value of KH = 1.10 must be used when the cover is made of essentially one-directional elements perpendicular to the joists or rafters (for example, timber decking).

Figure 4 Effect of system behaviour on β span relationship

No increase in the resistance factor (KH = 1.0) is allowed if the cover is made of essentially one-directional elements parallel to the joists or rafters. Thus, for a roof with main beams, purlins, and timber decking perpendicular to the purlins, a factor KH = 1.10 may be applicable for the purlin design but KH = 1.0 applies for the main beams. Other structural wood systems may exhibit load-sharing behaviour, but are used in a wider range of loading and component configurations. Thus, modification factors for a system of parallel chord trusses in a floor or for a system of pitched trusses in a roof that capture all the load-sharing effects are more difficult to determine. For these cases, the Standard includes a nominal 10% increase factor. For structural composite lumber products used in a load-sharing system, CSA O86 specifies a value of KH = 1.04. This increase is lower than that typically assigned for sawn lumber system to reflect the significantly lower variability in properties between parallel structural composite lumber component in a system. To qualify for this increase, structural composite lumber shall be part of the wood-framing system consisting of at least three parallel ed by transverse load-distributing elements capable of ing the design load and spaced not more than 610 mm on centre. The factor KH has been computed using short term strengths, since the duration of load adjustment (KD) can apply equally to the single member as well as the repetitive member system.

Clause 5.3.6 – Size factor, KZ Strength properties of wood structural vary depending on the member size and loading conditions. This phenomenon is referred to as a ”size effect”. In general, a larger volume of a member subjected to the design stress will tend to reduce the apparent strength. While a large member will have a higher capacity, size effects will reduce the rate at which the capacity increases with increasing member size. Further explanations for size effects can be found in the Commentary information related to specific materials: 6.4.5 for lumber, 7.5.6 for glulam and 15.3 for structural composite lumber.

Clause 5.5 – Serviceability requirements Clause 5.5 contains minimum serviceability requirements for general design. Acceptability of deflection criteria can depend on secondary effects such as long term loading or interaction with other parts of the building. Designers are advised to review additional guidance on deflection criteria such as provided for specific cases in Table 2.1 of the Wood Design Manual. Appendix A.5.4.2 of the Standard explains the relationship between single-member deflection criteria and the expected deflection performance of a wall, floor or roof system. Appendix A.5.4.5 was added to the 2001 edition of CSA O86 to provide basic information relating to vibration performance of wood-framed floor systems. Floor vibration is a serviceability limit state that reflects people’s satisfaction with the performance of a floor. People may consider floor performance to be unsatisfactory when: •

cabinet contents and materials on shelves or tables vibrate when someone walks across the floor,

•

a person walking across the floor is felt by another person sitting down, and

•

the floor appears to be bouncy when walking on it.

Point load deflection and natural frequency are two criteria that have been used to assess the vibration performance of wood frame floors. Depending on the span, construction and weight of the floor, one or a combination of the two criteria may be a better indicator of vibration performance and occupant satisfaction. On short-span, lightweight floors, each footfall by a walking person results in a relatively large change in deflection, with a small transient vibration. The transient vibration may or may not be sensed, depending on the amplitudes and damping. In conventional short span floors, point load deflection has been used to represent vibration performance. On long-span, heavier mass floors, which are stiffer systems, the deflection taking place under the action of an individual person is small in comparison to the transient vibration generated by a walking person. In some cases, natural frequency of the floor may assist in understanding vibration performance. It should be recognized that information on floor performance is difficult to analyze and interpret for reasons of measurement issues, subjectivity of opinions about acceptability, construction variables, information gaps and other factors.

Clause 5.4.6 – Building movements due to moisture content change Clause 5.4.6 was introduced in the 2014 edition of CSA O86 to accommodate the change made to NBC 2015 that allows wood construction of up to 6 storeys. Wood shrinks or swells due to moisture loss or gain when the moisture content is below the fibre saturation point. The considerations for differential movements caused by moisture content change becomes more critical for higher buildings due to the cumulative effect of movement. Anticipated building movements should be specified and proper design detailing should be provided to accommodate potential differential movements. Appendix A.5.4.6 of the Standard provides guidance on avoiding potential structural or serviceability problems related to wood shrinkage. See the Commentary on Clause A.5.4.6 for more information.

Clause 5.5 – Lateral brace force for wood truss compression webs New provision were introduced in the 2014 edition of the Standard to address lateral bracing for wood truss web loaded in compression. Designers have commonly assumed that the bracing force required to brace compression in metal plate connected wood trusses is 2 percent of the compression force applied to the web. This is a strength based model that ignores bracing stiffness and assumes that web are pin connected with the chord . However, the out-of-plane rotational stiffness of the metal plate connection plays an important role in estimating the bracing forces of truss . Analytical and experimental studies were conducted at the University of British Columbia to investigate the influence of metal plate connections on bracing requirements. The results of these studies showed that it is appropriate to use 1.25% of the axial compressive force of the member to resist out-of-plain deflection in axial compressive web where metal plate connections are used.

References: ATC Design Guide 1. Minimizing Floor Vibration. 1999. Applied Technology Council. Redwood City, CA. Prepared for ATC by D.E. Allen, D.M. Onysko and T.M. Murray. Barrett, J.D. 1996. Duration of Load—The Past, Present and Future. International COST 508 Wood Mechanics Conference, Stuttgart, . Canada Mortgage and Housing Corporation. 1999. Wood-Frame Envelopes in the Coastal Climate of British Columbia CWC and FPInnovations 2013. Vertical Movement in Wood Platform Frame Structures: Basics. Canadian Wood Council, Ottawa Canada and FPInnovations Vancouver, Canada CWC and FPInnovations 2013. Vertical Movement in Wood Platform Frame Structures: Movement Prediction. Canadian Wood Council, Ottawa Canada and FPInnovations Vancouver, Canada CWC and FPInnovations 2013. Vertical Movement in Wood Platform Frame Structures: Design and Detailing Solutions. Canadian Wood Council, Ottawa Canada and FPInnovations Vancouver, Canada CWC. 2001. Standard Practice Relating Specified Strengths of Structural to Characteristic Structural Properties, Canadian Wood Council, Ottawa, Canada. Dolan, J.D., Murray, T.M., Johnson, J.R., Runte, D. and Shue, B.C. 1999. Preventing Annoying Wood Floor Vibrations. Journal of Structural Engineering. American Society of Civil Engineers, Vol. 125, No. 1, pp. 19-24. Foschi, R.O., Folz, B., Yao, F. 1989. Reliability-Based Design of Wood Structures. Structural Research Series, Report No. 34, University of British Columbia, Vancouver, BC. Foschi, R.O. (1974). Load-slip characteristics of nails. Wood Science, 7(1), pp. 69-76 Foschi, R.O., Folz, B, Yao, F., and Zhang, J. (2007). RELAN – Reliability analysis software, V8.0 University of British Columbia, Vancouver Canada. Hu, L.J. and Tardif, Y. 2000. Effectiveness of Strong-back Wood-I-blocking for Improving Vibration Performance of Engineered Wooden Floors, . In Proceedings of the World Conference on Timber Engineering. Whistler, BC. Karacabeyli and Soltis. 1991. State-of-the-Art Report on Duration of Load Research for Lumber in North America. Proceedings of the 1991 International Timber Engineering Conference, London, U.K., Vol.4, pp 141-155. ISO Standard 2631-2, 2003. Evaluation of Human Exposure to Whole-Body Vibration – Part 2: Human Exposure to Continuous and Shock Induced Vibrations in Buildings (1 to 80 Hz). International Standards Organization. Madsen, B. 1971. Duration of Load Tests for Dry Lumber in Bending. Dept. Civ. Eng., University of British Columbia, Structural Research Series No.3, Vancouver, BC. Onysko, D.M., Hu, L.J., Jones, E.D. and Di Lenardo, B. 2000. Serviceability Design of Residential Wood Framed Floors in Canada. In Proceedings of the World Conference on Timber Engineering. Whistler, BC. Onysko D.M. 1988. Performance Criteria for Residential Floors Based on Consumer Responses. International Conference on Timber Engineering, Seattle, WA [Forest Products Society, V1, 1988 pp. 736-745]. Song X., F. Lam., H. Huang, and M. He (2010) Capacity of Metal Plate Connected Wood Truss Assemblies. Journal of Structural Engineering. ASCE. ASCE. 136(6): 723-730. Song X., F. Lam. (2009). Laterally braced wood beam-columns subjected to biaxial eccentric loading. Computers and Structures. 87:1058-1066

Song X., F. Lam. (2010). Stability capacity and lateral bracing requirements of wood beam-columns. Journal of Structural Engineering. ASCE. 136(2):211-218 Smith, I. and Chui, Y.H. 1988. Design of Lightweight Wooden Floors to Avoid Human Discomfort. Canadian Journal of Civil Engineering, pp. 254-262. Vanderbilt, M.D., Goodman, J.R., Bodig, J. 1974. A Rational Analysis and Design Procedure for Wood Joist Floor Systems. Final Report to the National Science Foundation for Grant GK-30853. Colorado State University, Dept of Civ. Eng., Fort Collins,CO. Xiaoqin L. and F Lam. University of British Columbia Report prepared for Truss Plate Institute of Canada (TPIC) under the guidance of the TPIC Technical Committee 25-February 2011 Wood, L. 1960. Relation of Strength of Wood to Duration of Load. U.S. Dept. of Agric., Forest Products Lab, Report No. 1916, Madison, WI.

Clause 6

Sawn Lumber

Clause 6.1 – Scope Design procedures and design values provided in the Sawn Lumber section apply to structural lumber meeting the requirements of the National Lumber Grades Authority grading rules (NLGA, 2010) and the requirements of CSA Standard O141, Softwood Lumber.

Clause 6.2 – Materials CSA O141 establishes basic provisions for lumber grading to be carried out under the supervision of grading agencies that are accredited by an independent body to ensure consistent accuracy of grading. In Canada, the accreditation is carried out by the Canadian Lumber Standards Accreditation Board. CSA O141 includes the principal grade classifications and sizes of softwood lumber for structural use. It also provides a common basis of classification, measurement, grading and grade-marking of rough and dressed sizes of various items of lumber, including boards, dimension and timbers. Hardwood species may also be classified, graded and grade stamped under the same provisions. CSA O141 identifies nominal and minimum sizes for structural lumber in the dry and green conditions. Dry lumber is defined as lumber seasoned to a moisture content of 19% or less at the time of surfacing.

Clause 6.2.1.3 – US lumber The equivalent species combination for Southern Pine was removed in the 2014 edition of the Standard since the design values in the US were being reassessed, and not available at the time this Standard was being published.

Clause 6.2.3 – Fingered lumber The NLGA Special Product Standards (SPS) for fingered lumber are intended to be used in conjunction with the NLGA Standard Grading Rules. Except as limited in Clause 6.2.3.2 or 6.2.3.3, design data specified in the Standard for dimension lumber are applicable to fingered lumber that has been grade-stamped in accordance with: NLGA SPS 1, Special Products Standard for Fingered Structural Lumber NLGA SPS 3, Special Products Standard for Fingered “Vertical Stud Use Only” Lumber, or NLGA SPS 4, Special Products Standard for Fingered Machine Graded Lumber. In the case of the SPS 1 and SPS 3 Standards, the lumber segments between the fingerts are visually graded in accordance with the NLGA grading rules for the grade indicated on the grade stamp. Near the fingerts, more restrictive visual limits are generally imposed. In the case of the SPS 4 standard, the lumber segments are required to meet the requirements of NLGA SPS 2, Special Products Standard for Machine Graded Lumber. Consequently, the lumber used in the manufacturing of finger-ed lumber has similar properties to that found in non-finger-ed lumber of the same size, species and grade. The SPS, fingert test requirements are selected to enable the same specified strength and stiffness as non-finger-ed lumber of the same grade and size to be assigned to the finger-ed lumber. Test methods (e.g. bending or tension tests) and target test load (e.g. minimum and 5th percentile finger t strengths) for samples of single fingerts are not only linked to the size, grade and species to be ed, but also take into the average fingert spacing. Fingerts used at lower average fingert

spacing need to achieve a higher 5th percentile strength level than the same fingerts used at higher average fingert spacing. In selecting the tests and developing the SPS, only some properties, such as bending strength, are directly tested. Others are established by correlation to the property monitored, or implied by the specification imposed on the adhesive (e.g. adhesive bondline performance). Because it is not practical to assess all possible performance attributes of non-fingered lumber under the standard specification, some end-uses that rely on implied properties may need further assessment (examples below). s familiar with the end use conditions should assess the suitability of the finger-ed lumber product for their application. The review can consist of a combination of analytical and empirical studies to determine if additional requirements or restrictions are necessary. Considerations include, for example: •

Interaction between chemical treatments to enhance wood properties and the adhesive used to bond the fingerts.

•

Connector-fingert interaction (particularly in the case of proprietary connectors) affecting either the fingert capacity and/or the connector capacity.

•

Assumed performance of finger-ed lumber under the service conditions prior to the product being put into service (e.g. transport, site storage, exposure during construction, etc.).

•

Stress condition unique to applications such as metal plated connected trusses, laminated glued lumber, and wood I-joist that can only be assessed empirically.

•

Lumber and component handling and storage procedures that are deemed to be acceptable based on past experience with non-fingered lumber.

In prefabricated assemblies or components where the fingered lumber used is designated as “DRY USE ONLY”, the assemblies or components should be of a type that is normally assumed to be for dry service conditions as defined in CSA O86. Examples of this include metal plate connected trusses or prefabricated wood I-joists. Because marks limiting the fingered lumber to dry service conditions may become illegible, use in assemblies or components where the end use service condition may change is not recommended. The design data may also apply to fingered visually graded lumber meeting the requirements of NLGA SPS 3 “Special Products Standard for Fingered ‘Vertical Stud Use Only’ Lumber,” subject to two conditions: 1.

the primary loading is in compression with only short-term bending or tension stresses, and

2.

the end use is protected from wet service conditions at all times, with equilibrium moisture content no greater than 19%, and where temperatures are not expected to exceed 50°C for an extended period of time.

SPS 3 adhesives are not required to meet exterior adhesive specifications, so SPS 3 studs are not permitted to be used in elevated moisture or temperature conditions. However, the adhesives used and samples of the product are subject to moisture exposure tests, which are intended to demonstrate resistance to the kind of moisture that might be encountered during construction or from accidental wetting. In typical wall assemblies under normal conditions, however, temperature and moisture content are not expected to reach these limiting conditions. The SPS 3 adhesive specifications also exclude the product from use where duration of bending or tension stresses could exceed short-term wind or seismic load conditions. NLGA SPS 4 finger-ed lumber may be bonded with adhesives meeting either CSA O112.9 or O112.10 or O112.7. NLGA SPS 4 finger-ed lumber bonded with adhesive meeting the specification of O112.10 are required to be marked “DRY USE ONLY”. Otherwise the adhesive used must meet the requirements of CSA O112.9 or O112.7.

Adhesives used in SPS 1, SPS 3 or SPS 4 fingered products can be further evaluated for extreme heat resistance for use in fire-resistance-rated assemblies. Fingered lumber marked “HRA” indicates that it has been manufactured with a “heat resistant adhesive” as demonstrated by a full-scale wall tested in accordance with ASTM D7374 - Practice for Evaluating Elevated Temperature Performance of Adhesives Used in End-ted Lumber. Although CSA O86 does not address fire protection requirements explicitly, it is assumed that such requirements are met.

Clause 6.2.4 – Remanufactured lumber Remanufacturing operations such as sawing lumber lengthwise (“ripping”) or reducing the member crosssectional area may have the effect of reducing the grade and thus lowering the structural properties of stress graded lumber. Lumber that is ripped or otherwise modified in a manner that may lower the grade must be regraded and grade stamped in accordance with the requirements of the National Lumber Grades Authority and CSA Standard O141 Softwood Lumber.

Clause 6.3 – Specified strengths The basis for lumber specified strengths has evolved from tests of small clear pieces of wood to full-sized lumber (“in-grade”) testing, where the latter method has been adopted internationally as the basis for derived engineering properties of structural lumber. In-grade test data has been used since the 1980 edition of CSA O86 as the basis for tension design values, since the 1984 edition for bending design values, and since the 1989 edition for compression parallel to grain design values. Tests of small clear wood are still used to determine other properties such as relative density, horizontal shear, and compression perpendicular to grain, since these properties are less dependent on wood characteristics such as knots and slope of grain. The in-grade lumber database was used to establish relationships between characteristic properties and specified strengths through reliability procedures. These relationships are documented in the CWC publication, “Standard Practice Relating Specified Strengths of Structural to Characteristic Structural Properties.”

Traditional Methodology Design values for lumber were developed traditionally from tests of small clear specimens free from strength-reducing defects such as knots, slope of grain, splits and holes. Small clear test data for Canadian softwood and hardwood species are summarized in Jessome (1977), and also in the American Society for Testing and Materials (ASTM) Standard D 2555, Standard Methods for Establishing Clear Wood Strength Values (ASTM, 2000). ASTM Standard D 245, Standard Methods for Establishing Structural Grades and Related Allowable Properties for Visually Graded Lumber, has provided the basis for deriving allowable lumber design properties from small clear wood data. These methods are still used for some types of visually graded structural lumber such as beams and stringers, and posts and timbers. Also, as noted above, some design values for minor properties of sawn dimension lumber are still based on small clear test data. Where the small clear wood method is used, design values in Canada have typically been based on the weakest species in a commercial species combination. Fifth percentile strength properties are calculated using the assumption that clear wood properties are normally distributed. These fifth percentile strength properties are modified to for differences between the small clear test specimen data and the strength and stiffness of full-size subjected to long term loads and in-service climate conditions. These resulting values are further reduced for effects of characteristics such as slope-of-grain, knots, splits and cracks to determine design values for individual grades and sizes of structural lumber.

CWC Full-Size Lumber Properties Program With the introduction of the limit states design philosophy, the CSA Committee for Engineering Design in Wood adopted the philosophy that design values for structural wood products should be based on fullsized structural tests of (“in-grade” testing) as used in building construction, wherever feasible. Bending and tension design values in the 1984 edition were derived from in-grade test data developed from studies conducted by Madsen and co-workers during the late 1970s (Madsen, 1992) on behalf of NLGA. Tension values based on this work were first implemented in the 1980 edition of the Standard. This first major in-grade testing program established bending and tension parallel to grain strength and mean modulus of elasticity data for the three major commercial species groups (D Fir-L, Hem-Fir and SP-F) using a proof loading approach; i.e., breaking only the weaker pieces. Following this, the Canadian lumber industry conducted a major research program through the Canadian Wood Council Lumber Properties Project for bending, tension and compression parallel to grain strength properties of 38 mm thick (nominal 2") dimension lumber of all commercially important Canadian species groups. The CWC Project was part of a cooperative project with the US industry. The basis for sampling and testing was coordinated with the North American In-grade Testing Committee recommendations that later evolved into ASTM consensus standards. Lumber samples in three sizes (38 by 89 mm, 38 by 184 mm and 38 by 235 mm) were selected across the Canadian growing regions for the three largest-volume commercial species groups: Spruce-Pine-Fir (S-P-F), Douglas Fir-Larch (D Fir-L) and Hem-Fir. Select Structural, No.1, No.2, No.3, as well as light framing grades, were sampled in bending. Select Structural, No.1 and No.2 grades were evaluated in tension and compression parallel to grain. Several lesser-volume species were evaluated at lower sampling intensities for two grades (Select Structural and No.2) and three sizes (38 by 89 mm, 38 by 140 mm and 38 by184 mm). After the test samples were conditioned to approximately 15% moisture content, they were sent to the Forintek Canada Corp. (now FP Innovations) in Vancouver, BC. Short term flexural stiffness properties, and ultimate bending, tension, and compression parallel to grain strengths were tested in accordance with ASTM Standard D 4761, Standard Test Methods for Mechanical Properties of Lumber and Wood-Base Structural Material. The dimension lumber database was examined to establish trends in bending, tension and compression parallel to grain property relationships as affected by member size and grade. These studies provided a basis for extending the results to the full range of dimension lumber grades and member sizes required in CSA O86. A detailed description of the research programs and a discussion of the results were presented in an ASTM Workshop on the North American In-grade Testing Program (FPRS, 1989). Further discussion of the Canadian testing program and its results appears in a book titled Canadian Lumber Properties (Barrett and Lau, 1994). Two other properties—compression perpendicular and shear—were not tested in this program. Compression perpendicular to grain design values are derived directly from the ASTM D 245 procedures. However, in the 1994 Standard, the relationship between compression perpendicular to grain and wood density was recognized in setting the design values (See commentary on Clause 6.5.7). Horizontal shear capacity of dimension lumber bending was traditionally based on ASTM D 245 and the assumption that every member contained a full-length split. During the in-grade program, actual crack length distributions were measured for all lumber evaluated in the in-grade testing program. The crack length distribution data was used in conjunction with a fracture mechanics failure model to study the reliability of dimension lumber subjected to bending loads (Lau et al, 1988). Although these reliability analyses were used to develop shear design values in 1989, the approach was further modified in the 2005 Supplement to the 2001 edition of the CSA O86 Standard (see Commentary on Clause 6.5.5).

Clause 6.3.1 – Visually stress-graded lumber The National Lumber Grades Authority (NLGA, 2010) Standard Grading Rules for Canadian Lumber classify visually graded structural lumber into specific categories defined by end use applications and member size. CSA O86 provides design values for three classes of visually graded lumber: Dimension Lumber, Beams and Stringers, and Posts and Timbers.

Tables 6.3.1A and B — Dimension lumber The NLGA grading rules further classify dimension lumber into four categories: Structural Light Framing, Structural Joists and Planks, Light Framing and Studs. Dimension lumber is surfaced to thicknesses from 38 to 102 mm and widths of 38 mm or more. Lumber is typically specified in metric actual dimensions (or imperial nominal dimensions). Table A 6.5.2 in CSA O86 lists green and dry sizes for dimension lumber and timbers. Relationships between categories, grades and member sizes are shown in Table 6.2.2.1 of the Standard. Within each category, permissible grade characteristics such as knots, slope of grain, splits, cracks and wane are defined by grading rule requirements (NLGA, 2008). Since grade characteristics vary with category as well as the grade, design values are also category- and grade-dependent. Specified strengths for typical grades of dimension lumber, given in Table 6.3.1A. of CSA O86, were based mainly on results from the CWC Lumber Properties Project. These strength values have been formatted in accordance with limit states design philosophy. The specified strength values for bending, tension, compression parallel to the grain, shear strength and modulus of elasticity were derived using reliability principles (Lau, 1988; Foschi, 1989). Specified strengths, except compression perpendicular to the grain, were modified to Standard Term duration of load. The specified strengths for the Standard Term duration, which applies for snow loads and occupancy loads, are 80% of the short term static test strengths obtained from full-size tests. Specified strengths for the species group identified as Northern Species were derived from the weakest species tested to provide a basis for design using any Canadian commercial species. The CWC in-grade testing program confirmed that design values for No.1 and No.2 grades of dimension lumber were not significantly different. These two grades were consequently assigned the same specified strengths. The testing programs also demonstrated that No.3 and Stud grades could be assigned the same specified strengths. Tabulated design values for dimension lumber (except light framing and machine stress-rated lumber), beams and stringers, posts and timbers were reduced to be applicable to a member width of about 300 mm. Specified strengths for other member widths are derived using the appropriate size factor (Table 6.4.5 in CSA O86), with the exception of the Light Framing grades. The Light Framing grades (i.e., Construction, Standard) apply only to lumber 89 mm or narrower, so the design values in Table 6.3.1B were tabulated on an 89 mm basis and size factors are not used except for compression properties. Sizes of Light Framing grades other than 38 by 89 mm were assigned different design values in previous editions of CSA O86, based on analysis of the small clear wood approach. In the 2001 edition of CSA O86, however, values for the other Light Framing grade sizes were changed to be equivalent to the 38 by 89 mm design values, on the basis of the relationships between in-grade data of various sizes and grades of lumber.

Table 6.3.1C —

Beams and Stringers

Timbers with a width more than 51 mm (2") greater than the thickness are classified as Beams and Stringers. Actual sizes of Beams and Stringers are typically 13 mm (1/2") less than the nominal imperial sizes. For example, a nominal 8 x 12" member would be graded in the Beams and Stringers category and have actual dimensions 191 x 292 mm (7-1/2" by 11-1/2"). Beams and Stringers design values are derived from the small clear specimen approach. Limited tests of full- size Beams and Stringers in bending demonstrated that the design values were conservative. Specified strengths for Beams and Stringers (Table 6.3.1C in CSA O86) were revised in 1989 to recognize the available test data for timbers (Madsen, 1982). The revised design values incorporate higher compression parallel to grain strengths and a size factor based on evaluation of the bending, compression and tension property relationships, as well as size effects derived from the dimension lumber program. Specified strengths were added for a No.2 grade in the Beam and Stringer category in 1989. Beams and Stringers are graded primarily for use in bending, assuming that the member will be loaded on the narrow face. Larger knots are permitted at along the centreline and at the ends of the wide face. When these are used on the flat, the specified strengths and modulus of elasticity must be reduced using the factors tabulated in the footnote to Table 6.3.1C. The tabulated values apply to dry service conditions and Standard Term load duration. Since these sections are typically supplied in the green condition and seasoning occurs rather slowly, caution must be exercised to prevent overloading in compression while the are in the green state. If overloading is anticipated prior to seasoning, the specified strengths should be modified to accommodate wet use conditions.

Table 6.3.1D —

Posts and Timbers

Posts and Timbers are graded primarily for use in compression but they may also be used in applications where bending and tension design stresses occur. These are essentially square cross sections where the maximum difference between the width and thickness is 51 mm (2"). Post and Timbers are graded so that the maximum permissible knot size may occur at any location across a face or along the length of the member. In rectangular sections the wider face determines the maximum knot size. As with Beams and Stringers, care must be taken not to overload these in compression prior to seasoning if specified strengths for the dry service condition are applied. Specified strengths for Posts and Timbers were also updated in 1989, considering the available full-size bending test data as well as the size effects and property relationships developed under the CWC Lumber Properties Project.

Clause 6.3.2 – Machine stress-rated and Machine evaluated lumber Property Classification Machine stress-rated (MSR) lumber is sawn lumber that has been evaluated by mechanical stress-rating equipment and meets the product requirements specified in NLGA Special Products Standard 2 (SPS 2), Machine Stress-Rated Lumber (NLGA, 2013). SPS 2 specifies requirements for visual characteristics, property requirements, quality control procedures and the grade marking requirements for MSR lumber. MSR lumber is distinguished from visually graded lumber in that each piece is non-destructively evaluated usually by a machine for modulus of elasticity (E) and each piece is grade marked to indicate the Fb-E rating.

Other properties may also be marked on the piece if part of the quality control process. Machine evaluated lumber (MEL) is also non-destructively evaluated by a grading machine, and conforms to requirements for MSR lumber with the exception that the process lower 5th percentile E value must equal or exceed 0.75 (rather than 0.82 for MSR lumber) times the characteristic mean E for the grade. The Fb-E grade designations for MSR lumber (e.g., 1650Fb-1.5E) were established to correspond to allowable design stresses for MSR lumber in imperial units (psi). The Fb-E grade descriptors were the allowable design values assigned to bending (psi) and modulus of elasticity (x 106 psi) in the working stress design format. In limit states design, the bending descriptor has an indirect relationship to the design values (specified strengths) for the grade. SPS 2 requires that qualification and quality control samples of MSR and MEL lumber meet minimum structural property requirements. For a standard grade, the property requirements for bending, tension and compression parallel to grain strength and modulus of elasticity are the same for all species. The property requirements for mean E, fifth percentile E and fifth percentile bending strength specified in SPS 2 must be qualified and checked using quality control programs based on full-size structural evaluations of samples selected from each days production. In some cases, tensile strength must also be part of the quality control.

MSR and MEL lumber properties In North America, machine rated lumber has been available only in 38 mm (nominal 2") thicknesses. The largest volumes of production occur in 89 and 140 mm widths. The Standard provides specified strength properties for 22 MSR Fb-E classifications and of these the 1450Fb-1.3E, 1650Fb-1.5E, 1800Fb-1.6E, 2100Fb-1.8E and 2400Fb-2.0E are produced in substantial quantities by Canadian mills. There are also 14 grades of MEL lumber in the Standard. Full-size tests have been conducted in bending and tension parallel to grain for 38 by 89 mm, SprucePine-Fir MSR lumber in the five commonly produced grades, as part of the CWC Lumber Properties Project. A similar test program was carried out by NLGA in 2002. These evaluations confirmed the property specifications for MSR lumber given in SPS 2 for Spruce-Pine-Fir in tension and bending, as determined by in-grade testing. SPS 2 specifies limitations on visual characteristics as well as minimum property requirements. Visual characteristics typically are limited to the No.2 visual grade level or better. In the higher grades, additional visual restrictions on edge knots and other edge characteristics are imposed to ensure that design values are attained. When resistance and duration of load factors were adopted in the Standard for sawn lumber based on reliability analysis in 1989, the specified tension strengths for MSR lumber were calibrated to maintain approximately the same sections as in previous editions of the Standard. This implementation of reliability analysis also resulted in increased factored bending resistance for roof and floor applications. The size effect factor for bending strength is Kz = 1.0 for MSR and MEL lumber. This is because the industry practice is to qualify and maintain machine grading parameters for MOE and bending strength that are independent of member width or species. Studies of tension to bending and compression to bending property relationships allow the tension and compression specified strengths to be defined in relation to the bending strength. In the 1994 edition of CSA O86, compression parallel to grain design values were modified in view of the in-grade compression/bending relationships determined for visually graded lumber, and limited data on MSR lumber. This information showed that compression parallel strength was under-rated for lower grades of machine rated lumber, and slightly over-rated for higher grades. In addition, the size factor for compression (KZc) was applied to MSR and MEL design formulae, since the specified compression strengths were tabulated on the same size basis as visually graded lumber strengths. Shear and compression perpendicular to grain specified strengths were derived using procedures adopted for visually graded lumber. Thus, these values are species-dependent and the specified

strengths are tabulated for the appropriate species in Table 6.3.1A of CSA O86. For Spruce-Pine-Fir machine graded lumber (all grades), and for Hem-Fir machine graded lumber of 10,300 MPa or higher E values, the compression perpendicular design values have been increased. Other grades and species may also be upgraded as data are made available. The basis for these increases relates to the higher density values for these grades as compared to visually graded lumber of the same species (See Appendix A 6.5.7 of the Standard for further information).

Clause 6.4 – Modification factors Clause 6.4.1 – Load duration factor, KD Effects of long term constant loads on time-to-failure behaviour of visually graded lumber have been studied extensively in the 1970s and 1980s. Results from these studies have been used to calibrate duration of load models (damage models), which can be used to predict member behaviour under typical load combinations encountered in buildings. Studies of full-size dimension lumber show that duration of load effects are relatively consistent across species, grades and member sizes. Based on these studies the CSA committee has concluded that previous duration of load adjustments for structural wood products, based on small clear specimens studies, required adjustment (See Commentary on Clause 6.3). From the 1989 Standard on, the duration of load factors were based on studies of full-size member behaviour and snow load, occupancy load and long term load models adopted for reliability studies of wood and other structural materials.

Clause 6.4.2 – Service condition factor, KS Studies of strength and stiffness changes of dimension lumber due to moisture content have been used to predict the effect of moisture content on lumber properties (Barrett and Lau, 1994). When calculating a member capacity or stiffness, increases in strength properties generally compensate for shrinkage. In fact, member capacities and stiffness are generally higher for dry service conditions than wet service conditions (Note: This may not be the case for very dry lumber - less than about 10% moisture content). The wet service condition factors take into both the shrinkage effect and the property differences between green and dry service conditions. The service condition factors in Table 6.4.2 of CSA O86 provide a convenient method of incorporating the effect of service conditions in design calculations while allowing calculations to be based on the product standard dry sizes.

Clause 6.4.3 – Treatment factor, KT Full-size tests have shown that strength and stiffness properties of 38 mm dimension lumber are reduced if the lumber is both incised and treated (Palka, et al, 1984; Palka, et al, 1992; and Lam and Morris, 1991). Preservative treatments themselves appear to have little effect on structural properties. However, when the 38 mm were incised and then treated and re-dried, significant structural property decreases occurred. Based on these full-sized tests the CSA Standard specifies that the modulus of elasticity and specified strength values of sawn lumber 89 mm or less in thickness must be modified for preservative treatment as shown in Table 6.4.3. Investigations of the influence of incising and treating on the strength properties of larger sections did not demonstrate any consistent reductions. In addition, some of the earlier work was done using incising methods that are no longer practiced. For dimension lumber, incising is done typically to a density of no more than 10,000 incisions per square metre, and a depth of no more than 10 mm. Evaluation of incisor methods summarized in Lam and Morris (1991), and Morris, Morrell and Ruddick (1994) in view of current industry practices led to a modification of the dry-service strength reduction factor from 0.7 to 0.75 in the 2001 edition of CSA O86.

Fire-retardant treatments also reduce strength properties of structural lumber. In past editions of CSA O86, the effect was identified as a 10% reduction. However, the chemical formulations used today are different from earlier formulations and behave differently, particularly at elevated temperatures. As a result, CSA O86 no longer includes specific treatment factors for fire-retardant wood, but continues to require that the effects of the treatment be considered in design. Information on appropriate adjustments should be obtained from the company that provides the treatment. CSA O86 requires that tests be carried out on the effects of fire-retardants, or any other potentially strength-reducing chemicals used with lumber, to determine an appropriate treatment factor. The tests must take into the service conditions of the treated lumber. In particular, elevated temperatures for prolonged periods are known to reduce the strength of fire-retardant treated wood. ASTM Standard D 5664 "Standard Test Method for Evaluating the Effects of Fire-Retardant Treatments and Elevated Temperatures on Strength Properties of Fire-Retardant Treated Lumber", provides a standard test protocol for evaluating the strength reducing effects of fire retardants. Three test procedures are given in ASTM Standard D 5664: •

•

•

Test Procedure 1 uses small clear specimens to evaluate the effects of fire-retardant treatments on bending, tension parallel, compression parallel, and horizontal shear properties at room temperature. These test results could be used to develop KT factors for treated wood used at room temperature. Test Procedure 2 uses small clear specimens to assess the effects of fire-retardant treatments on bending and tension parallel properties over the course of a prolonged exposure to elevated temperature. These test results could be used to develop KT factors for treated wood used at elevated temperatures. Treated wood used in roof framing would normally be exposed to high temperatures for prolonged periods of time. Test Procedure 3 is an optional test which uses full size test specimens to modify the results from Procedures 1 and 2 for size effects.

The test standard stipulates test methods and what data needs to be presented in the final report. Additional information on developing KT factors is given in ASTM Standard D 6841, “Calculating Design Value Treatment Factors for Fire-Retardant-Treated Lumber”.

Clause 6.4.4 – System factor, KH Wood structural systems such as floors, walls and roofs incorporating essentially parallel providing mutual due to the structural redundancy of the system, have a higher structural capacity than predicted from the single member analogue upon which most design equations are based. The interaction of in redundant floor and roof systems comprising materials and sawn lumber framing has been extensively studied (See Commentary on Clause 5). These studies have provided a better understanding of relationships between the stresses and deformations calculated using single member design models and those actually induced in systems. CSA O86 recognizes two different degrees of structural interaction (Case 1 and Case 2) for which factors are tabulated. System factors for Case 1 apply when the system consists of three or more essentially parallel spaced not more than 610 mm apart and arranged to mutually the load. Thus, Case 1 is analogous to the systems description adopted in earlier standards for which a so-called loadsharing factor would have applied. Case 1 applies to systems such as wood trusses where the relationships between member forces are complex, or to conventional framing applications such as tongue and groove decking, that do not meet the specifications of Case 2. For Case 1 the Standard permits a 10% increase in specified strengths in bending, shear, compression parallel and tension parallel to grain. System factors can be significantly greater than permitted in Case 1 for many common wood floor, roof and wall systems used in light frame structures. However, the magnitude of the system factor will vary with member size and spacing, the member stiffness and strength variability, the correlation between stiffness and strength properties, sheathing thickness, nail properties, nailing pattern and many other

factors. For the Standard, the CSA committee has chosen to adopt a system factor that can be applied over a wide range of structural configurations provided the structural system meets the requirements for a Case 2 system. For Case 2, different system factors are applied for visually graded and machine stress-rated lumber. The lower system factor for MSR lumber grades is a result of MSR lumber having lower variability in strength and stiffness properties than visually graded lumber.

Clause 6.4.5 – Size factor, KZ The dependence of apparent material strength on the member size and loading condition and the type of test specimen is classified as a size effect phenomenon. Size effects exist to a greater or lesser extent in all structural materials. Strength properties of sawn lumber vary depending on the member size and loading conditions. Full-size evaluations of 38 mm dimension lumber conducted at a constant length-to-width ratio typically show that wider have lower strengths. For constant width , strength tends to decrease as member test length increases. The magnitude of the size effect will usually vary directly with the variability of strength properties. For visually graded sawn lumber, where the natural property variability is high, the size effects are significant (Barrett and Lau, 1994; FPRS, 1989; Madsen, 1986 and 1992; Lam, 1987; Barrett, 1989). To maintain uniform safety levels across the range of member sizes and load cases it is necessary and convenient to adopt a reference strength for a particular member size, and then to use an adjustment factor for other member sizes in the design equation. Size factors for bending and shear depend on both the member width and thickness. For tension parallel to grain, the size factor depends only on the larger dimension (i.e., the member width). For compression parallel to grain, the size factor is a function of both column length and the member dimension in the direction of buckling (Clause 5.5.6 in CSA O86). In the 2001 edition of CSA O86, the size factor was also applied to 64 by 89 mm and 89 by 89 mm lumber of No.1 or No.2 grade. This reflected a change in grading practice for these sizes and grades of lumber, which were exceptions in earlier editions.

