AC14-0611-R1
#4
I-Joist Diaphragm Systems: Performance Trends
Observed with Full-Scale Testing
Ned Waltz, P.E., S.E.C.B., and Dr. J. Daniel Dolan, P.E.
Abstract
Pre-fabricated wood I-joists are routinely used to
construct roof and floor systems in modern lightframe wood construction. One of the primary functions of a light-frame floor or roof is to serve as a diaphragm that collects in-plane lateral load and transfers it to the shear walls and foundation elements.
Since the existing diaphragm design provisions contained within the model building codes are based on
a combination of testing and analysis conducted with
sawn lumber framing, designers often question
whether they can be reasonably applied to diaphragms framed with I-joists. This article summarizes
some of the rationalization and limitations associated
with I-joist framing for diaphragm construction. Performance trends and observations are presented
based on one manufacturer’s experience with 41 fullscale I-joist diaphragm tests conducted in accordance
with ASTM E455 (ASTM 2004).
Introduction
Modern light-frame roof and floor systems routinely combine pre-fabricated wood I-joists with
wood structural panel sheathing. A primary function
of a light-frame floor or roof is to serve as a diaphragm that collects lateral load and transfers it to
the shear walls and foundation. Diaphragm design
provisions for light-frame wood construction have
been successfully employed for decades and were
originally developed for lumber framing. Designers
often wonder if they are equally applicable to diaphragms framed with wood I-joists. The objective of
this article is to provide some insight into how shear
capacities are rationalized for I-joist diaphragms and
to summarize potentially useful trends observed with
full-scale testing.
Diaphragm Design
Diaphragms are typically modeled as deep, inplane beams. Perimeter framing is designed to act as
Summer 2010
tension/compression chords or struts. Wood structural panel sheathing combines with joists to serve as
the “web” that transfers shear. Joists provide out-ofplane stiffening and load transfer between discrete
panel sheathing elements. This makes sheathing-toframing attachment a critical element that often defines shear capacity of the assembly.
Wood structural panel diaphragms are normally
designed in accordance with provisions of the International Building Code (IBC) (International Code
Council 2009) and the Special Design Provisions for
Wind and Seismic (SPDWS) (American Wood Council
2008). Tables, like 2009 IBC Table 2306.2.1(1), were
first developed in the 1950s and have evolved over
time (Countryman 1952 and 1955, Peterson 1983,
Tissell 1967, Tissell and Elliott 2004). The original
rationalizations for these tables were based on an
analysis of sheathing and attachment schedules for
lumber-framed diaphragms with plywood panel
sheathing. They have been subsequently modified
based on results from a variety of full-scale test programs that introduced additional materials, failure
modes, and design considerations.
Original I-joist framing products used laminated
veneer lumber (LVL) or sawn lumber flanges with
thicknesses of 1.5 in. or greater. These thicknesses
were consistent with 2 in. nominal material typically
used to “block” a lumber diaphragm and exceeded
specified minimum fastener penetration requirements. Provided that the diaphragm configuration
adhered to the manufacturer’s fastener spacing recommendations to avoid splitting and the flange material could rationalize equivalent fastener performance
to an appropriate sawn lumber species, it was judged
that application of shear design tables like 2009 IBC
Table 2306.2.1(1) could be extended to these products.
As I-joist products are optimized, it has become
common to see LVL flange thicknesses less than 1.5
in. The industry has long recognized that reducing
flange thickness beyond a certain threshold has the
potential to adversely impact sheathing nail embed9
Table 1. Peak strength comparisons between similar full-scale diaphragm tests1, 2
3,4
Notes for Table 1:
All of the tests summarized in this table used I-joist framing with laminated veneer lumber (LVL) flanges. The “DF”
and “SP” species designations represent Douglas-Fir and southern pine, respectively.
2
All of the tests summarized in this table used either plywood or oriented strand board (OSB) sheathing materials.
3
Except as noted in Lines 14 and 15, all of these full-scale diaphragms had dimensions of 24 ft. x 24 ft.
4
The majority of the comparisons are based on the average of two tests. However, only a single diaphragm was tested
in the following cases: Line 3 Variables 1 and 2, Line 14 Variable 2, and Line 17 Variable 1.
5
Sheathing nails sizes were as follows: 6d - 0.113 x 2.0 in., 8d - 0.131 x 2.5 in., 10d - 0.148 x 3.0 in.
