If you’ve ever asked yourself “Why?
” about something
related to structural steel design or construction,
Modern Steel’s monthly Steel Interchange is for you!
Send your questions or comments to solutions@aisc.org.
Chevron Brace with Both Braces in Compression
I am designing the gusset for a chevron brace connection
with both braces in compression. The Whitmore sections
of the individual braces overlap. How should this condi-
tion be treated?
The condition with both braces in compression is addressed
in the Example 5.9 of AISC Design Guide 29: Vertical Bracing
Connections—Analysis and Design (a free download for members
at www.aisc.org/dg), though the Whitmore sections in the
example do not overlap. It should be noted that a few different
approaches are proposed for checking the stability of the gus-
set. Each is only a model considered to be reasonable by the
authors. In your case, you must determine a reasonable model
based on your own engineering judgment.
I imagine there could be many approaches one could take.
You could simply ignore the portion of the Whitmore section
that overlaps. You could perform some type of stress interac-
tion check. You could run a fine-element analysis. Personally,
I would likely be okay with the overlap in many instances for
a few reasons. First, when we check the Whitmore section, we
assume an even stress distribution along the Whitmore section
area which is established using the 30° angle. This was found
to give a good prediction of the peak stresses measured from
aluminum joint testing performed by Whitmore [Whitmore,
R.E. (1952), “Experimental Investigation of Stresses in Gusset
Plates,” Bulletin No. 16, Civil Engineering, The University
of Tennessee Engineering Experiment Station, Knoxville,
TN.]. Stress trajectories were plotted from the test data, and
they vary greatly along the Whitmore section. The stresses
were lower near the ends of the Whitmore section where the
overlap occurs in your situation, although connection con-
figurations could impact the stress distributions. Also, with the
braces both being in compression, I imagine the stress level
will be quite a bit lower than the yield strength of the plate.
Carlo Lini, PE
NDT and Special Inspection Waivers
Please confirm that when third party special inspections
are waived by the authority having jurisdiction over the
project, the NDT requirements in Chapter N of the Spec-
ification are also waived.
This is not correct. Section N7 clarifies the intent, stating:
“Quality assurance (QA) inspections, except nondestructive
testing (NDT), may be waived when the work is performed in a
fabricating shop or by an erector approved by the authority hav-
ing jurisdiction (AHJ) to perform the work without QA.” NDT
must be performed even when the QA inspections are waived.
Larry S. Muir, PE
 Modern STEEL CONSTRUCTION
steel
interchange
Channels Warped During Galvanizing
Several channels we are using on a current project have
warped significantly during galvanizing. It has been sug-
gested that the channels may have been a poor choice for
these members. Is there are validity to this suggestion?
Yes. ASTM A384: Standard Practice for Safeguarding Against
Warpage and Distortion During Hot-Dip Galvanizing of Steel
Assemblies recommends the use of symmetrical shapes and
singles out channels as a member type that typically requires
straightening after galvanizing. The April 2004 article “Galva-
nizing Tips” (available at www.modernsteel.com) reinforces
this point and provides other useful tips related to galvanizing.
In addition, the American Galvanizers’ Association sug-
gests a collaborative effort should be used to achieve the best
results: “The design of parts to be hot-dip galvanized is the
responsibility of the design engineer and the architect; how-
ever, when there is a part that has an asymmetric design the
galvanizer should let his customer know the part is very likely
to distort during the galvanizing process.”
Larry S. Muir, PE
Removal of Shim Stacks
For base plates that are shimmed and grouted, does AISC
consider it necessary to remove the shim plates and pack
grout in the voids left by the shims?
No, AISC standards do not require removal of the shim
stacks. Leaving the shims in place under the base plate is
common practice.
Ideally, column bases should be grouted as soon as possible
in construction when the axial load to the column is only a small
fraction of what the total anticipated final load could be. Done
properly, the base plates should be grouted before any concrete
is cast for the elevated floors when the only load delivered to
shim stacks is the weight of the bare steel frame and some con-
struction live loads. As additional load is added to the column,
the grout will then distribute the load to the foundation.
Axial compressive forces from the column can be assumed
to be evenly distributed as bearing forces on the shims and
non-shrink grout. Even if the shims were to start out taking the
majority of the load, the assembly will deform in a self-limiting
manner through localized yielding of the steel as the force-
distribution model assumed in sizing the base plate is attained.
Susan Burmeister, PE
 steel interchange
A More Efficient Approach to Uplift
I have designed a 30-ft-long W14×22 roof beam to resist
gravity loads. However, when we check the beam for
wind uplift, bottom flange bracing is required at the mid-
span. The W14×22 seems like a reasonable size for this
application, and I have seen it called out on other similar
projects without bottom flange bracing. Is there a method
that might permit me to omit the bracing?