Clause 6.5 – Strength and resistance

Clause 6.5.1 – General Design equations for bending moment resistance, shear resistance, simple compression and tensile resistance parallel to grain have been developed using reliability principles. (See Commentary on Clause 4 for a discussion of reliability analysis.)

Clause 6.5.2 – Sizes (See Table A 6.5.2 in CSA O86).

Clause 6.5.3 – Continuity This clause was intended to alert the designer to the fact that sawn timber classed as Beams and Stringers are graded differently in the middle one-third of the member length. The middle one-third is required to have smaller knots than the remaining two-thirds since the grading rule anticipates that the majority of these will be used in simply ed full-length applications. This does not affect most typical applications.

Clause 6.5.4 – Bending moment resistance The factored bending moment resistance Mr was derived following the same principles incorporated in earlier editions. Since design values were tabulated on the basis of Standard Term duration of load, the specified strengths are determined by reducing short-term characteristic strengths by a factor of 0.8. The resistance factor of 0.9 was derived based on reliability calculations and in-grade full-size test data.

Clause 6.5.5 – Shear resistance Design shear stresses for lumber in North America were derived from the provisions of ASTM standard D245, based on testing of small clear wood specimens. The primary reason for using this approach rather than full-size testing was the difficulty in obtaining pure shear failure for lumber under typical loading conditions. Although still based on the small clear wood approach, CSA O86 method for determining lumber shear values has been revised twice, first in 1989 and again in the 2005 Supplement to the 2001 edition of the Standard. During the In-Grade testing program in Canada, surveys were undertaken to determine the actual lengths of end splits in sawn , and shear design values were calibrated using a fracture mechanics model and reliability basis (Lau et al, 1988). Further studies on shear testing of full-size specimens of solid sawn and structural composite lumber (Lam and Craig, 2000; Lam, Yee and Barrett, 1997) provided a database for developing more accurate shear characteristic values. Analysis of this data and US Forest Products Laboratory shear data (Sanders, 1996) led to an adjustment factor of 0.74 to convert small clear wood test data to a full-size characteristic value basis assuming a representative centre-point bending configuration. Specified strengths were further reduced using factors for duration of load (0.8) and variability (0.97). The strength values for dimension lumber were also multiplied by 0.63 to adjust to a 286 mm depth basis. A size factor KZv is applied in design to for the change in member shear capacity with member size, providing increased shear resistance for smaller . Research at the University of New Brunswick (Smith and Springer, 1993; Smith et al, 1993) applied fracture mechanics principles to the problem of notched . Although notches impact the shear capacity of the beam, the strength is a function of the fracture toughness of the wood rather than the longitudinal shear strength. The model, which uses specific gravity to predict capacity of notched lumber beams, showed that the traditional shear calculation for notched beams tended to be unconservative in some cases; e.g., when the notch was fairly long. As a result, a revised design procedure for notching effects (Clause 6.5.5.3) was introduced into the Standard in 1994. This design procedure applies only to notches on the tension side at s, and is identified as the notched shear force resistance calculation (Fr), to distinguish it from the shear calculation (Vr). Both calculations are required for notched . The reduction in longitudinal shear resistance for notches on the compression side is captured by the use of net in the calculation of Vr. On the tension side, it is assumed that no notches occur on the interior of the tension face. A Figure (6.5.5.3.2) and Clause (6.5.5.3.3) were added in the 2001 edition to help clarify the calculation; i.e., depth and length of the notch, length of , and force components. In interpreting the notched shear force calculation for sloped , the reaction component normal to the member would generally be used.

The notched shear force calculation also has a separate service condition factor, since this limit state is more sensitive to moisture content effects.

Clause 6.5.5.1.2 Additional work by Smith (2015) determined that the effect of all loads including those near the , i.e. within a distance from a equal to the depth of the member, need to be included in the calculations for notch shear resistance. This was added to the 2014 edition of the Standard.

Clause 6.5.6 – Compressive resistance parallel to grain The load capacity of columns depends on their slenderness ratios Cc in both directions. Earlier versions of the Standard differentiated between short, intermediate and long columns depending on the slenderness ratio. Requirements for sawn lumber columns were modified in 1989 to incorporate research on column behaviour and resistance factors based on reliability analysis. Similar provisions were adopted for gluedlaminated timber columns in the 1994 edition. The variation in column capacity with slenderness is incorporated through the slenderness factor Kc, which is a continuous function of the slenderness ratio Cc. The slenderness factor Kc = 1.0 for Cc = 0, and decreases as Cc increases to approach the value corresponding to the elastic buckling load at high slenderness ratios. Kc is dependent upon the adjusted specified strength Fc, and the fifth percentile of modulus of elasticity E05. The factor Kc modifies a compression strength obtained from tests of where lateral deflection or buckling was prevented. The adjusted specified strength Fc must also be modified for effects of column dimensions using the factor KZc, based on tests of structural size that showed that specified strengths increased for narrower widths and shorter . Tests also showed that in long sawn lumber columns, overall buckling can be triggered by local crushing indicating that the KZc factor is applicable to long columns. Clause 6.5.6 was revised in 2009 to make it clear that compressive resistance should be checked in both directions, even if continuously braced in the “weak” direction, due to the KZc factor which will vary based on the dimension and associated length in the direction of buckling. The resistance factor = 0.8 has been selected to achieve adequate reliability for all combinations of species, grades and sizes of lumber. Reliability calculations are reported in Foschi (1989), and compared with test data from Buchanan (1984). The formula chosen for the Standard was a cubic Rankine-Gordon relation, based on Neubauer (1973) and Johns (1991). A nominal load eccentricity to represent random column deviations from straightness was assumed for reliability calculations. The load eccentricity was taken to be normally distributed (mean = 5% of member dimension in the direction of buckling, COV = 20%). Compression parallel to grain design values were increased in 2009 for lower grades of lumber (No.3 and Stud grades). These values were set equal to bending values in previous editions due to lack of compression data for lower grades, although existing compression-to-bending relationship data indicated that the ratios should be greater than unity for lower strength lumber. The property relationship data showed that the compression strength values for these grades could be increased by 15 to 30%, depending on the species. CSA O86 contains information on built-up columns (Clause 6.5.6.4), based on a design procedure described in Malhotra and Van Dyer (1978).

Clause 6.5.7 – Compressive resistance perpendicular to grain Specified bearing strengths applicable to the commercial lumber species groups were derived from ASTM procedures for clear wood compression perpendicular to grain strength. The strength tabulated for individual species in ASTM D 2555 corresponds to the average stress in compression perpendicular to grain at the proportional limit, as determined by the ASTM D 143 clear wood test procedure. This value was modified to a mean 1 mm (0.04") deformation limit using a formula in the Appendix of ASTM D 2555. In the 1994 edition, the relationship between mean compression perpendicular strength and mean ovendry wood density was used to establish a consistent basis for bearing strengths for various products (visually graded lumber, machine rated lumber, glued-laminated timber). This relationship is described in more detail in Clause A.6.5.7. The bearing design procedure was revised in the 1994 edition in three ways: 1. An additional check for “critical” bearing was added for cases where compression perpendicular to grain loads are applied to the top face of a member, within a distance “d” from the centre of the (see Figure 5). In critical cases, the stress can extend through the full depth of the member, in contrast with the general beam case where the bearing stress is typically limited to the outer fibres of the bending member. Due to the increased volume under stress, and the resulting increase in probability of deformation over time, the resistance for critical cases (Clause 6.5.7.3) is multiplied by a factor of 2/3, based on work at Forintek Canada Corp. (now FP Innovations) (Lum, 1994 and 1995). 2. A size factor for bearing KZ was introduced to for the effect of member orientation on resistance. In developing design values, the worst growth ring orientation (45°) was assumed. Used in a flat orientation, wood has improved crushing resistance compared with edge orientation, due partly to the likelihood that only some portion of the cross-section, if any, will be at the worst growth ring orientation (Lum, 1994 and 1995). 3. The duration of load factor was reintroduced into the bearing equation. This factor reflects the ability of wood to carry greater loads for shorter periods of time.

Figure 5 Compression perpendicular to grain procedure (sawn lumber, glulam)

In the critical bearing case, the bearing area may be calculated as an average of the top and bottom bearing surfaces (Clause 6.5.7.4). Where the critical bearing check controls and applications limit the increase in the length of the bearing (e.g. width of top wall plates), the critical bearing requirement can be met by increasing the length of the bearing under the load applied to the top edge of the beam. In application of the critical bearing formula (Clause 6.5.7.3), there is room for interpretation. For example, research shows that the toothed connector plates used to metal-plated wood trusses provide increased compression perpendicular to grain resistance at s (Bulmanis et al, 1983). The factored compressive resistance perpendicular to grain Qr is directly proportional to the bearing area except for cases where the bearing capacity may be increased by the bearing factor KB. The bearing factor adjusts the tabulated specified strengths to recognize the fact that bearing resistance at locations of wood to wood or metal to wood bearing may be substantially higher than the tabulated values under certain special circumstances as defined in the Standard.

Clause 6.5.8 – Compressive resistance at an angle to grain The compressive resistance at an angle to grain Nr is calculated using the compressive resistance parallel to grain Pr, the perpendicular to grain resistance Qr, and the Hankinson formula to derive Nr at any angle. The factored compressive resistance parallel to grain Pr is calculated using the design equation for a simple compression member and assuming a slenderness factor KC = 1.0. The expression for calculating the compressive resistance perpendicular to grain Qr has been modified as described earlier.

Clause 6.5.9 – Tensile resistance parallel to grain The tensile resistance parallel to grain Tr is calculated following procedures used since the 1984 Standard. The resistance factor of 0.9 has been selected based on reliability studies (Foschi, 1989). Specified tension strengths, like bending and compression strengths, apply to wide . For narrower widths the specified strengths are increased by the size factor KZt, and for very wide the specified strengths are decreased. Tension values for regular grades of MSR lumber are tabulated only for widths up to 184 mm. NLGA Standard SPS 2 “Special Product Standard for Machine Stress Rated Lumber” requires qualification and daily quality control for bending but not for tension strength. Because lumber wider than 184 mm has not been produced using SPS 2 and any production would likely be very limited in volume, verification of the standard tension assignments should be undertaken at the mill level. Note that tension testing is part of the daily qualification and quality control procedures for special MSR grades (bottom half of Table 6.3.2). In such cases tension design values are assignable to any width of lumber, and NLGA SPS-2 requires the Ft designation for the grade to be marked on the lumber. A footnote was added to the Table in the 2005 Supplement to the 2001 edition of CSA O86 to recognize that this same approach could be used to assign tension values to regular as well as special grades of MSR lumber.

Clause 6.5.10 – Resistance to combined bending and axial load This clause provides interaction equations to be satisfied when deg subjected to combined axial and bending loads. These interaction equations were revised in the 2009 edition based on research in Canada and elsewhere on the effects of combined bending and axial loading in typical wood . Prior to this revision, the interaction equations were based on a linear model originally proposed by Newlin and Trayer (1941).

Studies by Johns and Buchanan (1982) and Buchanan (1984, 1986) showed nonlinearity in the bendingcompression relationship due to elastic and plastic strength behavior of the bending compression interaction. Lumber fails in bending in a brittle manner (for example around knots). The addition of compression delays this brittle failure leading to increased member resistance under combined loading. Research showed that the ultimate failure condition followed a parabolic rather than a linear relationship for beam-columns with typical slenderness ratios (limited to 50 or less by Clause 6.5.6). The 2009 edition of CSA O86 introduced a non-linear equation with a squared axial (compressive) term. Reliability studies by Foschi (1989) have shown that, although the linear interaction equation is a more conservative approach, the non-linear equation provides more consistent reliability over a range of conditions. The revised equation provided for more efficient use of under combined bendingcompression stress. Note: the bending-tension relationship was not revised as tension combines linearly with the brittle failure modes in bending. Previous editions of CSA O86 have included two cases where the axial term was permitted to be squared to provide a more accurate solution. The first case was for wood trusses (6.5.13), and the second case was for preserved wood foundations (A.6.5.12). These cases have also been modified slightly to be consistent with the revised interaction equation in 6.5.10. The overall impact of the non-linear design equation is to reduce required sections. The change is more pronounced for taller studs and higher axial loads, and is offset in some cases by moment amplification effects (discussed below). A further 2009 revision to Clause 6.5.10 was explicit guidance on moment amplification due to axial loads in laterally loaded . For subjected to combined bending and axial compression, the total factored bending moment Mf was required to include the additional moment generated by the socalled P- Δ effect. Where a laterally deflected member is subjected to an axial factored compressive force Pf, then the moment generated within the member due to the axial load is the P-Δ effect. The moment amplification factor in Clause 6.5.10 is given as 1/(1-Pf/PE) where PE is the Euler buckling load, calculated using the 5th percentile modulus of elasticity value (E05) for the species and grade. This simplified approach is slightly more conservative than the detailed amplification approach, provided in Clause A.6.5.10 as an alternative. The more exact approach uses the product of axial load and lateral deflection calculated using serviceability load combinations, times the amplification factor to augment the applied moment. The simplified method is consistent with the approach used for other structural materials. Axial load in combination with biaxial bending is not addressed in the 2009 edition of the CSA O86 Standard. If the CSA O86 provisions are being extended to address axial load in combination with biaxial bending, it is recommended that the axial term not be squared and that the moment is amplified in both directions:

 Pf Mfx  1 +  Pr Mrx  1 − Pf  PEx

  Mfy +  Mry 

  1   1 − Pf  PEy 

   ≤1   

(Eq. 14)

Clause 6.5.11 – Decking The size of plank decking is based on lumber dimensions prior to milling. The lumber used for decking is milled to single tongue and groove or double tongue and groove. Typical milled sizes are shown below:

Note that 38 x 127 mm and 38 x178 mm decking is often surfaced to 36 mm thickness Decking may be installed as simple spans, controlled random span or continuous over two spans in accordance with the requirements of Clause 6.5.11. The controlled random pattern for plank decking is illustrated below.

Notes: 1. Random Pattern is not permitted for bridges. 2. The other methods, Simple Span and Two Span Continuous installation patterns, require that all butt ts occur over a .

Nailing arrangement of plank decking must conform to Clause 6.5.11 of CSA O86. The figure below illustrated the requirements in CSA O86 as well as typical industry practice to achieve diaphragm action.

Part 4 of the National Building Code specifies that the minimum specified concentrated roof live load on a roof area is 1.3 kN. In the 2010 edition of the National Building Code, the corresponding concentrated loaded area on the roof was reduced from 750 mm x 750 mm to 200 mm x 200 mm. Research at University of Ottawa (Rocchi and Doudak, 2014) has shown that there is load sharing between deck boards. For decking systems consisting of decking with widths of 140 mm or less, testing has shown that the National Building Code roof point load can be resisted by three or more courses of decking.

Clause 6.5.12 – Preserved wood foundations The provisions in Clause 6.5.12 that waived the use of the KT factor in the design of wood foundation studs was deleted in the 2001 edition. The resulting design for wood foundation studs is consistent with design of other wood systems in the Standard. Other changes to wood foundation design are given in Appendix A.6.5.12.6: squared term in the interaction formula, and specific guidance in the footnote regarding application of the duration of load factor.

Clause 6.5.13 – Sawn lumber design for specific truss applications Clause 6.5.13 was added in the 2001 edition to provide a more accurate methodology for the design of certain types of metal plate-connected wood trusses. Appendix A.6.5.13 provides detailed commentary on this section. The 2009 edition has revised the factored bending moment definition in 6.5.13.5 to be consistent with the general definition in Clause 6.5.10, for application in cases where mid- deflections are not limited.

References: Barrett, J.D. and H. Griffin. 1989. Size Effects for Canadian 2-inch (38 mm) Dimension Lumber. Proceedings of CIB Meeting. Berlin, East . Barrett, J.D. and W. Lau. 1994. Canadian Lumber Properties. Canadian Wood Council, Ottawa, ON. Bodig, J. and B. Jayne. 1982. Mechanics of Wood and Wood Composites. Van Nostrand Reinhold Company Inc. New York, NY. Buchanan, A.H. 1984. Strength Model and Design Methods for Bending and Axial Interaction in Timber . Ph.D. Thesis, Department of Civil Engineering, University of British Columbia. Buchanan, A. H. 1986. Combined Bending and Axial Loading in Lumber, ASCE Journal of Structural Engineering, Vol. 112, pp. 2592-2609. Bulmanis, N.S., H.A. Latos, and F.J. Keenan. 1983. Improving the bearing strength of s of light wood trusses. Canadian Journal of Civil Engineering, Vol. 10, pp. 306-312. CWC. 2001. Standard Practice Relating Specified Strengths of Structural to Characteristic Structural Properties, Canadian Wood Council, Ottawa, Canada. Forest Products Laboratory. 2010. Wood handbook-- Wood as an Engineering Material. Gen Tech Report FPL-GTR-190. Madison, WI: US Department of Agriculture, Forest Service, Forest Products Laboratory.. Forest Products Research Society (FPRS). 1989. In-Grade Testing of Structural Lumber. Proceedings of the ASTM Workshop on the North American In-Grade Testing Program. Forest Products Research Society. Madison, WI. Foschi, R.O. et al. 1989. Reliability-Based Design of Wood Structures. Structural Research Series, Report No. 34, Department of Civil Engineering, University of British Columbia. Jessome, A. 1977. Strength and related properties of woods grown in Canada. Forestry Technical Report 21. Eastern Forest Products Laboratory. Ottawa, ON (now FPInnovations) Fouquet. R.J-M. J D. Barrett. D.M. Ainsworth. 1986. Sustainability of fingered lumber for truss maufacture. National Research Council of Canada . Industrial Research Assistance Program, Project No. 6-1223-19s. Council of Forest Industries of British Columbia, Vancouver BC. Johns, K.C. 1991. A Continuous Design Formula for Timber Columns. Canadian Journal of Civil Engineering, Vol. 18, 617-623. Johns, K.C. and A.H. Buchanan. 1982. Strength of timber in combined bending and axial loading. Boras, Sweden: International Union of Forest Research Organizations, Proceedings IUFRO Wood Engineering Group S5.02. 343-368. Lam, F. 1987. Effect of Length on Tensile Strength of Lumber. Forintek Canada Corp.(now FPInnovations) Vancouver, BC. Lam F. and B. Craig. 2000. Shear strength of structural composite lumber. Journal of Materials in Civil Engineering. ASCE. 12(3):196-204. Lam, F. and P. Morris. 1991. Effect of Double-Density Incising on Bending Strength of Lumber. Forest Products Journal, Vol. 41, No.9, pp. 43-47. Lam F., H. Yee, and J.D. Barrett. 1997. Shear strength of Canadian softwood structural lumber. Canadian Journal of Civil Engineering. 24(3):419-430. Lau, W., J.D. Barrett and R.O. Foschi. 1988. Performance Factors for Lumber in Shear (Single-member, Short-term Strength). Reliability of Wood Structures Research Project. Report No. 9. Department of Civil Engineering, University of British Columbia. Lum, C. 1995. Compression Perpendicular-to-Grain Design in CSA O86.1-94. Report to Forestry Canada, No. 14. Project No. 1510K018, Forintek Canada Corp. (now FPInnovations), Vancouver, BC. Lum, C. 1994. Rationalizing Compression Perpendicular-to-Grain Design. Report to Forestry Canada, No. 13. Project No. 1510K018, Forintek Canada Corp. (now FPInnovations), Vancouver, BC. Madsen, B. 1992. Structural Behaviour of Timber. Timber Engineering Ltd. North Vancouver, BC. Madsen, B. and A.H. Buchanan. 1986. Size Effects in Timber Explained by a Modified Weakest Link Theory. Canadian Journal of Civil Engineering, 13:218-232. Madsen, B. and T. Stinson. 1982. In-grade Testing of Timber Four or More Inches in Thickness. Department of Civil Engineering. University of British Columbia. Malhotra, S.K. and D.B. Van Dyer. 1978. Rational Approach to the Design of Built-Up Timber Columns. Wood Science, Vol. 9, No. 4. Morris, P.I., J.J. Morrell and J.N.R. Ruddick. 1994. A review of incising as a means of improving treatment of sawnwood. International Research Group (IRG) on Wood Preservation. Document No. IRG/WP/94-40019. Stockholm, Sweden. National Lumber Grades Authority (NLGA). 2008. NLGA Standard Grading Rules for Canadian Lumber. National Lumber Grades Authority. Vancouver, BC.

National Lumber Grades Authority (NLGA). 2008. SPS 2 NLGA Special Product Standard For Machine Graded Lumber. Vancouver, BC. National Lumber Grades Authority (NLGA). 2008. SPS 1 NLGA Special Product Standard for Fingered Structural Lumber. Vancouver, BC. National Lumber Grades Authority (NLGA). 2008. SPS3 NLGA Special Product Standard For Fingered (Vertical Stud Use Only) Lumber. Vancouver, BC. Neubauer, L.W. 1973. A Realistic and Continuous Wood Column Formula. Forest Products Journal, Vol. 23, No. 3, pp. 38-44. Newlin, J. A., and Trayer, G. W., (1941) “Stress in Wood Subjected to Combined Column and Beam Action” Forest Product Laboratory Report No. 1311, Madison, Wis. Palka, L.C., J.D. Barrett, R.S. Smith and J.N.R. Ruddick. 1984. Effect of Incising and Pressure Preservative Treatments upon Selected Strength Properties of Commercial Dimension Lumber. Forintek Canada Corp. (now FPInnovations), Project No. 2-50-68584. Palka, L.C., P.I. Morris and D.C. Walser. 1992. Bending Strength Properties of Incised and Preservative Treated Commercial 2x4 Western Lumber. Forintek Canada Corp. (now FPInnovations), Vancouver, BC. Rocchi, K. and Doudak, G. (2014). "Effects of Load Sharing in Wood Tongue-and-Groove Plank Decking under Concentrated Loads." Journal of Performance of Constructed Facilities, Vol. 29, No. 3, June 2015. Sanders, C. 1996. The Effects of Testing Conditions on the Measured Shear Strength of Wood Beams. MS Thesis. Dept of Civil and Environmental Engineering, Washington State University, Pullman, WA. Smith, I., Kiwelu, H.M., Weckendorf, J. 2015. Explanation of tension-side notch design provisions for glulam in CSA Standard O8614. Canadian Journal of Civil Engineering, 42(10): pp. 787-796. Smith, I. and G. Springer. 1993. Consideration of Gustaffson’s Proposed Eurocode 5 Failure Criterion for Notched Timber Beams. Canadian Journal of Civil Engineering, Vol. 20, No. 6, Dec. 1993, pp 1030-1036. Smith, I., L. Poulin, L.J. Hu, and Y.H. Chui. 1993. Characterisation of Critical Reactions for Lumber Joists with an End Notch on the Tension Face. Paper prepared for CSA O86. U. of New Brunswick, Fredericton, N.B.

Clause 7

Glued-laminated Timber

Clause 7.1 – Scope Glued-laminated timber (glulam) is a manufactured product. As such it retains some of the basic characteristics of the parent lumber material from which it is created. This material is tested, graded and arranged to create an engineered material with characteristics that can be varied to suit the end purpose. Lumber meeting special requirements for moisture content and surfacing tolerance is stiffness rated, visually graded and assembled with adhesive into timbers that are longer, wider and more consistent in strength and stiffness than the parent material. The result may be straight or curved, with rectangular, square or I-shaped cross sections. Laminating stock of similar characteristics are assembled to produce of uniform strengths for specific types of end uses. Glulam has been widely used for many years to provide reliable structures in Canada and in other countries. Because the characteristics of the finished glulam member depend on quality of manufacture, the product must conform to CSA Standard O122, Structural Glued-Laminated Timber, and laminating plants must be qualified in accordance with CSA Standard O177, Qualification Code for Manufacturers of Structural Glued-Laminated Timber. Only material meeting the requirements of CSA O122 may be assigned the specified strengths given in Table 7.3 of CSA O86.

Clause 7.2 – Materials Glulam is widely available in both Douglas Fir-Larch or Spruce-Pine species groups. Douglas Fir-Larch has a brown to reddish colour with white sapwood. Spruce and lodgepole pine are light yellow brown to white in colour and jack pine is light orange brown to white. Spruce-Pine glulam is somewhat lighter and easier to handle and work in the field but is more subject to damage from slings and rough handling than the harder and denser Douglas Fir-Larch. Douglas Fir-Larch has a higher natural decay resistance than spruce or pine and is more suitable in moist service conditions. Under conditions requiring pressure treatment, Douglas Fir-Larch is typically chosen. Hem-Fir/D. Fir-L has similar characteristics to Douglas Fir-Larch because it utilizes Douglas Fir-Larch in outer laminations, but is available only on special order. This species combination was developed during a period when Douglas fir was in short supply and is not likely to be widely available unless these conditions recur. The standard depths for glulam vary in multiples of 38 mm for straight and for curved with radii greater than 10800 mm, or as small as 8400 mm if have tangent ends. The standard depths vary in multiples of 19 mm for radii between 3800 and 10800 mm, or as small as 2800 mm if have tangent ends. For radii smaller than these limits see Table A.7.5.5 of CSA O86. Tangent end means that a length of straight glulam is located between the cut end and the nearest tangent to the curve. For minimum radii this length must be at least 32 times the lamination thickness. Tangent ends are commonly used in curved , since the production of true curved ends requires special manufacturing techniques which increase the cost of the . Often a curved outer shape is achieved by trimming the member to the desired radius during finishing.

Clause 7.2.1 – Stress grades The various grades of glulam were developed to make the most efficient use of available material. The bending grade 20f-E or 24f-E is intended for use in simple beams or where high reverse moments do not occur (i.e., short cantilever spans). The 20f-EX or 24f-EX grade is intended for use in continuous beams, wind columns and arches where large moments can occur in either direction. Normally the 20f-E grade should be chosen for typical beam applications. The 24f-E grade requires more high-grade material and may command a price, which is usually only justified for very large where the high bending strength maintains practical sections, or architectural requirements demand smaller sizes. The 16c-E and 12c-E column grades are both widely available and suitable for axial stresses. The 18t-E grade is intended only for conditions where a high resistance to tension stress is required. Its greater content of high grade material will likely command a price.

Clause 7.2.2 – Appearance grades In glulam the second level of grades, appearance grades, are sometimes mistakenly related to strength. In fact, since appearance grade refers only to surface finish, any stress grade can be obtained in any desired appearance grade. The appearance grades are: •

Industrial Grade: This grade is intended for conditions where the timber is to be hidden from view or if appearance is of little importance. It receives no finishing other than planing and chamfering of the corners.

•

Commercial Grade: This grade is intended for use where appearance is important but economy is also a consideration. It would normally be used in cottages, commercial establishments and small recreational buildings such as gymnasia and swimming pools. This grade is planed, chamfered, all excess glue is removed and its surface is sanded.

•

Quality Grade: This grade is intended for use in buildings such as homes, schools and churches where appearance is of prime importance. It is planed, chamfered, and all open defects, slivers and checks are filled to give a smooth flat surface, and then it receives a high quality sanding.

In all of the grades, the natural characteristics of the wood are permitted to show. Clause 7.2.2.2 was revised in 2005 to advise the designer that textured finishes could reduce member structural properties. CSA O122 contains information on finished sizes and tolerances of glued-laminated timber.

Clause 7.3 – Specified strengths The specified strengths in bending, tension, and tension perpendicular to grain were calibrated to the values given in CSA O86-M84 (Working Stress Design). The decision to retain this calibration was based on the record of glulam structures over more than 50 years. Resistance to snow loading was used as the benchmark for calibration. In 2005, following a review of format conversion factors, the specified negative bending moments for some grades were slightly reduced.

The 2006 edition of the CSA O122 product standard introduced a modified lamination grade and grade combination, to improve use of the lumber resource and produce a more efficient 24f-E/EX Douglas FirLarch beam. The modified grades and layup were evaluated at the University of British Columbia through modeling studies (Lam and Mohadevan, 2007) and full-size testing of 48 beams confirmed the modeling. Based on this work, the CSA O86 modulus of elasticity for a 24f-E/EX Douglas Fir-Larch beam was reduced from 13,100 MPa to 12,800 MPa. The revisions did not require any change to bending strength or other design values. Although the revision was due to the introduction of the new 24f-E/EX Douglas Fir-Larch beam, the values also apply to beams made to the original 24f-E/EX beam combination. The specified strengths in longitudinal shear and compression parallel to grain are based on the results of reliability studies. (See the Commentary to Clauses 7.5.7 and 7.5.8). The specified strengths in compression perpendicular to grain are, in general, based on the values used for sawn lumber. (See the Commentary to Clause 7.5.9).

Clause 7.4 – Modification factors

Clause 7.4.1 – Load duration factor, KD The effect of the length of time that a load remains on a member has been shown to be similar for various sizes and species of wood. As such, the effect is considered to be a characteristic of the parent material, and the same duration of load factors were adopted for glulam as for sawn lumber.

Clause 7.4.2 – Service condition factor, KS The effect of moisture is considered to be a characteristic of the parent material independent of both the laminating process and the waterproof glue that is used in the process. This clause then, is similar to that for sawn wood products. Glulam is normally used in larger sections than sawn lumber. These sections contain frequent gluelines that are relatively impermeable to moisture. The effect of these factors in retarding moisture pick-up is recognized in Clause 7.4.2.2 by allowing the designer to choose a modification factor other than that given in Table 7.4.2 based on judgment of the moisture exposure and the specified protection methods. Caution should be exercised in coastal regions or microclimates where relative humidity may be high for prolonged periods. Untreated Douglas Fir-Larch glulam has demonstrated good performance under adverse moisture conditions, such as due to poor roof maintenance, provided that the moisture was not retained for long periods of time.

Clause 7.4.3 – System factor, KH The system factor is only rarely applicable to glulam structures as the are normally spaced farther than 610 mm apart.

Clause 7.4.4 – Treatment factor, KT Preservative treatments are not considered to have a deleterious effect on glulam provided that the are not incised. The incising process causes a reduction in the moment of inertia of the cross-section, and therefore can reduce the strength and stiffness of glulam . The magnitude of the incising effect depends on the incising depth, density, and tooth geometry used. Small beam sizes are more effected than large beam sizes. In the case of fire-retardant treatments, the previous practice of multiplying the specified strength values by a factor of 0.90 has been discontinued. The current edition requires that the treatment factor for fireretardant and other potentially strength-reducing chemicals be based on documented test results. It should be noted that the effect of fire-retardant treatment may vary, depending on the strength property tested and the treatment chemical used. For glulam , the effect of the treatment on the strength of the end-ts within the laminations should also be considered. The treatment of glulam with water-borne chemicals after gluing is not permitted. Water-borne treatments, followed by subsequent re-drying, cause large changes to the moisture content of the wood. Dimensional changes occur in response to these changes in moisture content. If glulam are treated with water-borne chemicals after gluing, these dimensional changes create both shear and tension perpendicular to grain stresses at the lamination boundaries, which may result in excessive checking or splitting of the member. Oil-borne treatments do not cause dimensional changes, and therefore this limitation does not apply to oil-borne chemical treatment.

Clause 7.5 – Strength and resistance Glulam’s strength is based to a considerable degree on the ability to place stronger laminations in areas where higher stresses will occur. This effect can be achieved most efficiently in beams, due to the high variation in stress intensity, provided that loading is in a plane perpendicular to the plane of the laminations. In beams loaded parallel to the plane of the laminations, the beam lay-up has little or no effect but there is still a considerable increase in strength due to the effect of laminating different pieces of lumber together, and load sharing forced by the glue bond. Although the true strength is considered to be higher, a conservative value equal to that of No.2 grade lumber is applied where glulam is stressed in bending by forces acting parallel to the laminations. (Note: this means that different specified strengths are applicable to the X-X and Y-Y axes in two-way bending design.)

Clause 7.5.6.3 – Lateral stability conditions Consideration of the effect of the lateral restraint of the compression face of glulam beams is separated into two parts. Where the beam is shallow (d/b < 2.5) or where the compression edge is restrained throughout its length by a rigid diaphragm, the uned length is considered to be zero. A suitable rigid diaphragm is generally considered to be formed by 64 or 89 mm decking in which the pieces are edge-nailed to each other, or by a diaphragm-forming sheathing such as plywood, waferboard or strandboard (OSB). Boards and 38 mm decking are not edge-nailed and are not considered to provide an adequate diaphragm unless they have diaphragm-forming sheathing nailed to the deck either directly or through strapping. Where rectangular glulam are carrying both axial and bending loads, the lateral restraint conditions default to the same rules as are applied to sawn timber. Where lateral restraint cannot be provided, a modification factor KL must be determined and applied to the moment resistance. The method used, the same as in previous standards, is based on the work of Hooley and Madsen (1964).

Clause 7.5.6.5 – Moment resistance

Clause 7.5.6.5.1 – General A size factor in bending was introduced in the 2001 edition. This was a significant change from previous editions where glulam bending strength values were not adjusted for size effects. In the 2014 edition of the Standard new size factors for glulam were introduced based on research conducted at the University of British Columbia (UBC). The research demonstrated that a revised glulam bending size factor, KZbg, can provide a more representative size factor for a larger range of beam sizes. The specified bending strength of glulam beams, is referenced to a beam volume of 130 mm x 610 mm x 9100 mm. Tests at UBC showed that the beam bending strength, Fb, is allowed to be increased for smaller beam sizes up to 30%. The upper limit of 1.3 for KZbg is based on the smallest beam size tested and reported by UBC – 130 mm x 152 mm x 2740 mm. As noted in Clause 7.5.6.5.1, the length “ L” to be used in the size factor equation is the length of the beam segment from point of zero moment to point of zero moment. This defines the length of the tension zone. For simple span beams, “L” is equal to the beam length. For beams with one or more points of inflection (typically multi-span and cantilevered beams) there exists more than one beam segment. In these cases “L” is less than the overall beam length. The moment resistance, as modified by KZbg, must be calculated for each beam segment and compared to the maximum factored moment within that segment. The size factor, KZbg, is not to be applied cumulatively with the lateral stability factor, KL. The size factor is associated with stresses on the tension side of the beam, whereas the lateral stability factor is associated with stresses on the compression side of the beam.

Clause 7.5.6.5.2 – Curvature factor One of the advantages of glulam is the ease with which it can be fabricated into curved shapes. The act of forming the wood to these shapes as well as the shapes themselves give rise to non-standard stress distributions, which must be considered in design. The first of these was quantified by Wilson (1939) and allows for the residual stresses induced when curving the individual laminations during fabrication. The modification factor KX allows for this effect.

Clause 7.5.6.5.3 – Double-tapered Instead of being curved during lamination, the surface of a beam may be sawn to provide a tapered or double tapered section for architectural or drainage purposes. This results in a member where the upper and lower surfaces are not parallel. The resulting reduction in bending capacity is ed for in Clause 7.5.6.5.3, which applies to both straight and curved .

Clause 7.5.6.6 – Radial resistance Another effect of curvature is the introduction of perpendicular to grain stress in service. Where bending moment tends to decrease curvature, the stresses are tension perpendicular to grain. The maximum value of these stresses are computed using methods described in Foschi (1970). Barrett, Foschi and Fox (1975) showed that the resistance of the member decreased with an increase in the volume of the curved tapered section of the member. This resulted in the size factor that appears in the formula for resistance to tension perpendicular to grain. For those , such as arches, where

long curved sections may either have no tangent or where the moment reverses at some point in the curve, the effective volume should be taken between points where the moment is equal to 85% of its maximum value. Although compression perpendicular to grain forces exist where the bending moment tends to increase the curvature, the capacity of wood in this direction is many times that for tension perpendicular to grain and the probability of an overstress is remote.

Clause 7.5.7 – Shear resistance In the 2014 edition of the Standard Clause 7.5.7.1 was introduced to clarify when it is necessary to include the effect of all loads acting within a distance from a equal to the depth of the member and when the loads within a distance from a equal to the depth of the member need not be taken into . The resistance to horizontal shear of glulam beams was based on the work of Foschi and Barrett (1975) who found that the resistance of Douglas fir varied depending on both the volume of the member and the loading pattern. This is the “detailed” method given in Clause 7.5.7.2 (a) and is the principal shear calculation method for glulam beams. For beams less than 2 m3 in volume, however, a simplified method may be used. The simplified method in Clause 7.5.7.2(b) is the same as that used for sawn lumber. The simplified method in Clause 7.5.7.2. (b) is also used for calculation of shear resistance in glulam other than beams. The detailed method results in a higher factored resistance than the simplified method for smaller ; so if the simplified method resistance of a beam less than 2 m3 in volume is found to be insufficient, either a greater section can be chosen to increase the resistance or a second check can be made using the detailed method. In the detailed method the total factored load on the beam is compared to the resistance calculated by a formula, which s for the influence of volume and loading pattern. In many cases, particularly for short heavily loaded beams, this method results in increased resistance. Large long-span beams however, will have lower resistance than would be given by the simpler formula. Based on the results of reliability studies (Foschi et al.,1989), the constant term in equation 7.5.7.2. (a) has been revised from “0.60” to “0.48”. This was done to achieve a reliability index β equivalent to that used for bending properties. As a result of the combined effects of volume and loading pattern, the principal of superposition does not apply when calculating the shear load coefficient (Cv). For those conditions not given in Tables 7.5.7.5.A to F, the shear load coefficient must be calculated according to the method of Clause 7.5.7.5. Prior to the 1994 edition of the Standard, the equation given in Clause 7.5.7.5(d)(ii) included adjustments for both impact and duration of load factors (normally applied to the design loads and specified strength values respectively). These adjustments resulted in the difference between the constant term in the equations for stationary and moving loads (i.e., 1.825 vs. 2.310 in editions of the Standard prior to 1994). The current clause has all “hidden” adjustments removed and therefore the equations for stationary and moving loads are identical. Table 7.5.7.3F was revised in 1994 to incorporate the above noted change and to include a number of vehicle live loads frequently used for bridge design. In the 2014 edition of the Standard, provisions for fracture shear resistance at a notch on the tension side at a were introduced in Clause 7.5.7.4. The notch factor is consistent with Clause 6.5.5 for sawn lumber, while the fracture shear resistance at notch is based on research by Smith (2015). The research by Smith (2015) also indicated that small notches create shear stress distributions in residual portions of above notches that are not parabolic. To recognize this, for glulam beam with tension side end notches, Clause 7.5.7.4.1 allows the longitudinal shear resistance of residual member above notch to be based on the full depth of the beam for tension side end notches meeting the limitations given in the

Clause. Additional information on fracture shear resistance at a tension side notch can be found in the commentary on Clause 6.5.5.