1
ment and split resistance to the point where diaphragm capacity may be influenced. International
Code Council Evaluation Service (ICC-ES) acceptance
criterion for I-joists defines that threshold as 1-5/16
in. (ICC-ES 2010) based primarily on sheathing nail
embedment calculations (APA 2007). I-joist products
with flanges thinner than 1-5/16 in. are subsequently
required to conduct “full-scale horizontal diaphragm
testing” to rationalize performance.
At present, at least two manufacturers have conducted full-scale diaphragm testing to investigate
performance of I-joist product lines with flange thicknesses less than 1-5/16 in. Both manufacturers have
developed related design recommendations and limitations that are included in their evaluation reports.
In general, the manufacturers have proven equivalence to a subset of the current diaphragm design
tables for sawn lumber. They generally do not permit
10
thinner flanged I-joist products to be used in the
highest load applications that require the closest
sheathing attachment schedules.
Full-Scale Testing
At present, Weyerhaeuser has conducted 41 fullscale tests on I-joist diaphragm systems framed with
I-joists with LVL flange thicknesses between 1-1/8
and 1-1/4 in. Figures 1 and 2 illustrate test conditions chosen to be consistent with requirements of
ASTM E455 (ASTM 2004) and benchmark testing
conducted with sawn lumber (Countryman 1955). A
variety of I-joist materials, sheathing products, diaphragm configurations, sheathing fasteners, and fastening schedules have been tested to verify shear
transfer and deformation performance capabilities.
Given the high cost and relatively low variability of
WOOD DESIGN FOCUS
Figure 1. Typical diaphragm test configuration (24 x 24 ft., Case 5 shown).
full-scale assembly testing, two replicates have typically been tested for each condition per the requirements of ASTM E455. Nearly all I-joist diaphragms
have been tested with dimensions of 24 ft. by 24 ft.
to focus on shear transfer capabilities and to correspond with a benchmark sawn lumber diaphragm
test database.
Regardless of configuration, observed diaphragm
behaviors are fundamentally consistent. Initially, diaphragms deform elastically as a deep beam without
perceptible relative movement between framing and
sheathing. At loads in excess of design loads, visible
Summer 2010
relative movement can be observed between adjacent
sheathing panels and between panels and framing.
Figure 3 illustrates these trends for a 2009 IBC Table
2306.2.1(1) “Case 5” diaphragm configuration that
aligns panel joints in both directions. Shear flow
causes panels at the reactions to rotate in opposite
directions towards the span centerline. The magnitude of this rotation is typically consistent between
diaphragm tension and compression chords.
Due to panel geometry, the observed movement
between adjacent sheathing panels is typically several times greater along long edge joints than at short
11
Figure 2. Illustration of test setup (24 x 24 ft., Case 5 shown).
end joints. The large relative movement between
panel edges creates a tension perpendicular-to-grain
splitting force across the framing at the fasteners for
adjacent panels. At panel end joints, it induces perpendicular-to-grain forces into the framing as panel
end joints rotate and the nails induce perpendicularto-grain prying forces into the framing. Ultimately,
these deformations lead to failure in some combination of panel buckling/crushing, sheathing nail withdrawal, framing splitting, and/or sheathing edge tear
out. The Case 1 diaphragms tested exhibited similar
behavior with the exception that panel bearing and
crushing were also observed between interlocking
panel rows. Figure 4 illustrates the mode of failure
observed for a “blocked” Case 1 diaphragm. However, as with the benchmark sawn lumber tests, the
dominant failure modes observed with I-joist diaphragms were tension perpendicular-to-grain fracture of the framing and sheathing nail withdrawal.
Sheathing related failure modes played a less significant role. Given that many of the potential diaphragm failure modes that limit capacity are not typically addressed by a connection analysis, the importance of test-based verification for diaphragm systems that depart significantly from the historical basis seems to be confirmed.
Performance Trends
The compiled database of full-scale I-joist diaphragm tests provides an opportunity to draw com12
parisons between similar test sets. While extrapolations beyond tested conditions should be approached
with caution, comparisons in Table 1 suggest trends
that could be useful to the designer:
Flange Species: Douglas-fir LVL flanged I-joists outperformed their southern pine counterparts in 4 out
of the 5 similar diaphragm configurations tested
(Lines 1-5). This trend contradicts what is expected
based on a sheathing fastener connection analysis
that assumes a higher specific gravity for southern
pine. This trend may be due to the difference in tension perpendicular-to-grain strengths or typical veneer thicknesses of the LVL fabricated with each species. How well these particular commercial species
combinations fit the specific gravity-based fastener
design models may also play a role. Regardless, it
highlights that an I-joist manufacturer needs to
evaluate the diaphragm performance of each primary
species used for flange material. It also suggests that
designers should avoid applying the diaphragm recommendations for one I-joist product to another.