Based on the scenario you have described, bracing of the bot-
tom flange might be needed. I have personally used bottom-
flange bracing on numerous projects where the wind uplift
pressures exceed the roof dead loads. In my experience, this is
not an uncommon practice.
However, you may be able to calculate enough extra capac-
ity for your beam if you take a closer look at the value of Cb used
in your analysis. The Commentary to Section F1 of the AISC
Specification provides some additional formulas that can be used
to calculate Cb for a roof beam subject to uplift loads, as shown in
Figure C-F1.5. This may increase the available strength enough
to eliminate the need for bottom flange bracing. It is certainly
worth investigating, especially for repetitive beam conditions.
Susan Burmeister, PE
Cambering Plate Girders and Heavy Beams
Can very large and very long beams, such as a 56-ft-
long W40×593, be cambered? Likewise, can a 56-ft-long,
50-in.-deep plate girder be cambered?
Many fabrication shops have the capability to camber typical
floor beams using a cold-bending operation (cold cambering).
If the machine capacity is exceeded, heat can be applied to the
member to reduce the yield stress. Because many bender-roller
companies have specialized, high-capacity equipment, it is often
more economical for the fabricator to sublet the cambering of
large beams. However, it is doubtful that a 56-ft-long W40×593
could be cambered by cold-bending or heat-assisted bending.
Another potential option is heat curving, which is a bend-
ing process that relies only on the application of heat in
specific patterns to induce curvature. This method is used pri-
marily used by fabricators for cambering and curving to very
large radii and for repairing damaged members. You should
contact a fabricator to get their advice on this method.
Generally, plate girders cannot be efficiently cold-bent about
the strong axis due to the high depth-to-thickness ratio of the
web. In most cases, cold bending would cause local web buckling
during the bending operation. The welding of the section also
would be a challenge, since curving means plastic deformation
and shear in the welds probably much greater than the design
anticipated for loads in service. It’s likely too that plate girders
would usually exceed the capacity of the available cambering
machine. Fortunately, there is another way. Plate girders are often
cambered by cutting the web to the desired curvature, and then
welding the flanges in place. This may be the best option.
Bo Dowswell, PE, PhD
DECEMBER 2016
Demand-Critical Welds on Seismic Projects
Must all welds on a seismic project meet AWS D1.8,
making them all demand-critical welds?
No. Your question indicates quite a bit of confusion about
the requirements and the terms used in AISC 341: Seismic
Provisions for Structural Steel Buildings. I will try to clarify the
requirements for you.
First, seismic effects must be considered for all projects. By
seismic project, I assume you mean a project that must meet
the Seismic Provisions.
The Seismic Provisions make several references to AWS
D1.8. Each of these applies only to welds within the seismic
force resisting system (SFRS). For example, Section A3.4a
states: “All welds used in members and connections in the
SFRS shall be made with filler metals meeting the require-
ments specified in clause 6.3 of Structural Welding Code—Seis-
mic Supplement (AWS D1.8/D1.8M).” Welds outside the SFRS
need not satisfy AWS D1.8.
Additionally, welds required to satisfy AWS D1.8 are not
necessarily demand-critical welds. Demand-critical welds are
a subset of the welds addressed in AWS D1.8. This is can be
seen in the User Note that accompanies Section A3.4a, which
states: “AWS D1.8/D1.8M subclauses 6.3.5, 6.3.6, 6.3.7 and
6.3.8 apply only to demand-critical welds.”
It should be noted that per Section A4, the engineer is
responsible for identifying the welds subject to requirements
beyond those in AWS D1.1 through “Designation of the
SFRS,” “Identification of the members and connections that
are part of the SFRS” and providing the “Locations of demand
critical welds.”
Larry S. Muir, PE
The complete collection of Steel Interchange questions and answers is available online.
Find questions and answers related to just about any topic by using our full-text search
capability. Visit Steel Interchange online at www.modernsteel.com.
Larry Muir is director of technical assistance and Carlo Lini is staff engineer—technical
assistance, both with AISC. Susan Burmeister and Bo Dowswell are consultants to AISC.
Steel Interchange is a forum to exchange useful and practical professional ideas and
information on all phases of steel building and bridge construction. Opinions and
suggestions are welcome on any subject covered in this magazine.
The opinions expressed in Steel Interchange do not necessarily represent an official
position of the American Institute of Steel Construction and have not been reviewed. It is
recognized that the design of structures is within the scope and expertise of a competent
licensed structural engineer, architect or other licensed professional for the application of
principles to a particular structure.