Clause 7.5.8 – Compressive resistance parallel to grain The calculation of compressive resistance parallel to grain for glulam columns uses the same design approach as sawn lumber columns (6.5.6). The design procedures in Clause 7.5.8 are based on research into the behaviour of wood columns (Johns, 1991) and full-scale column tests conducted by Forintek Canada Corp. (now FPInnovations) (Karacabeyli, 1992). The traditional “short”, “intermediate” and “long” column equations were replaced by a single continuous column formula. The slenderness factor KC which s for the effect of column slenderness on compressive resistance, was expressed in of a Cubic Rankine-Gordon relationship. An empirical curve shape parameter (the constant term of 35 in the denominator) was selected to make the CRG curve follow the Euler curve at high values of slenderness. In the calculation of the slenderness factor, the average MOE reported in Table 7.3 was reduced by a factor of 0.87 to calculate a 5th percentile property value, E05. The values for specified strength parallel to grain (fc) given in Table 6.3 were derived from the results of compression tests conducted on laterally restrained Spruce-Pine glulam columns (i.e. KC = 1.0). The corresponding values for Douglas fir glulam were based on data relating the compressive strength of Douglas fir to Spruce-Pine lamstock. Based on the results of reliability calculations, the committee choose a φ value of 0.8 to provide a satisfactory level of reliability. As part of the reliability analysis, the effect of member volume on compressive strength was ed for by means of a size factor, KZcg. This factor appears in the equations for both the factored compressive resistance and the slenderness factor, reflecting the fact that longer columns have lower compressive resistance due to the combined effect of buckling and reduced compressive strength. The formula for KZcg given in the Standard has been calibrated to a reference column size of 127-mm x 152-mm x 2540-mm. This corresponded to the column size tested by Forintek Canada Corp. (now FP Innovations) to establish the specified strength in compression parallel to grain. In the absence of test data for columns of lesser volume, the upper limit for KZcg has been changed from the value of 1.3 given in the 1994 edition of the Standard to the current value of 1.0. Although the reliability calculations included a nominal load eccentricity (to for out-ofstraightness) the designer is required to consider bending moments due to eccentrically applied loads when sizing glulam columns. In this case, the provisions of Clause 7.5.12 apply.

Clause 7.5.9 – Compressive resistance perpendicular to grain (bearing) Bearing design values assigned to sawn lumber have been adopted for use in Clause 7 with one exception. The value for Spruce-Pine glulam is about 10% higher than that used for S-P-F sawn lumber. Specified strength values in compression perpendicular to grain for both glulam and sawn lumber were derived from an equation which relates the compressive stress at a deformation of 1 mm to mean ovendry relative density (Clause A6.5.7). To be representative of the lumber grades used in glulam manufacture, a relative density value of 0.44 was chosen for Spruce-Pine glulam (Clause A12.1). Note that this differs slightly from the value of 0.42 used for S-P-F sawn lumber.

To reflect the similar calibration procedures used for sawn lumber, a φ value of 0.8 has been chosen. The equation given in Clause 7.5.9.2 is the same bearing resistance calculation found in previous versions of the Standard. The intent of the clause is to limit deformation at points of bearing. A “size factor” for bearing (KZ) has been added to for the effect of growth ring orientation on the specified strength in compression perpendicular to grain. Specified strength values given in Table 7.3 are based on a worst case orientation of 45°. Lumber loaded perpendicular to the wide faces (i.e. on flat) results in a random growth ring orientation and therefore higher bearing strength. In the case of glulam design, bearing stresses are usually applied perpendicular to the wide faces and KZ will generally be equal to 1.15. Vertically laminated beams may be an exception to this rule as the load is applied perpendicular to the narrow faces of the laminations. However, the growth ring orientation between lamination is likely to be just as random and, by the nature of the product, discontinuous between the glue lines. Although bearing resistance in accordance with Clause 7.5.9.2 has traditionally been considered a serviceability limit state, duration of load adjustments have been retained to compensate for deformations which may occur as a result of the application of long term loads. Clause 7.5.9.3 requires the designer to limit the bearing stress resulting from loads applied near the s to two-thirds of the value permitted for bearing on surfaces (Clause 7.5.9.2). Research has shown that such loads are transferred to the by compression perpendicular to grain stresses as opposed to shear stresses (Lum, 1994). The intent of the clause is therefore to protect against the occurrence of: a) rolling shear failures and b) excessive deformation resulting from the application of compression perpendicular to grain stress over a large volume of wood (see Figure 5). The value of F corresponding to compression face bearing from Table 7.3 must be used with Clause 7.5.9.3 to be consistent with the species combinations specified in CSA O122.

Clause 7.5.11 – Tensile resistance parallel to grain Tensile strength has been calibrated to resistance levels set in 1976. At that time there were indications that there was a volume effect in tension (i.e., reduced tension strength for of greater volume). To compensate, the design value for the gross section was decreased, but the design value for the net section was retained because the volume was too small to warrant an additional reduction beyond the reduction for stress concentration. As the specified resistance for the gross section is less than that for net section, two checks are required for tensile resistance (at net and gross section).

Clause 7.5.12 – Resistance to combined bending and axial load This clause provides interaction equations for deg subjected to combined axial and bending loads. The interaction equations were revised in the 2009 edition to be consistent with the sawn lumber provisions in Clause 6.5.10 (see commentary on Clause 6.5.10). The ratio of compression load to compression resistance is squared in the revised interaction equation, unlike previous versions. Also, the moment amplification factor is defined explicitly in the revised equation as 1/(1-Pf/PE) where PE is the Euler buckling load. This factor is slightly more conservative in most cases than the detailed amplification method that is provided in Clause A.6.5.10 as an alternative approach. The overall effect of the revised interaction equation, including the moment amplification factor, is generally a slight reduction in required glulam sections.

Connection design with glulam Connection design is covered in the section on fastenings but following are some principles that either apply mainly to glulam or deserve emphasis. Any problems with glulam are usually associated with connections. This is the point of interface with other , materials or climates. Connections can create discontinuities in protective coating. They can also be more rigid and unyielding than glulam and conduct heat more readily than the glulam. Deterioration should be avoided by good detailing of drainage, flashing and air barriers to either eliminate or limit with moisture-laden air. The large size of glulam makes shrinkage due to drying greater than the tolerances inherent in many connections. Designers should avoid connections with fastenings widely spaced perpendicular to the grain, especially where ed together with a more rigid steel plate, or a wood member stressed parallel to the grain. Instead, they should create a cluster of connections close together, and as near as possible to a bearing. If connections must be spaced out perpendicular to the grain, separate them so that they may move independently or design them so that the wood can shrink without seriously impairing their function. An example of the first would be three sets of shear plates at the peak of an arch (Figure 6) and of the latter, a vertical tie-down bolt at a beam seat. Most problems that have occurred with glulam have been caused by poor connection fit. The most common is a beam suspended above its seat by the connectors. This condition is initiated by poorly fitting connections and compounded by shrinkage during drying. The resulting loads on the connections cause splitting of the ends of the beam. This problem can be remedied by shimming and/or grouting of the beam seat.

Figure 6 Arch peak connection with three sets of shear plates spaced perpendicular to grain

Bolts and lag screws Where connections are between different materials supplied by different sources, the bolt holes should be drilled on site. Where this cannot be done, the connections and the glulam should both come from the same supplier and tolerance for misalignment should be provided. Lag screws are useful where inserted into deep but they must be installed by turning with a wrench. Lubrication with a liquid soap or other non-petroleum lubricant, proper pilot drilling and power wrenches can aid installation.

Split rings and shear plates are most useful when used in pairs on opposite sides of or in small groups. It is advisable to inspect these connections to ensure that the connectors were, in fact, installed. Specifications should require that shear plates be installed in the factory at the time of fabrication. Split rings and shear plates are very useful where two cut faces are subject to shear such as arch peak connections, or a wind column to a beam. When the shear resistance is in the plane of shear, the cut faces can be held together with a simple tension tie rather than a complex connection resisting the shear eccentrically. Shear plates are also useful in rod bracings to take the component of the force in the plane of the glulam.

Timber rivets Timber rivets, also known as glulam rivets, were developed more than 40 years ago. They have been the subject of continuous research and improvement with more applications under development. Some general cautions with respect to timber rivets: •

If possible, each job should use only one length of glulam rivet to avoid using the wrong length in a particular connection. Rivets must be driven with their wide face parallel to the grain. Care should be taken to ensure that the connection is properly placed, as a completed connection is extremely difficult to remove.

•

When driving timber rivets, safety glasses should be worn to guard against particles of the galvanizing and because the hardened rivet can break if struck wrongly. Rivets should be seated firmly but not hammered flush, which might allow them to pull through the plate. If rivet and holes are of the proper size, the rivet head will protrude from the plate (typically 3.2 mm) and there will be some yielding of the steel plate where the tapered rivet head s the plate (see Figure 12.7.1.1).

•

The effort of driving timber rivets can be eased greatly by the use of pneumatic or electric drivers or rivet guns.

Timber rivets have few fit-up problems, are very efficient and are readily field-inspected. They can be installed with an ordinary hammer and produce a connection with a neat textured patch instead of large bolt heads. They require little or no plant fabrication, thus lowering initial costs, and they greatly reduce site problems.

Truss plates Truss plates are currently not widely used for glulam since their capacity is much below that of these large . The development of innovative designs using these efficient connections may make them more feasible in the future. Nail-on plates are useful for attaching bracing or as tie down plates at columns.

Nails Nails are used in glulam construction to attach smaller to the glulam and, as such, the capacity is limited by the side .

References: ANSI/AFPA NDS 2005. National Design Specification (NDS) for Wood Construction. American Forest & Paper Association, Washington, DC. Barrett, J.D., R.O. Foschi and S.P. Fox. 1975. Perpendicular-to-Grain Strength of Douglas Fir. Canadian Journal of Civil Engineering, Vol. 2, No. 1, pp. 50-57. Foschi, R.O. et al. 1989. Reliability-Based design of Wood Structures. Structural Research series, Report No.34, University of British Columbia, Vancouver, B.C., pp 267-272. Foschi, R.O. 1993. Design recommendations for Timber Bridges, Report No.1. Report to CSA-S6 Committee. Foschi, R.O. and J.D. Barrett. 1975. Longitudinal Shear Strength of Douglas Fir. Canadian Journal of Civil Engineering, Vol. 3, No. 2, pp. 198-208. Foschi, R.O. and S.P. Fox. 1970. Radial Stresses in Curved Timber Beams. Journal of the Structural Division, Proceedings of American Society of Civil Engineers, Vol. 96, No. ST10. Hooley, R.F. and B. Madsen. 1964. Lateral Stability of Glued Laminated Beams. Journal of the Structural Division, Proceedings of American Society of Civil Engineers, Vol. 90, No. ST3, pp. 201-218. Johns, K.C. 1991. A Continuous Design Curve for Timber Columns. Canadian Journal of Civil Engineering. Vol. 18, pp. 617-623. Karacabeyli, E. 1992. Non-Residential Applications for Glulam. Report to Forestry Canada, Forintek Canada Corp. (now FP Innovations) Project No. 1510K014, Vancouver, BC. Lam,F and N Mohadevan, 2007. Development of New Constructions of Glulam Beams in Canada. International Council for Research and Innovation in Building and Construction, Working Commission W18-Timber. Lam, F. 2010. Size effect of bending strength in glulam beams. In: Proceedings of the CIB-W18 Meeting 43 Nelson New Zealand. Lum, C. 1994. Rationalizing Compression Perpendicular-To-Grain Design. Report to Forestry Canada, No. 13. Forintek Canada Corp. (now FP Innovations), Project No. 1510K018, Vancouver, BC. Wilson T.R.C. 1939. The Glued Laminated Wooden Arch. U.S. Department of Agriculture Technical Bulletin No. 691, pp. 118-119. Lam, F. 2013. Glulam Size Effect Adjustment Procedures for bending strength – impact study. Smith, I., Kiwelu, H.M., Weckendorf, J. 2015. Explanation of tension-side notch design provisions for glulam in CSA Standard O8614. Canadian Journal of Civil Engineering, 42(10): 787-796.

Clause 9

Structural s

Clause 9.1 – Scope The scope of this section includes Construction Sheathing OSB, unsanded Douglas Fir Plywood and unsanded Canadian Softwood Plywood. Information on the design of composite building components using either plywood or OSB structural s in combination with lumber or glued-laminated timber is found in Clause 10 of the Standard.

Clause 9.2 – Materials CSA O86 provides strength and stiffness values for three classes of structural s: unsanded plywood manufactured in accordance with CSA Standard O121 Douglas Fir Plywood (DFP); unsanded plywood manufactured in accordance with CSA Standard O151 Canadian Softwood Plywood (CSP); and Construction Sheathing OSB manufactured in accordance with CSA Standard O325, Construction Sheathing.

Clause 9.2.1 – Plywood The plywood product standards recognize two types of lay-ups: Standard Construction and Modified Construction. The standards also distinguish between sanded and unsanded grades. thickness, species requirements, minimum number of plies and limits for the thickness of individual plies are prescribed for both sanded and unsanded grades. CSA O86 provides strength values only for unsanded grades of Standard Construction. For Douglas Fir Plywood, CSA O121 requires that the faces and backs of a are Douglas fir. Veneer for inner plies can be any one of 21 listed species, including Douglas fir, western hemlock, and most spruce, pine and fir species in Canada. Most species that are allowed as inner plies for DFP are also allowed by CSA O151 as face or back plies for CSP. Balsam poplar, trembling aspen and cottonwood, three hardwood species, are restricted to use as inner plies in DFP and CSP. Revisions to CSA O121 in 2008 added combinations and changed the combinations for existing Standard Construction, leading to some reduction in section properties and lower tabulated capacities for unsanded Douglas Fir in CSA O86. Revisions were made to Canadian Softwood Plywood based on revisions to the CSA O151 Standard in 2009. Both standards define three grades of veneer (A, B and C). Grade A represents a high-quality surface and restricts any type of open defect in Douglas fir veneer to pin knots not more than 5 mm in diameter. There are also restrictions on the use of filler for splits, and the type of split and patches. These restrictions are relaxed for a B-grade veneer. C-grade veneer permits the presence of certain sizes of knots and knot-holes, which can be up to 50 mm in size measured across the grain. The strength values published in CSA O86 are for Sheathing Grade s based on lay-ups containing only C-grade veneers. These strength values can also be used safely for plywood grades of higher quality. Since the effectiveness of plywood can only be maintained through long-lasting bonding of veneers, CSA O121 and CSA O151 require that test specimens be subjected to moisture conditioning cycles, then tested in shear. If all or most of the shear failure takes place in the wood, the plywood meets the bond requirements.

The product standards require that s be marked in a specific manner. All plywood s conforming to either CSA O121 or CSA O151 are marked legibly and durably to designate the manufacturer, the bond type (EXTERIOR), the species (DFP) or (CSP), the standard (CSA O121 or CSA O151) and the grade. Unsanded plywood of Standard Construction will show the full name or an abbreviated grade mark (shown in Table 2).

Table 2 Grade marking for standard constructions of unsanded plywood

Abbreviated grade marks CSA O121

CSA O151

Douglas Fir

Canadian Softwood

Grade name

Plywood (DFP)

Plywood (CSP)

Select Tight Face

SELECT TIGHT FACE

–

SELECT TF

–

SEL TF

–

Select

SELECT

SELECT

Sheathing

SHEATHING

SHEATHING

SHG

SHG

Clause 9.2.2 – Construction sheathing OSB The 2001 edition of the Standard added provisions for design values and procedures for OSB s certified to meet the requirements of CSA Standard O325, Construction Sheathing. CSA O325 is a performance standard for wood-based s such as plywood, or OSB or waferboard, which are qualified for use in typical floor, roof and wall sheathing applications. This standard was first published in 1988 and has been referenced in Part 9 of the NBC since 1990. s are stamped with marks denoting the span rating and an appropriate end use, as opposed to a nominal thickness stamped on plywood or other OSB s, such as those certified to CSA Standard O437. There is a relationship between a given mark and its corresponding predominant nominal thickness, although other thicknesses can be produced for a given mark, as explained in the Appendix. Besides the mark, the Building Code also requires that the grade stamp indicates that the s are suitable for exterior exposure.

Clause 9.3 – Specified strength capacities The strength and stiffness of structural s are given in of capacities for specific thicknesses rather than in of unit stresses. This obviates the need for publishing section properties for s and the designer is able to identify the thickness required. Strength and stiffness are given in of capacities per unit width of , with the exception of planar shear which is given in MPa.

Clause 9.3.1 – Plywood The specified capacities for nominal plywood thicknesses are based on the “worst case” allowed by the manufacturing standard. A worst case is assumed to have the weakest species for each property, the combination of veneer thicknesses resulting in the lowest possible section property, and a near minimum thickness. The product standards O121 and O151 were revised to provide more and altered layup combinations. Among the changes were reductions in minimum ply thicknesses and tolerances, and inclusion of asymmetrical lay-ups. Plywood capacities in CSA O86 were decreased in 2009 to reflect the effect of these changes on section properties of Standard s. Capacities for plywood thicknesses ranging from 22.5 to 31.5 mm were added in 2001. These thicker plywood s provide more options for combinations of plies. As a consequence there is a greater range of strength and stiffness capacities for these thicker s that are available to the designer. The specified strength capacities for unsanded DFP and CSP were based on the results of in-grade tests of plywood by the Council of Forest Industries of British Columbia (now CertiWood Technical Centre). (Parasin, A.V. et al 1985; Smith, G.R. 1974). These tests represented the strength of over 1700 s. Tests were carried out to determine the ultimate strength of plywood in bending, tension, compression, shear-through-thickness, planar shear and bearing. Modulus of elasticity and modulus of rigidity were also determined for all properties, except for planar shear and bearing. For each species, thickness, and orientation, the data were reduced to unit strengths for individual plies stressed in a direction parallel to the grain. These unit strengths were then combined for the various thicknesses, species and orientations of commercial s. The fifth percentile exclusion values were calculated as a point estimate using the normal distribution assumption [x05 = xmean (1 - 1.645 COV)], where xmean equals the mean of the observed strength values. Based on an examination of full data distributions from various sources, the following nominal coefficients of variation (COV), (shown in Table 3), were used for the calculation of the fifth percentile exclusion value x05 of plies.

Table 3 Nominal coefficients of variation COV

Strength property

COV (%)

Modulus of Rupture, 3-, 4-plies 0° DFP

25

Modulus of Rupture, all other lay-ups and species

20

Tension, single ply

40

Tension, multiple plies

30

Compression

20

Shear-Through-Thickness

15

Planar Shear

20

The resulting fifth percentile values were modified for load duration by a factor of 0.8 to establish the specified strengths. The effect on strength of combining various species in one plywood lay-up was considered by Stieda (1974).

Studies by Wilson (1976) indicated that the strength of small, clear plywood bending specimens was 50 to 60% greater than that of large-size, in-grade bending specimens. Later studies by Bier (1984b) indicated that plywood manufactured from clear veneer differed little in strength regardless of specimen size. It must therefore be concluded that the size effect for bending strength observed by Wilson was primarily caused by the presence of defects in large-size s.

Since the plywood strength values in the Standard were based on the results of tests with ingrade material, these strength values are characteristic for large specimen sizes. Wilson (1977) observed that 300 mm wide in-grade bending specimens were approximately 11% lower in strength than 900 mm wide specimens. This observation can probably be explained by the fact that grading rules for knots in plywood are not related to the size of the plywood. A knot with a given diameter therefore will have a greater effect on a narrow strip of plywood than on a fullwidth . Clause 9.3.2 – Construction sheathing OSB Specified capacities for the strength and stiffness of sheathing and single floor grades of Construction Sheathing OSB are given in Table 9.3C. Minimum nominal thicknesses range from 9.5 to 28.5 mm. For each nominal thickness, specified strength and stiffness capacities are given for two directions. The 0˚ direction in Table 9.3C refers to the direction of the alignment of the strands in the face layers or major axis of a . For a 1220 by 2440 mm , this usually will be in the longer dimension. The 90˚ orientation then refers to the direction of the shorter dimension, or minor axis. Two different sets of values are given for s marked for sheathing or single floor applications. Sheathing grade s are often identified with a combination of roof and floor end uses. Shear-through thickness strengths and capacities are identical in the 0˚ and 90˚ directions. s certified to CSA O325 were sampled from a national selection of Canadian OSB mills, nondestructively evaluated for bending stiffness at the Alberta Research Council and ranked to determine the lower half of the sample test s. These identified samples were then tested by Forintek Canada Corp. (now FP Innovations) and the Alberta Research Council (now Alberta Innovates – Technology Futures) in accordance with accepted CSA and ASTM test methods for engineering strength and stiffness properties in bending, axial tension and compression, shear in-plane and through-thickness. Test results were then adjusted to climatic conditions of 20 ˚C and 80% RH, followed by reliability normalization procedures. These adjusted values were then compared to the OSB design values published in the U.S. by APA, the Engineered Wood Association in their publication TN375B, after a soft conversion to limit states design format and appropriate moisture conditions. With the exception of axial stiffness values, the tabulated values for Construction Sheathing can be traced back to the APA values. For axial stiffness values, the test data were less than the values published by APA. This changed in 2009 when axial stiffness values for Construction Sheathing OSB were increased, based on a review of data including APA mill tests of EA perpendicular (90°) in representative OSB s. The Appendix to CSA O86 contains additional material about the marks and equivalent thicknesses for s conforming to CSA O325. Results of the test program are summarized in confidential Structural Board Association reports by Forintek Canada Corp. (now FP Innovations) (Karacabeyli and Lum, 1998 and 1999) and the Alberta Research Council (Wasylciw, 1998)

Clause 9.4 – Modification factors Strength and stiffness properties of structural s, like those of all other wood-based building products, are affected by the length of time a load is applied to a structural component, moisture content and chemical treatment. Other factors may also apply, depending on the application.

Clause 9.4.1 – Load duration factor, KD Experimental studies on the effect of constant loading on the bending strength of waferboard were carried out at Forintek (now FPInnovations) (Palka et al, 1990) and on plywood at the US Forest Products Laboratory (Laufenberg et al, 1994). OSB has evolved from the waferboard product that was manufactured in the eighties and tested in the studies reported by Palka. In the absence of additional data on OSB, a conservative approach was to base duration of load factors on the waferboard research. The results on waferboard were analyzed by Foschi (1993). He concluded that for dry conditions the same duration of load factors that are used for lumber and plywood (Table 5.3.2.2 of the Standard) could also be used for waferboard and OSB, with one exception. The duration of load tests indicated that the waferboard specimens were sensitive to intermittent high temperatures or high humidity conditions (Palka, 1990), even though they were protected from direct exposure to moisture. Based on these results for waferboard, the Standard prescribes a reduced duration of load factor for long term loading on OSB. For loads in excess of 50% of the OSB capacity and where the s are subjected to intermittent high temperatures of up to 65°C and high humidity conditions, the duration of load factor for long term loading has been set at 0.45, instead of the 0.65 used for other wood products. To clarify how such climatic conditions affect OSB, the Structural Board Association commissioned a creep-rupture study, which was carried out at a number of sites in North America. (Tichy, 1997) In a separate study conducted by Forintek Canada Corp. (now FP Innovations) for the SBA, it was also established that the load duration factors in CSA O86 were applicable to CSA O325 OSB s in dry service conditions (Karacabeyli, 1998).

Clause 9.4.2 – Service condition factor, KS The strength of wood is affected by moisture content. There have been a number of studies on the relationship between moisture content and strength of Douglas Fir Plywood using full-size specimens. The results of these studies have been compiled and analyzed by Palka (1982). In general, the strength of plywood increases as moisture content decreases from 30% to a maximum strength at about 10%. For compression strengths, the maximum occurs at very low moisture content, and for shear-throughthickness and stiffness, the maximum was observed at approximately 5% moisture content. Detailed studies of the effect of wet service conditions on the strength and stiffness of OSB have not yet been carried out. The Standard’s provisions therefore apply only to OSB used in dry service conditions. Plywood strength capacities published in the Standard were based for the most part on strength tests carried out at 15% moisture content and the published values are intended for use at that moisture content. Field observations indicate that, for a given temperature and relative humidity, the equilibrium moisture content of plywood is lower than that of lumber. For example, at 20˚C dry-bulb temperature and 80% relative humidity, the equilibrium moisture content of lumber will be about 16% whereas that of plywood will be about 14%. For wet service conditions (moisture content greater than or equal to 30%) the Standard provides two reduction factors: 0.80 for specified strengths and 0.85 for specified stiffness and rigidity.

Clause 9.4.3 – Treatment factor, KT Plywood can be chemically treated to improve resistance to decay or to fire. Preservative treatment must be done by a pressure process, in accordance with CSA O80 standards. There were a few early studies of the effect of preservative treatments on the strength of plywood. Wise (1952) reported that creosote does not weaken wood. Countryman (1957) studied Douglas Fir Plywood and concluded that treatment with regular Wolmann salts affected strength properties only slightly. Also, field experience with known preservative treatments appeared to be satisfactory. A treatment factor of KT = 1.0 has therefore been assigned to preservative treated plywood. On the other hand, some fire-retardant treatments used in North America have resulted in significant reductions in the strength of plywood, particularly in locations having elevated temperatures (LeVan, 1989). There have been studies of the strength-reducing effect of fire retardants (King, 1961), but with changes in the chemical composition of fire retardants, these studies are no longer applicable. CSA O86 requires that tests be carried out on the effects of fire retardants, or any other potentially strengthreducing chemicals used with structural s, to determine an appropriate treatment factor. Standards on “Evaluating the Mechanical Properties of Fire-Retardant Treated Softwood Plywood Exposed to Elevated Temperatures” (ASTM D 5516), and for “Calculating Bending Strength Design Adjustment Factors for Fire-Retardant Treated Plywood Sheathing” (ASTM D 6305) have been developed by ASTM (See also commentary on Clause 6.4.3). There have not yet been reported studies on the effects of chemical treatments on structural properties of OSB.

Clause 9.4.4 – Stress t factor, XJ The stress-t factor XJ is intended for structural s used in I-beams, box beams or stressed-skin s. The factor therefore is also referenced in Clause 10.3 of CSA O86. Only scarf-ts manufactured with resorcinol or phenol-resorcinol adhesives meeting the requirements of CSA O112.7 and having scarf slopes ranging from 1:12 to 1:4 are included in these provisions. Since it is conceivable that a designer might want to use scarf-ed s for uses other than Ibeams, box-beams or stressed skin s, Tables 9.4.4.1 and 9.4.4.2 are included in the general provisions of Chapter 9. Research work at the University of New Brunswick developed information that led to the derivation of appropriate values for stress t factors for glued OSB scarf and butt ts. Table 9.4.4.1 contains values for OSB scarf ts, and Table 9.4.4.2 contains values for OSB to OSB and OSB to lumber splice ts (Chui and Ni 2001).

Clause 9.4.5 – Factor KF for preserved wood foundations The end use factor KF for preserved wood foundations applies only to bending strength, and to the planar shear resistance associated with bending. This factor calibrates to the field performance of plywood in wall systems of preserved wood foundations, and is restricted to plywood spans of not more than 815 mm. Normally, plywood s are ed at spacings of not more than 410 mm for these uses. Provisions are included only for plywood, as this is currently not an application suitable for OSB.

Clause 9.5 – Resistance of structural s Structural reliability of plywood was analyzed and reported by the Council of Forest Industries of British Columbia, (CertiWood Technical Centre) (Parasin, 1988), and of design-rated OSB by Foschi (1992), reported by the Structural Board Association. Reliability considerations were included as well in the testing program on Construction Sheathing OSB (Karacabeyli, 1999). The basic reliability approach described in the Commentary on Clause 3 was used for s. Distributions of strength and stiffness values were examined and it was determined that a normal distribution would best represent the available plywood data, whereas for OSB a two-parameter Weibull distribution was chosen. The available plywood strength data had a coefficient of variability (COV) ranging from 15 to 20%, based on the full distribution. For OSB, the target COV was also about 15%. The reliability index β for a given loading condition is a function of the probability distribution for that load condition, the type of strength distribution for the structural , the coefficient of variation for that distribution and the resistance factor φ. The resistance factor φ chosen for structural s in CSA O86 is 0.95 for all properties except one. The exception is plywood of 3- and 4-ply construction, where stressed in tension perpendicular to the orientation of the face grain. For this particular case, the Standard requires a φ of 0.6. The specified capacities for structural s given in CSA O86, when multiplied by φ, result in an average reliability index of 2.6 to 2.8. Foschi (1992) has shown that, while the reliability index β for specific loading conditions is slightly different for plywood than for design-rated OSB, the average of these reliability coefficients for residential, office and snow loads across Canada are the same. Lum (1999) reported that previous analyses have shown that the use of the Vancouver snow load consistently provided a lower or more conservative estimate of the structural reliability, in evaluating the lower tail of the strength distribution with the limited test data available for construction sheathing OSB. Part 4 of the National Building Code specifies that the minimum specified concentrated roof live load on a roof area is 1.3 kN. In the 2010 edition of the National Building Code, the corresponding concentrated loaded area on the roof was reduced from 750 mm x 750 mm to 200 mm x 200 mm. CSA Standard O325 “Construction sheathing”, provides performance requirements for structural wood s that include minimum concentrated static loads on rated roof sheathing. The minimum required concentrated static loads on rated roof sheathing result in stresses that are approximately eight times higher than the National Building Code stress for concentrated roof live loads. The National Building Code concentrated roof live loads will not govern the design of roof sheathing where roof sheathing thickness is properly installed and meets the requirements of the “Roof Sheathing Selection Tables” in the Wood Design Manual. For specific point load designs or non-uniform loading for roof or floor sheathing, design should be in accordance with Clause 9 of the Standard.

Stiffness Elastic properties of structural s in the Standard are given in of stiffness (EI) for a cross section based on a 1 mm wide strip. For plywood in bending, this stiffness is equal to the product of the modulus of elasticity (E) of the individual plies parallel to the fibre and the second moment of area (I) of all effective plies parallel to the fibre. The 1984 allowable stress version of CSA O86 (now withdrawn) gave the modulus of elasticity and the second moment of area for the various plywood thicknesses and constructions as separate values. The limit states design version of the Standard gives the bending stiffness in the two directions, for each plywood thickness and construction.

For OSB the bending stiffness is defined as the product of the applied bending moment and the corresponding radius. Similarly the Standard provides a stiffness for axial loading, which consists of the product of modulus of elasticity and the effective cross-sectional area A. For plywood, the effective cross-sectional area is based on the plies oriented in the same direction as the applied forces. For OSB, it is the total cross section. For shear displacements due to shear-through-thickness stresses, the Standard provides a rigidity value Bv, which is calculated as the product of the modulus of rigidity for shear-through-thickness and the total thickness of the . The units of the tabulated bending stiffness are N·mm2/mm; for axial and shear-through-thickness stiffness, the units are N/mm. Studies carried out at COFI (Parasin, 1983) indicated that for plywood the modulus of rigidity for planar shear is about 1 to 2% of that for shear-through-thickness. A rough estimate of this modulus of rigidity for plywood can be obtained by dividing the tabulated shear-through-thickness rigidity by the corresponding thickness and using 1% of the resulting modulus of rigidity. No serviceability design equations for structural s are provided in CSA O86. However, deflections and other deformations can be calculated by the usual engineering equations together using the mean stiffness values given in Tables 7.3A to C. For serviceability limit states (deflections and shear deformations) a resistance factor of 1.0 tends to result in a reliability index of about 2.0.

Clause 9.5.1 – Stress orientation Both plywood and OSB are orthotropic materials, i.e. with different strength and stiffness properties in the two principal directions. For plywood, these differences are the result of placing veneers during the manufacture of the s in such a way that the fibre orientation in at least one of the veneers is at right angles to that in an adjacent veneer. For OSB, the surface layer strands are aligned in the same direction as the , whereas inner layer strands are aligned at right angles or randomly. As a result of this orientation of veneers or strands, two distinct strength axes can be identified in these s. The orientation with the greater unit stiffness and unit strength in bending has been designated as the major axis of the structural , that with the lesser stiffness and strength as the minor axis. The strength values given in CSA O86 are for directions parallel and perpendicular to the major strength axis of the structural s. The Standard does not provide information on the strength of s in any other directions. Kollmann (1951) reported results of plywood tensile strength tests at intermediate angles to the face plies. According to the results of these tests on birch plywood, the minimum strength occurred where the face grain orientation was approximately at 45° to the direction of the applied tensile stresses. These results were confirmed by Bier (1984a) for softwood plywood. Similar results would be expected if OSB were to be tested in directions at an angle to the major axis.

Clause 9.5.4 – Planar shear For planar shear (or rolling shear, as it is sometimes called) the Standard provides strength values for two cases: bending and direct shear stress. The first case results from transverse shear forces in a that is subject to bending (Stieda, 1983). For this case the Standard provides a transverse shear capacity vpb for each thickness in N/mm of width.

The second case arises where s are subject to in-plane forces (where forces are transferred from one member to another, such as through a glued gusset plate). To determine the minimum plate size the designer needs to know the shear capacity vpf for such a gusset plate. The critical plies for planar shear stresses are always oriented at 90° to the in-plane shear stresses. The unit strength in planar shear is dependent on the thickness of plies perpendicular to the stress. For plywood lay-ups where the ply thickness at the critical location can equal or exceed 4.2 mm, the unit strength is 0.55 MPa. If the lay-up only allows plies less than 4.2 mm in thickness, tests show that a higher planar shear strength can be used. For plywood, the Standard provides separate planar shear strength values in the two principal directions, and the designer also has to consider the number of plies for the plywood thickness to be used. The specified planar shear strength for Douglas Fir Plywood (DFP) and Canadian Softwood Plywood (CSP) are identical, since most species that can be used in the core of a DFP are also permitted in the core of a CFP . For construction sheathing OSB, in-plane shear values vary slightly from one mark to another as well as for the two principal directions. OSB tables also have values for bending planar shear capacities.

Clause 9.5.8 – Compression perpendicular to grain The tabulated values for compression perpendicular to grain were updated to reflect decisions made during the previous code cycle and were revised for both plywood and OSB. (Tables 9.3A, B, C).

References: APA, the Engineered Wood Association, 1995, Design Capacities of APA Performance Rated Structural-Use s, Technical Note N375B Bier, H. 1984a. Strength Properties of Pinus Radiata Plywood at Angles to Face Grain. New Zealand Journal of Forestry Science, 14(3):349-367. Bier, H. 1984b. Pinus Radiata Plywood: Influence of Width and Loading Method on Bending Properties. New Zealand Journal of Forestry Science. 14(3):400-403. Chui, Y. H. and Ni, C. 2001. Influence of splice length on strengths of glued spliced OSB ts in tension. Forest Products Journal. 51(2):23-28 Countryman, D. 1957. Effect on Plywood Strength of Preservative Treatment. Douglas-Fir Plywood Association Laboratory Report 74. Foschi, R.O. 1993. Duration of Load Factors for OSB. Report prepared for Structural Board Association, Willowdale, ON. Foschi, R.O. 1992. Reliability-Based Performance Factors for OSB. Report prepared for Structural Board Association, Willowdale, ON. Karacabeyli, E, 1998, CSA O325 OSB Design Values: Long Term Performance, Forintek Canada Corp. (now FP Innovations). Confidential report prepared for the Structural Board Association Karacabeyli, E. and Lum, C., 1999, CSA O325 OSB Design Values: Strength and Stiffness Capacities (Final Report), Forintek Canada Corp. (now FP Innovations). Confidential report prepared for the Structural Board Association Karacabeyli, E. and W. Deacon. 1993. Summary of Short-Term Test Results for Waferboard and Oriented Strandboard. Report prepared for Structural Board Association. Forintek Canada Corp. (now FP Innovations), Vancouver, BC. King, E.G., Jr. and D.A. Matteson, Jr. 1961. Effect of Fire-Retardant Treatment on the Strength of Plywood. American Plywood Association Laboratory Report 90. Kollmann, F. 1951. Technologie des Holzes und der Holzwerkstoft. Springer Verlag, Berlin, p.692. Laufenberg, T., L.C. Palka and J.D. McNatt. 1994. In Consolidation of Information of Strength and Creep of Oriented Strandboard (OSB) s. Report prepared for Canadian Forest Service. Forintek Canada Corp. (now FP Innovations)Canada Corp,Vancouver, BC. LeVan, S. and M. Collet. 1989. Choosing and Applying Fire-Retardant-Treated Plywood and Lumber for Roof Designs. US Forest Service, FPL General Technical Report, FPL-GTR-62.