LVL Veneer Thickness: The Line 6 comparison illustrates what a relatively subtle difference in I-joist
product composition can have on capacity. Diaphragms framed with LVL flanges that had the same
species and grade but used a slightly thicker veneer
peel had about 15% less capacity. As with the last
item, this would seem to confirm that diaphragm performance is product dependent. It further emphasizes
that extrapolation of performance recommendations
WOOD DESIGN FOCUS
Figure 3. Typical diaphragm
movement
mechanisms
(Case 5 shown).
between seemingly similar manufacturers and products should be avoided.
Blocking Quality: Line 7 illustrates the influence
that blocking selection can have on capacity. Even
with I-joist materials taken as being a constant between tests, use of low specific gravity blocking material (0.45 vs. the intended 0.50) reduced diaphragm
capacity by about 15%. This shows that selection of a
blocking material is likely as important as selection of
a joist and should be consistent with the design assumption. For example, avoid using spruce-pine-fir
blocking if Douglas-fir diaphragm design values are
targeted.
Flange Width: Comparisons on Lines 8-11 illustrate that, as with sawn lumber, wider framing results in increased capacity. This is consistent with the
code design provisions and can be attributed to the
fact that wider framing tends to reduce splitting and
can provide for increased edge distance and staggered nail patterns that provide better load transfer.
Nail Size: Line 12 illustrates that a relatively intense 10d sheathing nail pattern resulted in a 16%
increase in capacity relative to using 6d sheathing
nails. In contrast, corresponding shear design capacities for lumber diaphragms increase by about 70%.
This highlights the importance of flange splitting and
the need for the manufacturer to address the resulting capacity limitations in their design guidance.
Also, since the design recommendations provided by
Summer 2010
the manufacturer are typically governed by “worst
case” conditions that involve larger diameter fasteners, there is likely some relative conservatism associated with smaller diameter fasteners. As with most
wood connections, it is another example where
more/larger fasteners do not necessarily improve
performance if they lead to splitting.
Diaphragm Case: The design codes permit six different diaphragm configurations to be constructed.
It is not intuitively obvious that they are all equal
when it comes to I-joist diaphragm performance.
The majority of tests summarized by Table 1 were
undertaken using the “Case 5” configuration illustrated by Figures 1-3 that aligns panel joints in both
directions. Since splitting was a primary concern, this
was judged conservative because it maximized the
number of fasteners and requires the full lateral load
to be transferred through the I-joist flanges. The Line
13 comparison confirms that assumption relative to
the “Case 1” configuration that requires fewer fasteners and interlocks the panels. It also suggests that
there are some relative benefits for the designer to,
whenever possible, favor specification of diaphragm
“cases” that use interlocking panels and fewer fasteners to transfer shear.
Diaphragm Size: The 24 ft. by 24 ft. diaphragm
size used for most of the testing summarized by Table 1 was chosen to promote shear distortion in a
condition that combined full-size sheathing elements
13
with at least 2 interior panel joints in each direction.
It also corresponded with a benchmark database for
sawn lumber (Countryman, 1955). As suggested by
Lines 14 and 15, testing other sizes may result in
slightly different answers. This highlights the importance for the manufacturer to evaluate a configuration that encourages realistic stress flows through the
system if design values are being developed. The designer should also specify products that have been
rationalized accordingly.
Fastener Type: Lines 16 and 17 provide some insight into the relative influence of fastener selection.
Eight penny (8d) ring shank (0.120 in. diameter) and
8d common (0.131 in. diameter) nails are assumed
to provide equivalent performance for unblocked diaphragms in some prescriptive situations. The fullscale tests of Line 16 suggest that the smaller diameter ring shank nail actually out-performed the larger
diameter common nail. Line 17 provides a similar
comparison for a proprietary fastener that claims superior diaphragm performance for some configurations based on small-scale fastener testing and analysis. In reality, the proprietary fastener performed
about the same as the smaller diameter ring shank
nail. Extra withdrawal and lateral resistance doesn’t
necessarily translate into improved diaphragm performance if an alternative failure mode not addressed by the fastener review, such as framing splitting, governs. The designer should be cautious when
specifying proprietary fasteners that claim diaphragm
performance improvements that have not been verified against all failure modes possible in a full-scale
diaphragm.