If you have a question or problem that your fellow readers might help you solve, please
forward it to us. At the same time, feel free to respond to any of the questions that you
have read here. Contact Steel Interchange via AISC’s Steel Solutions Center:
866.ASK.AISC • solutions@aisc.org
 If you’ve ever asked yourself “Why?” about something
related to structural steel design or construction,
Modern Steel’s monthly Steel Interchange is for you!
Send your questions or comments to solutions@aisc.org.
Headed Stud Anchor Diameter for Developing Flexural Strength of Spliced
Composite Beams
What should be taken into account for selecting headed
stud anchor diameter for composite steel beams? Are there
any limitations on using ¾-in.- or ½-in.-diameter studs
welded through metal deck to create composite action?
There are some considerations in selecting headed stud
anchor (stud) diameter and a few limitations that are indepen-
dent of diameter.
Size selection: Per the AISC Specification (a free download
at www.aisc.org/specifications) Section I3.2c (1)(2), studs
shall be ¾-in. or less in diameter. Also, per Section I8.1, the
diameter of the stud shall not be greater than 2.5 times the
thickness of the beam flange unless the stud is welded directly
over the beam web. Per Section A3.6, headed studs shall con-
form to AWS D1.1, which only addresses studs ½-in. in diam-
eter and larger, which therefore defines the lower size limit.
Those are the specific code provisions pertaining to diameter
limitations. Beyond that, it becomes an engineering assess-
ment by you as to what is the more economical or practical
solution to transfer the shear for your specific beam.
The cross-sectional area of a ¾-in. diameter stud is more
than double the area of a ½-in. diameter stud. Since stud shear
capacity is directly proportional to stud cross-sectional area, you
will need more than double the quantity of ½-in. diameter studs
to provide the same strength as ¾-in. diameter studs. You could
confirm with a local steel fabricator or erector in your area, but
I would expect that the labor cost of installing a larger quantity
of smaller-diameter studs would exceed any cost benefit associ-
ated with using smaller-diameter studs unless you have a floor
system that uses small beams and doesn't demand very much in
the way of shear transfer between the steel and concrete.
With respect to welding through deck, ICC-ES report
ESR-1094 provides some good information. This may not be
the only report available. Section 4.1 provides some guidelines
for when studs can be welded through two layers of deck-
ing versus only one layer. This report does not distinguish
between various stud diameters in defining the limitations on
welding through deck.
Studs should not be welded through coated sheet metal
other than typical steel decking. Welding through other mate-
rials, such as galvanized architectural flashing materials and
even paint on the beam flange, can introduce contaminants
into the weld that could affect the weld performance. Standard
structural composite decks have controls in place to limit the
potential of contaminants from the coatings.
Susan Burmeister, PE
 Modern STEEL CONSTRUCTION
steel
interchange
Wide-Flange Members
Specification Section J6 has always vexed me because it
seems impossible to satisfy the requirement for groove-
welded splices. By my calculation, for a W36×160 the
weld-access hole reduces the flexural strength of the
member to less than 70% of the flexural strength of the
member without the weld access hole. My calculation is
based on the moment being resisted only by the flanges,
which are governed by yielding on the area of the flange.
The shear strength is likewise reduced due to the pres-
ence of the weld access hole to 85% to 90% of the shear
strength of the member. How does the spec intend for
the designer to "develop the strength" of the shape?
Relative to the flexural strength, the situation is similar to
directly welded beam-to-column moment connections. The
argument is sometimes made that one cannot develop the
strength of the beam by connecting only the flanges while
at the same time reducing the overall area by including weld
access holes. An explanation is provided in the May 2012 arti-
cle “Developing Mp” (available at www.modernsteel.com).
Though I am not aware of a document that addresses shear
in this manner, a similar argument could be made. In fact, I
suspect you make a similar assumption all of the time without
giving it a second thought. It is not common for engineers to
perform a yielding check on an uncoped beam with a bolted
connection. Even a net area check is not typically considered
necessary, since the flanges will tend to prevent such a failure.
The only check commonly made is gross shear (yielding)
based on the full area of the web. Regardless of the size of the
weld access holes, they certainly remove less area than a full
depth bolted connection.
It is also important to note Specification Section J6 is not
requiring the connection to develop the actual strength of the
member or the expected strength of the member (as we might
in seismic design) but rather the “nominal strength of the
smaller spliced section.”
The case of the expected strength is an interesting one.
The welded unreinforced flange-welded web moment con-
nection in AISC 358 Prequalified Connections for Special and
Intermediate Steel Moment Frames for Seismic Applications (a free
download for members at www.aisc.org/seismic) has been
shown through physical tests not only to develop the nomi-
nal strength of the beam but also to develop the actual beam
strength and force hinging of the beam outside the connec-
tion. This is accomplished even with the larger weld access
holes required in AWS D1.8.