Lucuik, M. 1992. Engineering Properties of OSB: Phase-2, Verification of Draft Industry Standard. Report prepared for the Structural Board Association. Forintek Canada Corp. (now FP Innovations), Ottawa, ON Lum, C. and Karacabeyli, E., 1998, CSA O325 Design Values: Summary Report for the Engineering Properties Tests, Forintek Canada Corp. (now FP Innovations). Confidential report prepared for the Structural Board Association Lum, C., 1998, Planar and Through-the-thickness Shear tests on Oriented Strand Board (Data Report), Forintek Canada Corp. (now FP Innovations). Confidential report prepared for the Structural Board Association Palka, L.C. and B. Rovner. 1990. Long-Term Strength of Canadian Commercial Waferboard 5/8-inch s in Bending. Report prepared for Forestry Canada. Forintek Canada Corp. (now FP Innovations),Vancouver, BC. Palka, L.C. 1982. A Proposed Model of the Effect of Moisture Content on the Mechanical Properties of Sheathing-Grade Douglas-Fir Plywood in Structural Use of Woods in Adverse Environments, edited by R.W. Meyer and R.M. Kellogg, Van Nostraad, New York, pp. 100-116. Parasin, A.V. 1983. Strength Properties of 9.5 mm 3-Ply and 15.5 mm 5-Ply Western White Spruce Sheathing Grade Plywood. Council of Forest Industries of British Columbia (now CertiWood Technical Centre), Report 130. Vancouver, BC. Parasin, A.V. and C.K.A. Stieda. 1985. Recommendations for Allowable Stresses for Canadian Softwood Plywood. Council of Forest Industries of British Columbia (now CertiWood Technical Centre), Report 131. Vancouver, BC. Parasin, A.V. 1988. Structural Reliability Analysis of Plywood. Council of Forest Industries of British Columbia (now CertiWood Technical Centre), Report 144. Vancouver, BC. Smith, G.R. 1974. The Derivation of Allowable Unit Stresses for Unsanded Grades of Douglas Fir Plywood. Council of Forest Industries of British Columbia (now CertiWood Technical Centre), Report 105. Vancouver, BC. Stieda, C.K.A. 1974. Bending Strength and Stiffness of Multiple Species Plywood. Proceedings CIB W18, Paper 3-4-2. Stieda, C.K.A. 1983. Planar Shear Capacity of Plywood in Bending. Proceedings CIB W18, Paper 16-4-1. Stieda, C.K.A. and A.V. Parasin. 1985. Recommendations for Allowable Stresses for Douglas Fir Plywood. Council of Forest Industries of British Columbia (now CertiWood Technical Centre),Report 134. Vancouver, BC. Tichy, R., 1997, Creep and Creep Rupture of OSB, TMI, Inc., Confidential report prepared for the Structural Board Association Wasylciw, W. and Mirbach, C., 1998, SBA Mill Testing: Phase I and Addendum, Alberta Research Council, Confidential report prepared for the Structural Board Association Wasilciw, W. and Mirbach, C., 1998, SBA Mill Testing: Phase 2, Alberta Research Council, Confidential report prepared for the Structural Board Association Wilson, C.R. 1976. Comparison of the Size and Type of Specimen and Type of Test on Plywood Bending Strength and Stiffness. Proceedings CIB W18, Paper 6-4-2. Wilson, C.R. and A.V. Parasin. 1977. Comparison of the Effect of Specimen Size on the Flexural Properties of Plywood Using the Pure Moment Test. Proceedings CIB W18, Paper 7-4-4. Wise, L.E. and E.C. Jahn. 1952. Wood Chemistry, Vol. 1, 2nd Edition. American Chemical Society Monograph. Reinhold Publishing Corp. New York.

Clause 10

Composite Building Components

Clause 10.1 – Scope Lumber and materials can be ed together to form new building components. Such components can be constructed either by nailing or by gluing. This Clause covers only glued composite components that are not manufactured by a proprietary process. Proprietary structural wood products are covered by Clause 15 and 16 of this Standard. Fabrication specifications should stipulate the method by which pressure should be applied to hold the lumber and material in place until the adhesive has set. This is often done by using nails or screws. Any shear resistance that might be developed by the nails or screws cannot be considered in the design, since the glue t is relatively rigid. Information on the strength of nailed-only composite building components is available from such organizations as CertiWood Technical Centre. The Standard provides for design of two specific types of composite building components; glued web beams and stressed skin s. Since the design equations are largely based on engineering mechanics, these equations can be applied equally to building components containing plywood or OSB webs or face skins. This edition of the Standard was modified to allow the use of OSB as sheathing material in composite building components covered by this clause.

Clause 10.2 – Materials The components, which can be lumber and/or glulam, plywood and/or OSB, must be glued using either resorcinol or a phenol-resorcinol resin meeting requirements set out in CSA O112.7 or adhesives conforming to CSA Standard O112.9. CSA O112.7 has been used to evaluate phenol resorcinol formaldehyde (PRF) and resorcinol formaldehyde (RF) adhesives for many years for wood assemblies used in severe moisture conditions. Advances in adhesive technology to deal with higher production rates and market requirements, such as appearance, emissions and environmental concerns, has resulted in a wider range of innovative adhesives that target the same service conditions as PRF or RF adhesives. CSA O112.9 is a performance-based standard for adhesives used for structural wood products exposed to exterior exposure and is independent of the adhesive chemistry.

Clause 10.3 – Stress-t factor XJ To provide continuity in standard size plywood or OSB s used in composite components, it becomes necessary to end- s for constructions more than 3.05 m long, and often for s more than 2.44 m long. However OSB s can be specified in long lengths up to 7.3 m by special order. Clause 10 has specific requirements for the construction of such end-ts between structural s. Stress-t factors (XJ) are provided to modify the strength capacities of s. Since scarf-ts could occasionally be used for applications other than stressed-skin s or box- and I-beams, the stresst factor is found in Clause 9 of the Standard.

Clause 10.5 – Plywood and OSB web beams

Clause 10.5.1 – General Beams having an I- or a box-section can readily be constructed with plywood or OSB and lumber. Where the meets the lumber, the must be glued to the lumber over its whole length. Such composite building components have been used successfully for many years in Canada and other countries. More recently prefabricated wood I-joists have become common, consisting of a web (plywood, or most frequently oriented strandboard), flanges (lumber or laminated veneer lumber), and proprietary ts in the webs and between the web and flanges. These products are generally used in repetitive member applications, such as joists or rafters. These products are covered under Clause 15 and 16 of CSA O86, and not under Clause 10.

Clause 10.5.2 – Effective stiffness The design of the box- or I-beam is based on the assumption that cross sections that are plane before bending remain that way as bending takes place, so that strains at any given cross section are directly proportional to their distance from the neutral axis of the beam. The design also assumes elastic behaviour, so that stresses at any given distance from the centroid of the cross section are proportional to the modulus of elasticity of the materials present at that location. In the case of asymmetrical cross sections, the stiffnesses of the various parts of the composite can affect the location of the neutral axis. Calculations of bending moment resistance, shear resistance and deflection are simplified by the concept of effective stiffness of the beam (EI)e. The effective stiffness is expressed in of the axial stiffness Ba given in Tables 9.3A to 9.3D of the Standard, and the flange stiffness. The flange stiffness, of course, can readily be calculated from the dimensions and the elastic properties of the lumber flanges.

Clauses 10.5.3 to 10.5.5 – Bending and shear resistance As in previous issues of CSA O86, the factored bending moment resistance is based on either the compression or tension capacity of the section. Web shear-through-thickness and flange-web shear are calculated using the effective stiffness. In the case of flange-web shear, it is assumed that the strength of the glue bond between the lumber flange and the web exceeds that of the wood itself. Therefore, the limiting capacity is either the planar shear strength of the or the shear strength of the wood. The resistance factor for composite beams is based on those resistance factors appropriate for the various materials used in the beams. Where three or more beams meeting the requirements of this Clause are used parallel to each other and spaced not more than 610 mm apart to the applied loads, a systems factor KH of 1.10 applies (Case 1). It must be noted, however, that the systems factor KH introduced in Clause 10.5 can only be used to modify the specified compressive and tensile strength of the lumber flanges of box- or I-beams.

Clause 10.5.6 – Deflection The relatively thin webs of box- and I-beams generate shear stresses that are greater than those in lumber or glued laminated wood beams of rectangular cross section and comparable height. These stresses result in shear deformation of the web and, as a consequence, in deflection (Newlin, 1924; Cziesielski, 1965; Stieda, 1967; Booth, 1977). This shear deflection must be added to deflection caused by the applied bending moment.

Deflection due to bending moment is calculated in the usual manner using the effective stiffness (EI)e of the beam. Shear deflection is based on specified stiffness, rigidity and a shear coefficient XS. This shear coefficient is given for a range of flange-depth to beam-depth ratios and web-thickness to beam-width ratios and is reproduced from Payne (1969). Until 1984 the Standard only provided for calculation of shear deflections for three load cases. Since that time, the equation for this calculation has been generalized to include any conceivable load case for simply ed beams.

Clause 10.5.7 – Lateral Stability of web beams The expressions for the bending resistance of composite beams given in CSA O86 are based on the assumption that the beam will only deflect in the plane defined by the applied loads. For certain crosssectional dimensions, however, lateral deflections and rotations can take place. In the case of box- and I-sections, the various combinations of web and flange dimensions make it impractical to reduce the results of the theory of elasticity to simple design rules. However, the Standard does provide some basic rules for the degree of lateral required, depending on the ratios of moments of inertia about two axes. If these rules are not applied, the designer must carry out a detailed analysis. Stieda (1968) has formulated a finite element solution for such an investigation. The lateral stiffness required for the analysis of lateral stability can be calculated readily for I- and box-beams. For plywood box sections, numerical values of the torsional rigidity for a range of cross-sectional dimensions have been calculated using a finite element approach (Stieda, 1969).

Clause 10.6 – Stressed-skin s

Clause 10.6.1 – General Floor and roof constructions consisting of longitudinal ribs of lumber with plywood or OSB glued to the top and bottom of the lumber are referred to as stressed-skin s. These s usually are 1.22 m wide, the width of the sheet. Ribs commonly are of standard lumber sizes and can be spaced to suit the requirements of the design. Ribs and flanges form an integral part of the cross section. The engineering theory of bending assumes that plane cross sections remain plane during bending. This is not the case with stressed-skin s. During bending the between the ribs will deform. As a result, at some distance away from the ribs, the will be stressed less than over the longitudinal ribs. This phenomenon is known as shear lag and was investigated in detail by Foschi (1969b). At first, shear lag was considered in design by calculating an effective width as a function of plywood thickness. Foschi (1970) showed that the ratio of rib height to rib spacing, and rib height to length, had a far greater effect on shear lag than did plywood thickness. The 1976 edition of CSA O86 therefore introduced a geometry modification factor for bending. More detailed methods of design of stressed-skin s have also been proposed by Mazur (1968), Kuenzi (1976) and Smith (1979), and reviewed by Booth (1976). Three of the modification factors given in Clause 8 are not material specific. They are geometric reductions for geometry, shear modification and shear section, and apply to plywood and OSB alike.

Clause 10.6.2 – Effective stiffness Effective stiffness is calculated to for asymmetry in the construction.

Clauses 10.6.3.1 to 10.6.3.7 – Bending and shear resistance To determine moment carrying capacity of a stressed-skin , three strengths have to be considered separately: that of the tension flange, the compression flange and the lumber web. The capacity is calculated in the usual manner and then multiplied by the geometry reduction factor XG to for the effect of shear lag. The compression flange also has to be checked to ensure that no buckling will take place. This phenomenon has been studied by Foschi (1969a). For a uniformly distributed load, the Appendix includes a buckling equation based on Foschi’s work. Since there was no information available regarding a buckling coefficient for OSB, Clause10.6.3.4 was rewritten to avoid buckling of the compression flange, while the existing clause was moved to the Appendix and restricted to plywood skins. To determine the shear capacity of a stressed-skin , three locations have to be checked: the lumber at the neutral axis of the , the attachment of the plywood flange to the lumber web on the compression side and on the tension side. In these calculations, corresponding (matched) values of Qw, Ba and y must be used. The factored shear resistance based on the shear strength of the lumber is calculated as in Clause 6. When calculating the shear resistance of the glued attachment between and lumber, both the shear strength of the lumber and that of both flanges have to be considered together with their appropriate resistance factors. The values of yt or yc corresponding to the value of Ba must be used. The Standard assumes that the bond between and lumber is adequate and that failure will occur in the wood or but not in the adhesive holding the two together. As a result of shear lag, the actual shear at the lumber- interface is less than would be calculated using the engineering theory of bending (Foschi, 1969b). CSA O86 therefore allows an increase in the factored shear resistance. This increase is a function of the area between lumber and and the clear spacing of the ribs. In the case of lumber shear strength at the rib-flange interface the shear modification factor has been set to 2. This last increase is not based on any consideration of t geometry, but simply reflects the fact that shear stresses published for lumber are greatly reduced from test values to allow for possible checking at the neutral axis of lumber. The strength modification factors for duration of load KD, service conditions KS and chemical treatment KT are the same as those for plywood, OSB or lumber respectively. Other applicable modification factors are the stress t factor XJ, the geometry reduction factor XG and the shear modification factor XV, discussed in Clause 10.5.

Clause 10.6.3.8 – Deflection Shear lag in stressed-skin s reduces stiffness (Foschi, 1969a and b). The geometry reduction factor s for this effect, resulting in a small increase in the calculated deflection.

References: APA, the Engineered Wood Association, 1990, Design and Fabrication of Plywood Stressed-Skin s, Supplement 3, Plywood Design Specification (Revised 1996) Bach, L and Cheng, J, 1990, Full scale tests of OSB faced stressed skin s, in proceedings 1990 Annual Conference and 1st Biennial Environmental Specialty Conference, Canadian Society for Civil Engineering Booth, L.G., C.B. Pierce and M.K. Surendranath. 1976. A Comparative Study of Three Methods of Deg Plywood Stressed Skin s. International Union of Forest Research Organizations, Wood Engineering Group. Proceedings of Meeting in Blokhus/Denmark; v.1. Booth, L.G. 1977. Shear Deflections of Box and I-Beams Formed from Flanges and Webs with Different Bending and Shear Moduli. Journal of the Institute of Wood Science 7(6-42):37-44. COFI. 1981. Nailed Plywood Beams. Council of Forest Industries of British Columbia. Vancouver, BC., pp. 8. Cziesielski, E. 1965. Ermittlung des Schubwiderstandes Symmetrischer I-und Hohlkastenquerschnitte Sowie Symmetrischer und Unsymmetrischer Winkelquerschnitte. Bautechnik, 42(7):232-237. Foschi, R.O. 1969a. Buckling of the Compressed Skin of a Plywood Stressed-Skin with Longitudinal Stiffeners. Western Forest Products Laboratory (now FPInnovations), Canadian Forestry Service, Publication No. 1265. Foschi, R.O. 1969b. Stress Distribution in Plywood Stressed-Skin s with Longitudinal Stiffeners. Western Forest Products Laboratory (now FPInnovations), Canadian Forestry Service, Publication No. 1261. Foschi, R.O. 1970. Rolling Shear Failure of Plywood in Structural Components. Western Forest Products Laboratory (now FPInnovations), Canadian Forestry Service, Information Report VP-X-67. Vancouver, BC. Kuenzi, E.W. and J.J. Zahn. 1975. Stressed-Skin Deflections and Stresses. USDA Forest Service, Research Paper FPL 251. Mazur, S.J. 1967. Shear Strain Energy and Shear Deflection Constants. Studies in Structural Engineering, No. 1. Nova Scotia Technical College, Halifax, NS. Newlin, J.A. and G.W. Trayer. 1924. Deflections of Beams with Special Reference to Shear Deformations. USFPL Report No. 1309. Reaffirmed March 1956. Raadschelders, J and Blass, H, 1995, Stressed Skin s, Lecture B10, STEP/Eurofortech - An initiative under the EU Comett Programme Smith, I. 1979. Analysis of Plywood Stressed Skin s with Rigid or Semi-Rigid Connections. CIB W18, Paper No. 11-4-1. Stieda, C.K.A. 1967. A Shear Stiffness Factor for Plywood Box Beams. Forest Products Laboratory (Canada). Information Report VP-X-31. Vancouver, BC. Stieda, C.K.A. 1968. The Lateral Stability of Timber Beams and Arches. Doctoral Thesis, University of London. Stieda, C.K.A. 1969. A Finite Element Solution for the Torsional Rigidity of Box Beams. Forestry Branch, Canada. Departmental Publication No. 1255.

Clause 11 Lateral-Load-Resisting Systems Clause 11.1 – Scope Design procedures for shearwalls and diaphragms were first introduced into CSA O86 in 1989. Additional procedures, supplementary to those in Division B of the National Building Code of Canada, were introduced for seismic design in 2009. The title of this section of CSA O86 was changed from “Shearwalls and Diaphragms” to its present title in 2009 when the scope was expanded to cover other aspects of structural wood design under lateral loads including wind, seismic and earth pressure loads. Clause 11 focuses on the two major sources of lateral forces: wind and seismic (earthquake) loads. In wood structures, the lateral forces are transmitted through diaphragms such as roofs and floors to the vertical lateral load resisting system (shearwalls in the case of wood-frame construction), and in turn to the lower storeys and the foundation. Note that shearwalls in O86 refer to lateral-load-resisting stud wall systems. The most common design method for diaphragms is based on the assumption that the diaphragm resists the lateral load in bending, where the edge perpendicular to the load act in tension and compression, and the horizontal shear is resisted by the sheathing between the edge . This design approach is sometimes referred to as the girder analogy (Applied Technology Council, 1981). The sheathing acts as the "web" to resist the shear in diaphragms, and the edge act as the "flanges" to resist induced tension or compression forces. For shearwalls, which may be considered to be deep vertical cantilever beams, the "flanges" are the end studs, subjected to tension and compression while the sheathing serves as the "web" to resist the shear. It is important that all elements, including the "flange" , connections at intermediate floors, and the connection to the foundation be detailed and sized for the induced forces. The provisions of CSA O86 apply to the engineering design of wood lateral-load-resisting systems. For housing and small buildings, Part 9 of the NBC provides prescriptive lateral-load-resisting measures for non-engineered buildings in high wind and seismic load zones. In addition, CWC’s “Engineering Guide for Wood Frame Construction” is referenced in Part 9 and provides lateral load design principles and details to address the transition between prescriptive and fully engineered construction.

Clause 11.2 – Materials Materials that can be used for the construction of diaphragms and shearwalls are referenced in the other clauses of the Standard. For products, this includes all structural types referenced in Clause 9 for plywood (CSA O121 Douglas Fir Plywood, CSA O151 Canadian Softwood Plywood, and Construction Sheathing OSB in accordance with CSA Standard O325). In addition, it is acceptable to use gypsum wallboard conforming to Type X (fire-rated) in ASTM Standard C1396. The majority of OSB s produced in Canada are Construction Sheathing s manufactured to CSA Standard O325. Construction Sheathing OSB s were tested in shearwalls at Forintek Canada Corp. (now FPInnovations) and at APA- the Engineered Wood Association. They were found to provide similar level of performance as nailed shearwalls sheathed with plywood s in dry service conditions, and consequently are recognized in CSA O86 by allowing the use of CSA O325 OSB in nailed shearwalls or diaphragms. Additional information can be found in the Commentary on Clause 9 and the Appendix of the Standard. In the 2014 edition of the Standard, new provisions were introduced to allow proprietary structural composite lumber that has been qualified for a specific grade for use in shearwalls with full-scale cyclic testing.

Clause 11.3.1.1 – Standard methods In previous edition of O86, the standard method was based on testing whereas in O86-14 the standard methods are mechanics-based, and the diameter of nails used in shearwall and diaphragm construction is limited to 3.66 mm or smaller to limit splitting in framing . This restriction does not apply to shearwall capacity derived from full-scale testing under 11.3.1.2. Alternative methods.

Clause 11.3.2 – Resistance to overturning The lateral load capacities of shearwalls depend on the amount of uplift restraint available to resist the overturning moment. If hold-downs are used or factored dead loads are sufficient to prevent overturning, every nail along the perimeter of the s contributes to the lateral capacity and the maximum lateral load capacities can be reached. However, if no hold-downs are used and factored dead loads are not sufficient to prevent overturning, nails along the bottom plate at the tension end are also required to resist the uplift forces due to overturning moment, therefore the lateral load capacity of the shearwall is reduced. A methodology that s for the partial or full restraint against overturning in shearwall segments was implemented in the 2001 edition of the Standard. Further details are given in the Commentary on Clause 11.4.5.

Clause 11.3.3 – Shearwalls with segments The lateral load capacity of a shearwall is determined as the sum of the lateral load capacities of fullheight shearwall segments. This applies to shearwalls with openings, and also to shearwalls with different construction (dissimilar materials, thickness or nail spacing) along the length of the shearwalls. For shearwalls with openings, the lateral load capacities of the segments vary depending on whether holddowns are applied around openings. For simplicity, the lateral load capacity of sheathing above and below openings in shearwalls are neglected in assessing the shearwall capacity. For shearwalls of different construction along the length of the shearwalls, the uplift restraint contribution from adjacent shearwall segments is also neglected.

Clause 11.3.4.2 – Two-sided shearwalls Tissel (1993) showed that where a shearwall is sheathed with the same type of wood-based s on both sides, the load carrying capacity of the wall is approximately twice the capacity of a wall sheathed on only one side. To develop this capacity, sheathing should be installed with staggered ts to ensure that the close nail spacing around the perimeters do not fall on the opposite edges of the same stud member; or, if sheathing is nailed to both sides of the wood without staggered ts, it should be attached to 3 inch nominal thickness lumber (or two 2x4 adequately fastened together) to provide more nail edge distance in the framing member. Karacabeyli and Ceccotti (1996) tested comparable shearwalls that were sheathed with a) OSB on one side only; b) gypsum wallboard (GWB) on one side only; and c) OSB on one side and GWB on the other side. They concluded that, up to a 30 mm horizontal displacement level, the load-displacement curves of single sided OSB sheathed wall (Case “a”) and single sided GWB sheathed wall (Case “b”) could be superimposed to determine the lateral resistance of a Case “c” wall. They also reported that the combined OSB and GWB sheathed wall had greater ultimate lateral load resistance than the single sided OSB sheathed wall, although the ductility was reduced.

The above test results suggested that the lateral load capacities were cumulative for same or different sheathing applied on both sides of a shearwall. Summing shear capacities of wood-based s and gypsum wallboard was recommended under the following conditions: a) a lower ductility-related seismic force modification factor (Rd = 2) is used to for the reduced ductility; b) application is limited to platform frame construction only; and c) In the case of 4-storey buildings, the percentage of storey shear force resisted by GWB walls is limited so that the length of shearwalls with wood-based sheathing in the first two storeys is similar to what it would be in a seismic design with an Rd=3 without the contribution from the gypsum wallboard walls ed for. Note: More discussion on GWB can be found under Commentary on Clause 11.5.4.

Clause 11.4.1 – Fastener spacing factor, Js In the 2014 edition of CSA O86 the fastener spacing factor, Js, was introduced with the mechanics-based approach, which was used to determine the lateral load capacity of shearwalls and diaphragms. The fastener spacing factor was developed based on the study by Ni et al. (2012) to take into the reduced lateral load capacities of shearwalls and diaphragms at close fastener spacing. With close fastener spacing, the possibility of splitting of framing and probability of fasteners hitting weak spots of framing is increased, consequently reducing the lateral load capacities of shearwalls and diaphragms. Note that this factor applies to shearwalls and diaphragms with equal fastener spacing along the edges. Where unequal fastener spacing along the edges is used, analysis should be carried out to further reduce the lateral load capacity. Note: More discussion on the mechanics-based approach can be found under Commentary on Clause 11.5.

Clause 11.4.2 – Fastener row factor for blocked diaphragms, Jf The fastener row factor for blocked diaphragms, Jf, was introduced with the mechanics based approach in the 2014 edition of CSA O86. It is based on the findings and recommendations by Countryman (1952) and Tissell and Elliott (2000). The lateral load resistance shall be reduced by 11% when 38 mm thick framing member is used. This reduction also applies to 2 rows of fasteners in 64 mm thick framing member and 3 rows of fasteners in 89 mm thick framing member. Note: More discussion on the mechanics based approach can be found under Commentary on Clause 11.5.

Clause 11.4.3 – Strength adjustment factor for unblocked diaphragms, Jud The strength adjustment factor for unblocked diaphragms, Jud, was introduced with the mechanics-based approach in the 2014 edition of CSA O86. The shear strength of unblocked diaphragms can be derived from the shear strength of a blocked diaphragm with fasteners spaced at 150 mm on centre along edges and 300 mm on centre along intermediate framing . The strength adjustment factor for unblocked diaphragms was introduced to maintain the same ratio as 2009 edition when comparing unblocked diaphragms with the reference blocked diaphragm. Note: More discussion on the mechanics based approach can be found under Commentary on Clause 11.5.

Clause 11.4.4 – Strength adjustment factor for unblocked shearwalls, Jus In some applications such as exterior wall sheathing, the wood-based s may be applied to the wall frame without blocking at all of the edges. In the 2001 edition of the Standard, provisions were added to allow horizontally sheathed unblocked shearwalls based on studies at APA (Tissell, 1990) and Forintek Canada Corp. (now FPInnovations) (Ni et al, 2000). The studies showed that the specified shear strength for an unblocked shearwall can be expressed as a percentage of the specified shear strength of a reference blocked shearwall with 150 mm perimeter nail spacing, based on the following observations: a) For 406 mm stud spacing, the average capacity of unblocked shearwalls with 150 mm nail spacing at ed edges and 300 mm nail spacing at intermediate studs was approximately 60% of the average capacity of blocked shearwalls with 150 mm perimeter nail spacing and 406 mm stud spacing. b) For both 406 mm and 610 mm stud spacing, the average capacity of unblocked shearwalls with 150 mm nail spacing at ed edges and 300 mm nail spacing at intermediate studs was approximately 80% of the average capacity of unblocked shearwalls with 150 mm nail spacing both at ed edges and at intermediate studs, and c) For 100 mm nail spacing at ed edges and at intermediate studs, unblocked shearwalls with 406 mm stud spacing had approximately 150% of the capacities of unblocked shearwalls with 610 mm stud spacing. This indicated that the capacities of unblocked shearwalls were inversely related to stud spacing. Strength adjustment factors for unblocked shearwalls, Jus, were introduced in the 2001 edition based on the above findings. The design approach was initially limited to 2.44 m high walls. In the 2009 edition of the Standard, the allowable height of the unblocked shearwall was extended to 4.88 m based on further research (Mi et al, 2006) that showed the same adjustment factors for unblocked shearwalls could be applied conservatively to 4.88 m tall walls. When the Jus factors were considered, the load displacement and energy dissipation responses of the unblocked walls were comparable to corresponding blocked shearwall performance. This work also broadened application to both vertically and horizontally sheathed walls with different sheathing patterns. As a cautionary measure, the unblocked shearwall provisions were limited to a maximum aspect ratio of 2:1.

Clause 11.4.5 – Hold-down effect factor for shearwall segments, Jhd Hold-downs are mechanical connections that transfer overturning forces from chords of shearwalls to lower-storey shearwalls or the foundation below. Where hold-downs are not used, overturning tension forces may be transferred through the sheathing on shearwalls; however, this reduces the shear capacity of the shearwall segment. The reduction in shear capacity is calculated using the Jhd factor. Prior to 2001, CSA O86 required all shearwalls to be designed and detailed to transfer forces around the openings and did not provide any guidance on the design of shearwalls without hold-downs at openings. The design capacity of the shearwalls was assumed to be the sum of the capacities of all of the shearwall segments that did not contain openings. This design procedure is maintained in the present edition, but another procedure is provided for design of shearwalls without hold-downs around openings. Sugiyama (1981) proposed an empirical equation to calculate the lateral load carrying capacity and the stiffness of shearwalls without hold-downs around openings. This empirical equation forms the basis of the perforated shearwall (shearwalls with openings) design method which appears in other codes and standards, such as U.S. Building Codes and the Wood Frame Construction Manual for One- and TwoFamily Dwellings (WFCM, 2001). Although Sugiyama's empirical equation was found to predict conservatively the load carrying capacity of the shearwalls with openings (Johnson and Dolan, 1996), it did not provide a design procedure for calculating a clear load path for shearwalls between storeys.

Forintek Canada Corp. (now FPInnovations) developed a mechanics-based design method for shearwalls with openings (Ni et al, 1998, Ni and Karacabeyli, 2000). This method ed for the overturning restraints at the ends of shearwall segments, and provided a load path design procedure for shearwalls between storeys. The method was compatible with the traditional engineering approach for calculating the load-carrying capacities of shearwalls with openings. Comparison between the Forintek Canada Corp. (now FPInnovations) method and test results from Dolan and Heine (1997a, 1997b) showed that the method conservatively predicted the wall capacities.

Clause 11.4.5.1 – Shearwall segments with hold-downs Shearwall segments with hold-down devices designed to resist the factored uplift forces at the bottom and at the top of the segments were assigned a Jhd factor of 1.0. A steel threaded rod is an example of a holddown device. For shearwall segments with factored dead loads that are sufficient to resist overturning, Jhd is also unity.

Clause 11.4.5.2 – Shearwall segments without hold-downs The background information behind the mechanics-based approach for this clause was developed by Ni and Karacabeyli (2000). The uplift restraint force, P, was defined as consisting of a) the resultant force from upper storeys (negative if it is a net uplift force); and b) forces due to dead weight in that storey. An example is given in Figure 11.4.5.2 in the Standard. Where factored dead loads are not sufficient to prevent overturning but there is a positive factored uplift restraining force at the top and at the bottom of the segment, and no hold-downs are used either at the bottom or at the top of the segment, anchorage is required on the bottom plate, within 300 mm from both ends of the shearwall segment, to transfer the uplift force to the top plate of the shearwall below or to the foundation. In such cases, Jhd is calculated to be less than unity. The factored shear resistance of the shearwall must be determined for lateral loads in opposite directions.

Clause 11.4.5.3 – Shearwall segments with hold-downs only at the bottom of the segment The top of a shearwall segment in the top storey in a building is not subject to overturning forces. For storeys below the top storey, overturning forces from the upper storeys can create uplift at the top of the segment. A hold-down is required at the top of the shearwall to resist this uplift; or the shearwall sheathing is required to transfer the tension force, thus reducing the shear capacity of the shearwall segment. In such cases, where a hold-down is used at the bottom but not at the top of the segment, Jhd is calculated to be less than unity. The factored shear resistance of the shearwall is determined for lateral loads in both directions.

Clause 11.4.5.4 – Shearwall segments with hold-downs only on one side For a shearwall segment with hold-downs only on the right or left side of the segment, the Jhd factor is calculated based on the following: a) Clause 11.4.5.1 if the hold-down is designed to resist all the uplift forces due to overturning; or b) Clause 11.4.5.2 if the segment does not have a hold-down connection on the tension side of the segment; or c) Clause 11.4.5.3 if a hold-down is used only at the bottom of the tension side of a lower-storey segment, with uplift at the top of the segment.

The factored shear resistance of the shearwall is determined for lateral loads in opposite directions.

Clause 11.4.5.5 – Conditions for shearwall segments with Jhd < 1.0 Additional requirements in this Clause are intended to limit the application of provisions in Clauses 11.4.5.2., 11.4.5.3., and 11.4.5.4. to the configurations of the full-size tests.

Clause 11.5 – Strength and Resistance Horizontal Distribution of Shear In accordance with ASCE 41-13 Seismic Evaluation and Retrofit of Existing Buildings, diaphragms shall be classified as flexible when the maximum in-plane deflection of the diaphragm is more than twice the average drift of vertical elements of SFRS of the associated storey below. Furthermore, diaphragms shall be classified as rigid when the maximum in-plane deflection of the diaphragm is less than half the average drift of vertical elements of SFRS of the associated storey below. The flexible diaphragm assumption simplifies the analysis and allows lateral forces to be distributed to the vertical elements (mainly shearwalls) of the lateral load resisting system by the tributary area method. In this case, the diaphragm can be visualized as a single-span beam ed on rigid s. Another method of distributing loads to shearwalls uses a continuous beam analogy. Although rotation of the diaphragm may occur because lines of vertical elements have different stiffnesses, the diaphragm is not considered stiff enough to redistribute lateral forces through rotation, and consequently torsional effects are neglected. Analytical methods relating to rigidity, rotation and torsional behaviour of diaphragms affect the way loads are assumed to be distributed to ing elements, including shearwalls. For rigid diaphragms, the loads are assumed to be distributed according to the relative stiffness of the shearwalls. Torsional response of certain structures due to eccentricity (centre of mass does not coincide with centre of rigidity) at a given level can be very important in this case. In reality, most wood diaphragms behave as semi-rigid elements, neither fully flexible nor fully rigid. To for this fact, SEAOC (1997) published an “Envelope Approach” that suggested two analyses be performed: the first analysis used the traditional flexible diaphragm assumptions, and the second analysis was based on rigid diaphragm assumptions. The lateral load-resisting system could then be designed for the most severe forces produced by either assumption. It is important to note, however, that because of the difficulties in determining the stiffness of the various elements, the use of rigid diaphragm assumptions may not be significantly better than the traditional flexible diaphragm assumptions. For the design of mid-rise wood frame construction, it is recommended when the stiffness cannot be accurately determined, a designer should consider adopting an envelope approach. In the APEGBC Technical and Practice Bulletin (APEGBC, 2015), it is recommended that all walls should be designed for the envelope forces of the two diaphragm assumptions if the force in any wall is different by more than 15% due to the change in flexible and rigid diaphragm assumptions.

Distribution to shearwall segments After the applied loads are distributed to lines of shearwalls, the next step is to determine how to distribute the forces to shearwall segments within each line. One method in common use is to distribute the forces based on relative capacity. Some practitioners also distribute lateral forces to the shearwall segments in proportion to the lengths of the wall segments, such that each segment carries the same unit load. This assumes wall segment stiffnesses are proportional to the wall segments’ lengths. This is a reasonable assumption for a line of shearwall segments with similar configuration and ductility as they will have the same displacements at a given load level.

On the other hand, for a line of shearwall segments with different configurations (for example with or without hold-downs, different unit load per nail due to sheathing, nail sizes or spacing), the assumption that stiffness is proportional to the wall length may not be valid. It would be preferred to determine the distribution of load to the wall segments on the basis of the stiffness of segments. Clause 11.7 covers the calculation of deflection of shearwall segments and diaphragms, making it possible to distribute the shear force to the wall segments based on relative stiffness. The stiffness of each segment is determined using the specified shear force distributed to this segment divided by the corresponding deflection. This approach requires an iterative process to arrive at a solution where all the shearwall segments have the same deflection at the load level being considered. This distribution method results in higher loads in stiffer segments.

Clause 11.5.1 – Shear resistance of nailed shearwalls The specified shear strengths of shearwalls in CSA O86 were traditionally soft converted from allowable shear forces published in the U.S. codes. The 1994 edition of O86 provided shearwall design values for a limited combination of thicknesses and nail diameters. In later editions, the specified shear strengths were expanded to cover a wider range of combinations of nail diameters and thicknesses. The allowable shear force for a shearwall in working stress codes was approximately equal to the average ultimate load carrying capacity of a tested shearwall divided by a load factor of 3. These values were confirmed through shearwall tests conducted by APA (Adams, 1987; Tissell, 1990). Because the allowable shear forces had been in place for 30 years, they were used as a basis to develop specified shear strengths for CSA O86. Specified strengths for shearwalls in 1989 edition of CSA O86 were derived from allowable stresses using a factor of 1.863. This conversion was based on the premise that the resistance to wind design loading in 1990 should be the same as for previous editions of the National Building Code of Canada. Except for a few modifications, the same approach was followed to derive the specified shear strength values for later editions of CSA O86, until a mechanics based approach was introduced in 2014 edition of the Standard. The mechanics-based approach to the calculation of factored shear resistance for shearwalls is based on the work of Ni, Chui and Karacabeyli (2012). The factored shear resistance for shearwalls constructed with wood based s is calculated as the lower of the resistance based on the sheathing-to-framing connection, and buckling. For nailed wood based shearwall, the lateral strength of sheathing-to-framing connection along edges is determined in accordance with Clause 12 and modified by several adjustment factors. The fastener spacing factor, Js, is to take into the reduced lateral load capacity of shearwalls at close fastener spacing. The factor for shearwall and diaphragm construction, JD, is to take into consideration the averaging effects of nail t capacities around the perimeter. The strength adjustment factor, Jus, is to accommodate the reduced lateral strength of unblocked shearwalls. The number of shear plane factor, ns, addresses the double shear in mid- shearwall construction. The lateral strength of sheathing-to-framing connection is also modified by the hold-down factor. A possible failure mode in shearwall and diaphragm under lateral load is buckling of sheathing s before the ts along the edges reach their capacities. The buckling strength is calculated using a model proposed by Dekker et al. (1978) and based on the assumption that the is simply ed on four edges and subjected to a uniform shear stress along the edges. This simplification ignores the contribution of the intermediate stud and is therefore conservative.

Clause 11.5.2 – Shear resistance of nailed diaphragms A mechanics-based approach to the calculation of factored resistance for diaphragms was introduced in the 2014 edition of the Standard based on the work of Ni et al. (2012). The specified resistance for diaphragms constructed with wood based s is calculated as the lower of the resistance based on the sheathing-to-framing connection, and buckling. This is consistent with the procedure for shearwalls.

However, different modification factors are used to modify the lateral strength of sheathing-to-framing connection for diaphragms. The fastener row factor for blocked diaphragms was introduced to adjust the sheathing-to-framing connection capacity for diaphragms with multiple rows. The adjustment factor for unblocked diaphragm was developed to maintain the same ratio as previous edition of the Standard when comparing the lateral capacity of unblocked diaphragms with reference blocked diaphragms.