Stiffness Observations
In some cases, a designer will also need to predict
diaphragm deformation. For sawn lumber diaphragms, this is typically done using either the traditional “4-term” or simplified “3-term” diaphragm deflection equations (American Wood Council, 2008).
Calculation procedures developed for sawn lumber diaphragms also provide a reasonable means of
predicting I-joist diaphragm deformation in the design range. Figure 5 illustrates a comparison between
calculation methodologies and the measured behavior for a Case 5 I-joist diaphragm configuration that
conservatively combined large diameter fasteners
with a tight spacing that tends to promote splitting.
For this example, observed performance reasonably
approximated modeled deformation predictions
based on the tested Case 5 configuration. It should be
noted that the actual deformations are less than deformations predicted using the default apparent
shear stiffness term in SDPWS. The SDPWS default is
conservatively based on Case 1-4 diaphragm configu14
Figure 4. Failure modes — framing splitting from
panel prying (Case 1 shown).
rations which have fewer nails (e.g. greater load per
nail) at panel edges than the Case 5 diaphragm configuration tested.
One of the benefits of testing diaphragms with a
1:1 aspect ratio is that it focuses the test on the shear
strength and deformation of the assembly. A downside is that deflection measurements are small. For
example, at a load and resistance factor load for the
configuration illustrated, the disparity between the
measured deformation and the Case 5 predictions
was about 1/16 in. This absolute differential arguably falls below the reasonable precision of the full
scale test method and highlights that the absolute
magnitudes of deformation should be considered
when interpreting the accuracy of a predictive model.
Conclusions
Subject to the manufacturer’s recommendations,
pre-fabricated wood I-joists can be used for diaphragm construction. However, the performance of
an I-joist diaphragm assembly will be dependent on
the specific I-joist product used and its relevant attributes (i.e. flange geometry, material, species, veWOOD DESIGN FOCUS
Figure 5. Deflection predictions.
neer thickness for LVL flanges, etc.). Few I-joists can
serve as a direct substitute for sawn-lumber framing
in the full range of applications addressed by building code diaphragm design provisions. The manufacturer’s sheathing nail spacing and diaphragm design
recommendations should always be considered as
part of the design process.
References
APA-The Engineered Wood Association. 2007. TT061A: 1-5/16 in. Thick I-Joist Flanges and Diaphragm Nail Penetration. Tacoma, WA.
American Society for Testing and Materials. 2004.
E455: Standard Test Method for Static Load Testing
of Framed Floor or Roof Diaphragm Constructions
for Buildings. West Conshohocken, PA.
American Wood Council. 2005. National Design Specification for Wood Construction. Washington, DC.
American Wood Council. 2008. Special Design Provisions for Wind and Seismic. Washington, DC.
Countryman, David. 1952. Laboratory Report 55: Lateral Tests on Plywood Sheathed Diaphragms. Douglas Fir Plywood Association. Tacoma, WA.
Countryman, David. 1955. Laboratory Report 63a:
1954 Horizontal Plywood Diaphragm Tests. Douglas Fir Plywood Association. Tacoma, WA.
Summer 2010
International Code Council. 2009. International
Building Code. International Code Council. Country Club Hills, IL.
International Code Council Evaluation Service. 2010.
AC14: Acceptance Criteria for Pre-Fabricated Wood
I-Joists. Whittier, CA.
Peterson, J. 1983. Bibliography on Lumber and Wood
Panel Diaphragms. Journal of Structural Engineering. 109(12):2838-2852.
Tissell, J.R. and Elliott, J.R. 2004. Report 138: Plywood Diaphragms. APA-The Engineered Wood Association. Tacoma, WA.
Tissell, J.R. 1967. Laboratory Report 106: 1966 Horizontal Plywood Diaphragm Tests. American Plywood Association. Tacoma, WA.
Ned Waltz, PE, SECB, Senior Engineer, Product
Evaluation, Weyerhaeuser Company, Boise, ID. Contact
him at ned.waltz@weyerhaeuser.com. J. Daniel Dolan,
P.E., Professor and Director of Codes and Standards,
Composite Materials and Engineering Center, Washington State University, Pullman, WA. Contact him at
jddolan@wsu.edu.