Larry S. Muir, PE
 steel interchange
Tributary Length for Prying Action
The article “A Slightly Longer Look at Prying” (an online
supplement to the July 2016 article “A Quick Look at Pry-
ing,” available at www.modernsteel.com) states that the effec-
tive width, p, for prying action can be conservatively taken as
3.5b but cannot exceed the spacing between the bolts. This
conflicts with Part 9 of the Manual, which indicates that p is
limited to twice b. Can the effective width exceed 2 times b?
Yes. The statement in the article is based on the upcoming
15th Edition Manual, which will revise the default effective
width from p = 2b to p = 3.5b. The new default value is based
on guidance provided by the South African Institute of Steel
Construction that was evaluated by the AISC Manual Commit-
tee and deemed to be adequate. The new assumed distribution
angle is 60°, which is conservative but not as conservative as the
assumed 45° angle used in the 14th Edition Manual.
It should be noted that the 2p limit was not intended to be a
requirement. Even though it is not stated, it was only a recom-
mendation. The recommendation was established because it was
brought to the attention of the Manual Committee that there
was a wide range of assumed tributary lengths being used in
practice. It was felt that the Manual should provide guidance. As
is often the case when engineers are forced to provide guidance,
the first pass was conservative. Given the lack of data available
at the time, the committee felt that the 2b guidance was a safe
lower bound. With a closer look at the South African Institute
of Steel Construction data the 3.5b limit was adopted. Note that
the Manual also allows that a larger tributary length may be jus-
tified based upon testing or rational analysis.
Carlo Lini, PE
Fatigue and Removal of Backing for Fatigue
We have received shop drawings for a steel structure with
moment frame connections using complete joint penetration
groove welds. It is not a high-seismic project but we do have
fatigue design considerations. The contractor has indicated on
the shop drawings that backing bars will be used. We requested
that the backing be removed in our review comments. The
contractor is asserting that this is an unusual requirement and
is treating this as a change in the contract. We believe that since
the structure is subjected to fatigue, the contractor should be
required to remove the backing at no additional cost. How is
this situation treated in AISC documents?
This situation is not directly addressed by any AISC docu-
ment. However, it is addressed in AWS D1.1, which is adopted
by reference in Section J2 of the Specification.
Clause 2.17.2 of AWS D1.1 addresses backing and directly
addresses the removal of backing, which often can generally be
left in place. The treatment of backing is tied to fatigue consid-
erations as you indicate. Clause 2.17.2.1 requires the engineer
to provide the fatigue stress category in the contract drawings. If
you provide the applicable fatigue stress category in the contract
documents and AWS D1.1 Clause 2.17.2 requires removal for
that fatigue stress category, backing removal is required. Other-
JANUARY 2017
wise, adding a requirement to remove backing with comments
during shop drawing approval or by RFI response may repre-
sent a change to the contract. Section 4.4.3 and 9.3 of the AISC
Code of Standard Practice (a free download from www.aisc.org/
specifications) addresses revisions to the contract documents.
Carlo Lini, PE
Short-Headed Stud Anchors
Chapter I of the AISC Specification requires that “stud shear
connectors, after installation, shall extend not less than 1½
in. above the top of the steel deck.” I have an existing building
and the original design documents indicate the shear studs
extend only 1 in. above the steel deck. When calculating the
composite strength of this member, is there a reduction fac-
tor that can be used to account for shorter stud?
AISC does not have sufficient information to make a recom-
mendation about the performance of composite flexural members
when the stud projection above the deck flutes is less than 1½ in.
Therefore, the Specification does not provide a reduction factor
for use with the current equations. You would have to use your
own engineering judgment. The parameter limitations noted in
current Section I3.2c were established to ensure beam designs
are performed within the margins of the available research data,
largely summarized in the first quarter 1977 Engineering Journal
article “Composite Beams with Formed Steel Deck” (available for
free to AISC members at www.aisc.org).
Provisions for composite members with formed steel deck
did not appear in the AISC Specification until 1978. At that
time, the Specification required the same 1½ in. projection per
the research in the above-mentioned article. However, Sec-
tion 1.11.6 stated: “When composite construction does not
conform to the requirements of Sects. 1.11.1 through 1.11.5,
allowable load per shear connector must be established by a
suitable test program.” This may have permitted a designer to
use a shorter stud projection if they had access to some other
test data in order to establish their shear connector strength.
Susan Burmeister, PE
The complete collection of Steel Interchange questions and answers is available online.
Find questions and answers related to just about any topic by using our full-text search
capability. Visit Steel Interchange online at www.modernsteel.com.
Larry Muir is director of technical assistance and Carlo Lini is staff engineer—technical
assistance, both with AISC. Susan Burmeister is a consultant to AISC.