Clause 11.5.3.4 – Sheathing fastening High capacity shearwalls and diaphragms have been introduced to the 2014 edition of O86. Where the fastening in a shearwall is a double shear t with a in the middle and framing as outside t , the high capacity shearwall could be achieved. Where there are multiple-rows of fasteners in diaphragm sheathing, the high capacity diaphragm could be achieved. In late 1990s, the Mid- shearwall system was developed at Forintek (now FPInnovations) by redeg the ts between sheathing and framing . Figure 7 illustrates a cross section of standard shearwall and that of a mid- shearwall system that uses the same 38 × 89 mm lumber studs.

Figure 7 Cross-section of a standard shearwall (top) and a mid- wall (bottom) with two exterior s

The superior performance of mid- shearwall under lateral loading is attained through the following means (Varoglu et al., 2006): 1. A wood based is used at the centre of the wall to provide the lateral resistance of the wall without increasing the nominal thickness of the wall. Nails connecting this to the framing work in double shear as illustrated in 2. Figure 8 (or in triple shear if additional exterior wood based sheathing is used), providing the increased lateral load-carrying capacity. 3. Studs in mid- wall system are placed at a 90 degree rotated position relative to those in standard shearwalls. Sheathing material is fastened to the wide face of the studs instead of the narrow face of the studs in case of standard walls. This increases the lateral load capacity of the mid- wall by providing more edge distance for fasteners on the perimeter of the sheathing s placed in the mid plane and at the exterior face of the wall. Increased edge distance reduces the possibility of nail tear out failures and makes it easier for framers to nail the sheathing to the studs.

4. The heads of nails are kept away from the surface of the mid-; consequently nail pull through failure at the mid is physically prevented. Also, poor construction practices such as over driven sheathing nails and nails going through the s missing the studs are minimized.

Figure 8 Nailed ts working in single shear in a standard shearwall and in double shear in a mid- wall

Stud or Plate

Sheathing

89 mm Grain direction

Nail in single shear

38 mm

38 mm

Stud or Plate

Nail in double shear

To ensure that the mid- shearwalls are able to reach their shear capacities, the following construction and detailing requirements are recommended: 1. According to Clause 12.9.2.2 of CSA O86, the minimum thickness of the center member should be at least eight times the nail diameter. However, test results showed that mid- walls with 12.7 mm thick had equal or better performance than a double-sided shearwall (Varoglu et al., 2006). As a result, it is recommended that the wood-based s in the center of the wall should have a minimum thickness of 12.7 mm. 2. Framing and the mid-s should be nailed together using nails with adequate penetration. According to Clause 12.9.2.2 of CSA O86, the minimum penetration length into the third member should be at least five times the nail diameter. 3. Framing and the mid-s should be nailed together with the same nail spacing in all locations. Two rows of fasteners around perimeters can be used. 4. The vertical framing of mid- shearwall should be checked for buckling. Additional studs may be necessary to resist the high compression forces and prevent buckling in the end studs, as shown in Figure 9.

Figure 9 A detail of an end stud in a mid- wall that includes the added buckling stud

Buckling Stud

End stud with an 5. Where the dead loads are notadded sufficient to prevent buckling stud overturning moment, hold-down connections at the end of the mid- shearwall should be designed to resist the uplift forces. To prevent premature end stud failure, it is suggested that continuous steel rods, in lieu of hold-downs, be used to restrain the end studs against uplift. This idea was proved effective during the cyclic testing of mid- walls where two steel rods of 16 mm diameter were used to connect the top and bottom plates at each end of the test wall. Test results showed that these walls exhibited good ductility and the premature end stud tension failure was eliminated (Varoglu et al., 2007). 6. To prevent detachment of the studs at the ts during reversible earthquake loading, additional fasteners not subjected to double shear should be installed to connect the pair of the studs. This can be accomplished by fasteners with adequate penetration equally spaced along the length of the studs.

Figure 10 Detail of studs in a mid- shearwall where the two s meet

One way to increase the capacity of diaphragms is to increase the number of fasteners per linear length. This can be achieved through the use of multiple rows of fasteners to prevent lumber from splitting. APA – The Engineered Wood Association has carried out test program to study the performance of diaphragms with multiple rows of fasteners. The test results in the APA Research Report 138 (Tissell and Elliott, 2000) show that the lateral load capacity of the diaphragm can be increased if multiple rows of fasteners are used along the sheathing perimeter. Based on the test results, design values of diaphragms with multiple rows of fasteners have been included in Table 2306.3.2 of IBC since 2003. With multiple rows of fasteners, minimum of 64 mm or 89 mm wide framing need to be used at ading edges and boundaries. Alternatively, built-up framing consisting of two or three 38 mm wide framing connected to transfer the factored shear force could be used. To ensure the performance of diaphragms with multiple rows of fasteners, the number of fastener lines and the spacing between fasteners in each line and at edge margins should not be less than the specifications provided in Figure 11. A gap of not less than 2 mm shall be left between ading edges. Reduce spacing between lines of nails as necessary to maintain minimum 9 mm fastener edge margins, minimum spacing between lines is 9 mm.

Figure 11 Nailing requirements for diaphragms with multiple rows of fasteners

Clause 11.5.3.5 – Additional construction details Splitting of the bottom plate of the shearwalls has been observed in tests as well as in structures subjected to earthquakes. Splitting of plates can be caused by the rotation of individual sheathing s inducing upward forces through the nails at one end of the and downward forces at the other end. With the upward forces from the nails and a downward force from the anchor bolt, the sill plate acts like a cantilever beam and cross-grain bending stresses are introduced. Splitting can be prevented by use of large, sufficiently stiff plate washers. In the case of close nail spacing, to guard against the potential to split studs particularly where two s meet, it is required that for 50 mm nail spacing and for 75 mm spacing with 3.66 mm nails, framing consist of at least 64 mm thick lumber or two 38 mm framing properly connected to transfer the factored shear force, with staggered ts.

Clause 11.5.4 – Nailed shearwalls using gypsum wallboard While the lateral load resistance of gypsum wallboard was not recognized in early CSA O86 provisions, some building codes have long recognized the use of gypsum wallboard to resist lateral loads induced by earthquake or wind. Where gypsum wallboard is used in combination with structural wood-based s in the same shearwall or parallel shearwalls, a ductility-related force modification factor, Rd, equal to 2 for the entire building was found to be appropriate for deg wood frame residential buildings (Ceccotti and Karacabeyli, 2002). A gypsum-sheathed shearwall option was introduced into CSA O86 in 2001 for cases where shearwalls are sheathed with both gypsum wallboard and wood-based s. At the same time, it was noted that the R factor was lower for this option, since shearwalls sheathed with wood-based s alone are better at dissipating energy under earthquake loading. This was modified slightly in 2005 after the NBC began characterizing the seismic structural force resisting system by the ductility-related force modification factor, Rd (2 for gypsum and wood-based s in combination, and 3 for wood-based s alone). In Table 11.5.4 of CSA O86, specified shear strength values are provided for gypsum wallboard shearwalls with nail spacing 200 mm, 150 mm and 100 mm on the edges. The use of this design procedure for shearwalls of gypsum wallboard is restricted to Type X (fire rated) wallboard conforming to ASTM Standard C1396, and also to platform frame construction where the height of a storey does not exceed 3.6 m.

Clause 11.5.5 – Nailed shearwalls and diaphragms using diagonal lumber sheathing Diagonal lumber-sheathed shearwalls represent an older type of construction, with design provisions based originally on monotonic racking tests. These provisions were modified in 2009 based on more recent cyclic shearwall research at Forintek Canada Corp. (now FPInnovations (Ni and Karacabeyli, 2007). The specified shear strength of shearwalls with double-layer diagonal sheathing was reduced from three times to two times the shear strength of single-layer shearwalls, to reflect the results of testing and analysis. Further 2009 revisions provided additional guidance on how to address forces at boundary and at corners of boundary elements in single-layer lumber shearwalls and diaphragms. Loading of singlelayer shearwalls and diaphragms creates additional forces that need to be addressed in design. These forces occur in the boundary parallel to the interior framing (studs in shearwalls or joists in diaphragms). The normal load component, acting perpendicular to the boundary member, creates a moment, Mf that must be ed for in design of the boundary member. The boundary load component also causes the corners of the shearwall or diaphragm to separate. The boundary frame in single-layer diagonally sheathed shearwalls and diaphragms shall be designed to resist the separation force Pc.

Figure 12 Factored bending moment and factored mutual separation force in a single layer diagonally sheathed shearwall segment

Clause 11.6 – Detailing Structures without a shearwall on one side Provisions for lateral resistance of buildings that are open on one side were removed in the 2001 edition, pending development of more comprehensive shearwall design provisions. However, the information from the previous edition may be useful in some applications and is provided here for reference purposes. Designers may also wish to use more sophisticated models to assess torsional effects on these types of structures. Structures meeting the limitations below were permitted to be designed without a shearwall of one storey on one side. The diaphragm ed by the remaining walls was required to be sheathed with diagonal boards or plywood or OSB. Otherwise, structures having one side without a shearwall was permitted if it could be demonstrated that all lateral forces could be safely transferred to the ground and that no excessive deformations of the structure would occur as a result of torsion.

Limitations on length of uned edge and depth-to-width ratio of diaphragms a) Except as given in (b) and (c), the depth of the diaphragm normal to the open side of the structure should not exceed the lesser of 7.5 m and two thirds of the diaphragm width. b) For one storey wood-framed the depth normal to the open side should not exceed the lesser of 7.5 m and the diaphragm width. c) Where calculations show that diaphragm deflections can be tolerated, the depth normal to the open side may be increased, and the depth-to-width ratio may be increased to i) 1.5:1 for diaphragms using single layer diagonal sheathing; or ii) 2:1 for diaphragms using nailed double layer diagonal sheathing, or nailed plywood or OSB sheathing. The intent of these recommendations were to restrict building depth to 7.5 m for structures with one open side only in cases where lateral deflections were not calculated. If the lateral deflections were calculated

and shown to be acceptable, and provided that the building conformed to specified aspect ratios, no restrictions on the absolute depth were intended.

Clause 11.7 – Deflections of shearwalls and diaphragms The deflection provisions for shearwalls and diaphragms were revised and moved into the body of the Standard from the commentary in 2009, after revisions to the NBC raised the need for deflection calculations.

Clause 11.7.1– Deflection of shearwalls Deflection of single-storey blocked and unblocked shearwall segments may be calculated in accordance with Clause 11.7.1.2 and 11.7.1.3. For multi-storey shearwalls, the multi-storey effects on shearwall deflection shall be considered. One methodology based on engineering mechanics with the assumption that the shearwall is cantilevered from its base and stacked for the full height is provided in the Appendix Clause A.11.7.1. See Commentary on Clause A.11.7.1 for more information. The deflection formula for single-storey blocked shearwalls was developed by APA based on engineering principles (APA 2002). It provides a reasonable estimate of shearwall deflection due to loads applied in the range of factored shear resistance. The formula consists of four components contributing to the shearwall deflection: deflection due to shearwall bending, shear, fastener slip, and anchorage deformation. The calculation of anchorage deformation, da, for blocked shearwalls is not explicitly provided in the Standard. The term, da, s for the total slip of the wall anchorage system, including oversized holes for fasteners, slip due to fastener deformation, anchor bolt deformation and elongation of the bracket. Manufacturers of hold-down connections typically provide literature for hold-down connector elongation and/or hold-down assembly displacement based on the capacity of the hold-down. Published values for hold-down assemblies are likely to provide more realistic da values but would not consider shrinkage or crushing. Some engineers choose to adjust the published deformation values based on the actual forces in the hold-down. Crushing of the plate and shrinkage of the wood may also contribute to da but may be mitigated using shrinkage compensating systems. Although the formula was developed originally for blocked shearwalls with hold-downs at both ends, recent research (Ni and Karacabeyli, 2004) has shown that the formula can also be adapted for unblocked shearwalls, and shearwalls without hold-downs. For unblocked shearwalls, the deflection is calculated as follows: •

Calculate deflection based on a blocked shearwall segment with 150 mm nail spacing at edges according to Clause 11.7.1.2.

•

Divide the deflection calculated according to Clause 11.7.1.2 by the appropriate J us factor in Table 11.4.4.

For shearwalls without hold-downs, the deflection due to anchorage details, da, may be calculated using the da formula provided in Clause 11.7.1.2. which is based on force equilibrium taking into the nail deflection at the tension end of the wall. The formula is not applicable to shearwalls sheathed only with gypsum wallboard or lumber, but it can be applied to shearwalls with wood-based s on one side and gypsum wallboard on the other side. Test results have shown that the lateral load resistances for shearwalls are cumulative in the elastic range, where either the same or different sheathing is applied on both sides (Karacabeyli and Ceccotti 1996). As a result, shearwalls with wood sheathing on both sides can be treated as two separate walls, and the deflection can be determined based on shearwalls sheathed with wood-based s on one side only, after the load distribution between the two sides (s) is determined.

Deflection of a single-storey shearwall with wood-based s on one side and gypsum wallboard on the other, subjected to factored design shear at the top of the wall, v, can be determined using the deflection formula from Section 11.7.1.2 where the factored design load resisted by the side of the wall with woodbased s only, vp, is substituted for v (Ni and Karacabeyli, 2004). The distribution of v to both sides of the wall are determined by the relative stiffness of the wood-based and gypsum wallboard.

Clause 11.7.2 – Deflection of wood diaphragms The lateral deflection formula for simply ed diaphragms was also revised and moved into the body of the Standard in 2009. The first part of the equation represents the deflection due to bending, assuming that only the flanges carry the bending moment. The second part of the equation represents the shear deformation of the s forming the web of the diaphragm. The third part represents the relative movement of s due to the deformation of nails along boundaries. A fourth term was added to the formula in 2009 to estimate the contribution of chord-splice slip to the total diaphragm deflection. The formula is applicable to blocked diaphragms uniformly nailed throughout, but is not applicable to unblocked diaphragms.

Clause 11.8 – Seismic design considerations for shearwalls and diaphragms Clause 11.8 was added in the 2009 edition of the Standard for consistency with the requirements and design philosophy of the NBC. The design requirements were developed based on the concept of capacity based design philosophy and taking into the current practices adopted by the design community, lessons learned from seismic performance of wood-frame structures during earthquakes and observations from full-scale shake table tests. According to capacity based design principles, structural collapse is prevented by ensuring that the inelastic behaviour is confined to the structural elements that can dissipate energy inelastically without significant loss of strength, while all other elements are designed to be stronger to allow this mechanism to develop (Popovski et al., 2009). Research in this area is ongoing. Wood shearwalls and diaphragms resist lateral forces through the ductile yielding of nails that connect sheathing to framing , as demonstrated during quasi-static cyclic and shake-table tests. Review of the past seismic performance of platform wood-frame structures across North America (Rainer and Karacabeyli, 1999) showed that the objectives (i.e. prevention of structural collapse to achieve life safety) of the seismic design have been achieved and the very few observed failures were mainly related to the soft-storey mechanism that developed at the first storey of multi-storey buildings. Except as identified in Clause 11.8.7, Clause 11.8 applies to structures in locations of moderate or high seismic loads; i.e., where IEFaSa(0.2) is 0.35 or higher.

Clause 11.8.1 – General Most building codes contain force modification factor(s) that for the ductility, over-strength and energy dissipating characteristics of the structural system under earthquake loads. In Part 4 of the National Building Code of Canada (NBC), force modification factors or “R” factors are provided for various types of lateral force resisting systems in steel, concrete, masonry and timber. The R factors for timber structures were first implemented in the 1990 edition of NBC and were initially selected so that designs of lateral force resisting systems would be consistent with the observed performance history. In the 2005 NBC, the R factors were revised to ductility-related force modification factors (Rd) and overstrength-related force modification factors (Ro). The Ro factor for wood shearwalls is 1.7, in recognition of the reserve strength of shearwall design, and is reduced slightly (1.5) for braced or moment-resisting timber frames. The Rd factor is 3.0 for typical wood shearwalls sheathed with woodbased s, and is reduced to 2.0 for wood shearwalls sheathed with gypsum wallboard s in

combination with wood-based s. The Rd factor for braced or moment-resisting timber frames varies from 1.5 to 2.0 depending on the ductility of the system. A methodology for assessment of seismic design parameters for a wood-frame shearwall system was developed by Ceccotti and Karacabeyli (2002). This methodology consisted of a test program of shearwalls and the application of non-linear time history analyses to a 4-storey wood-frame building designed to resist the seismic requirements for Vancouver, BC. In the analysis, 28 selected earthquake accelerograms were scaled upwards until the maximum Peak Ground Acceleration (PGA max) brought the shearwall to a “near collapse” state. All of the recorded values of PGA max for the test configuration were found to be greater than the “design” PGA, indicating the adequacy of the current design procedures (Rd=3) for the shearwall system. These results were ed by the historical performance of platform-frame wood construction in past earthquakes (Rainer and Karacabeyli, 1999). The influence of gypsum wallboard on the behaviour of shearwalls was also evaluated in the Ceccotti and Karacabeyli study, resulting in the force modification factor “Rd=2” for walls composed of a mixture of wood-based and gypsum s. The effect of the flexible floor diaphragm model was considered separately and found to have a relatively small effect on the ultimate capacity of the building analyzed. With the introduction of the mechanics based approach for determining the shear resistance of shearwalls and diaphragms to the 2014 edition of the Standard, the nailed sheathing-to-framing connection shall governs the resistance for seismic design and the nails are required to fail in the modes where the plastic hinge(s) form to ensure sufficient ductility.

Clause 11.8.2 – Shearwall hold-downs and shear transfer connections To ensure adequate performance of the structural system during an earthquake, and following the capacity design principles, the design loads on critical system components and force transfer elements are increased. Based on the test data from quasi-static tests on shearwalls, as well as from the numerous shaking table tests around the world, the current design practice for anchor bolts and hold-downs has provided satisfactory wall performance. To further reduce the likelihood of failures in critical force transfer elements, Clause 11.8.2 requires that anchor bolts, inter-storey connections resisting seismic shear forces, and hold-downs resisting seismic uplift forces be designed for seismic loads that are at least 20% greater than the force that is being transferred. The intent of Clause 11.8.2 is to ensure that the desired ductile nail yielding is achieved throughout the structure without any failure in the hold-downs and shear transfer connections.

Clause 11.8.3 – Over-capacity of wood-based seismic force resisting system (SFRS) This clause identifies over-capacity requirements for the vertical SFRS in structures with three or more storeys, to address the concern of soft-storey mechanism developing at the lower storey. The overcapacity coefficient, Ci, of the 2nd storey of the vertical SFRS is required to be between 0.9 and 1.2 times the coefficient of the 1st storey of the vertical SFRS. This is to ensure that non-linear deformation begins to develop in the 2nd storey soon after it begins in the 1st storey, and that the response and mode of vibration is consistent with design assumptions. The concern with significant over-capacity in the 2nd storey was that it might not yield under an earthquake load, causing a soft-storey effect at the first storey. On the other hand, if the 2nd storey was significantly weaker, it could start to yield without transferring the forces to the 1st storey, creating a softstorey effect at the second storey. One could check the remaining Ci+1/Ci ratios along the remaining storeys of the building and apply the same procedure especially if the building has up to six storeys. Results from dynamic analyses on four storey platform frame buildings suggest that the first two storeys contribute the most to the seismic response and energy dissipation of the building.

Clause 11.8.4-11.8.6 – Wood diaphragms where IEFaSa(0.2) is 0.35 or higher Loads on diaphragms are required to respond to the over-capacity of the ing SFRS. The overcapacity requirements for diaphragm design depend on the type of SFRS, as described below (see also Figure 13 for an overview of the section): •

Where a wood diaphragm is ed on wood shearwalls, the seismic design force in the diaphragm at each storey “i” is obtained by multiplying the factored seismic force at storey “i” by an over-capacity coefficient CDi that is equal to the lesser of 1.2 or Ci. Note that a wood shearwall may consist of a wood stud system sheathed with wood-based s, either alone or in combination with GWB s.

•

Where a wood diaphragm is designed to yield before the SFRS, and is ed on an SFRS of reinforced concrete, masonry, steel, or wood-based structural systems other than wood shearwalls, the diaphragm shall be designed for a seismic force that is calculated using the lesser of: the RdRo of the SFRS or RdRo = 2.0

•

Where a wood diaphragm is designed not to yield and is ed on a wood-based SFRS other than wood shearwalls, the seismic design force in the diaphragm is multiplied by an over-capacity coefficient CDi equal to 1.2 times the ratio (Ci) of the factored resistance of the SFRS to the factored seismic shear load in the applicable storey.

•

Where a wood diaphragm is designed not to yield and is ed on a non-wood-SFRS, the seismic design force in the diaphragm is multiplied by the over-strength coefficient for the SFRS, determined on the basis of capacity design in accordance with the applicable material design standard.

For diaphragms that are designed not to yield, it is not required for design forces on the diaphragm to exceed the value of the factored seismic force at the storey determined using RdRo = 1.3.

Figure 13 Decision path for diaphragm design where IEFaSa(0.2) is 0.35 or higher

Clause 11.8.7 – Structures in low seismic zones For structures in low seismic zones, the only additional requirements are as follows: •

Diaphragm force increases are calculated using an over-capacity coefficient (ratio of the factored resistance of the SFRS to the factored seismic shear load) equal to 1.0. For wood diaphragms ed on wood shearwalls, this results in no increase in the diaphragm force.

•

Diaphragm chords, drag struts, connections and other load transfer elements in structures in low seismic zones are required to be designed using a force that is 20% higher than the seismic design load on the diaphragm. Similarly, connections and drag struts that transfer shear forces between the segments of the vertical SFRS and the diaphragm are to be designed for seismic loads that are at least 20% greater than the shear force that is being transferred. Parts of the diaphragm around wall offsets are to be designed for seismic loads that are at least 20% greater than the seismic design load of the offset SFRS. Forces used for the design of force transfer elements need not exceed the forces determined using RdRo = 1.3.

The suggested process of determining the seismic shear forces on shearwalls and diaphragms for a structure with more than 3 storeys can be summarized as follows: •

Determine the seismic storey shear forces Fi according to Articles 4.1.8.11 and 4.1.8.15 of the NBC. Note that the NBC specified diaphragm forces may be governed by other distributions (e.g. V/n, where n is number of storeys);

•

Obtain the cumulative factored seismic shear forces Vfi at each storey i as a sum of the factored seismic forces above that storey;

•

Design the shearwalls at each storey of the structure using the cumulative factored seismic loads and calculate the actual factored shear resistance of each storey, Vri;

•

Calculate the over-capacity coefficient for each storey Ci as a ratio of Vri and the factored load for that storey Vfi;

•

Determine the over-capacity coefficients CDi for diaphragms at each storey i as the lesser of Ci and 1.2;

•

Determine the seismic design loads for the diaphragms at each storey by multiplying the factored seismic shear load of each storey Fi by the seismic coefficient CDi;

•

Check the ratio of C2/C1 to be within the limits of: 0.9 < C2/C1 ≤ 1.2. If the ratio is not within the specified limits, either reduce or increase the actual resistance in one of the stories (i.e. Vr1 or Vr2).

Clause 11.8.8 – Seismic design requirements for shearwalls using gypsum wallboard In the 2014 edition of the Standard, additional restriction were introduced to limit the use of gypsum wallboard in combination with wood-based structural s. Research (Ceccotti and Karacabeyli, 2000 and 2002) showed that gypsum wallboard reaches its ultimate displacement at around 1% drift ratio which is far below the inter-storey drift limit of 2.5% for a building of normal importance. The CUREE experimental research of seismic performance of gypsum walls (McMullin and Merrick, 2002) showed that the maximum strength of gypsum walls occurred in a range of drift ratios between 0.75% and 1.25%. Based on test results on gypsum walls under monotonic and cyclic loading, it may be reasonably assumed that the drift level of 1% is a conservative estimate of the drift level and above this level gypsum wallboard should be assumed to be ineffective in providing lateral resistance.

Previous research (Ceccotti and Karacabeyli, 2000 and 2002) to justify the Rd factor of 2.0 when gypsum wallboards are used in combination with wood based s and the maximum percentage of total seismic forces resisted by gypsum wallboard in each storey is based on 4-storey buildings and below. Without further research it is required that in 5 and 6 storeys of wood frame construction the contribution of gypsum wallboard shall not be ed for in seismic resistance calculation.

Clause 11.8.9 – Load bearing walls constructed with gypsum wallboard only For load bearing walls sheathed with gypsum wallboard only, it is of concern that above a certain drift ratio the gypsum wallboards may become ineffective in providing lateral to the studs and preventing studs from buckling in the weak axis. It may be reasonable to assume that up to 1% that the gypsum wallboard may be considered effective, but above 1% the gypsum wallboard should not be assumed to be effective in providing lateral to the studs. In this case, the studs shall be designed to either resist the axial load effects given by Case 5 load combination in Table 5.2.4.1 (1.0D+1.0E+0.5L+0.25S), assuming the weak axis buckling is not prevented by the sheathing, or alternatively provide a secondary system for bracing the studs. One method to consider would be to design a secondary blocking system to prevent weak axis buckling.

References: Adams, N.R., 1987. Plywood shearwalls. APA Research Report 154, P.O. Box 11700, Tacoma, Washington. APA, 2002. Shear Wall Deflection and Predictive Equations. Research Report T2001L-65. APA – The Engineered Wood Association, Tacoma, WA. APEGBC, 2009 revised 2015. Technical and Practice Bulletin: Structural, fire protection and building envelope professional engineering services for 5 and 6 storey wood frame residential building projects (mid-rise buildings). Association of Professional Engineers and Geoscientists of British Columbia, Burnaby, BC, Canada, Applied Technology Council, 1981. Guidelines for the Design of Horizontal Wood Diaphragms. Berkely, Ca., U.S.A. Ceccotti, A., Karacabeyli, E., 2000. Dynamic analysis of nailed wood-frame shearwalls. 12th World Conference of Earthquake Engineering. Auckland, New Zealand. Ceccotti, A., Karacabeyli, E., 2002. Validation of Seismic Design Parameters for Wood-frame Shearwall Systems. Canadian Journal of Civil Engineering 29: 484-498. Countryman, David. 1952. APA Report 55. Lateral Tests on Plywood Sheathed Diaphragms. (Out of Print) APA – The Engineered Wood Association, Tacoma, WA. Dekker, J., Kuipers, J. and van Amstel, H. P. 1978. Buckling strength of plywood – Results of tests and design recommendations. Heron, 23(4):1-59 Dolan, J.D., Heine, C.P., 1997a. Monotonic tests of wood-frame shearwalls with various openings and base restraint configurations. VPI&SU Report No. TE-1997-001, Department of Wood Science and Forests Products, Virginia Polytechnic Institute and State University, Blacksburg, Virginia. Dolan, J.D., Heine, C.P., 1997b. Sequential phased displacement cyclic tests of wood-frame shearwalls with various openings and base restraint configurations. VPI&SU Report No. TE-1997-002, Department of Wood Science and Forests Products, Virginia Polytechnic Institute and State University, Blacksburg, Virginia. Johnson, A.C., Dolan, J.D. 1996. Performance of long shearwalls with openings. International Wood Engineerings Conference. New Orleans. Vol 2: p337-344. Karacabeyli E., Ceccotti A., 1996. Test Results on the Lateral Resistance of Nailed Shear Walls. Proceedings of the Int. Wood Engineering Conference, New Orleans, US, Vol.2, 179-186. Mi, H.Y., Ni, C., Chui, Y.H., Karacabeyli, E., 2006. Racking Performance of Tall Unblocked Shearwalls. ASCE Journal of Structural Engineering, Volume 132, Issue 1, pp. 145-152. McMullin, K.M., Merrick, D., 2002. Seismic Performance of Gypsum Walls: Experimental Test Program. CUREE Publication No. W15. Ni, C., Karacabeyli, E., Ceccotti, A., 1998. Design of shearwalls with openings under lateral and vertical loads. International Conference on Timber Engineering, PTEC’99 Rotorua New Zealand.

Ni, C., Karacabeyli, E., 2000. Effect of overturning restraint on performance of shearwalls. World Timber Engineering Conference, Whistler, British Columbia, Canada. Ni, C., Karacabeyli, E., 2004. Deflection of Nailed Shearwalls and Diaphragms. Proceedings of 8th World Conference on Timber Engineering, Lahti, Finland. Ni, C., Karacabeyli, E., 2007. Performance of Shearwalls with Diagonal or Transverse Lumber Sheathing. ASCE Journal of Structural Engineering, Volume 133, Issue 12, pp. 1832-1842. Ni, C., Chui, Y.H., Karacabeyli, E. 2012. Mechanics-based approach for determining the shear resistance of shearwalls and diaphragms. Proceedings of World Conference on Timber Engineering, Auckland, New Zealand. Popovski M., Karacabeyli, E., Ni C., Lepper P., Doudak, G. New seismic design provisions for shearwalls and diaphragms in the Canadian Standard for Engineering Design in Wood, Proceedings of the 42nd CIB-W18 Meeting, Zürich, Switzerland, 2009. Rainer, J.H., and E. Karacabeyli. 1999. Performance of Wood-frame building construction in earthquakes. Special Publication SP40. Forintek Canada Corp. (now FPInnovations) Sugiyama, H., 1981. The evaluation of shear strength of plywood-sheathed walls with openings, Mokuzai Kogyo (Wood Industry), Vol.36, Vol.7. Tissell, J.R., 1990. Structural shearwalls. APA Research Report 154, P.O. Box 11700, Tacoma, Washington. Tissell, J.R., Elliott, J.R. 2000. Plywood diaphragms. APA Research Report 138. P.O. Box 11700, Tacoma, Washington. Varoglu, E., Karacabeyli, E., Stiemer, S., and Ni, C. 2006. The Mid- wood shearwall system: concept and performance in monotonic and cyclic testing. Journal of Structural Engineering, American Society of Civil Engineers. Vo. 132, No. 9. pp. 1417-1425. Varoglu, E., Karacabeyli, E., Stiemer, S., Ni, C., Buitelaar, M., and Lungu, D. 2007. Mid- wood shearwall system: performance in dynamic testing. Journal of Structural Engineering, American Society of Civil Engineers. Vol. 133, No. 7. pp. 1035-1042. WFCM, 2001. Wood Frame Construction Manual for One- and Two- Family Dwellings, American Forest and Paper Association, Washington, D.C.

Clause 12

Connections

Clause 12.1 – Scope Clause 12 applies to split ring and shear plate connectors, bolts, dowels, drift pins, lag screws, timber rivets, truss plates, nails, spikes and wood screws. Other types of fastenings, including proprietary products, are used in wood design and construction. Joist hangers and truss plates are two examples addressed in this clause. Designers can also manufacturers of proprietary products for design information. Failure modes for wood connections are complex and vary from connection to connection. The Standard provides information on the number of fastenings required and the placement of individual fasteners in the connection.

Clause 12.2 – General requirements

Clause 12.2.1.4 – Shear depth Where a wood member is loaded at an angle to grain, the load results in a shear force component causing tension stress perpendicular to the grain of the member. Where the load is imposed by a fastening group, this stress is presumed to be greatest at the extreme boundary of the fastening group. Clause 12.2.1.4 therefore defines an “effective shear depth” de that must be used instead of the total member depth d when checking for member shear resistance. The dimension de is the perpendicular distance measured from the extremity of the fastening or group of fastenings to the loaded edge of the member.

Clause 12.2.1.5 – Service condition factor, KSF Table 12.2.1.5 lists the moisture service condition factors KSF for fastenings. The Table identifies two aspects of connection resistance that may be affected by moisture service conditions. First, wood tends to be weaker when wet, so wood connections used in wet service conditions generally have less resistance than those used in dry service conditions. Second, where fasteners are installed in wood that will dry in service, or where the moisture conditions change seasonally in service, shrinkage damage may occur. The effects of this shrinkage damage are included in the service condition factor. The service condition factors for timber rivets were modified in the 1994 edition of the Standard, reflecting data collected in the Forintek Canada Corp. (now FP Innovations) timber rivet research project (Karacabeyli et al., 1995). The KSF value for rivets installed in seasoned material and used in wet service conditions was reduced to 0.8 from 0.85, and values for rivet connections assembled in unseasoned material were added. Rivet withdrawal in wet service conditions should be avoided. In the 2014 edition of CSA O86 Changes were made to Table 12.2.1.5 of the Standard to align with the moisture content definitions in CSA Standard O141 – Softwood Lumber: Dry lumber – lumber that at the time of manufacture has a moisture content of 19% or less. Green lumber – lumber that at the time of manufacture has a moisture content greater than 19%.

This change does not affect the embedment strengths specified in the Standard. The current embedment strengths are based on an average moisture content of 15% which is consistent with the CSA O86 definition of dry service condition: Dry Service Condition – a climatic condition in which the average equilibrium moisture content of solid wood over a year is 15% or less and does not exceed 19%.

The diagrams shown below illustrate the connection details in Table 12.2.1.5.

A: External steel splice plates – load parallel to grain A: External steel splice plates – load perpendicular to grain

A: Internal steel splice plates – load parallel to grain A: Internal steel splice plates – load perpendicular to grain

B: Wood splice plates – load parallel to grain

B: Wood splice plates – load perpendicular to grain

B: Wood splice plates – Two rows parallel to grain not more than 127 mm apart with a common wood splice plate – load parallel to grain Max. 127 mm

B: Wood or steel splice plates – Multiple rows with separate splice plates for each row – load parallel to grain

Prior to the 2014 edition of the standard, the connection service condition factors – KSF - were applied to all connections. In the 2014 edition, brittle failure modes were introduced into the calculations for bolt and dowel connections. An experimental study conducted at Université Laval (2010) demonstrated that KSv and KSt for lumber are more appropriate than KSF to determine the resistance of row-shear and group tear-out failures in bolted or dowel connections. The use of tight-fit dowels with large steel plates in service conditions with significant moisture variations should be avoided. Where this condition cannot be avoided, the KSF values for drift pins and lag screws should apply.

Clause 12.2.1.6 – Load duration effects The cumulative time that a load is applied to a wood connection affects both the ultimate strength and the connection deformation (slip). Traditionally the ultimate strength of wood connection is assumed to have the same duration of load behaviour as wood itself, and the fastener parts (usually steel) are assumed to have no duration of load effects. Most load duration research has been done on sawn lumber. To be consistent with sawn lumber and other wood-based materials, the same load duration definitions and factors given in Table 5.3.2.2 of the Standard are applied to connection design. Load duration factors are unchanged from the 1989 edition of the Standard. In the 1989 edition there were major changes to the Connections section, as well as other sections, reflecting changes to load duration adjustment factors. These changes were explained in the 1990 commentary (Turnbull, 1990).

Clause 12.2.1.7 –Treatment factor, KT Some chemical fire retardants have been shown to cause wood strength loss, thus affecting the ductility and strength of connections in wood-based materials. CSA O86 requires that tests be carried out on the effects of fire retardants, or any potentially strength-reducing chemicals used with wood-based materials to determine an appropriate treatment factor. Additional information on the effects of fire-retardant treatment can be found in Clauses 5.3.4., 6.4.3 and 7.4.4 of the Standard and the corresponding Commentary sections.

Clause 12.2.2.3 – Group of fasteners The total strength of a connection with a small group of fasteners has been assumed traditionally to be the product of the strength of a single fastener times the number of fasteners. This assumption of equal loads on each fastener was the basis of early design standards. However, tests on large, heavily loaded connections have shown that each connection can have a total load capacity significantly lower than the strength of one fastener times the number of fasteners. A theoretical analysis by Lantos (1969), based on assumed elasticity of timbers end-connected by wood or steel side under tension, showed that the end fasteners in each row could carry above-average loads whereas the middle fasteners could carry below-average loads. This was confirmed by researchers such as Cramer (1968), and empirically by comparison of calculated values with the results of tests. Tables 12.2.2.3.4A and B list factors to be used for determining the resistance of connections made with two or more shear plates, split rings or lag screws. The group effects for other connections such as bolts and timber rivets are incorporated in the design equations. Relatively ductile ts made with truss plates, nails or spikes are considered less vulnerable to this effect. For these fastenings, group factors are not applied.

Clause 12.2.2.4 – Washers Clause 12.2.2.4 requires washers between wood and the head (or nut) of a bolt, or lag screw, unless a steel strap or plate is used. Standard cut washers are sufficient with bolts and lag screws used alone for lateral loads, but when used with split ring or shear plate connectors they require bigger, thicker washers to draw the connectors into full wood . Otherwise the standard washer may deform, leading to non-uniform bearing stresses and crushing of the wood. Table 12.2.2.4.1 specifies the minimum special washer sizes to be used with various applications. Malleable iron washers, ogee washers or steel plate washers are recommended for special applications.

Clause 12.2.2.5 – Net section In adjacent rows, bolt holes and connector grooves must be spaced so that stress concentrations in the wood do not further reduce member strength, or else the holes or grooves must be considered to occur at the same net section. Clause 12.2.2.5 defines how far fastenings in adjacent rows must be set apart in order not to be considered at the same critical section.

Clause 12.3 – Split ring and shear plate connectors

Clause 12.3.1 – General Shear plates and split rings are circular steel connectors that are installed in specially cut wood grooves to provide a large diameter connection near the wood surface. Design data apply to connectors dimensioned as in Table 12.3.1.A and placed in grooves accurately machined in the wood as dimensioned in Table 12.3.1.B (Note that some dimensions for fastenings are given in imperial units to reflect industry practice). Split ring connectors (2-1/2" or 4" diameter) are manufactured from hot-rolled carbon steel SAE 1010, meeting the requirements of the Society of Automotive Engineers Handbook. Shear plate connectors may be either pressed steel or malleable iron. The steel shear plates (2-5/8" diameter) are manufactured from hot-rolled carbon steel SAE 1010. The manufacturing reference for malleable iron shear plates (4" diameter) is to ASTM Standard A47, Grade 32510 or ASTM Standard A47M, Grade 22010 in CSA O86.