15
AC14-0611-R1
#4
May 20, 2011
Mr. Jason V. Smart
Senior Evaluation Specialist
ICC Evaluation Service, Inc.
900 Montclair Road, Suite A
Birmingham, AL 35213
Re:
Proposed Revisions to the Acceptance Criteria for Prefabricated Wood I-joists, Subject AC140611-R1 (JS/KS)
Dear Mr. Smart:
We have reviewed the proposed revisions to AC14 and would like the committee to consider the following
comments and suggested modifications.
1) Remove portions of 2.3.2 and relocate to 2.3.2.2 (new) and 2.3.2.3 (new).
2.3.2 Full Scale Diaphragm Testing: I-joists with flange depths of less than 1-5/16 inches (33.3
mm)require full-scale horizontal diaphragm testing to justify their lateral load performance. Test
proposals shall be submitted to ICC ES for approval in advance. Any proposed diaphragm test
program shall conservatively bracket the range of permitted applications in terms of flange material,
flange dimensions, allowable lateral load, joist spacing, sheathing thickness, nail diameter/type, nail
spacing, blocking inclusion, and diagram “case” configuration as described by 2006 IBC Table
2306.3.1. The full-scale diaphragm test program shall include the following minimum elements:
2) Add 2.3.2.1 and group together requirements that are specific to the test method.
2.3.2.1 Test Method:
2.3.2.1.1 Diaphragms shall be tested as a simple-span diaphragms using ASTM E455 as the
reference test standard. Loads may be applied using either two- or four -point loads equally spaced
along the compression chord.
2.3.2.21.2 Test diaphragms shall have a1:1 aspect ratio and a minimum size of 24 feet by 24 feet (7.3
m by 7.3 m).
2.3.2.5 1.3 At least two replications shall be tested for each diaphragm configuration included in the
test program.
2.3.2.6 1.4 The permitted allowable shear (pounds per foot) seismic design load derived from the test
shall be the average maximum reaction achieved in all diaphragm tests of the same configuration,
divided by the diaphragm depth in feet, divided by a factor of 2.8.
Page 2
May 20, 2011
3) Add 2.3.2.2 and group together requirements that are specific to specimen design. Revise, expand,
and clarify 2.3.2.7 with 2.3.2.2.3 (new), 2.3.2.2.4 (new), and 2.3.2.2.5 (new).
Reason: Manufacturers may be compelled to evaluate a proprietary joist, fastener, or sheathing
material in the Case 1 configuration in order to evaluate the failure mechanism of their product in its
primary end-use condition. However, limited research has shown that in blocked diaphragms a Case 5
configuration may be a slightly (7% difference) more conservative configuration to test for qualification
purposes. While this difference is within the margin of error for diaphragm testing the suggested
revisions below recognize that Case 5 may be the conservative configuration for blocked diaphragms.
2.3.2.2 Specimen Design:
2.3.2.3 2.1 I-joist materials shall be used for the floor joistsframing members and parallel closure at the
diaphragm edges. Blocking used in the field of the diaphragm and rim board material used for
perpendicular closure shall be consistent with that permitted in the application.
2.3.2.4 2.2 The I-joist framing materials members may be reinforced as necessary to get the
concentrated test loads into and out of the test diaphragm. However, the framing reinforcement must
be consistent for all tests, and shall not contribute favorably to the sheathing-to-flange connection,
and shall notor increase the shear strength or stiffness of the constructed diaphragm by 2% or more.
This shall be confirmed by running at least one unsheathed diaphragm test as outlined in Section 9.1
of ASTM E455. The unsheathed diaphragm configuration used for the confirmation test shall be
compared against the sheathed diaphragm with the lowest capacity in the test program.
2.3.2.2.3 Equivalence for all unblocked diaphragm configurations may be established by testing the
unblocked Case 1 configuration with the highest desirable allowable shear for the targeted nominal
width of framing, as described by Table 2306.2.1(1) of the 2009 IBC or Table 2306.3.1 of the 2006
IBC.
2.3.2.2.4 Equivalence for Cases 1 through 4 blocked diaphragm configurations may be established by
testing the blocked Case 1 configuration with the highest desirable allowable shear for the targeted
nominal width of framing, as described by Table 2306.2.1(1) of the 2009 IBC or Table 2306.3.1 of the
2006 IBC.
2.3.2.2.5 Equivalence for all blocked diaphragm configurations may be established by testing the
blocked Case 5 configuration with the highest desirable allowable shear for the targeted nominal width
of framing, as described by Table 2306.2.1(1) of the 2009 IBC or Table 2306.3.1 of the 2006 IBC.