Steel Interchange is a forum to exchange useful and practical professional ideas and
information on all phases of steel building and bridge construction. Opinions and
suggestions are welcome on any subject covered in this magazine.
The opinions expressed in Steel Interchange do not necessarily represent an official
position of the American Institute of Steel Construction and have not been reviewed. It is
recognized that the design of structures is within the scope and expertise of a competent
licensed structural engineer, architect or other licensed professional for the application of
principles to a particular structure.
If you have a question or problem that your fellow readers might help you solve, please
forward it to us. At the same time, feel free to respond to any of the questions that you
have read here. Contact Steel Interchange via AISC’s Steel Solutions Center:
866.ASK.AISC • solutions@aisc.org
 If you’ve ever asked yourself “Why?” about something
related to structural steel design or construction,
Modern Steel’s monthly Steel Interchange is for you!
Send your questions or comments to solutions@aisc.org.
Composite Beams
A project on which we are installing shear studs specifies
a composite steel system comprised of 2 in. concrete over
3 in. metal deck. Headed anchor studs, ¾ in. in diameter,
are specified and noted to be a minimum of 1.5 in. above
the deck and ½ in. below the top of the concrete. In an
ideal situation, this can theoretically be achieved with 47⁄8-
in. studs that achieve 4½ in. of finished length. However,
this only occurs where studs are installed through metal
deck and 3⁄8-in. burn-through is theoretically achieved. At
girders parallel to deck direction where the stud attaches
directly to the girder flange, the theoretical burn-through
is 3⁄16 in. and thus the finished length is 411⁄16 in. Both con-
ditions run a high risk of being exposed when typical fab-
rication tolerances are considered (crown-up fabrication)
even if there is no camber required. Section I3.2c of the
AISC Specification has the following requirements: 2 in.
minimum slab over deck, 1.5 in. minimum length above
metal deck and ½ in. minimum of concrete cover to sur-
face. Are there permitted deviations to this rule? Are two
different stud lengths required in this situation?
The system you have described satisfies the requirements of the
AISC Specification but, as you’ve noted, does not allow much
room for tolerance. The specific provision in Section I3.2c(1)(2)
states: “Steel headed stud anchors, after installation, shall extend
not less than 1½ in. above the top of the steel deck and there
shall be at least ½ in. of specified concrete cover above the top
of the steel headed stud anchors.” There are a couple of nuances
within the wording here that are worth pointing out.
First and foremost, the 1½ in. minimum stud projection
above the deck is structurally more important to the perfor-
mance of the system than the ½ in. clear cover over the top.
Purely from a strength perspective, the concrete cover over the
top of the stud provides no recognized additional capacity. In
the above referenced language, the phrase “specified concrete
cover” was carefully chosen and deliberated over within the
technical committee that maintains this section of the Specifica-
tion. The intent is to ensure that designers specify a minimum
of ½ in. concrete coverage to account for some of the field
inaccuracies, but it was recognized that the in the final, as-built
condition, the coverage could be less. The Commentary to this
section of the Specification discusses ways an engineer can miti-
gate the potential for exposed studs in their slab system which
are obviously more critical in a thin-slab system.
So, to answer your first question, it is acceptable to
encroach into the ½ in. cover if necessary, but the 1½ in. mini-
mum stud projection should be maintained.
As to whether or not two different stud lengths are
required, I think that is a question that should be posed to
the engineer of record. If the design specifies a uniform slab
 Modern STEEL CONSTRUCTION
steel
interchange
thickness of 2 in. everywhere, regardless of whether or not the
final floor is level, then I would be inclined to say two different
stud lengths are not necessary. However, if the design specifies
a level floor finish then it is possible that if beam cambers do
not come out, you could have exposed studs and that extra 3⁄16
in. of stud length could become very important.
Susan Burmeister, PE
Web Compactness for Singly
Symmetric I-Sections
I am designing a singly symmetric I-shaped member
in flexure. The plastic neutral axis for this section falls
within the compression flange resulting in a negative
value for hp/2. How can I determine whether the web is
compact, non-compact or slender? Note that if the web is
not compact, then Section F4 of the Specification applies
and since λp is equal to λr the denominators in Equations
F4-9b and F4-16b become zero—again resulting in a
result that is difficult to interpret.
Table B4.1b of the AISC Specification applies to compression
elements of members subject to flexure. If hp/2 is within the
flange, then, under a plastic stress distribution, the web is in
tension and therefore doesn’t need to be classified. If hc /2 is not
within the flange, then, under elastic stress, some portion of the
web will be subjected to a linearly varying compression load.