Clause 12.3.3 – Distance factors The location and relative position of connectors in affects connector resistance. Factors which must be considered in the geometry of a t made with split ring or shear plate connectors are: •

the angle of load to grain

•

the thickness of the to be connected

•

whether one or two faces of the member is to receive connectors

•

the distance between the connector and the edge of the wood member measured perpendicular to the grain

•

the distance between the connector and the end of the wood member measured parallel to the grain

•

the centre-to-centre spacing of connectors

•

whether the connector is in side grain or end grain

•

the group effect as discussed in Clause 12.2.2.3

Scholten (1944) established the original spacing requirements for shear plates and split rings. Erki and Huggins (1983) conducted tests on shear plates in glulam billets with steel side plates. These tests did not completely confirm the Lantos (1969) analytical models used to develop the group modification factor as given in Tables 12.2.2.3.4. To compensate for these results, the required loaded end distance for less than 130 mm thick was increased in 1984 to reduce the probability of tensile failure in multiple-connector ts. A further modification in 1989 specified that the increased end distance applied to 130 mm or thicker with 2 shear planes, or 65 mm or thicker with one shear plane in a t. In the 1995 edition, Table 12.3.3B was modified to reflect research on end distance effects in compression (Quenneville, Charron and Van Dalen, 1993). This research indicated that the end distance did not affect the connection capacity for loaded in compression. Based on this research, the JC factor in Table 12.3.3B is always 1.0 for loaded in compression. Minimum end distances for in compression are 70 mm for 2-1/2" split rings or 2-5/8" shear plates and 90 mm for 4" split rings or shear plates. For thinner , the end distances increase to 105 and 135 mm respectively. The research indicates that these end distances are conservative for the thinner . In the 2001 edition of CSA O86, the requirements for end distance in tension were modified in light of more recent research (Quenneville and Sauvé, 1998) that showed the strength of these connectors in loaded in tension parallel-to-grain, when related to the end distance, was not further influenced by the number of shear planes. The end distance requirements are now as they were in the 1984 edition of O86.

Clause 12.3.6 – Lateral resistance Scholten (1944) tested split ring and shear plate connectors in several species using three-member ts. His data showed that strengths parallel and perpendicular to grain were proportional to the applicable bearing strength of the wood. To make use of this data, mean regression lines were developed for each loading mode and each connector type and size. In the 1984 limit states design edition of CSA O86, parallel to grain connector capacities were determined by calculating lower fifth percentile values from the mean ultimate strength regressions, then dividing by 1.6 to adjust to the 1984 “normal load duration” strengths (Anonymous, 1982). In 1989, the duration of load adjustment was changed (Turnbull, 1990).

Perpendicular to grain compressive connector resistance values were calibrated directly from the Douglas fir values in earlier working stress standards (Anonymous, 1982). These resistance values are based on proportional limit strengths rather than ultimate strengths. The calibrated connector values were then adjusted for other species in proportion to compression perpendicular to grain strength ratios. Table 12.3.6C gives maximum values for shear plates in case the t is limited by the shear strength of the bolt.

Clause 12.4 – Bolts and dowels

Clause 12.4.1 – General Design requirements for bolted wood connections were revised significantly in 2009 to reflect an updated approach that included separation of design requirements for ductile and brittle failure modes. The scope was also broadened to include doweled connections in a broad range of applications. Since 1989, bolt design in CSA O86 has been based on ductile failure modes using modified European Yield Model equations. Prior to the 2009 edition, the Standard addressed brittle failure of the wood with multiple-bolted ts with group reduction factors based on t test data. The 2009 edition introduced a mechanics-based method to assess brittle wood failure modes associated with a group of bolts. The brittle failure modes include row shear and group tear-out under parallel-tograin loading, and splitting under perpendicular-to-grain loading. These modes are in addition to the ductile provisions in earlier editions. Other changes that were also made in 2009 are discussed below.

Clause 12.4.1.2. – Holes Because bolt holes are pre-drilled in wood before the ts are assembled, some over-sizing of the holes is required to allow insertion of the bolts without damage. The upper limit on hole oversize prevents excessive slip, and the lower limit prevents assembly damage. No over-sizing is identified for typical dowel connections since they are set tightly into prebored holes.

Clause 12.4.1.3 – Multiple Connections of 4 or more are designed by assuming each member is part of a 3-member connection.

Clause 12.4.2. – Material Design information in the Standard is limited to connections that use a metallic bolt or dowel whose material properties are referenced by a Canadian material design standard. Clause 12.4.4.3.3.3 includes yield strength properties for bolts and dowels conforming to the requirements of ASTM Standard A307, as well as a general formula to address other bolt and dowel specifications. The member thickness needs to be defined when calculating lateral resistance. In a connection with two wood , the member thickness is defined as the thickness of the thinnest wood member in the connection, or—if the bolt or dowel is recessed into the wood member—the smallest bearing length. Clause 12.4.2.4 covering wood to concrete connections, was first added to the 1994 edition of the Standard and applies to the design of sill plate and anchor bolt connections. Studies on wood-to-concrete bolted connections loaded parallel-to-grain (Stieda et al., 1997) and on connections loaded perpendicularto-grain (Quenneville and Mohammad, 1999), showed that it was conservative to use an embedment

strength of 125 MPa for concrete and an assumed thickness equal to the penetration of the anchor bolt. This design method is independent of a check on concrete or masonry capacity for the connector, using the appropriate concrete or masonry design codes. Figure 14 illustrates the interpretation of Clause 12.4.2.4.

Figure 14 Member thickness for anchor bolt connections

Clause 12.4.3 – Placement of fasteners in connections Fastener placement provisions were reassessed in 2009 based on the revised design approach considering both yielding and wood failure modes. The minimum bolt spacing, end and edge distances were originally based on work by Trayer (1932). More recent research quantified the effects of connection configuration on the resistance of multiple-bolted connections (Quenneville and Mohammad, 2000; Mohammad and Quenneville, 2001). Spacing between bolts in a row and end distance are now direct inputs into the wood failure design calculations. Minimum spacings of rows and of fasteners within rows remained nearly the same as in previous editions, with only a slight adjustment to perpendicular-to-grain fastener row spacing. The minimum loaded end distance parallel-to-grain was relaxed slightly since the design resistance can be adjusted to fit the selected parameters. Clause 12.4.3.3 prohibits use of a “single steel splice plate” when the distance between the outermost rows of bolts exceeds 125 mm. This is intended to apply to parallel-to-grain connections. Alternatively, the designer can specify two or more narrow parallel splice plates so that swelling and shrinking of the wood will not introduce stresses across the grain.

Clause 12.4.4 – Lateral resistance The 2009 bolt design procedure requires the designer to address the following applicable design cases, each of which represents more than one mode of failure: •

Yielding failure modes (parallel to grain, perpendicular to grain and at an angle to the grain),

•

Parallel-to-grain wood failure modes (row shear, group tear-out, net tension)

•

Perpendicular-to-grain wood failure modes (splitting, net shear); and

•

Combined parallel- and perpendicular-to-grain wood failure modes for loaded at an angle to the grain.

Yielding design applies to all bolted connections. Parallel-to-grain design cases include spliced with wood or steel side plates. Perpendicular-to-grain examples include bolted hangers, vertical truss webs and ledgers. Yielding failure modes achieve the highest lateral bolted t capacity up to the point where the fastener yields and/or the wood fails in bearing at the t. The next highest parallel-to-grain capacity is reached where a connection fails in “row shear” wood failure mode. The least efficient parallel-to-grain case is the “group tear-out” wood failure mode.

Clause.12.4.4.3 – Yielding resistance Multi-mode yielding equations (the so-called European Yield equations) were first introduced into CSA O86 in 1989. The yield equation resistances were derived from a method proposed originally by Johansen (1949) and subsequently refined by Larsen (1973). Larsen’s prediction equations assumed that a dowel connection (including bolts) could fail by full bearing (i.e., crushing) in one or more wood (for bolts with low length to diameter ratios), or by bending of the bolt (for bolts with high length to diameter ratios) accompanied by local wood crushing near the member interfaces. In the 1989 edition of CSA O86, lateral resistance values were based on modified forms of Larsen’s equations by Whale et al. (1987) and presented in tabular format in the body of the Standard. In the 1994 edition, the Whale et al. (1987) equations were presented directly in the Standard. Also, there were revisions to the embedding strengths of wood, the bolt yield strength in bending and the bolt-bearing strength of steel side plates. One of the results of the separation of yielding failure from wood failure in the 2009 design revisions was an increase of the resistance factor for yielding failure, φy, from 0.7 to 0.8. The revised value reflected the ductility as well as the lower variability of bearing resistance (5 to 10% CoV) compared with wood failure modes. For single shear connections, the minimum of equations a, b, d, e, f, and g is used to determine unit yield resistance. For double shear connections, unit yield resistance is calculated as the minimum of equations a, c, d or g. The lateral yield resistance is proportional to unit yield resistance, the total number of fasteners in the t, and the number of shear planes. These equations have not changed from previous editions, except for minor revisions; i.e., incorporation of the 0.8 factor used to convert the f1 to a Standard Term duration of load basis directly in the embedment strength (it was separate in previous editions). In the 2009 edition, the wood embedment strengths were modified for angle-to-grain loading using the Hankinson formula as well as material modification factors prior to being input into yield equations. This is more representative of material behaviour, compared with the earlier approach. The embedment strength for steel is defined as 3 times the ultimate tensile resistance, multiplied by the ratio of the steel resistance factor (0.67) to the resistance factor for yielding failures (0.8). The embedment strength of the steel plate was modified so that the bolt-to-steel plate bearing resistance

obtained in O86 would yield identical results to the ones in S16.1 for the same failure mode. The designer still has the responsibility to check the other steel plate failure modes using S16.1. The yield equations do not consider wood splitting and shear failure modes formerly addressed by group and spacing factors: JG, JL and JR. These factors were conservative estimates based on research by Massé et al. (1988) and Yasumura et al. (1987) and were replaced by the failure-specific design equations for wood shear, tear-out and splitting modes discussed below.

Clause 12.4.4.4 – Parallel-to-grain row shear resistance Row shear failure is the brittle fracture of a wood plug along a row of bolts under tension loading parallelto-grain (see Figure 12.4.4.1). The resistance factor for wood failure modes φw is 0.7. In the design process, row shear resistance is calculated for each fastener row “j” in each member “i,” and the minimum row shear resistance is used to determine the member resistance. The overall t resistance is the lesser of the sum of the side plate resistances or the main member resistance. Since the minimum row resistance is taken for each member, the overall resistance can be conservatively estimated using the shortest row. The shear resistance of each row in each member, defined as PRij, is given by the equation: PRij = 1.2 fv Kls t nc acr i

(Eq. 15)

To for the two shear-resisting planes on each side of a bolt row, the shear resistance would be based on 2 times the member thickness and the connection length, without any further reduction. A 0.6 factor is applied to the row shear resistances resulting in the 1.2 factor in the equation above. Specified longitudinal shear strengths fv are given in the appropriate tables of the Standard. Specified shear strengths for sawn lumber (Clause 6.5.5) are derived from 5th percentile values reduced by factors for size and load duration effects, and further reduced to calibrate to full-size lumber test data. These values include consideration of the grading rule allowances for small checks and splits. The 0.6 factor applied to the shear resistance converts the specified shear strength for beams to a shear strength suitable for bolted connection design and is consistent with connection test data (Quenneville and Mohammad, 2000; Jorissen, 1998). In cases of large splits or checks near a t, connections may require reassessment and repair. Kls is a factor to adjust resistances for the more severe case, where a wood side member is loaded on one surface. This case was assigned a Kls value of 0.65, based on the work of Mohammad and Quenneville (2001). The critical length acri is the minimum continuous wood length resisting shear along a given row; i.e., the loaded end distance aL or the row spacing SR. In the case of connections with sloping ends, where each row may have a different end distance, or connections with different numbers of fasteners in a row, the design method allows the calculation of different row shear resistances. It should be noted that there is no design advantage to having different end distances and in-row bolt spacing since the lowest of the two variables govern row shear failure.

Clause 12.4.4.5 – Parallel-to-grain group tear-out resistance Group tear-out failure is the brittle fracture of a wood block around the outside of a connection with more than one row of bolts under tension loaded parallel-to-grain (see Figure 12.4.4.1). The resistance factor for wood failure modes φw is 0.7. In design, the group tear-out resistance in each member “i” is determined as the sum of the average row shear resistance for the first and last row, and the tensile resistance of the net area in the connection. The overall group tear-out resistance of the connection is the lesser of the sum of the tear-out resistance of the side or the tear-out resistance of the main member in the connection. This approach has been corroborated with experimental results Mohammad and Quenneville (2001). For the row shear component of the equation, the resistance is divided in half because the group tear-out mode involves only the outermost failure planes in the group. The resistance for the two outer rows is averaged in this calculation. The critical perpendicular net area, APGi, is the minimum wood section resisting tear-out of a group of bolts in a given member; i.e., the area between the outer two rows (1 and nR) in mm2 excluding all bolt holes. Note that the group tear-out failure mode does not normally occur where the member is in compression.

Clause 12.4.4.6 – Net tension resistance As in previous editions of CSA O86, connection design includes a check on tension parallel-to-grain resistance of wood at net cross section (see also 6.5.9 and 7.5.11).

Clause 12.4.4.7 – Perpendicular-to-grain splitting resistance Splitting failure is the brittle fracture of wood across a connection with bolts under tension loading perpendicular-to-grain (see Figure 12.4.4.1). The resistance factor for wood failure modes φw is 0.7. In the design process, the splitting resistance in each member “i” is determined using a fracture mechanics-based formula taken from Eurocode 5. The overall splitting resistance of the connection is the lesser of the sum of the splitting resistance of the side in the connection or the splitting resistance of the main member. This approach was corroborated by experimental results from Canadian sources (Mohammad and Quenneville, 2001; Massé, Salinas and Turnbull, 1988) and European (Leijten and Jorissen, 2001). Research in this area is ongoing. A note in the Clause 12.10.2.1.4 refers designers to the check net section shear based on the dimension de shown in Figure 12.2.1.4. This check is similar to design for splitting resistance in that it uses the effective depth perpendicular-to-grain to determine member resistance; however, it uses material-specific shear values while splitting design is applied identically to all wood species and product types. Net shear design is used to determine if the member is adequate to resist the beam shear at the connection location and protects bending from shear failures induced by stresses at the section weakened by the connection. Design for splitting protects against connection failure due to cracks forming in the wood around bolts.

Bolted connection detailing observations The bolt design procedures reinforce the importance of connection detailing in optimizing the resistance of the connection. Row shear design, for example, provides no advantage to a connection with long end distances and narrow parallel-to-grain spacing between bolts, since the smaller of the two distances will govern. Where connections are governed by the row shear failure mode, increasing the number of bolts in the row could change the mode to group tear-out. This will reduce the effectiveness of the bolts used in the connection, unless the connection is reconfigured. For perpendicular-to-grain loading, splitting resistance is sensitive to the distance of bolts from the loaded edge of the member. For fasteners loaded perpendicular to the grain, capacity is maximized where the effective depth of the member is maximized. It is always recommended to avoid eccentricity loading through bolt design; e.g., by ensuring that side wood in three-member connections are the same thickness.

Clause 12.5 – Drift pins Drift pins were first introduced in 1989. Since drift pin lateral resistance is based on bolt formulas, design provisions for drift pins were modified in accordance with changes to bolt design in the 2009 edition. Design information in the Standard is based on 16 to 25 mm (5/8" to 1") diameter drift pins made from round mild steel conforming to ASTM Standard A307 or CSA Standard CAN3-G40.21. Drift pins are round steel dowels that do not have heads and are not threaded. The ends are tapered or shaped so that the pin may be easily driven into pre-bored holes with minimum damage to the wood. Drift pin holes are pre-bored with the aligned and held in place. With hole alignment thus assured, the hole diameter is slightly undersized to give a tight fit when the pins are driven in. This provides friction, drawing the tightly together during assembly. The holes are between 0.8 mm and 1.0 mm smaller than the diameter of the drift pin. Drift pins must be slightly shorter than the combined depth of each pair of to be pinned. This prevents interference between the pins in successive layers as they are tightened together. Drift pin size and spacing rules are based on a 90° t using four 20 mm pins to secure pairs of 200 mm wide timbers, a typical cribwork t as shown in CSA O86 Figure 10.5.5.

Clause 12.5.6 – Lateral resistance With the addition of wood failure modes to bolt design, this failure mechanism was addressed directly in drift pin connections. As a conservative measure, due to uncertainties associated with field installation, the lateral resistance for drift pins was set at 0.6 times that of bolts. The size and placement provisions for drift pins are also different from bolt design provisions.

Clause 12.6 – Lag screws

Clause 12.6.1 – General Lag screws (or lag bolts) are generally used where through-bolts are not practical. Lag screws may be loaded laterally or in withdrawal, in side or end grain, with wood or steel side plates. End-grain loading in withdrawal is discouraged because of the inherent weakness and susceptibility to splitting of this type of t. ANSI/ASME B18.2.1 is the North American standard for lag screw dimensions. This standard does not address requirements for steel properties of lag screws. In the 2014 edition of CSA O86 a new provision was introduced that requires that the steel used for lag screws must meet or exceed the properties of SAE J429 Grade 1. Product evaluation report for compliance to CSA O86 (reliability-based for limit states design) may be available from CCMC for proprietary self-drilling screws. Otherwise, the design engineer may consider test results based on the ICC-ES AC233 as ing evidence for the use of such alternate fastener with the design guidelines given in 12.6.

Clause 12.6.2 – Placement of Lag Screws in ts In general, the same minimum spacing, end distances, edge distances and net sections for bolts apply to lag screw ts. A note was introduced in the 2014 edition of the standard that allows 10 mm and smaller diameter lag screws, loaded primarily in withdrawal, to be inserted without a lead hole where the relative density is less than or equal to 0.5. This provision is ed by lag screw research, Newlin and Gahgan (1938), and by recent studies on self-drilling screws by Blass et al (2006) Pirnbacher et al (2009) Abukari et al (2012), Kennedy et al (2014), and by field experience. The report by Newlin and Gahgan (1938), as well as most of the other studies, are based on withdrawal tests of only the threaded shank without studying the effects potential splitting and long-term performance of the unthreaded shank without predrilling. The intent of the note is to allow installation of lag screws without lead holes for the threaded shank, but to keep the lead hole requirements for the unthreaded shank. In previous editions of the Standard it was required that the threaded portion of the lag screw be inserted by turning the screw with a wrench; not by driving. In the 2014 edition of the Standard other types of tools, including power tools, would be allowed. In all cases, the lag screws must be inserted by turning, not by driving. Over tightening the lag screw will reduce the withdrawal resistance of the screw. If the lag screw is over tightened it will not have the withdrawal resistance specified in the Standard.

Clause 12.6.3 – Penetration of lag screws The maximum penetration lengths of lag screws were originally recommended by Newlin and Gahgan (1938). Increasing the length of penetration of a lag screw beyond a critical point does not increase either the strength or stiffness of the connection. Clause 12.6.3.2 gives the maximum lengths of penetration which are allowed to be used in the calculation of withdrawal resistance. The penetrations are a function of shank diameter and species. For lag screws loaded laterally, the same limits on penetration into the main member may be used. In the 2001 version of O86, a clause was added to impose a 5d lower limit on the penetration of lag screws. It is important that the tensile strength of fasteners at the net (root) section is not exceeded under loads of short or long duration. Clause 12.6.5.2 was introduced in the 2014 edition of the Standard requiring that the tensile resistance of the lag screw have a minimum specified strength of 410 MPa.

Clause 12.6.5 – Withdrawal Resistance In previous editions of the Standard, the basic resistance values for lag screws loaded in withdrawal from side grain were developed using proprietary unpublished data from tests on five species (redwood, white pine, Douglas fir, southern yellow pine and white oak) with seven sizes of lag screws and six replications for each combination. This information was presented in table format in the Standard (Anonymous, 1982). For the 2014 edition of the Standard, a literature review was undertaken to compare different equations used to determine the withdrawal resistance of lag screws. Five different equations, converted to the CSA format were considered:

NDS Working Stress in CSA format NDS Limit States converted to CSA format McLain (1997) Working Stress in CSA format McLain (1997) Limit States converted to CSA format Morrison et al (1982) specified

𝑃𝑃𝑟𝑟𝑟𝑟 = 52 𝑑𝑑𝐹𝐹 𝐺𝐺 2 𝐿𝐿𝑡𝑡 (𝑁𝑁)

𝑃𝑃𝑟𝑟𝑟𝑟 = 62 𝑑𝑑𝐹𝐹 3/4 𝐺𝐺 3/2 𝐿𝐿𝑡𝑡

𝑃𝑃𝑟𝑟𝑟𝑟 = 59 𝑑𝑑𝐹𝐹 0.82 𝐺𝐺 1.77 𝐿𝐿𝑡𝑡 𝑃𝑃𝑟𝑟𝑟𝑟 = 87 𝑑𝑑𝐹𝐹 0.61 𝐺𝐺 1.35 𝐿𝐿𝑡𝑡

𝑃𝑃𝑟𝑟𝑟𝑟 = 88 𝑑𝑑𝐹𝐹 0.75 𝐺𝐺 1.5 − 56

A review of the equations showed that the McLain 1997 Working Stress format converted to the CSA format was the most approriate equation to use for lag screws. In the 2014 edition of the Standard, the same equation is used for the withdrawal resistance of lag screws and wood screws.

Clause 12.6.6 – Lateral resistance In early Canadian limit states design Standards (1984, 1989) lag screw ts were governed by either strength or serviceability limit states, with the serviceability limit state used to prevent slips exceeding 0.8 mm at specified loads. In the 1994 edition of the Standard, the European Yield Model, (Larsen, 1973, Whale et al., 1987) was adopted for lag screws and the slip criteria was moved to the Appendix. The European Yield Model provides an estimate of the ultimate capacity of a lag screw connection. In the 2001 edition of the Stanard, a factor was added to the Pr and Qr equations to take into the effect of reduced penetration and resulting reduced capacity. The full lateral load capacity is obtained for a penetration of 8d and is reduced by 37.5% at a penetration of 5d. Also, the embedment strength of the steel side plate has been modified so that the lag screw-to-steel plate bearing resistance obtained in O86 yields identical results to the ones in S16.1 for the same failure mode. The factor of 3 used to determine the plate embedment strength is taken from CSA S16. The designer still has the responsibility to check the other steel plate failure modes using S16. Lateral deformation of a lag screw connection under specified loads can still be estimated using Appendix A.12.6.3.3. In cases where stiffness must be estimated for analysis, or where stiffness should be controlled, the designer should use the Appendix. As well, research (Newlin and Gahagan 1938) has indicated that lag screws loaded laterally tend to reach a point of permanent deformation at about 0.8 to 1.0 mm. This limit should be avoided to prevent permanent deformation. The same yield equations (Whale et al., 1987) are used for lag screws and bolts. Lag screws are only considered for single shear connections and the equations are presented accordingly. The wood embedding strength is the same as for bolts.

Lateral slip resistance values (in Appendix A.12.6.3.3) were based on mean published values by Newlin and Gahagan (1938) and unpublished test data. The tests indicated that slip resistance values were proportional to shank diameter, depth of penetration in the main member and wood relative density. Unpublished tests indicate that steel side plates increase t stiffness by a factor of 1.5 over wood side plates. Therefore a JY factor of 1.5 is applied when checking lateral slip resistance parallel to grain for ts with steel side plates.

Clause 12.7 – Timber rivets (also known as glulam rivets) Timber rivets are high strength steel nails originally developed to connect glued-laminated timber through pre-drilled steel plates. Timber rivet ts offer flexibility of design and efficiency of load transfer in wood. In the 1994 edition, additions to the timber rivet clause included the following: design values for use in solid sawn timber, withdrawal values, and design values for use with plates thinner than 4.8 mm. These additions to rivet design values were based on extensive tests carried out by Karacabeyli et al. (1995). Both lateral and withdrawal tests were undertaken for a variety of species under different service conditions.

Clause 12.7.1 – General The rivets have a flattened-oval shank section and a wedge-shaped head. Dimensions are standardized as illustrated in Figure 12.7.1.1. When driven through the pre-drilled steel plate and into the timber beyond, the tapered heads wedge tightly in the holes. This prevents any movement of the heads within the plate, thus increasing both strength and stiffness of the t. The flattened shank section is 3.2 by 6.4 mm and must be driven with the wider dimension parallel to the wood grain. This minimizes wood splitting and optimizes the lateral resistance, whether loads are parallel or perpendicular to grain. Timber rivets are supplied hot-dip galvanized, whereas the plates are usually custom-made and may be galvanized depending on the service conditions. Clause 12.10.7.1.4 states that the minimum thickness of steel side plate to be used with timber rivet connections is 3.2 mm. This is reduced from the previous minimum of 4.8 mm. Side plates thinner than 6.3 mm result in less fixity and consequently rivet bending capacity is reduced. The research (Karacabeyli et al., 1995) has confirmed that the side plate factors of 0.9 and 0.8 are applicable to 4.8 and 3.2 mm plates respectively. Corrosion may be a consideration where thinner plates are used (the minimum thickness of un-protected steel plates subjected to corrosion is 4.5 mm). Although the Standard does not provide a limit on the maximum plate thickness, plates that are too thick decrease the penetration of the rivet into the wood member and reduce the wood capacity of the connection. Table 12.7.2.3 has been calculated assuming a side plate thickness of 6.3 mm. These tabular values could be reduced, particularly for smaller rivets, if thicker side plates were considered. Glued-laminated timber is manufactured using dry wood. Therefore, it is unlikely that timber rivets will be installed in wet glulam material, reducing the possibility of splitting in service at timber rivet connections. Moisture content of solid sawn material is not as consistent, and it has somewhat decreased resistance to splitting, leading to some precautionary clauses in the Standard. Clause 12.7.1.5 includes a note with guidance on installation procedures for timber rivet connections. Rivets should be installed progressively from the outside of the connection to the interior of the connection. Experience has shown that this procedure helps to prevent splitting. Clause 12.7.1.9 includes provision for the seldom-used condition where rivets in a two-sided connection (plates on both sides of a member) are long enough to penetrate beyond the midpoint of the member. In this case, rivets on both sides are required to be spaced apart as though they were all driven from one side. Capacities are calculated on the more conservative basis of the spacing between rivets from both

sides at the centre of the member. This spacing requirement was adopted in lieu of specific research information. In no case is the rivet penetration permitted to exceed 70% of the wood member thickness. Clause 12.7.1.10 was added to minimize the possibility of splitting where timber rivets are installed in wet wood. Tests (Karacabeyli et al., 1995) indicated that splitting was not a problem for rivets installed in green material that was subsequently allowed to dry, where the width of the rivet group perpendicular to grain was restricted.

Clause 12.7.2 – Lateral resistance Design of timber rivet ts was based primarily on Foschi and Longworth (1975). The theory describes and quantifies two main modes of failure, in the rivets and in the wood. The ultimate capacity is limited either by the rivet shank bending combined with localized wood crushing at the rivets, or by the wood strength (shear or tension) surrounding the rivet group. Due to this variety in failure modes, the design of timber rivet ts is a trial-and-error process, optimizing the ts in relation to the sizes of the timbers and the forces to be resisted. In the case of the rivet bending failure mode, the formulae for basic parallel and perpendicular design resistance values PY and QY were obtained by dividing the mean theoretical values from Foschi and Longworth’s work by 1.6 to for variability of rivet yielding values. Because the original rivet steel specification had later been modified to eliminate the possibility of hydrogen embrittlement in environments of high temperature and humidity, a slightly softer rivet resulted. Thus the capacities governed by nail bending were multiplied by a factor of 0.88. These values were further modified in 1989 to accommodate load duration calibrations (Turnbull, 1990). In the 1994 Standard, the KSF and KT factors were applied to rivet capacity. In the 1989 edition, Karacabeyli and Fraser (1990) compared the lateral strengths of timber rivets in Spruce glulam versus Douglas fir glulam, loaded parallel and perpendicular to grain. Based on this data, a species factor “H” was adopted in the 1989 Standard to accommodate timber rivet design in SprucePine glulam. In 1994, the “H” factor was renamed the material factor, since it was expanded to include the various species groups of sawn timbers, as well as glulam. Material factors for timbers were established based on tests in Douglas fir, Hem-Fir and Spruce-Pine-Fir solid sawn (Karacabeyli et al., 1995). To simplify the design, the same “H” factor is applied to all strength criteria parallel and perpendicular to grain. Due to the complexity of equations for lateral strength resistance parallel to grain, the basic unfactored values for a wide range of rivet lengths, rivet spacings and sizes of rivet groups are given in Table 12.7.2.3 of the Standard. For values beyond the range of the table, see Appendix A.12.7.2.3. In the 1989 and earlier editions, tables included both rivet and wood capacity values parallel to grain. In 1994, this approach was changed so that the values in Table 12.7.2.3 were based on wood capacity only; rivet capacity must be calculated separately, providing a more accurate result.

Clause 12.7.3 – Withdrawal resistance Design of timber rivets for withdrawal loads was added in the 1994 edition. Design values are based on Karacabeyli et al., (1995). Rivet withdrawal tests were undertaken to investigate the effect of species, condition of wood when fabricated, density and waiting period before testing. Tests were for short duration loads and limited to material that was dry at the time of testing. The design values for timber rivets loaded in withdrawal are conservative estimates of the data, and are limited to dry service conditions for Short Term and Standard Term load durations.

Withdrawal values are a function of the number of rivets in the connection and the length of penetration of the rivets. For simplicity, only two unit withdrawal values are given: one for glulam and one for sawn timber. The values are independent of the species used. The withdrawal values were adopted for both short and standard duration loads; therefore, the duration factor is not applicable to these design values.

Clause 12.8 – Truss plates

Clause 12.8.1 – General Truss plates are toothed metal-plate connectors used to assemble pieces of dimension lumber into roof and floor trusses. Truss plates are produced by punching light gauge galvanized steel (normally 16, 18 or 20 gauge) so that teeth protrude from one side. The pattern of teeth in the plate varies by manufacturer. In the 2014 edition of Standard, new provisions were introduced to allow truss plates to be used in structural composite lumber to structural composite lumber connections and sawn lumber to structural composite lumber connections. Truss plates must be produced from galvanized sheet steel conforming to the standards specified in 16.4.1.2. Because the sheet steel used for truss plates is relatively thin, corrosion of the plates could reduce its performance. The Standard contains restrictions on use in corrosive environments and, in particular, prohibits use in the highly corrosive environment where fire retardant treated wood is exposed to moisture. Truss plates are installed at every connection of a light frame wood truss. Truss plate sizes vary considerably depending on the loads which must be transferred at the connection. Truss fabrication is carried out at a plant where the truss plates are embedded into the lumber with either a hydraulic press or a roller.

To meet the design requirements of the Standard, the plates are to be installed in the following manner: •

the connections are tight fitting with identical truss plates placed on opposite faces so that each plate is directly opposite the other

•

the truss plates are not deformed during installation

•

the teeth are normal to the surface of the lumber

•

the teeth are fully embedded in the wood so that the plate is tight to the wood surface; however, the plate is not embedded into the lumber deeper than half of the plate thickness

•

the lumber beneath the plate complies with Appendix G of TPIC-2011.

The last bullet refers to an Appendix in the Truss Plate Institute of Canada’s Truss Design Procedures that provides minimum quality control requirements for lumber covered by truss plates. Truss plates must be tested in accordance with CSA Standard S347, Method of Test for Truss Plates used in Lumber ts. Test results for commercially used plates are listed in the Registry of Product Evaluations published by the Canadian Construction Materials Centre (CCMC), Institute for Research in Construction, Ottawa. The evaluation reports also include resistance values derived from the tests. The resistance per tooth varies little within a given manufacturer’s range of plate sizes, and usually the larger plates are tested, resulting in a generally conservative strength value per tooth for the smaller plates.

Clause 12.8.2 – Design Due to the proprietary nature of truss plates, specific resistance values for truss plates are not given in the Standard. Instead, the Standard outlines procedures for deriving resistance values based on required tests. The resulting resistance values are normally used by the truss plate manufacturer who usually has the responsibility for truss design. Truss plates transfer forces in tension or shear or both. Those forces may act in various directions depending on the geometry of the t. Truss plates are also used in compression connections, mainly to keep the truss together and in alignment during handling, erecting and for stress reversal situations. The main compression forces are transferred by wood-to-wood bearing for which tightly fitted connections are essential. With truss plate connections that are relatively short in the direction of the load, failure occurs when the teeth yield and then withdraw from the wood. Failures of this type are governed by the “lateral resistance of the teeth”. With longer plates, the failure mode can change to tension failure across the most critical net section of the plate. This failure mode is resisted by the “tensile resistance of the plates”. A third failure mode involves shear of the plate, usually indicated by diagonal buckling of the plate followed by tooth withdrawal or tensile tearing of the steel along the shear line. This is governed by the “shear resistance of the plate”. CSA O86 gives procedures for converting the test results into design values. In accordance with the Standard, truss plates must be designed so that the following criteria are satisfied: 1. Factored ultimate lateral resistance of the teeth, Nr ≥ factored lateral load Nf 2. Factored tensile resistance of the plate Tr ≥ factored tensile force Tf 3. Factored shear resistance of the plates Vr ≥ factored shear force Vf 4. Lateral slip resistance Nrs ≥ specified lateral load Ns

Clause 12.9 – Nails and spikes This clause provides lateral and withdrawal resistance values for common nails, spiral nails and spikes described in CSA Standard B111. Other nail-type fastenings not so described (such as nails of different dimensions) may also be used if ing data are available. In the 2009 edition of CSA O86, lateral design of nailed connections was converted from a tabular approach to the same European Yield Model equation approach that is used for other dowel-type connectors (bolts, lag screws and wood screws).

Clause 12.9.2 – Connection configuration The spacings for round wire nails were based on Ramos (1960), and crack propagation experiments by Lau and Tardiff (1987). These minimum requirements were established to limit strength reductions as a result of splitting of the wood. The tabulated minimum nail spacings parallel to grain were set at double the maximum crack lengths caused during driving, as observed by Lau and Tardiff (1987). Different species were assigned different spacing requirements, based on the experimental results. Because very dry wood has a greater tendency to split when nails are driven into it, spacing requirements for nails and spikes are provided for connections in solid wood at a moisture content greater than 10%. This is not intended, however, to restrict application to products such as plywood or OSB that are typically lower in moisture content than solid wood at the same ambient conditions. Figure 12.9.2.1 of the Standard illustrates how the spacing rules are applied to wood-to-wood connections connected at various angles, and shows how extra nails may be added on the diagonals between nails (Clause 12.9.2.1).

Clause 12.9.2.2 was revised in 2009 to provide more complete information on member and penetration thicknesses for single- and double-shear connections. Clause 12.9.2.3, which relaxed the minimum side plate thickness requirement for plywood or OSB compared to lumber, was dropped in the 2009 edition because the minimum side member thickness in 12.9.2.2 was revised for all wood materials, and side member thickness is input into the yield equations.

Clause 12.9.3 – Connection design In the first Canadian LSD Standard (1984), nail connections were designed for strength and serviceability limit states, with the serviceability limit state intended to prevent slips exceeding 0.4 mm at specified loads. The serviceability requirements were removed in the 1989 edition of the Standard. The design approach was based on yield theory, i.e. connection strength estimation based on the yielding moment of the nail and embedding strength of the wood, although it did not include EYM equations for different yield failure modes. Concurrently with the removal of the serviceability requirements in 1989, a formula adapted from a Eurocode 5 was added as Appendix A.12.9.3.2 to permit designers to estimate the in-service deformation of nailed ts. The formula was limited to specified loads of up to 1/3 of the ultimate load. Designers are advised to use this in cases where stiffness must be estimated for analysis, or where t slip should be controlled. When the EYM equations were added in the 2009 edition, the embedment strengths were generally consistent with other fastener types with some further adjustments as discussed in the next section.

Clause 12.9.4 – Lateral resistance At the same time that the EYM equations were introduced in CSA O86, a higher resistance factor (φ) was adopted for lateral resistance of nailed connections. The resistance factor was raised to 0.8 to reflect ductile failure and to be consistent with the approach used for the yield equations in other parts of the Standard. Introduction of the yield equation approach in Clause 12.9 completed the conversion of design for doweltype fasteners to a common theoretical basis, using the same approach as for bolts and screws. At the same time, there was a need to minimize changes to traditional nailed connection design, based in particular on experience with Douglas fir ts. To minimize differences with traditional designs for nailed connections, the concept of the so-called ‘string’ effect under lateral load was incorporated in the calculation method. The string effect is an additional lateral strength component provided by the withdrawal resistance of the fastener shank in the wood when the fastener undergoes bending deformation. In Eurocode 5 the string effect is ed for by increasing the lateral resistance of a nail connection by a component equal to a portion of the withdrawal strength of the nail. A different approach was adopted for the 2009 edition of the Standard. In the 2009 edition of the Standard, the wood embedment strength of the wood member where a plastic hinge is formed is increased. The embedment strength is higher where the failure mode is in fastener yielding compared to where the failure mode is in wood bearing. This increase is a function of the wood specific gravity, on which the withdrawal strength of a nail depends, and was calibrated to connection test data. To that end, the f3 embedding strength, incorporating the string effect, is used in place of the f2 in equations (d), (e) and (g) of Clause 12.9.4.2. This approach leads to a simpler design calculation than that required by Eurocode 5.