2.3.2.7 Diaphragm approvals shall be limited to the 2006 IBC Table 2306.3.1 “case” configuration
tested.
Exception: Any 2006 IBC Table 2306.3.1 “case” configurations shall be permitted in application when
a Case 5 diaphragm configuration is used for testing.
Page 3
May 20, 2011
4) Add 2.3.2.3 and group together requirements that are specific to member design. Revise and clarify
2.3.2.8 with 2.3.2.3.1 (new) and 2.3.2.3.2 (new).
Reason: Structural Composite Lumber (SCL) flange material is evaluated in accordance with ASTM
D5456 and AC47. Within these documents the variables to be considered for qualification and ongoing
quality control testing is clearly defined. Specifically, the evaluation of connection performance is
outlined in section 3.2 of AC47 which states, “Connection tests, in addition to those required in
Sections 3.2.1 and 3.2.2 of this criteria, must be conducted on each product having different wood
species. In addition, each grade shall be tested unless recognition of fasteners is based on the tests of
the lowest grade.” In addition, X1.3.9 of ASTM D5456 requires the evaluation of material splitting when
combinations of fastener size, spacing, and penetration are aligned in a row. The suggested revisions
below recognize the critical variables that are specific to a full-scale diaphragm test and complement
existing evaluation procedures.
2.3.2.3 I-Joist Member Design:
2.3.2.3.1 The I-joist member with the minimum desired flange depth and that product’s minimum width
shall be used to establish equivalency.
2.3.2.3.2 The I-joist member with the minimum desired grade and/or equivalent specific gravity for
connector design, as defined in section 3.2 of AC47 for both nail withdrawal and dowel bearing in the
Y-L and Y-X directions, shall be used to establish equivalency.
2.3.2.8 The applicant shall prove, by limited full scale diaphragm testing, that the i-joist flange material
used for the test program conservatively represents the range of materials that will share the same
design value for a given diaphragm configuration in application. The governing joist material shall be
selected for the test program. This comparison shall be made by testing single replications of the
diaphragm configuration with the highest corresponding design load. Variables to be considered in
this screening include, but are not limited to: flange thickness, wood species used for the flange
material, flange density, flange stiffness, flange width, veneer peel thickness for I-joists that use
laminated veneer lumber flanges.
Page 4
May 20, 2011
Thank you for this opportunity to comment on the proposed revision.
Sincerely,
Guy T. Anderson, P.E.
Engineering Manager, Codes and Standards
Boise Cascade Engineered Wood Products LLC
Joe Kaiserlik, P.E.
Product Development Manager
Georgia-Pacific Wood Products LLC
Borjen (“B.J.”) Yeh, Ph.D., P.E.
Director Technical Services Division
APA – The Engineered Wood Association
Phil Vacca, PE
EWP Sr. Engineer
Louisiana-Pacific Corp.
Dave Anderson, P.E.
EWP Senior Engineer
Roseburg Forest Products
AC14-6011-R1
#4
May 20th, 2011
Jason V. Smart
Senior Evaluation Specialist
ICC Evaluation Service, LLC
900 Montclair Road, Suite A
Birmingham, AL 35213
Subject: ICC­ES AC14­0611­R1
Dear Mr. Smart,
We have reviewed the proposed revisions to AC 14. As the proposed revisions provide a
standard set of provisions for diaphragm testing, Simpson Strong­Tie would like the ICC
committee to consider the following comments.
Section 2.3.2.7: The proposal, as written, requires the testing of each case configuration as
tested unless a Case 5 configuration is tested. The Case 5 condition is not a common case
condition that is seen in the field. Additionally, the panel configuration as it relates to the
framing may cause a failure mode that is not as prominent as the other cases. Case 1 is a very
common case used in the field. Therefore, consideration should be made to allow testing in the
unblocked and unblocked Case 1 condition and allow the other appropriate cases be bracketed
by that testing.
Thank you for your consideration of these comments and revisions.
Please email me at akhachadourian@strongtie.com or call me at (925) 560­9022 with any
questions or comments you may have.
Sincerely,
Simpson Strong­Tie Co., Inc.
[Electronic]
Aram Khachadourian, P.E.
R&D Engineer, SST
1 of 1
Simpson Strong-Tie Co., Inc.