In such a case, the magnitude of the compression stress will be
relatively small when the section is elastic. As more and more of
the section is strained beyond the elastic limit, the length of web
in compression will decrease. Both of these trends tend to indi-
cate that the stability of the web will not be a concern.
There are several possible approaches. First, the limits
could be calculated based on Case 15, the doubly symmetric
case, with the length of the web, h, assumed to be hc. I believe
this would be a conservative approach. The coefficient of λr is
the same for the doubly symmetric and singly symmetric cases.
Now consider the calculation of λp. If the equation for Case 16
is applied to a doubly symmetric I-shape hc /hp is 1.0. A reason-
able value for the shape factor of a rolled wide flange is 1.12.
This value produces a coefficient of 3.77—pretty close to the
coefficient for Case 15, 3.76. So Case 16 produces about the
same result as Case 15 assuming the same parameters.
There are two ratios that determine the value of λr for Case
16. The first is hc /hp. For a case like yours, with the larger flange
in compression this ratio will always be greater than one. A
negative value for hp does not make sense physically relative
to checking the stability of the web. However, as hp approaches
zero, it can be seen that the value for hc /hp becomes very large.
This again tends to indicate that buckling of the web becomes
less and less of a concern. The other factor is related to the
shape factor, Zx/Sx, which is obviously in the same proportion
 steel interchange
as Mp/My. Up to a shape factor of about 2 the denominator will
be less than one, tending to increase the coefficient. Beyond
this shape factor, the coefficient will begin to decrease. Why
should this be? The greater the shape factor, the more inelastic
deformation will be required to fully yield the section. In other
words, the demand becomes greater and greater. Also, at a shape
factor of about 2, the shape is likely moving from a singly sym-
metric I-shape to something approaching a tee. It is interesting
to note that there is no case addressing the web of a tee with the
flange in compression. To me, this is another indication that at
this extreme the stability of the web is not a concern.
This condition will be addressed in the Commentary to the
2016 Specification. The following statement has been added:
“In extreme cases where the plastic neutral axis is located in
the compression flange, hp = 0 and the web is considered to be
compact.” This corresponds to the logic above.
If the web is compact, then Section F3, not Section F4,
applies and a zero will not appear in the denominator.
I believe it is always appropriate (necessary!) to exercise
engineering judgment. It is especially critical to do so when
addressing conditions at the fringes of those considered in the
Specification. It seems there are two different extremes that can
cause the plastic neutral axis to be located in the compression
flange. One would be where the compression flange is very
clearly compact—i.e., it is very thick and relatively narrow.
In such a case, it would seem the assumption that the web is
compact is uncontroversial. At the other extreme, where the
compression flange is very thin but very wide, I would be hesi-
tant to treat the condition using Case 16. The distribution of
stress typically assumed when calculating hc and hp might not
be appropriate when the effective flange consists of a very thin
but very wide element.
Larry S. Muir, PE
Not Qualified vs. Not Approved judgment and knowledge.
in ASTM F3125
The new ASTM F3125, which consolidates the previous
ASTM A325, A490, F1852 and F2280 standards, indicates
in Table A1.1 that F1136 coatings are not approved for use
with twist-off bolts (Grades F1852 and F2280). It is my
understanding that this indicates that these coatings are
prohibited for use with twist-off tension-control bolts. Some
vendors state that these bolt-coating combinations are not
prohibited. What is the intent?
You are referring to an ASTM standard. Therefore ASTM
would be the appropriate source for an interpretation. I will,
however, provide my own opinion.
F3125 provides two different descriptions: not approved
and not qualified. These terms are defined in the standard:
➤ “Not qualified” in Table A1.1 means that a particular
coating has not been qualified and accepted by ASTM
committee F16 for use on 150 ksi/1040 MPa bolts.
➤ “Not approved” in Table A1.1 means that a particular coat-
ing was not approved for a particular bolt style or grade in
the individual standard prior to combination into F3125.
NOVEMBER 2016
The reason for the different designations may not be
immediately clear, since both would seem to discourage the
use of the coating with the fasteners listed. However, the
Annex also states:
“Coatings listed in this Annex for 150 ksi/1040 MPa
bolts have been qualified and approved where indicated
for use with 150 ksi/1040 MPa strength bolts. For
use on 150 ksi/1040 MPa bolts, other coatings must
be qualified in accordance with IFI 144. Hydrogen
embrittlement testing required by IFI 144 shall be per-
formed in accordance with F1940 for internal hydrogen
embrittlement and F2660 for environmental hydrogen
embrittlement.”
A footnote to the table in the Annex states:
“Other metallic and nonmetallic coatings may be used
on 120 ksi/830 MPa minimum tensile fasteners upon
agreement between the purchaser and user. Perfor-
mance requirements shall be specified by the purchaser
and agreed to in writing. Coatings for 150 ksi/1040
MPa bolts must be qualified. See A1.1.”