Nailed connection resistances were confirmed by test results for a range of wood density and fastener sizes (Smith et al 2001). The resulting factored resistances for wood-to-wood connections fall within a range of 20% of previous values, and are close to previous values for connections with steel side plates thicker than 20 gauge steel. For light gauge steel side plate, t resistances are slightly reduced from previous values. Embedment values for structural (plywood or OSB) side plates were adopted from European research (Whale et al, 1987). Where resistance values in earlier editions of CSA O86 were based on assumptions about member thicknesses, and referenced to the weakest member in the connection, these assumptions were no longer necessary using the equation-based approach. Provisions for double shear were also no longer required since they were accommodated directly in the equation-based approach. For nails or spikes installed in side grain, connection resistance is assumed to be the same whether loaded parallel or perpendicular to the grain. Where nails are installed in end grain, only 67% of the side grain resistance is specified. Toe-nails are partially resisted in the end grain of the wood. Therefore, they have only 83% of the lateral resistance of a connection installed in the side grain. Factors for service conditions are ed by research by Brock (1957) and by Welchert and Hinkle (1966). There are two other factors specific to nail and spike connections: JB for nail clinching and JD for diaphragm construction. In a two-member nailed connection, lateral resistance may be increased by 60% if the nail is clinched on the far side. (Crandall, 1959). The factor for nail clinching does not apply to three member connections, since the increase in resistance is covered by the JS factor. The calculated values in the Standard are near-minimum resistance values. In diaphragm construction, many nails are used and the effects of variability of ultimate resistance are reduced. Therefore the resistance of nails used in diaphragms may be increased by 30%.

Clause 12.9.5 – Withdrawal resistance The driving of a nail jams bent or broken wood fibres between the nail and other fibres. The fibres are forced aside as the nail enters the wood, exerting a squeezing action on the nail. It was commonly accepted that most of the withdrawal resistance of a nail was due to the jamming and wedging action of bent or broken fibres. Frictional withdrawal resistance depends on the bearing of these fibres on the nail surface. Research has shown that the withdrawal resistance of a nail is a function of the length of penetration into the main member, the diameter of the nail and the specific gravity of the wood. With time, the displaced fibres around a driven nail relax and the withdrawal resistance is greatly reduced. This effect is more pronounced for nail connections subject to long term loads and cyclic moisture conditions. Therefore, withdrawal loading for nails is only permitted for short duration loads (wind and earthquake). An equation was introduced in the 2001 edition to relate the nail withdrawal resistance to the nail diameter and the material’s relative density (RD), based on McLain’s model (McLain, 1997). The test strengths were described using a normal distribution (Smith, 2000) and the average COV were assumed to be 30%. With these assumptions, the factor to convert the mean strength to 5th percentile characteristic strength was (1-1.645 x 0.3) = 0.506.

Therefore, the characteristic property, yw, was determined from the following: yw = 0.506 x (1.15 / 1.25) x 0.62 x 56.9 x d0.82 x RD2.2 = 16.4 x d0.82 x RD2.2

(Eq. 16)

The factor of 1.15/1.25 was used to adjust the design value for duration of load. A delay between nailing and testing reduces the withdrawal resistance of smooth-shank nails. The effect of delay varies greatly between and within species. Data by Mack (1960) showed a 30% loss of withdrawal resistance after a delay of four months. Clause 12.9.5.2 provides a factor of 1.6 for improved withdrawal resistance when the nail is effectively clinched. This was adopted from Crandall’s (1959) tests on clinched nails loaded laterally. It was assumed conservatively that nail clinching would be at least as effective in withdrawal as his tests showed for lateral strength. Nails driven into end grain are not considered to resist any loads in withdrawal. Toe-nails are partially resisted in the end grain of the wood. Therefore, they have only 67% of the withdrawal resistance of a connection installed in the side grain. The effect of seasoning in service is critical for nails loaded in withdrawal. As wet lumber dries, the fibres surrounding the nail relax, showing up to 75% loss of withdrawal resistance (Mack, 1960). Perkins (1971) found the same effect for nailed ts subject to continual wetting and drying. Prior to 2001, Table 12.2.1.5 included a service condition factor of only 0.25 for nails in withdrawal following assembly with unseasoned wood. As it is generally accepted now that the moisture change factor contains the influence of delay loading, the KSF value was adjusted upward in 2001 from the previous value of 0.25 to 0.4 (0.4 = 0.25 / 0.62) in Table 12.2.1.5. Nailed connections in a continuously wet environment showed withdrawal strengths approximately 60% of those in a dry environment (Hellawell, 1965).

Clause 12.10 – Joist hangers A clause on joist hangers was added in the 1998 Supplement to CSA O86. Joist hangers are light-gauge steel framing connectors typically used to transfer load from a joist to a beam. Clause 12.10 is similar to Clause 12.8 on truss plates, in providing minimum requirements for determining design values for use with this Standard. In the 2014 edition, Clause 12.10 was modified to allow lag screws and wood screws to be used for joist hanger connections.Test results for hangers are listed in the Registry of Product Evaluations published by the Canadian Construction Materials Centre, Institute for Research in Construction, Ottawa.

Clause 12.11 – Wood screws

Clause 12.11.1 – General Provisions for design of wood screws were added to the 2009 edition of CSA O86, based on the same European Yield Model equation design approach used for bolts, lag screws and nails.

Clause 12.11.2 – Connection configuration Wood screw design data were predicated on connection configuration assumptions including lead holes, location distances, and minimum fastener penetration and member thickness requirements. The location, penetration and thickness provisions are the same as for nailed connections.

Clause 12.11.3 – Connection design Like nail design, lateral design of wood screws is assumed not to vary with load angle to grain. This reflects the smaller diameters of the wood screws specified in CSA O86.

Clause 12.11.4 – Lateral Resistance The embedment strength of the main member varies depending on the EYM failure mode. The embedment strength is higher where the failure mode is fastener yielding (f3), compared to where the failure mode is in wood bearing (f2). The resistance was compared with test results over a range of density and fastener sizes. The adjustment to fastener yielding strength is the same as for nails and represents the so-called “string effect” discussed in 12.9.4. Although the design method is the same as for nails, wood screws are expected to have more of a string effect than nails due to the effect of the threads in the t. The resistance factor for lateral resistance of wood screws is 0.8, reflecting the yield failure mode and consistent with resistance factors for other fasteners (Chui 2000).

Clause 12.11.5 – Withdrawal resistance When wood screws were introduced in the 2009 edition of the Standard, design for wood screw withdrawal resistance was limited to short-term wind or seismic loads. Clause 12.11.5 was revised in 2014 to allow wood screws loaded in withdrawal to be designed for short, standard and long-term loads. The resistance factor for withdrawal resistance is 0.6, reflecting the high variance in the mode. The withdrawal resistance is higher for wood screws than nails reflecting the better holding capacity of wood screws loaded in withdrawal. Wood screw design also includes a requirement to check the head pull-through failure mode. This can occur when the force on the screw head exceeds the resistance of the light gauge steel, structural wood or lumber side plate. The resistance factor for pull-through resistance is 0.4, reflecting the high variance of the mode of failure (Mohammad and Chui 2006). When driving screws with a power tool, it is easy to overdrive the screws into the wood. Care should be taken not to overdrive wood screws as this may significantly reduce the withdrawal capacity of the screw.

References: Anonymous. 1982. Improved Design of Fastenings in Timber Structures Based on Limit States Design Procedures. Unpublished report prepared by Morrison, Hershfield, Burgess & Huggins, Ltd., and sponsored by the Canadian Forestry Service. Abukari, M.H., M. Coté, C. Rogers, and A. Salenikovich. 2012. Withdrawal resistance of structural screws in Canadian glued laminated timber. In: Proceedings of the 12th World Conference on Timber Engineering, July 16-19, 2012, Auckland, New Zealand. Blaß, H. J., I. Bejtka T., Uibel. 2006. Tragfähigkeit von Verbindungen mit selbstbohrenden Holzschrauben mit Vollgewinde. Universitätsverlag Karlsruhe Brock, G.R. 1957. The Strength of Nailed ts. Forest Products Research Bulletin No. 41, Department of Science and Industrial Research, Charles Horisi, 5-11 Regent St., London, England. BS EN 1995-1-1:2004+A1:2008 Eurocode 5. Design of timber structures. General, Common rules and rules for buildings. Canadian Constructions Materials Centre. Registry of Product Evaluations. Institute for Research in Construction, NRC, Ottawa, Canada.Chui, Y.H. 2000. Design Provisions for Wood Screw ts – A Preliminary Proposal. Prepared for the Fastening Subcommittee of CSA O86. Cockrell, R.A. 1933. A Study of the Screw-holding Properties of Wood. Technical Bulletin 44, New York State College of Forestry. Cramer, C.O. 1968. Load Distribution in Multiple-Bolt Tension ts. Journal of the Structural Division, American Society of Civil Engineering, ST-5, 5939: 1101-1117.

Crandall, C. 1959. Allowable Loads for 8d Nails. Forest Products Journal. 9(8): 250-262. DeGrace, R.F. 1986. Commentary on CSA Standard CAN3-086.1-M84 Engineering Design in Wood (limit states design). Canadian Standards Association. Special pub. 086.1.1-M1986, Rexdale, ON. Erki, M.A., and M.W. Huggins. 1983. Load Capacity of Row and Shear Plate Connectors. Journal of Structural Engineering, American Society of Civil Engineers, No. 12, Vol. 109. Foschi, R.O. and J. Longworth. 1975. Analysis and Design of Griplam Nailed Connections. Journal of the Structural Division, American Society of Civil Engineering. 101(ST12): 2537-2555. Fairchild, U. 1926. Holding power of screws. U.S. Dept. of Commerce, Technologic Papers of the Bureau of Standards, No. 319. Gov. Print. Office, Washington, D.C. Gehloff, M. 2011. Pull-out resistance of self-tapping wood screws with continuous thread. Thesis submitted in partial fulfillment of the requirements of the degree of Master of Applied Science in the Faculty of Graduate Studies. University of British Columbia, Vancouver, Canada. Hellawell, C.R. 1965. Mechanical Fasteners and Timber ting in New Zealand from a Laboratory Viewpoint. International Symposium on ts in Timber Structure. Timber Research and Development Association. ICC-ES. 2012. AC233. Acceptance criteria for alternate dowel-type threaded fasteners. www.icc-es.org. Johansen, K.W. 1949. Theory of Timber Connectors. Publications of the International Association of Bridge and Structural Engineering. No. 9: 249-262. Bern, General Secretariat. Jorissen, A. 1998. Double Shear Timber Connections with Dowel Type Fasteners, PhD thesis, Delft University Press, The Netherlands, 264 p. 1998. Karacabeyli, E. and H. Fraser. 1990. Short-term strength of glulam rivet connections made with spruce and Douglas fir glulam, and Douglas fir solid timber. Can. J. Civ. Eng. 17, 166-172. Karacabeyli, E., H. Fraser and W. Deacon. 1995. Lateral and withdrawal load resistance of glulam rivet connections made with sawn lumber. Unpublished report, Forintek Canada Corp. (now FP Innovations), Vancouver, BC. Kennedy, S., A. Salenikovich, M.H. Abukari, C. Rogers, M. Mohammad, W. Munoz. 2013. Withdrawal resistance of lag screws. Progress report. Université Laval (available upon request). Legras, B., Salenikovich, A., Mohammad, M., and Quenneville, P., 2010. Effect of moisture on the performance of bolted connections in timber structures. World Conference on Timber Engineering, Riva del Garda, Italy. Lantos, G. 1969. Load Distribution in a Row of Fasteners Subjected to Lateral Load. Wood Science, No.3, Vol. 1, pp. 129-136. Larsen, H.J. 1973. The Yield Load of Bolted and Nailed ts. Proceedings, International Union of Forestry Research Organization meeting – 5th Congress. South Africa. Larsen, H.J. 1977. K.W. Johansen’s Nail Tests. Bygningsstatisk Meddelels No. 48: 9-30. Lau, P.W.C. and Y. Tardiff. 1987. Cracks Produced by Driving Nails into Wood Effects of Wood and Nail Variables. Unpublished progress report, Building Systems Dept., Forintek Canada Corp. (now FP Innovations), Ottawa, ON. Leijten, A.J.M., and Jorissen, A. 2001. Splitting strength of beams loaded by connections perpendicular to-grain, model comparison, CIB-W18 meeting Proceedings, paper 34-7-1, Venice, Italy. Mack, J.J. 1960. Grooved Nails. Australian Timber 26(8): 43-44, 46, 50, 131. Massé, D.I., J.J. Salinas and J.E. Turnbull. 1988. Lateral Strength and Stiffness of Single and Multiple Bolts in Glued-laminated Timber Loaded Parallel to Grain. Unpublished contract. No. C-029, Eng. and Stat. Res. Centre, Research Branch, Agriculture Canada, Ottawa, ON. McLain, T.E. 1997. Design axial withdrawal strength from wood: II. Plain-shank common wire nails. Forest Products Journal, 47(6):103-109. McLain, T.E. and J.D. Carroll. 1990. Combined load capacity of threaded fastener- wood connections. J. Structural Eng. 116(9): 2419-2432. Mohammad, M. and Chui, Y.H. 2006. Derivation of Code Requirement to Prevent Head Pull-Through Failure of Wood Screws. Unpublished Report to CSA O86 Committee. Mohammad, M. and Smith, I. 1993. Load-Slip Response of Nailed OSB-to-Lumber ts as Influenced by Moisture Conditioning. Proceedings, Annual Conference, Canadian Society for Civil Engineering. Mohammad, M. and Quenneville, J.H.P. 2001. Behaviour of wood-steel-wood bolted glulam connections. CJCE, 28, pp. 254-263 Newlin, J.A. and J.M. Gahagan. 1938. Lag Screw ts: Their Behavior and Design. U.S. Dept. of Agriculture Technical Bulletin No. 597 Perkins, P.H. 1971. Nail Withdrawal Resistance in Plantation Red Pine Grown in Indiana. Forest Products Journal 21(6): 29-32.

Pirnbacher, G., R. Brander, G. Schickhofer. 2009. Base parameters of self-tapping screws. Paper 42-7-1 In: Proceedings of the CIB-W18 Meeting Forty-Two, Duebendorf, Switzerland. Simpson Strong-Tie. 2006. Proprietary test reports. Courtesy of Simpson Strong-Tie. Smith, I., Craft S.T. and Quenneville, J.H.P. (2001) Design capacities of ts with laterally loaded nails. CJCE, 28, pp 282-290 Quaile, A.T. and F.J. Keenan. 1979. Truss Plate Testing in Canada: Test Procedures and Factors Affecting Strength Properties. Proceedings, Metal Plate Wood Truss Conference, Forest Products Research Society. Quenneville, J.H.P., J.G.A. Charron and Van Dalen, K.. 1993. Effect of End Distance on the Resistance of Split Ring Connectors in Timber ts Loaded in Compression. Canadian Journal of Civil Engineering. 20(5). Quenneville, J.H.P and Mohammad, M.. 1999. The Anchorage Strength of Bolted Wood-to-Concrete Connections Loaded Perpendicular-to-Grain. Royal Military College of Canada, report to the Canadian Wood Council, 1999.Quenneville, J.H.P. and Sauvé, G. 1998. Strength and Behaviour of Full-Scale Split-Ring Connections. 5th World Conference on Timber Engineering, Montreux, Switzerland. Quenneville, J.H.P. and Mohammad, M. 2000. On the failure modes and strength of steel-wood-steel bolted timber connections loaded parallel-to-grain. CJCE, 27, pp. 761-773. Ramos, A.N. 1960. Spacing of Sixpenny and Eightpenny Wire Nails in Douglas Fir Multi-nail ts. U.S. Department of Agriculture, Forest Products Laboratory. Report No. 2155. Madison, WI. Scholten, J.A. 1944. Timber Connector ts, Their Strength and Design. U.S. Department of Agriculture Bulletin 865. Smith, I., 2000. Some Nail Withdrawal Data and an Interpretation of it. CSA O86 Technical Committee 44th Annual Meeting Book. Stieda, C.K.A., Karacabeyli, E., Fraser, H. and Deacon, W. 1997. Lateral Strength of Bolted Wood-to-Concrete Connections, Forintek Canada Corp. (now FP Innovations) Project No. 1565K242, Vancouver, BC. Trayer, G.W. 1932. Bearing Strength of Wood Under Bolts. U.S. Department of Agriculture, Technical Bulletin No. 332. Turnbull, J.E. 1987. Fastenings. Unpublished committee paper C-034, Eng. and Stat. Res. Centre, Research Branch, Agriculture Canada, Ottawa, ON. Turnbull, J.E. 1990. Commentary on Clause 10 of CSA Standard CAN/CSA-086.1-M89 Engineering Design in Wood (Limit States Design). Wood Design Manual (1990 edition). Canadian Wood Council, Ottawa, ON. USDA Forest Serv., Forest Prod. Lab. Wood Handbook. Wood as an engineering material. FPL-GTR-190. Madison, Wisconsin. Welchert, W.T. and C.N. Hinkle. 1966. The Effect of Change in Moisture Content on the Strength of Nailed Wood ts. Trans. of the American Society of Civil Engineering. 9(6): 774-6-781. Whale, L.R.J., I. Smith and H.J. Larsen. 1987. Design of Nailed and Bolted ts Proposals for the Revision of Existing Formulae in Draft Eurocode V and the CIB Code. International Council for Building Research Studies and Documentation Working Commission W18A. Paper No. 20-7-1. Yasumura, M., T. Murota and H. Sakai. 1987. Ultimate Properties of Bolted ts in Glued-laminated Timber. Report to the Working Commission W18 – Timber Structures. Dublin, Ireland.

Clause 15

Proprietary Structural Wood Products—Design

Clause 15.1 – Scope Clause 15 is intended to apply to proprietary structural wood products that meet the requirements of an appropriate consensus standard. Proprietary products covered under Clause 15 are manufactured under a quality assurance program supervised by an independent third-party certification organization (CO) to give assurance that the manufacturer has met all appropriate requirements of Clause 15 and the applicable referenced standard(s), with the intent of providing evidence that design values are being maintained in on-going production. Clause 15 was updated in 2009 to separate provisions intended for use by a building designer or project engineer from those used mainly by a manufacturer’s engineer. The latter information was moved to Clause 16. Although building designers are not expected to be familiar with all aspects of the derivation of design values as outlined in Clause 16, they should be aware that Clauses 15 and 16 were developed together to provide a confidence level consistent with the intent of Part 4 of the National Building Code of Canada and with this Standard with respect to strength, serviceability, and reliability. Clause 15 was revised to make frequent reference to the “Product Evaluation Reports,” which are defined as third-party reports confirming that the design values were derived in compliance with Clause 16 and the NBC. In Canada, such reports are found in the Registry of Product Evaluations published by the Canadian Construction Materials Centre, Institute for Research in Construction, Ottawa. Following is a brief summary of items considered essential requirements for the proprietary products in Clause 15. Note that Clauses 15 and 16 refer to a number of standards that will not be familiar to most building designers, but must conform to the following minimum requirements to be referenced in CSA O86: •

the standard is derived based on consensus through a recognized standards writing organization,

•

the standard has a consistent methodology for the development of structural capacities used by the manufacturers of a given type of proprietary structural wood product, and

•

the standard content to include as a minimum: ˗ ˗ ˗ ˗

scope, definition, and references, qualification procedures and methods for development of structural capacities and or design values, minimum requirements of an appropriate "in house" manufacturing standard and quality assurance program, in association with a qualified third party CO, and a system for re-evaluation of recommended structural capacities.

Clause 15.2 – Prefabricated wood I-joists Clause 15.2 contains engineering design requirements and appropriate adjustment factors for prefabricated wood I-joists for which design values have been determined in accordance with ASTM Standard D 5055. The Note to Clause 15.2.1 brings the designer’s attention to the need for a certification organization’s mark on the I-joists. Certification Organizations as defined in Clause 2 of the Standard are accredited by the Standards Council of Canada to provide certification services in Canada.

Clause 15.2.2 – Modification factors

Clause 15.2.2.1 – Duration of load, KD The duration of load factors, KD, in Table 5.3.2.2 of the Standard are applicable for use with the specified strengths of prefabricated wood I-joists.

Clause 15.2.2.2 – Service condition factor, KS The specified strengths and stiffness of prefabricated wood I-joists are applicable for use under dry service conditions only. It is generally recommended that these products not be used in other than dry service conditions. In the rare cases where these products are required for use in other environments, the specified strengths and stiffness must be re-evaluated for both short and long-term effects of the particular environmental conditions. Environmental conditions which result in high or fluctuating moisture content may affect material and product quality and may result in strength and stiffness adjustments different from those suitable for sawn lumber. Wet service condition use should be considered only when appropriate KS values for the individual prefabricated wood I-joist components may be determined from the clauses of this Standard, or when appropriate factors for specified strengths and stiffness are provided based on the documented results of tests that take into the effects of time, temperature, and moisture content.

Clause 15.2.2.3 – Treatment factor, KT The specified strengths and stiffness of prefabricated wood I-joists are applicable for use with untreated products only. Pressure impregnating chemical treatments may affect material quality and result in the need for strength and stiffness adjustments different from those suitable for sawn lumber. Pressure impregnating chemical treatments should be permitted only when treatment adjustments for specified strengths and stiffness are based on the documented results of tests that take into the effects of time, temperature and moisture content.

Clause 15.2.2.4 – System factor, KH The system factor for prefabricated wood I-joists was revised to 1.0 (no increase) in the 2005 Supplement No.1 to CSA O86. Formerly the allowable bending resistance increases for load sharing use were derived taking into consideration the coefficient of variation (COV) of the stiffness of various flange materials. The original theory justifying this type of increase was based on the relative stiffness of the and positive correlation between bending strength and stiffness, with the assumption that as stiffness COV decreases so would the load sharing. That is, as stiffness COV tends to zero, lack of differential deflection decreases load sharing. Further examination of the underlying theory of repetitive member effects and comparison against test data did not factors greater than unity across the entire range of joist types and joist depths. Part of the reason for this was the lower stiffness variability between wood I-joists compared to between sawn lumber joists. At the same time, revision to specified tensile strength parallel to grain through addition of the length-adjustment factor (14.2.3.1) removed some of the conservatism in bending moment resistance of wood I-joists. Therefore, KHB was set to 1.0.

Clause 15.2.3 – Strength and resistance

Clause 15.2.3.1 – Bending moment resistance This clause and subsequent clauses on resistance include references to the listing of proprietary product design values in the CCMC Registry of Product Evaluations. These values are often published as factored resistance values for the purpose of simplicity, but the method and calculations are based on CSA O86 design procedures and principles.

Clause 15.2.3.2 – Lateral stability factor, KL Prefabricated wood I-joists require lateral at points of bearing to prevent lateral displacement and rotation. Additionally, all compression edges require lateral such as that provided by the permanent attachment of floor or roof sheathing in conformance with the applicable building code or, for other floor or roof decking products or systems, as determined by appropriate design procedures. For other applications such as long cantilevers or areas near intermediate bearing s for continuous spans, lateral stability requirements should follow the recommendations of the manufacturer and should be determined on the basis of appropriate analytical and engineering principles and/or documented test data. When performing lateral stability calculations, only the compression flange material and geometric properties should be used.

Clause 15.2.3.3 – Notches Notches, holes, or cuts in the flanges of prefabricated wood I-joists will result in strength loss. Such notches, holes, or cuts are only permitted on the basis of the documented results of tests that take into the effects of such notches, holes, or cuts.

Clause 15.2.3.5 – Reaction resistance Adequate bearing performance of prefabricated I-joists is dependent on more than perpendicular to grain bearing capacity of the material and the ing structure. Earlier editions of ASTM Standard D 5055 included an Annex with a prescribed method for evaluating reaction capacity, including qualification tests for the determination of the required bearing length. The 2008 edition of the ASTM D 5055 formalized the testing configuration for end bearing and intermediate bearing for the qualification testing at the manufacturer’s specified minimum bearing length. In the 2014 edition of CSA O86, the same reliability normalization factors in Table 14.2.3.2 were adopted for shear and bearing reactions based on their similarity in test methods and mechanical behaviours. A new clause was introduced in the 2014 edition of the Standard that provides the designer with the equation for determining the end bearing resistance or intermediate bearing resistance.

Clause 15.2.3.6 – Web openings, bearing length and web stiffener requirements Web openings Web openings (holes) are allowed only when appropriate documented test data is available from the manufacturer. ASTM Standard D 5055 provides details for determination of the effects of web openings.

Web stiffeners Web stiffeners are elements, generally sawn lumber, plywood, oriented strandboard (OSB) or structural composite lumber that are oriented vertically in pairs and attached to opposite sides of the web member between the top and bottom flanges. They are generally installed with nails or staples loaded in double shear, with a minimum gap of 3 mm between the end of the web stiffener and the unloaded or nonbearing flange. If they are cut too long and require a force fit, the inherent prying action can fracture the critical glue t between the web and the flange. No gap is allowed between the end of the web stiffener and the loaded or bearing flange. Web stiffeners can be critical to the bearing capacity of a prefabricated wood I-joist and, when required, must be installed as per the manufacturer's instructions, which should be based on the testing required by ASTM Standard D 5055. Web stiffeners perform two main functions: (1) they transfer a portion of the load from the web into the top of the bottom flange, reducing the load on the web-to-flange t; and (2) they serve the implied function of reinforcing the web against buckling. Without careful testing and evaluation, web stiffeners should not be used for vertical load transfer through the prefabricated wood I-joist at the bearing location, since critical tolerances that are required for adequate performance are more stringent than can be expected under the best shop or field conditions. ASTM Standard D 5055 provides the basis for determination of web stiffener requirements for bearing capacity for use with prefabricated wood I-joists. Special details such as concentrated loads may require web stiffeners at special locations. Loads may also be ed by connection to the web or on bottom flanges. The specific detail requirements should be based on documented test results.

Clause 15.2.4 – Serviceability limit states The requirements for serviceability limit states in Clauses 5.1.3 and 5.4 of this Standard applicable to sawn lumber and glued-laminated timber are also applicable to prefabricated wood I-joists. Maximum deflection limits are specified by code for all structural products. ASTM Standard D 5055 establishes requirements for the determination of the stiffness and evaluation of the short term dry service creep characteristics of prefabricated wood I-joists. It also stipulates that any formula used for stiffness capacity calculations must accurately predict the combined effects of bending and shear deformation. Two commonly used methods of determining shear deflection are detailed in Clause 10.5.6 of CSA O86 and in Western Forest Products Laboratory (now FPInnovations), Vancouver, BC, Information Report VP-X-31, A Shear Stiffness Factor for Plywood Box Beams, by C.K.A. Stieda. Appropriate deflection formulas for both, bending and shear are available from the manufacturers. With very few exceptions, prefabricated wood I-joists are noncambered products. Use of these products in roof applications should be limited to sloped applications. A minimum slope of 1:50 is recommended.

Clause 15.2.5.2 – Joist hangers and other framing connectors Although there are numerous joist hangers and connectors available that appear compatible with prefabricated wood I-joists, some are not. Hangers developed for conventional lumber or glued-laminated timber often use large nails spaced in a pattern that can split prefabricated wood I-joist flanges and web stiffeners. In addition to physical fit and load capacity, joist hanger selection for prefabricated wood I-joists should include consideration of nail length and diameter, nail location, joist bearing capacity, and the composition of the ing member. Because of the large and varied number of connector products available, the manufacturers of prefabricated wood I-joists must be consulted regarding the performance compatibility of joist hangers and other framing connectors with their products (See also Commentary on Clause 12.10).

Clause 15.3 – Structural composite lumber products Clause 15.3 contains engineering design requirements and appropriate adjustment factors for structural composite lumber (SCL) for which characteristic values have been determined in accordance with ASTM Standard D 5456. SCL is defined as including laminated veneer lumber (LVL), parallel strand lumber (PSL), laminated strand lumber (LSL), and oriented strand lumber (OSL), The Note to Clause 15.3.1 brings the designer’s attention to the need for a certification organization’s mark on the SCL products. Certification Organizations as defined in Clause 2 of the Standard are accredited by the Standards Council of Canada to provide certification services in Canada.

Clause 15.3.2 – Modification factors

Clause 15.3.2.1 – Load duration factor, KD The load duration factors, KD, in Table 5.3.2.2 of the Standard are applicable for use with the specified strengths of structural composite lumber products provided that the appropriate testing demonstrating the validity of the factor for use with the applicable structural composite lumber product has been conducted. The performance, in particular creep, under sustained loading is used to determine the equivalency of structural composite lumber products to sawn lumber in of duration-of-load behaviour.

Clause 15.3.2.2 – Service condition factor, KS The specified strengths and stiffness of structural composite lumber products are applicable for use under dry service conditions only. It is generally recommended that these products not be used in other than dry service conditions. Where these products are required for use in other environments, the specified strengths and stiffness must be modified by appropriate adjustments on the basis of the evaluation of the short and long-term effects of the particular environmental conditions. Environmental conditions which result in high or fluctuating moisture content may affect material and product quality and may result in strength and stiffness adjustments different from those suitable for sawn lumber. Wet service condition use should be considered only when appropriate factors for specified strengths and stiffness are provided based on the documented results of tests that take into the effects of time, temperature, and moisture content. This clause notes the need for testing and documentation of strength reduction factors for adhesive as well as wood product properties if other conditions are contemplated,

Clause 15.3.2.3 – Treatment factor, KT The specified strengths and stiffness of structural composite lumber products are applicable for use with untreated products only. Pressure impregnating chemical treatments may affect material quality and adhesive performance, and result in the need for strength and stiffness adjustments different from those suitable for sawn lumber. Pressure impregnating chemical treatments should be permitted only when treatment adjustments for specified strengths and stiffness are based on the documented results of tests that take into the effects of time, temperature and moisture content.

Clause 15.3.2.4 – System factor, KH The system factor, KH, for structural composite lumber products, was set at 1.04 to be consistent with the repetitive member factor Cr as per National Design Specification for Wood Structure (NDS) for Structural Composite Lumber.

Clause 15.3.2.5 – Size factor in bending, KZb The size factor in bending, KZb, is based on the provisions of ASTM Standard D 5456. The volumetric adjustments in Clauses 15.3.2.5 and 15.3.2.6 are intended to be performed by the manufacturer in determination of published and code evaluated properties and are not intended for general design use by a designer. Appropriate volumetric parameters, design information and limitations applicable to specific proprietary structural composite lumber products should be listed in the Registry of Product Evaluations, published by the CCMC.

Clause 15.3.2.6 – Size factor in tension, KZt The size factor in tension, KZt, is based on the provisions of ASTM Standard D 5456. Also see the second paragraph of the Commentary for Clause 15.3.2.5.

Clause 15.3.2.7 – Lateral stability factor, KL The lateral stability factor is identified as being the same as for sawn lumber.

Clause 15.3.3 – Strength and resistance

Clause 15.3.3.2 – Notches Notches, holes, or cuts in structural composite lumber products will result in strength loss. Such notches, holes, or cuts are only permitted on the basis of the documented results of tests that take into the effects of such notches, holes, or cuts.

Clause 15.3.3.3 – Shear resistance Loads within a distance equal to depth d of the are considered in shear design for structural composite lumber (SCL) unless the SCL manufacturer has ing information that would allow loads within a depth d of the to be ignored. Designers are advised to check with the product manufacturer regarding neglecting loads within a depth d of the for shear design. Presently there is no information available to the use of a size factor in shear, KZv; therefore a value of 1.0 is used for structural composite lumber. Where shear is evaluated in full-size , it is based on testing of deeper to ensure that no depth effect is required.

Clause 15.3.3.4 – Compressive resistance parallel to grain Presently there is no information available to the use of a size factor in compression parallel to grain, other than KZc = 1.0.

Clause 15.3.3.9 – Resistance to combined bending and axial load This clause provides interaction equations to be satisfied when deg subjected to combined axial and bending loads. The interaction equations were revised in the 2009 edition to be consistent with the sawn lumber provisions in Clause 6.5.10 (see commentary on that clause). The ratio of compressive load to compressive resistance is squared in the revised interaction equation. Also, the moment amplification factor is defined explicitly in the equation as1/ (1-Pf/PE) where PE is the Euler buckling load. This factor is slightly more conservative in most cases than the detailed amplification factor that is provided in Clause A6.5.10 as an alternative approach. The general effect of the revised interaction equation including the moment amplification factor is a slight reduction in required SCL sections.

Clause 15.3.4 – Serviceability limit states The requirements for serviceability limit states in Clauses 5.1.3 and 5.4 of this Standard applicable to sawn lumber and glued-laminated timber are also applicable to structural composite lumber products. Maximum deflection limits are specified by code for all structural products. With very few exceptions, structural composite lumber products are non-cambered. Use of these products in roof applications should be limited to sloped applications. A minimum slope of 1:50 is recommended.

Clause 15.3.5 – Connections for structural composite lumber Although there are numerous connectors available that appear to be compatible with structural composite lumber, some are not. Hangers and other connectors such as split rings that have been developed for sawn lumber or gluedlaminated timber can split structural composite lumber products in certain orientations, such as parallel to the gluelines, resulting in unacceptable or low performance. Manufacturers of structural composite lumber products must be consulted regarding the performance compatibility of connectors with their products. The performance of dowel type connections such as nails, bolts and lag screws with structural composite lumber products is generally consistent with the performance of these connectors with sawn lumber and glued-laminated timber for the same density of material. However the specific product manufacturer should be consulted for appropriate guidance.

Clause 16 Proprietary Structural Wood Products—Materials and Evaluation As discussed under Clause 15, CSA O86 was revised in 2009 to separate information required by building designers from that used by manufacturers’ engineers to develop design values. The latter information was moved to the new Clause 16. Clause 16 also covers similar information on truss plates (Clause 16.4) and joist hangers (Clause 16.5) since these clauses also address evaluation of proprietary materials through testing. Design values applicable to prefabricated wood I-joists and structural composite lumber products, as defined in CSA O86, are based on the capacities determined by the procedures of ASTM Standard D 5055 and ASTM Standard D 5456 respectively. Design values for truss plates and joist hangers are based on capacities determined from the procedures of CSA S347 and ASTM Standard D 7147 as modified by CSA O86, respectively. It is the responsibility of the producer to determine design values. Judgement is required particularly when assessing design values from qualification tests. Design values need to consider potential low-line lot capacities to avoid marginal application performance or uneconomical reject rates in the quality assurance program. Clause 16 is also intended to be a reference to help designers understand the origin of manufacturer’s proprietary design values. The designer is not normally expected to calculate the proprietary design values using the equations provided herein. Proprietary design information including specified strengths and stiffness, application details and requirements, and installation guidelines are available from the manufacturer.

Clause 16.2 — Prefabricated wood I-joists

Clause 16.2.1 — General ASTM Standard D 5055 has been adopted for reference as an appropriate standard for establishing the requirements for prefabricated wood I-joists and their structural capacities. Some exceptions and additional requirements are noted in CSA O86. Note that the Clause also indicates that a CO mark should appear on the I-joists.

Clause 16.2.2 — Materials Flange materials for wood I-joists are referenced to Clause 6 for sawn lumber and Clause 7 for gluedlaminated timber, as well as to ASTM D 5055 for structural composite lumber and lumber not conforming to Clause 6. The product conformance standards referenced in Clause 16.2.2.2 establish the basic requirements for products for use as structural sheathing; these products are also generally acceptable as prefabricated wood I-joist web s. However, it is recommended that s used for the webs of prefabricated wood I-joist be qualified and monitored through the I-joist manufacturer’s quality assurance program under the supervision and approval of the qualified third party CO.

Clause 16.2.2.3 — Adhesives Traditional resorcinol and phenol-resorcinol adhesives meeting CSA 0112.7 or CSA O112.6 have been used successfully in severe moisture conditions and are accepted for use in structural applications. The advent of new adhesive types and processes has led to new standards to evaluate adhesives against universal criteria, regardless of adhesive or process. In CSA O86 Supplement No.1 (2005) the use of alternative adhesives in wood I-joists was referenced to CSA standard O112.9, based on the common

criteria for adhesives that structural wood adhesives need to be suitable for exterior or wet services conditions. Some of the more recent adhesive classes and processes have been shown to work in typical applications but may not perform as well in extreme environments. For this reason, CSA O112.10 was developed to provide criteria for adhesives to be used in structural wood products in “limited” moisture exposure conditions. Limited moisture exposure means that a wood product is to be used in protected conditions but should be capable of withstanding several wetting and drying cycles if necessary. Therefore, the test requirements for CSA O112.10 are less severe than for CSA O112.9 but are considered acceptable for products such as wood I-joist, which are intended for use under dry service conditions. One of the criteria for adhesives used in wood I-joists is durability under exposure to elevated temperatures. This is addressed in elevated temperature creep tests in CSA O112.9 and O112.10, as well as the heat durability test in ASTM D 5055.