5956 W. Las Positas Boulevard, Pleasanton, CA 94588
Phone: 925.560.9000 Fax: 925.847.1605 www.strongtie.com
AC14-0611-R1
Stanley Fastening Systems
Briggs Drive
East Greenwich, RI 02818
Robert J. Leichti, Ph.D.
Manager Product Compliance,
Hand Tools & Fastening
#4
T: 401-471-4166
F: 401-884-2485
E : Robert.Leichti@sbdinc.com
May 16, 2011
Mr. Jason Smart
ICC- Evaluation Service
900 Montclair Road, Suite A
Birmingham, AL 35213
RE: AC14-0611-R1
Dear Mr. Smart;
Staff has proposed revisions to the requirements for diaphragm testing of thin-flange wood Ijoists and the subject has been placed on the agenda for the upcoming Criteria Development
Hearing. It seems that diaphragm testing requirements have become an issue in several ACs.
Some clarifications are needed for test standards and analysis requirements in AC14, Section
2.3, and at the same time, some of the proposed revisions to AC14 are concerning in technical
justification and cost-benefit outcomes.
I had an opportunity to review the editorial revisions being proposed by the APA and their
members. Their revisions to the AC14 organization should be implemented.
The impact of the proposed revisions proposed in Section 2.3 related to size of test specimens
and numbers of tests could be far reaching even though the technical basis lacks justification.
My comments focus on three aspects of the proposed revisions: (1) diaphragm size and
boundary conditions, (2) numbers of tests, and (3) other considerations.
Issues related to test specimen size and boundary conditions
The paper by Waltz and Dolan (2010) is the most recent publication on wood-frame diaphragms
and is unique in that it reports information on wood I-joist diaphragms. Waltz and Dolan justify
the 24x24-ft dimensions based on the benchmark testing by Countryman (1955), which used
similar size, aspect ratio, and boundary conditions. However, the bibliography by Peterson
(1983) complied for the ASCE Committee on Wood and published in the Proceedings of the
Structural Division show that Countryman (1966) also conducted diaphragm test using other
sizes and reported using aspect ratios that ranged from 1:1 to 3.5:1 (APA, TR138). In fact, a lot
of the early diaphragm testing (1945-1975) was done on much larger dimensions (60 ft lengths
in some cases) and nearly all were in aspect ratios other than 1:1 (1:1 to 5:1). In the 1970s,
Johnson and others at Oregon State University specifically examined the effect of diaphragm
scale. Their work may still be available through the College of Forestry Publications Office, but
many of those test reports from the early 1970’s are out of print and not easily available. The
work by Dolan and Waltz shows that the effect of size and aspect ratio seem immaterial to I-joist
Mr. Jason Smart
RE: AC14-0611-R1
Page 1 of 4
diaphragms when they compared the performance of 24x24-ft and 12x24-ft diaphragms loaded
as simply supported beams and concluded that using a size other than 24x24-ft “may result in
slightly different answers.”
The test methods of ASTM E455 include both simply supported tests and cantilever tests. The
test method used by AISI in S907-08 is a cantilever test (minimum dimensions are 12x12 ft
where rectangular up to 1:1.33 is permitted).
The proposed use of test method of ASTM E455 treats the diaphragm as a deep beam on simple
supports. This analogy works well because we can manage engineering analysis using simple
mechanics theory. A cantilever beam analogy also works well as a simple mechanics model.
However, a diaphragm on a concrete foundation is not subject to those boundary conditions
and even the boundary conditions of the framed second-floor diaphragm differs from the model
in that the diaphragm is sitting on springs (the shear walls) rather than a pin and roller, and the
end rotations are constrained by the supported walls. The actual deformation of a framed
second-level diaphragm probably fits between the simply supported deformations and
cantilever deformations because the bottom plate of an attached wall when fully fastened acts
to stiffen the boundaries through partial composite action. The wall plates have a moment of
inertia that is greater than a 2x12 rim board in the plane of bending and also act to constrain the
rotation of boundary sheathing panels. No evidence is given by Waltz and Dolan or others that
indicates the cantilever method or the simply supported beam method yields a preferable or
objectionable outcome.
Issues related to numbers of required tests
ASTM E455 acknowledges that diaphragm tests are large and expensive and exhibit low
variability. Waltz and Dolan noted these same features. For that reason ASTM E455 permits
testing a single replication when multiple configurations are being evaluated. The object is to
use dissimilar constructions as the basis for rational engineering assessment. Consider, for
example, a situation where four diaphragm tests are conducted without replication: the analysis
would examine whether the tests are consistently below the Code values, whether they are
consistently above the Code values, or whether they are effectively equivalent to Code values.