So the requirements for F1852 and F2280 are different. The
standard does not prohibit any coating to be used with F1852
“upon agreement between the purchaser and user” with “per-
formance requirements…specified by the purchaser and agreed
to in writing.” For F2280, coatings must be must be qualified.
The difference seems to involve hydrogen embrittlement.
Galvanizing of A490 bolts (150 ksi) has been prohibited for some
time. This is because the process can lead to hydrogen embrittle-
ment, which can lead to failure of the bolt and is therefore a
safety concern. The same concerns do not exist for 120 ksi bolts.
I take “not approved” as meaning that this combination
has not been explicitly considered, but there is no reason to
believe there is an inherent safety concern. Therefore, if you
are going to do it you are on your own, relying on your own
I take “not qualified” as meaning that there are known safety
concerns with this combination, and it should not be used.
Larry S. Muir, PE
The complete collection of Steel Interchange questions and answers is available online.
Find questions and answers related to just about any topic by using our full-text search
capability. Visit Steel Interchange online at www.modernsteel.com.
Larry Muir is director of technical assistance at AISC. Susan Burmeister is a consultant to AISC.
Steel Interchange is a forum to exchange useful and practical professional ideas and
information on all phases of steel building and bridge construction. Opinions and
suggestions are welcome on any subject covered in this magazine.
The opinions expressed in Steel Interchange do not necessarily represent an official
position of the American Institute of Steel Construction and have not been reviewed. It is
recognized that the design of structures is within the scope and expertise of a competent
licensed structural engineer, architect or other licensed professional for the application of
principles to a particular structure.
If you have a question or problem that your fellow readers might help you solve, please
forward it to us. At the same time, feel free to respond to any of the questions that you
have read here. Contact Steel Interchange via AISC’s Steel Solutions Center:
866.ASK.AISC • solutions@aisc.org
 If you’ve ever asked yourself “Why?” about something
related to structural steel design or construction,
Modern Steel’s monthly Steel Interchange is for you!
Send your questions or comments to solutions@aisc.org.
Eccentricity at Axially Loaded
Beam-End Connections
Figure 1
Shop-welded, field-bolted beam-to-beam double-angle
connections must transfer both axial and shear end reac-
tions. In Figure 1 the beams are non-composite and the
entire axial force must be transferred through the con-
nections. The connections will be designed assuming that
only the top three rows of bolts, which are common to
both connections, will transfer the axial force.
Does the eccentricity of the axial load from the beam
centerline to the centerline of the bolt group need to
be considered in the design of the connection? Does an
eccentricity need to be considered when evaluating the
coped section?
Axially loaded double-angle connections are typically designed
without considering an eccentric moment. The rotational
stiffness of the connections is typically much lower than the
stiffness of the beam. Therefore, the beam will carry almost
all of the moment due to the eccentricity, and it is common
to assume that the beam resists the entire moment. It is also a
good idea to use the maximum number of bolt rows that will
fit into the web when resisting axial end reactions, as this will
tend to minimize whatever eccentricity does exist.
Because floor systems are usually modeled with the beam
elements at the same elevation, the eccentricity between the
axial load and the beam centroid is often neglected.
I would analyze the coped section assuming an eccentricity
relative to the axial load equal to the distance between the cen-
troid of the bolt group and the centroid of the coped section. For
axial loads in tension, the resulting moment opposes the moment
caused by a downward vertical beam shear. I would also locate
the bolt-group centroid as close to the beam centroid as practi-
cal, typically using the maximum number of bolt rows that will fit
into the web. In practice, the small eccentricity that might exist is
sometimes neglected based on engineering judgment.
 Modern STEEL CONSTRUCTION
steel
interchange
As you stated, the bottom row of bolts on the right side
of the connection should be neglected relative to the transfer
of the axial load. However, all of the bolts will participate in
transferring shear.
Other approaches are possible and contract-specific
requirements could be imposed, but the comments above
reflect what I understand to be common practice.
Bo Dowswell, PE, PhD
Cambering of Cantilevered Beam Framing
Continuously Over Column
I have a beam framing con-
tinuously over a column simi-
lar to the condition shown in
Figure 2-2a of the 14th Edi-
tion AISC Manual (available
at www.aisc.org/publications);
see Figure 2. In my case, the
right-hand side of the beam
cantilevers 17 ft beyond the
column, and the left-hand side
is a 10-ft back span. The beam
is a W18 and the column an
HSS4×4. I wish to put a cam-
ber in the cantilevered section
such that its end will be 3/4 in.
higher than the elevation at the supported when erected.