Clause 16.2.3 — Specified strengths and modulus of elasticity ASTM Standard D 5055 allows for the determination of shear capacities of prefabricated wood I-joists on the basis of prescribed qualification testing. Moment capacities are determined by one of the three following methods: (1) analytically, using the code listed axial capacity parallel to grain of flange material, (2) analytically, using the flange material axial capacity statistically determined on the basis of tests and (3) empirically, using the statistically determined moment capacity on the basis of prescribed full-scale tests. Additionally, ASTM Standard D 5055 includes requirements for the verification of the theoretical design capacities of the prefabricated wood I-joist. ASTM Standard D 5055 does not define characteristic values for prefabricated wood I-joist strength properties. Clause 16.2.3.7 provides guidance to convert the design values developed in accordance with ASTM Standard D 5055 to appropriate characteristic values for use with this Clause. The reliability normalization factors, Kr, in Table 16.2.3.2 were based on and derived in accordance with CWC Publication “Standard Practice Relating Specified Strengths of Structural to Characteristic Structural Properties” providing equivalent target reliabilities, on a consistent basis, for derivation of specified strength’s from characteristic values, for all products. A reliability index of 2.8 was used for all properties shown Table 16.2.3.2, with the exception of prefabricated wood I-joist shear which was based on a reliability index of 2.6. A scaling factor of 0.8, used to convert from test term to standard term, was applied to all properties listed in Table 16.2.3.2 with the exception of prefabricated wood I-joist shear which are based on a scaling factor of 0.7. For the calibration, a snow-loaded roof with snow load-to-dead load ratio of unity was selected. The calibration resulted in no change in allowable span, although there was an increase in design strength capacity for floors and some decrease in many snow-loaded roofs. This was judged appropriate as floors were generally deflection limited and considered more reliable than roofs. The reliability normalization factors adjust values to the target reliability, but also for differences in scaling of capacities. The scaling differences, embodied within the ASTM reduction factors for allowable stress design, were preserved to continue to relate the specific test data basis to capacity basis, for all wood products. In 2009, Table 16.2.3.2 was revised to include a new approach for determining the shear strength of structural composite lumber based on full-scale member testing. The Kr factor for SCL shear strength based on small shear block (Case 1 in Table 16.2.3.2) testing includes a 2/3 stress concentration reduction factor, while the Kr factor for full-scale (Case 2) testing does not. The full-scale member testing is judged to be more representative of the end-use configuration than the small shear block test method.

In the 2014 edition of the Standard, new requirements were introduced to address the end and intermediate reaction capacities for I-joists, and the same reliability normalization factors in Table 16.2.3.2 as that for shear were adopted for bearing reactions based on their similarity in test methods and mechanical behaviours. The bending moment resistance for wood I-joists is calculated using the flange material’s specified strength parallel to grain, defined as the lesser of the specified tension strength multiplied by the lengthadjustment factor and size factor, or the specified compression strength. The length-adjustment factor, introduced in 2005, takes into stress distribution effects due to typical load configuration, flange test gauge length and properties, and is explained further in Clause A.16.2.3.1. The I-joist section analysis is based on a truss analogy that includes the net flange section multiplied by the distance between flange centroids, but no contribution from the web. The net effect of the length-adjustment factor was relatively minor, because a) the system factor was reduced to unity (15.2.2.4), b) qualification and quality control testing is required for I-joists and flanges, and c) many applications are governed by deflection criteria. Capacities increased slightly for some Ijoists with sawn lumber flanges.

Clause 16.3 — Structural composite lumber products

Clause 16.3.1 — General ASTM Standard D 5456 has been adopted for reference in CSA Standard O86 as an appropriate standard for establishing the requirements for structural composite lumber products and their structural capacities.

Clause 16.3.2 – Adhesives and binder systems Traditional resorcinol and phenol-resorcinol adhesives meeting CSA O112.7 or CSA O112.6 have been used successfully in severe moisture conditions and are accepted for use in structural applications. The advent of new adhesive types and processes has led to new standards to evaluate adhesives against universal criteria, regardless of adhesive or process. See the Commentary on Clause 14.2.2.3 for further information. Another means of demonstrating equivalent performance is set forth in CSA Standard O437.2, “Evaluation of Binder Systems for OSB and Waferboard,” which provides a method of evaluating the performance of a binder system to be used in the manufacture of an OSB product in comparison to the performance of the same product manufactured using a known phenol formaldehyde resin adhesive. To demonstrate equivalence, the procedures of CSA Standard O437.2 need to be modified with respect to conditioning, aging and other appropriate durability issues and to be representative of the geometry of the wood furnish used in specific structural composite lumber products.

Clause 16.3.3 - Specified strengths and moduli of elasticity ASTM Standard D 5456 specifies the determination of characteristic values of structural composite lumber products. See the Commentary on Clause 16.2.3 for additional information on specified strength derivation in CSA O86.

Clause 16.3.3.3 – Specified shear strength, fv ASTM D5456 was revised to permit use of full-scale testing to determine shear strength properties. Accordingly CSA O86 was revised to accommodate this approach through the Kr factors for shear strength (see Commentary on Clause 16.2.3.2).

Clause 16.3.3.5 - Specified compression strength perpendicular to grain, f The characteristic value in compression perpendicular to grain as determined by ASTM Standard D 5456 is the average value at 1.0 mm deformation for compression perpendicular to grain from test results. To adjust from characteristic value to specified strengths, a factor of 1.09 (= load factor of 0.868 divided by φ of 0.80) was used. The load factor of 0.868 represents exact agreement with working stress design at a snow load to dead load ratio of 4.0, when the traditional adjustment factor of 1.67 for compression perpendicular to grain is properly ed for. This approach is similar to that adopted for sawn lumber in this Standard.

Clause 16.4 — Truss plates The material evaluation provisions for metal truss plate connecters were moved to Clause 16.4 from Clause 12.8 in the 2009 edition.

Clause 16.4.1 — General Truss plates are evaluated in accordance with CSA Standard S347, and manufactured with galvanized sheet steel with minimum properties as specified in Clause 16.4.1.2.

Clause 16.4.2 — Strength resistance of truss plates As noted in 12.8, truss plates are tested in accordance with CSA Standard S347, Method of Test for Truss Plates used in Lumber ts. Test results for commercially used plates are listed in the Registry of Product Evaluations published by the CCMC. The evaluation reports also include resistance values derived from the tests. In the 2014 edition of the CSA S347 Standard, changes were made to allow truss grade structural composite lumber (SCL) to be used in truss applications based on the CCMC Guide for Evaluation of Truss-Grade Laminated Strand Lumber – Metal Truss Plate Connector Design Values. Table 1 of the 2014 edition of the CSA S347 Standard, was revised to be based on “Oven-dry relative density”. This change is consistent with Table A.12.1 of CSA O86 that provides mean oven-dry relative density values for visually graded lumber, machine stress rated lumber and machine evaluated lumber. The restrictions on yield strength for steel used in testing were moved from CSA O86 to the 2014 edition of CSA S347 in order to have all testing requirements in the same location. New test requirements were included in the CSA S347 Standard to ensure that moisture content changes in SCL connections do not result in strength reductions greater than the moisture content related strength reductions found in sawn lumber connections.

CSA S347 2014 requires four types of truss plate tests: •

Lateral resistance of the teeth,

•

Tensile strength of the plate,

•

Shear strength of the plate, and

•

Ultimate tensile strength of the plate material.

To determine lateral resistance of the teeth, plates are tested with the primary axis of the plate oriented parallel and perpendicular to the direction of load, and loaded parallel and perpendicular to the grain for a total of four different test configurations. For design, intermediate orientations of plate to load are obtained by linear interpolation; for intermediate orientation of grain to load, the familiar Hankinson formula is used (see Clauses 12.8.3.2 and 12.8.4.2), as with other fastenings. The calculations for ultimate lateral resistance in clause 16.4.2.1 have been revised from previous editions of the Standard and are now based on average test values, with adjustment factors for variability and a 0.8 factor to adjust from test load duration to standard term loading. A moisture response factor, which was embedded into the previous design procedures has been shown separately for transparency, and to generally calibrate the calculations to the previous standard. The effect of moisture change after pressing of truss plates is discussed in Quaile and Keenan (1979). To evaluate the tensile strength of the plate, six specimens are tested; three specimens have the plate length parallel to the loads and three specimens have the plate length perpendicular to the load. The plates are required to have dimensions that would induce a tensile failure in the plate rather than a tooth withdrawal failure. The plates are also required to be positioned so that the minimum cross-section of the plate is centred over the t between wood even if this requires that the plate be positioned asymmetrically. For the “shear strength of the plates”, the test standard requires that the plates be tested at angular increments of 30°, plate-axis-to-shear-line. At each shear angle, the average of these replications is corrected for the ultimate tensile strength of the plate material. The test standard further requires that the plates be large enough to ensure a plate shear failure along the t rather than a failure in the teeth. A number of other test conditions are specified (i.e., lumber density, species, moisture content, over- or under-pressing, time from fabrication to testing, end and edge distances, plate size). Quaile and Keenan (1979) have discussed the effects of these factors and the reasons for the manner of their specification in the Standard. Design values are determined on the same basis as tension values. A new requirement was introduced in the 2014 edition of the standard to establish a modification factor for roller pressing that is to be applied to lateral resistance tests conducted on connections assembled using a hydraulic press.

Clause 16.4.3 – Lateral slip resistance Slip resistance is a mandatory part of truss plate design. Lateral slip resistance is based on tests carried out in accordance with CSA S347. In the 2014 edition of CSA O86, modification factors for roller pressing and moisture response in sawn lumber truss connections were introduced. Where roller pressing is used, the roller press modification for slip must be applied.

Clause 16.5 — Joist hangers Material evaluation provisions for joist hangers were moved to Clause 16.5 from Clause 12 in the 2009 edition. In the 2014 edition, Clause 16.5 was revised to conform to ASTM Standard D7147, Standard Specification for Testing and Establishing Allowable Loads of Joist Hangers.

Clause 16.5.4 – Corrected ultimate load of joist hangers The correction factor for steel strength has been revised in the 2014 edition to reduce the conservatism in the previous editions for situations where a steel failure is not observed during testing.

References: ASTM D6555. 2000. Evaluating system effects in repetitive-member wood assemblies. American Society for Testing and Materials, West Conshohocken, PA, USA. CWC. 2001. Standard Practice Relating Specified Strengths of Structural to Characteristic Structural Properties, Canadian Wood Council, Ottawa, Canada. Quaile, A.T. and F.J. Keenan. 1979. Truss Plate Testing in Canada: Test Procedures and Factors Affecting Strength Properties. Proceedings, Metal Plate Wood Truss Conference, Forest Products Research Society. Stieda, C.K.A. 1967. A Shear Stiffness Factor for Plywood Box Beams. Western Forest Products Laboratory (Canada) (now FPInnovations). Information Report VP-X-31. Vancouver, BC.

Appendix Clause A.5.3.5 – System modification factor, KH For further information on system modification factor, KH refer to the Commentary Clause 5.3.5.

Clause A.5.4.6 – Building movements due to moisture content change For wood materials, movement is primarily related to shrinkage or swelling caused by moisture loss or gain when the moisture content is below the fiber saturation point (FPS). The fibre saturation point is the moisture content at which wood cell walls are fully saturated but below the point where there is free water in the cell cavities. Drying below this fibre saturation point then removes water from the cell walls and the cell wall shrinkage results in changes in member width and thickness. The fibre saturation point varies with species and can be dependent on the method of measurement. For sawn lumber it is typically assumed to be between 25 and 30% moisture content. More specific information can be found in the Wood Handbook and related literatures (Forest Products Laboratory, 2010 and Bodig, 1982). When wood dries from its green condition, little or no shrinkage occurs until the moisture content falls below the FPS. Below the FSP, wood shrinks when its moisture content is reduced and swells when its moisture content is increased. Wood shrinks most in the tangential direction and about half as much in the radial direction. As lumber is sold as a mixture of species and grain orientation, for purposes of design, it is usually assumed that perpendicular-to-grain dimension will change approximately 1% for every 5% change in moisture content. In the parallel-to-grain direction, the dimensional change is approximately one fortieth of the dimensional changes perpendicular to grain and can be ignored in most low-rise buildings. Most of the building movement due to shrinkage occurs in horizontal such as floor joists and wall plates. However, the parallel-to-grain dimension change in studs should be considered in design of 5-6 storey buildings. In mid-rise wood frame buildings, it is particularly important to design to accommodate the expected amount of shrinkage because the overall magnitude of the dimensional change can be accumulative and larger than typically experienced in low-rise wood frame buildings (APEGBC, 2015). Dimension lumber is generally dried before being shipped to s. Lumber designated as “S-GRN” (Surfaced Green) is at a moisture content greater than 19% at the time of surfacing. Lumber designated as “S-Dry” (Surfaced Dry) indicates that the lumber was surfaced at a moisture content (MC) of 19% or less. Dimension lumber installed at 19% MC will typically dry and shrink before reaching the Equilibrium Moisture Content (EMC). Glulam typically has a moisture content between 11 and 15% at the time of manufacturing. products such as plywood or OSB, and engineered wood products such as I-joists and Structural Composite Lumber tend to have lower moisture content in the range of 6 to 12 % at the time of manufacturing. Reducing the initial moisture content of wood greatly reduces the amount of wood movement that may occur during and after construction. It is very important to implement provisions that will protect dry wood from sources of water during construction and in service. It is recommended that all wood components within a floor system or connecting system should be made of materials with similar shrinkage properties, unless the potential differential movement is calculated and details are provided to accommodate it (APEGBC, 2015).

Designers should consider design details to accommodate wood movement: Example 1: In a well-ventilated roof space, the temperature differences and the consequent relative humidity differences between the top and the lower chords of the roof truss may cause lower chords to shrink and top chords to expand in winter. This will force the peaks of the trusses upward and consequently cause cracks between the ceiling and the partitions. This can be prevented by eliminating ceiling fasteners within 300 mm of the partitions, and by coupling the ceiling to the partitions at their juncture so that the trusses can move upwards without breaking the t between the partition and ceiling. Example 2: Any shrinkage of the wood along the embedded length of metal fasteners causes their heads to rise above the wood surface while forcing the tips slightly deeper into the wood. The initial and final moisture contents of the wood, and the depth of fastener penetration, are the principal factors in determining the amount of outward movement. Wood shrinkage can cause nail popping in gypsum wallboard finishes. Using nails with annular grooves or wood screws is recommended to prevent nail popping. Example 3: Where mixed systems are used, such as wood frame structures incorporating horizontal wood components connecting to vertical structural of steel, concrete, or wood-based columns or posts, the potential differential movement should be calculated and designed for. Where wood frame structure needs to be tied to the concrete stairwells/elevator shafts, vertical slotted connections can be used to accommodate the differential movement.

Clause A.6.5.7 – Compression perpendicular to grain and density Some manufacturers of machine graded lumber establish lumber density on a mill-specific basis using qualification tests and daily quality control. The density values may be higher than the generic density values identified in Appendix A.12.1, and the resulting compression perpendicular to grain values based on the CSA O86 relationship with density will also be higher. In these cases, the “specific gravity”, which is numerically equivalent to the mean oven-dry relative density, is marked on the lumber. Clause A.6.5.7 identifies the approach used to relate compression perpendicular to grain strength values to mean ovendry relative density in CSA O86.

Clause A.6.5.10 – Detailed formula for combined bending and compressive loads A new interaction equation was included in the 2009 edition of the Standard. In Clause 6.5.10 the ratio of factored compressive load to compressive resistance is squared, while it was not in previous editions. The moment amplification factor is defined explicitly in the body of the Standard as 1/(1-Pf/PE) where PE is the Euler buckling load. This simplified approach is slightly more conservative in most cases than the detailed amplification factor that is provided in Clause A.6.5.10 as an alternative approach.

Clause A.11.7 – Fastener deformation, en, for shearwall and diaphragm deflection calculations In the 2014 edition of the Standard a simplified formula was introduced that can be used to estimate the deformation of nails at edges of shearwalls and diaphragms. This formula is based on the approximate nail slip equation for wood-based sheathing nailed to green lumber that were developed by APA and published in Report 138 “Plywood Diaphragms”, and divided by 2 to convert to dry lumber. The simplified formula also incorporated the 20% increase in nail slip for sheathing s that do not meet the US Structural 1 requirements.

Clause A.11.7.1 – Deflection of shearwalls in multi-storey buildings The deflection calculation for multi-storey shearwalls was introduced in Appendix A.11.7.1 in the 2014 edition of the Standard. It is a purely mechanics based approach and is appropriate for a light frame wood shearwall cantilevered from its base and stacked for the full height of the building. This methodology follows the same engineering principle as the deflection equation in Clause 11.7.1.2 for a single-storey shearwall but is expanded to take into multi-storey effects. Similarly, it is comprised of four components, i.e. the deflection due to bending, shear, and nail slip and anchorage system elongation. Among the four components, shear and nail slip represent the shear deformation within one storey and will not affect the drift in an adjacent level above. However, the deflection due to bending and wall anchorage system elongation, which represent the flexural behavior, will result in a rotation at the top of each wall segment, which in turn will cause an increase in drifts and deflections for each storey above (APEGBC 2015; Leung 2009; Newfield et al. 2013; Newfield and Wang 2016). This methodology is intended to provide a conservative estimate of lateral deflections and inter-storey drifts. For flexible structural systems such as wood-based shearwall systems deformations are likely to govern the performance and therefore underestimating the deflections and interstorey drifts would be problematic. This mechanics based approach is appropriate for determining the deflection of a cantilevered single wall acting as a component of the building. There are other factors, such as out of plane diaphragm stiffness and system effects, that could potentially reduce building deflections but their effects are not currently well understood and therefore not addressed in the standard. Wall stiffness’s are used for distributing forces to the walls if the rigid diaphragm assumption is used. Wall stiffness’s based on deflection may differ if the Clause 11.7.1.2 equation is applied to each level or the A.11.7.1 equation used; however the relative stiffness between various walls and thus the force distribution are fairly consistent regardless of the method used (Newfield, G. and Wang, J.B., 2016). Wall stiffness’s may also be used for calculating building periods using mechanics-based design approaches such as the Rayleigh Method. It is recommended that the deflection formula provided in A.11.7.1 should be used when checking building deflections and interstorey drifts for mid-rise wood frame buildings. This should lead to a reasonable distribution of forces for a SFRS of wood-based shearwalls and should provide a conservative approach for determining deflections and drifts. Designers should use their judgement in determining the appropriate stiffness’s to be used for distributing forces to walls based on the relative walls stiffness’s or when calculating building periods used to determine seismic forces.

References: APEGBC Technical and Practice Bulletin: Structural, fire protection and building envelope professional engineering services for 5 and 6 storey wood frame residential building projects (mid-rise buildings). 2009 revised 2015. Association of Professional Engineers and Geoscientists of British Columbia (APEGBC), Burnaby, BC, Canada. Leung, T. 2009. Uncertainties in Lateral Load Design in Wood Engineering Newfield, G., Ni, C., Wang, J.B. 2013. A mechanics-based approach for determining deflections of stacked multi-storey wood-based shear walls. FPInnovations, Vancouver, B.C. and Canadian Wood Council, Ottawa, Ont., Canada. Newfield, G., Wang, J.B. 2016. A comparative analysis of three methods used for calculating deflections for multi-storey wood shearwalls. Proceedings of the World Conference on Timber Engineering 2016, Vienna, Austria.

Annex B

Fire resistance of large cross-section wood elements

Building code regulations require that key building assemblies exhibit sufficient structural fire resistance to allow time for occupants to escape, firefighters to carry out their duties and to minimize property losses. The intent is to compartmentalize the structure to prevent the spread of fire and smoke, and to ensure structural adequacy to prevent or delay collapse. In Canada, the fire-resistance rating of a building assembly has traditionally been assessed by subjecting a replicate of the assembly to a standard fireresistance test, CAN/ULC- S101“Fire Endurance Tests of Building Construction and Materials”. The standard test requires a wall or floor assembly, or a structural element such as a column or beam, to be exposed to a severe fire in which the temperature of the fire gases increases over time following a specified temperature-time curve. The test standard also requires the assembly to be loaded if it is to be loadbearing. In North America, assemblies are typically loaded to their full design load determined based on strength (as opposed to serviceability criteria such as deflection). This ensures that the fire-resistance rating obtained for a particular assembly is appropriate for use in any building independent of the load conditions (assuming the structure satisfies the applicable structural design requirements). The standard fire-resistance test entails three failure/acceptance criteria. Firstly, the ‘structural criterion’ must be met: the assembly must the applied load for the duration of the test. Secondly, the ‘insulation criterion’ must be met: the assembly must prevent the temperature rise on its unexposed surface from being greater than 180°C at any location, or an average of 140°C measured at a number of locations. Lastly, the ‘integrity criterion’ must be met: the assembly must prevent the age of flame or gases hot enough to ignite a cotton pad. The time at which the assembly can no longer satisfy any one of these three criteria defines the assembly’s fire-resistance rating. For structural elements, however, such as beams and columns, only the structural failure/acceptance criterion is applied. The National Building Code of Canada (NBC) currently provides three explicit methods within Division B to demonstrate that large cross-section wood meet various levels of fire-resistance requirements within the code: •

Meeting the minimum cross-sectional dimensions for the classification of ‘heavy timber construction’;

•

Calculating the fire-resistance rating using the equations in Appendix D of the NBC (D-2.11.), which are currently restricted to glulam in bending or compression; and

•

Conducting a fire-resistance test in accordance with the standard test method CAN/ULC-S101.

Annex B of the CSA O86 Standard provides a calculation methodology intended to provide an alternative means to determine the fire-resistance rating of a structural wood element. It was added in the 2014 edition of the Standard as an informative Annex, since Part 3 of the NBC does not currently reference it.

Clause B.1.4 Since the intent of Annex B of the Standard is to determine the fire-resistance rating of a structural wood member without having to run a test in accordance with CAN/ULC-S101, the calculation methodology is simply designed to predict the result of the standard test with some degree of conservatism. Therefore, the structural load to be used in the calculations should be that which would be applied in the standard test. Typically, in standard tests, the superimposed load applied is what is called the full specified load condition. This is the largest specified load that is likely to occur in practice. Since design is based on a factored load such as 1.5L + 1.25D, the specified load, L + D, will be less than the factored load, and, therefore less than the factored resistance of the member. If one assumes that the design will be based on the factored load condition of 1.5L + 1.25 D, one can calculate the largest specified load condition which is likely to occur if an assumption is made as to the ratio between the live and dead loads. For lightframe wood construction, this ratio is often assumed to be 4:1 and for mass timber construction this ratio is often taken as 1:1. By using the full specified load condition in a standard test, the test results can be used with no restrictions based on load. If the standard test is conducted with a superimposed load which

is less than the full specified load condition, a restriction is placed on using the resulting fire-resistance rating to those instances in design when the specified load is less than that applied during the test. A significant benefit of the calculation methodology in Annex B is that the calculations are conducted for a structural member in which the loads are known. Therefore, the actual specified loads can be used to calculate the fire-resistance rating. This is akin to conducting a standard test for each specific structural member of a building in which the specified loads used in design are applied during the test. Benefits of calculating the fire-resistance using Annex B over the standard test include: •

The actual specified loads are known, allowing for a more economical means to satisfy the fireresistance requirements of the NBC;

•

The actual span of a beam or height of a column is used opposed to a standard span or height being tested, and then applied to unlimited spans and heights in practice; and

•

While the high cost of conducting standard tests makes the assessment of a variety of options difficult during testing, designs can be easily optimized by varying sizes and methods of protection when using a calculation methodology.

Clause B.1.5 A study completed by O’Neill (2009) at the University of Canterbury in New Zeeland found that LVL beams screwed or nailed together separated during fire exposure due to shrinkage of the outermost surface, leading to charring on multiple surfaces and contributing to structural failure earlier than would be expected if the built-up beams had remained in full . While methods can be used to ensure that the remain in full , such as some adhesives or fully-threaded screws, more research is necessary to be able to specify generalized minimum criteria.

Clause B.2.1 – General While many studies have been conducted on the charring rates and fire resistance of solid-sawn and glued-laminated timber beams and columns, there have been limited studies on the proprietary products referred to as structural composite lumber (SCL). A study by White (2006) found that of the 14 SCL products tested, all performed within the expected range of results for sawn lumber and glued-laminated timbers. Tests included exposure to the standard temperature-time curve to assess the charring rate using both a small-scale, and later, an intermediate-scale furnace. Additionally, 13 tests were conducted in an intermediate-scale furnace under a tensile load. This extensive study was used to modify the American wood design standard, the NDS, to permit the fire-resistance calculations to be applied to SCL products. The Standard references product standard ASTM D 5456 for structural composite lumber (SCL). The 2012 edition of the standard revised the definition of SCL to include bamboo-based products (laminated veneer bamboo). The fire resistance and charring rate information assessed to date for SCL products is limited to wood-based products. More research is required related to fire resistance and charring of bamboo-based products to be able to provide specific guidance for them in Annex B.

Clause B.3.2 – Resistance factor, φ Since the calculation methodology for fire resistance is not based on limit states design, but is intended to predict the result of a fire-resistance test, it is not appropriate to apply a resistance factor. Using a value of 1.0 simply allows the same equations to be used in ways the designer is familiar with and which are already published within the standard. Using a value of 1.0 is common practice for fire design as described by Buchanan (2001) and is specified in Eurocode 5: Part 1-2.

Clause B.3.3 – Load duration factor, KD The use of the short term duration of load factor is suggested based on the rationale that the fire exposure and resulting maximum stress condition in the reduced cross-section of the member is an event that lasts on the order of a few minutes in a fire-resistance test. While it is possible that a higher modification factor could be justified based on the short time frame of a fire test, further research would be required in order to ensure the overall calculation method continues to provide reasonable predictions when compared to experimental data.

Clause B.3.4 – System factor, KH As discussed under Commentary information 5.3.5, the system factor s for the design of a member in isolation when in actual fact it is part of a larger system, it experiences composite action between member and sheathing through fasteners, and the low probability that the weaker member in a system is placed in a location of high stress. Since under fire design, the specified strength is modified by the specified strength adjustment factor, Kfi, in order to predict the mean strength of the member, the application of the system factor is no longer applicable using the values specified. Also, in most cases, application of a system factor is not applicable for large cross-section considered in Annex B. Further, any system behavior that may be relied upon in regular design may be compromised during exposure to fire.

Clause B.3.5 – Size factor, KZ In general, the specified strength will increase as the member size decreases. Additional explanation on size effect can be found in the Commentary information 5.3.6 on size factor. While it is likely that an argument could be made that, during a fire, the reduction in the cross-sectional area of a member would result in an increase in the size factor to be used in calculating the residual capacity, no research has been conducted to or quantify such an increase. Since the calculation method is somewhat calibrated to test data, it is possible this effect may be implicitly considered in the overall method. The American wood design standard, the NDS as well as the Eurocode 5: Part 1-2 also uses this same conservative approach.

Clause B.3.9 – Specified strength adjustment factor for fire design, Kfi The aim of the calculation method is to predict the result of a standard test (in contrast to the rest of the Standard, which is based on limit states design). Therefore, it is necessary to adjust the specified strengths in the Standard to the mean value by multiplying by a factor for fire design, Kfi. In order to determine an appropriate value for Kfi irrespective of member size or grade, some simplifications are necessary. By assuming a normal distribution, the mean strength can be predicted from the specified strength by estimating the coefficient of variation (COV). For example, the value of Kfi is taken as

𝐾𝐾𝑓𝑓𝑓𝑓 =

1 1 − (1.65)(𝐶𝐶𝐶𝐶𝐶𝐶)

where the COV was estimated as 20% for solid sawn timber, 16% for glulam and 12% for structural composite lumber. While the simplified method described above to determine the appropriate factor to adjust the specified strength to the mean strength may be an oversimplification, the resulting adjustment factors have been used in the verification of the methodology. Therefore, in the future, if changes are to be made to the specified strength adjustment factors, a comparison of calculated fire resistance to fire test data should be conducted to ensure overall accuracy of the procedure in Annex B remains.

This same philosophy has been applied in CSA-S850-12 “Design and assessment of buildings subjected to blast loads.” The values presented in B.3.9 are slightly different from those in the blast standard since the values for solid wood in the blast standard are intended to modify strength values for lightframe lumber while the values in this Annex are intended to modify larger cross-section timbers. Lightframe lumber and large cross-section timbers will have different coefficient of variations and therefore the modification factor to adjust the specified strength to the mean strength will vary.

Clause B.4.2 – Design charring rate In order to ensure that the charring rates are representative of those expected in the standard fire exposure, it is necessary that the wood member remain thermally thick with respect to the one dimensional heat transfer on the exposed surfaces. In a study by Janssens and White (1994), the temperature rise beyond the char front in the non-charred wood extended approximately 35 mm. Therefore, by restricting the minimum dimension of the residual cross-section to 70 mm, the heat transfer, and, therefore the charring rate is likely to remain constant when exposed on each side. Once the temperature at the centre of the cross-section begins to rise, the charring rate will begin to increase due to heating from multiple surfaces.

Clause B.4.5 – Corner rounding When large timber beams or columns are subjected to fire on multiple sides, some rounding of the corners occurs due to the heat transfer from the ading surfaces. This effect has implicitly been included in the calibration of the nominal char rate specified in Table B.4.2.

When explicitly considered, the section modulus, Sr, of the reduced cross section can be approximated by

𝑆𝑆𝑟𝑟 = 𝑏𝑏𝑓𝑓 𝑑𝑑𝑓𝑓 2 ⁄6 − 0.215𝑟𝑟 2 𝑑𝑑𝑓𝑓 Where bf is the residual width of the beam, df is the residual depth of the beam and r is the radius of the charred corner (Buchanan, 2001). The radius of the charred corners has been determined to be approximately equal to the char depth by Hadvig (1981). In another study, Majamaa (1991) proposed a radius equal to 80% of the char depth, suggesting the assumed radius used in B.4.5 may be slightly conservative.

Clause B.5.1 – Zero-strength layer depth The zero-strength layer is subtracted from the residual cross-section to for the loss of strength and stiffness in the heated zone beyond the char front due to thermal degradation. The depth of 7 mm was proposed by Schaffer et al. (1986) based on the temperature profile beyond the char front and the loss of strength and stiffness of the wood as a function of temperature. Subsequently, the depth of 7 mm for the zero-strength zone has been adopted by the Eurocode 5, Part 1-2. The American wood design standard, the NDS, uses a slightly different approach to for the loss of strength of the heated wood beyond the char front. The NDS uses a 20 percent increase in the non-linear char depth to for the loss of strength of the heated wood. It may be of little surprise that the two methods of ing for the loss of strength of the heated wood, 7 mm and 20 percent of the char depth, provide similar results after one hour of fire exposure since both methods have been validated/calibrated using the same experimental test results.

Clause B.6.3 – Modification of specified strengths An explanation on the modification of the specified strength to the mean strength can be found in the Commentary information B.3.9 on Specified strength adjustment factor for fire design.

Clause B.6.4 – Modulus of elasticity Similarly to the adjustment of the specified strength to the mean strength, the modulus of elasticity (MOE) used should be the mean value so as to predict the actual stiffness of the structural member. The mean value of the MOE is published in the Standard for solid sawn timbers and glued-laminated timbers and is identified by the symbol E (in contrast to the 5th percentile value identified by the symbol E05).

Clause B.7 – Fire-Resistance Rating The fire-resistance rating determined in Clause B.7 is based on that fire-resistance rating that would be expected if a replicate of the member or assembly was tested according to CAN/ULC-S101 at the same loading. For example, when a 60-minute fire-resistance rating is sought, the resistance of a beam determined according to the proposed Annex B methodology after 60 minutes of standard fire exposure must be greater than the specified gravity loads (e.g. D+L). Typically, if the fire-resistance rating is to be checked at a specific time, such as after one hour of fire exposure, the resistance of the member can be calculated at that time. This structural resistance at a specific point in time can then be compared to the specified load effects to determine if the structural member has sufficient strength to the superimposed loads. Alternatively, a spreadsheet is often used to determine the actual time at which the member fails to the specified gravity loads, which allows optimization of the design or the use of additional measures, such as the application of gypsum board, thereby changing the fire-resistance as needed to meet or exceed the target fire-resistance rating.

Clause B.8.1 – Gypsum board The times associated with gypsum board protection in Clause B.8.1 are based on research completed at the U.S. Forest Products Laboratory by White (2009). The experiments were conducted on Douglas-fir laminated veneer lumber (LVL) beams and Douglas-fir glued-laminated beams in tension and compared to the structural fire resistance with and without the protective covering. The research found that one layer of 15.9 mm Type X gypsum board added, on average, 33 minutes to the fire resistance of beams in tension while two layers added 72 minutes, on average. The 30- and 60-minute additions to the fire resistance are viewed as conservative values. The 15-minute addition attributed to a single layer of 12.7 mm Type X gypsum board covering is based on a very conservative basis since the time is essentially equal to the finish rating of the membrane protection (time in a standard fire test at which the temperature at the backside of the gypsum board reaches approximately 200°C). The time attributed to 12.7 mm Type X gypsum board also does not for the additional protection provided by the gypsum board by its remaining in place for an extended period of time after the wood begins to char. As test data becomes available in the future, the time may be updated to reduce the level of conservatism.

Clause B.8.2 – Gypsum board fasteners The minimum penetration of gypsum board fasteners is required to ensure the fasteners do not fail prior to the gypsum board, so that the gypsum board will stay in place after the wood has commenced charring. Once the char depth has exceeded the fastener penetration, it is likely the fasteners will fail to the gypsum board. The requirements in Clause B.8.2 for screw length and spacing are based on the fastening used in the tests completed by White (2009).

Clause B.8.3 – ts The requirements in Clause B.8.3, for finishing of gypsum board ts and staggering of gypsum layers when using multiple layers is standard practice in specifications such as fire-resistance listings or generic fire-resistance listings in the building code.

Clause B.9 – Connections Connections are an important part of the fire resistance of wood structures. One of the major challenges in deg connections for fire resistance is the reliance on steel plates and fasteners. Due to the relatively high thermal conductivity of steel compared to wood, the less exposed steel a connection exhibits the better. Fire resistance tests conducted at Carleton University by Peng (2010), on wood-steelwood and steel-wood-steel ts loaded in tension confirmed that connections with a significant amount of exposed steel in the form of steel plates or exposed steel bolt heads and nuts tend to have a lower level of fire resistance. Further research at Carleton University is investigating the performance of beamto-column connections. Research in this area is also currently being conducted at a number of universities in Europe and New Zealand. Further research and analysis is necessary to properly address the topic in the Standard. It is important that the consequence of failure be considered in the design of connections in mass timber structures. For instance, a tension connection in the bottom cord of a glulam truss could become the critical link in determining the fire resistance for a floor or roof structure, while a beam-to-column connection in which the beam is bearing on the top of the column, such as in the top storey of a structure, could pose little risk due to the inherent load transfer between the . Membrane protection as described in Clause B.8.1 can also be applied to connections to increase the fire-resistance rating.

Clause B.10.3 – Butt-ted The method specified in Clause B.10.3 addressing butt-ted wood decking is identical to the method specified in AFPA’s Technical Report 10 (1999) (currently, AWC’s TR10, 2014) and was subsequently validated by Richardson and Batista (2001).

Clause B.10.4 – Unexposed surface protection In addition to ing the superimposed load, fire-resistance-rated floor assemblies must also satisfy the criteria aimed at preventing fire spread to the floor above for the duration of the fire-resistance rating. These criteria include preventing the age of hot gases and preventing excessive temperature rise on the unexposed (upper) surface. Therefore, in addition to calculating the resistance of the wood decking to the specified gravity loads, steps must be taken to prevent the age of fire through the decking. Clause B.10.4 ensures that an additional layer is present on the upper surface of the floor/roof deck to satisfy the two additional failure modes for fire resistance of a floor/roof assembly. The first two prescriptive solutions are based on those required in the National Building Code (Article 3.1.4.7, Sentence (5)), while the subsequent two are based on common generic concrete-based toppings used in generic fire-resistance-rated light-frame floor assemblies provided in Appendix A of the National Building Code (Appendix A, A-9.10.3.1)

References: AFPA. 1999. Calculating the fire resistance of exposed wood . Technical Report No. 10. American Forest and Paper Association, Washington DC, USA. ASTM. 2013. Standard Specification for Evaluation of Structural Composite Lumber Products (ASTM D5456). West Conshohocken, PA. American Wood Council. 2012. National Design Specification for Wood Construction (NDS). American Wood Council, Washington, DC. American Wood Council (AWC). 2014. TR10: Calculating the Fire Resistance of Exposed Wood . Technical Report No. 10. American Wood Council, Washington, DC. Buchanan, A.H. 2001. Structural Design for Fire Safety. West Sussex, England: John Wiley & Sons Ltd. CEN. 2004. Eurocode 5 – Design of Timber Structures, Part 1-2: General – Structural Fire Design (1995-1-2:2004). European Committee for Standardization, Brussels, Belgium. CSA. 2012. Design and assessment of buildings subjected to blast loads (CSA-S850-12). Canadian Standards Association, Mississauga, Canada. Hadvig, S. 1981. Charring of Wood in building Fires, Report. Laboratory of Heating and Air Conditioning, Technical University of Denmark. Janssens, M.L. and White, R.H. 1994. Temperature profiles in wood exposed to fire. Fire and Materials, 18, 263-265. Majamaa, J. 1991. Calculation models of wooden beams exposed to fire. VTT Research Note No. 1282. Technical Research Centre of Finland. O’Neill, J.W. 2009. The Fire Performance of Timber-Concrete Composite Floors. Master’s Thesis, University of Canterbury, New Zeeland. Peng, L. 2010. Performance of Heavy Timber Connections in Fire. PhD Thesis, Carleton University, Canada. Richardson, L.R. and Batista, M. 2001. Fire Resistance of Timber Decking for Heavy Timber Construction. Fire and Materials, 25, 21-29. Schaffer, E.L., Marx, C.M., Bender, D.A and Woeste, F.E. 1986. Strength validation and fire endurance of glued laminated timber beams (FPL-RP-467). US Forest Product Laboratory, Madison, WI, USA. ULC. 2014. Standard Methods of Fire Endurance Tests of Building Construction and Materials (ULC S101-14). Underwriters’ Laboratories of Canada, Scarborough, Ont., Canada. White, R. 2009. Fire Resistance of Wood with Directly Applied Protection. Proceedings of the Fire and Materials 2009 Conference, San Francisco, California, USA, 26-28th January, 2009. White, R. 2006. Fire Resistance of Structural Composite Lumber Products (FPL-RP-633). US Forest Product Laboratory, Madison, WI, USA.