A test of one configuration needs some replication, but with multiple configurations used in
comparison to code values, the different and non-replicated tests form a body of evidence that
that when used in a pairwise comparison to the Code values is as effective from a statistical
perspective as using means of two tests to make the same pairwise comparison. Ultimately, the
analyst has to interpret limited pairwise comparisons with few degrees of freedom.
Other considerations
(1)
The AC14 revisions are written in terms of “floor” diaphragms (Section 2.3.2.3). No
mention is made of roof diaphragms and whether the testing for floor diaphragms is applicable
to roofs (flat or pitched). A revision to the AC should specifically state that the results of
diaphragm tests are applicable to floor and roof diaphragms.
(2)
One should search the literature and anecdotal reports of structural inspections in the
aftermath of natural hazard events to find instances where floor and roof diaphragms have
exhibited failures in building structures that resemble the failures that we are looking at in
AC14; it is possible that no instances are reported. Even in the NDS seismic tests of the six-story
Mr. Jason Smart
RE: AC14-0611-R1
Page 2 of 4
wood-frame structure conducted in the summer of 2010, no diaphragm failures or distress was
observed even though the building experienced a MCE.
(3)
For lack of time, my remarks have not addressed the issues surrounding laboratory
expense, cost (and consumption) of materials, and time to conduct testing that is necessary to
address the many variables that are raised in the Staff proposed revisions. In seismic/wind force
resisting systems of wood-frame structures, most of the deformation occurs in the vertical
systems, so a positive cost-benefit ratio for a test program with replicated tests using a 24x24 ft
specimen is difficult to imagine.
(4)
A small-scale method of evaluating the relative diaphragm capacity as a function of
construction materials and failure mechanics related to sheathing rotations and interference like
that of a full-scale diaphragm is not difficult to imagine. It should be raised to the ASTM
Committee on Wood where it logically would be investigated and developed.
Synopsis -- What do we know about the fundamentals of testing wood-frame diaphragms?
• Evidence by Waltz and Dolan (2010) indicate that full-scale assembly testing
demonstrated “relatively low variability.” Low variability has been shown for shear
walls and assemblies in general.
• Evidence from the APA (TR138) indicates that sheathing strength is fully developed with
10d common nails and even 8d common nails (or equivalent for some sheathing
products), and as a result, in assemblies where the sheathing strength is fully developed,
failure is forced into the framing members. As a result, I-joist diaphragm performance is
enhanced to a point by using more smaller, ductile fasteners than fewer bigger
fasteners.
• The effect of configuration is difficult parse out of the data because the data are sparse.
Waltz and Dolan (2010) indicated that “behaviors are fundamentally consistent” for
different diaphragm cases. The dominate failure modes for I-joists as reported by Waltz
and Dolan paralleled those with sawn lumber as reported by Countryman – “tension
perpendicular to the grain fracture of the framing and sheathing nail withdrawal.” These
are the result of sheathing panel rotations in the plane of the diaphragm. No
information has been presented showing that either the simple beam method or the
cantilever test method produces “better results.”
• The effect of diaphragm dimensions is not obvious. Waltz and Dolan (2010) conclude
that “testing other sizes may result in slightly different answers.” If the work on
diaphragm scale by Johnson et al from the 1970s can be found, it will provide more
important information. However, in the absence of other information, it appears that
test specimens with planer dimensions greater than12 x 12 ft are unnecessary.
• Framing member dimensions affect diaphragm performance whether using sawn
lumber or I-joists – width and depth of sawn framing and I-joist flanges affect diaphragm
load capacity.
In summary, the AC14 proposed revisions to required specimen dimensions and numbers of
tests need further study and open discussion before the ES Committee can make a rational
assessment.
Mr. Jason Smart
RE: AC14-0611-R1
Page 3 of 4
The indulgences of the Committee and the readers of this letter are genuinely appreciated as I
commented on behalf of Stanley Fastening systems, LP and Stanley Black & Decker, Inc.
Sincerely,
STANLEY BLACK & DECKER
[Electronic]
Robert J. Leichti
Manager Product Compliance, Hand Tools and Fastening
Mr. Jason Smart
RE: AC14-0611-R1
Page 4 of 4