I am aware that there are issues with providing camber
for cantilevered or moment-connected beams. However,
is it feasible to camber a cantilevered beam framing con-
tinuously over a column?
No. It is generally not feasible to camber a cantilevered beam
framing continuously over a column.
When designers call for a simply supported steel beam to be
cambered, the steel fabricator applies a load or heat to the beam
to introduce a permanent deformation in a roughly parabolic
shape with the apex at mid-span. But based on the geometry
you’ve described, this “conventional” method of introducing
camber into a member does not seem like it would be appropri-
ate for your condition. Additionally, since your member is only
27 ft long, it may not be a candidate for cambering depending
on the fabricator's cambering method or equipment. Typically,
it is not recommended to camber members less than about 30
ft long because most cambering equipment is not configured to
accommodate shorter members.
In general, when engineers do specify camber for a canti-
levered beam, it is provided in a manner that is different than
the way we typically think of camber being introduced into a
simply supported steel beam. The beam itself usually remains
a straight element, and the beam is fabricated so the erec-
tor can simply install the beam so that the tip of the beam is
 steel interchange
higher than the beam elevation at the connection by the speci-
fied amount. Since the beam isn’t being bent, you do not have
the same physical constraints on how long a member needs to
be before it can be “cambered.”
An alternative solution might be to bevel cut and splice
the member after it passes over the column, which will cre-
ate a change in slope at the cantilevered portion to provide
a specified top elevation at the cantilever tip. However, this
is not an inexpensive approach and it should be weighed
against increasing member sizes or using other methods to
mitigate the effect of the anticipated deflection at the tip of
the cantilever. An alternative that some fabricators prefer is to
make a V-shaped cut in the member, leaving one flange intact
and then bending and welding the member into the kinked
geometry. If you choose to kink the beam, I would suggest you
consider annotating your drawings and labeling the elevation
difference as something other than “camber.”
You might also consider splitting the beam and running
the column through the joint. This is typically the better
option for wide-flange columns, but is also a possibility for
your HSS column, and in fact may be the most economical
solution. If you choose this option, note that the 2016 AISC
Code of Standard Practice (ANSI/AISC 303-16, available at
www.aisc.org/specifications) contains new treatment of
preset requirements at the ends of cantilevers that will help
you with your goal; see Section 3.1.
Susan Burmeister, PE
Slip-Critical “Bolts”
What are the differences between slip-critical Class A
bolts and slip-critical Class B bolts, and how should they
be indicated in shop and erection drawings?
There is no such thing as a slip-critical bolt, a Class A bolt, a
Class B bolt or a bearing bolt. The same bolt can be used in
slip-critical joints with either Class A or Class B faying sur-
faces. In fact, the same bolt can be used with either slip-critical
or bearing-type joints. The difference between a slip-critical
joint and a bearing-type joint is that a slip-critical joint resists
movement of the plies through friction, and a bearing-type
joint resists movement between the plies through bolt shear
and bearing at the plies. The Class A and B designations refer
to the surface preparation required.
Section J3.8 of the AISC Specification for Structural Steel
Buildings (ANSI/AISC 360-16, available at www.aisc.org/
specifications) defines Class A surfaces as “unpainted clean mill
scale steel surfaces or surfaces with Class A coatings on blast-
cleaned steel or hot-dip galvanized and roughened surfaces” and
Class B surfaces as “unpainted blast-cleaned steel surfaces or
surfaces with Class B coatings on blast-cleaned steel.”
The detailer must properly indicate on the shop drawings
the required surface preparation at the slip-critical joints, as
this affects the strength of these joints.
Slip-critical joints also need to be pretensioned, and this must
be conveyed in some manner in the documents related to the bolt
APRIL 2017
installation, either the shop or erection drawings. Indicating that
the joints are slip-critical is sufficient to ensure pretensioning.
Carlo Lini, PE
Average Versus Peak Shear Stress
When applying Chapter G of the Specification to wide-
flange beams, 0.6Fy is the shear yielding stress, and Aw is
the area of the web. Some textbooks appear to indicate
that a uniform distribution of shear stress can be assumed
because the maximum web shear stress does not differ
much from the average web shear stress.
However, for a rectangular plate the ratio of peak to
average stress is 1.5, which does not seem insignificant.
Should the peak stress be used when designing rectangu-
lar plates to resist shear?
You are correct that the difference between the average shear
stress and the maximum shear stress in a wide-flange section is
relatively small with τpeak /τaverage equal to about 1.15. However,
this is not the reason the Specification is based on the average
stress. The stresses above assume an elastic distribution of
stress, which does not represent the true failure condition of
the element. Instead, the Specification is based on an inelastic
distribution of stress, which will be uniform.
A similar situation exists related to flexure. If an elastic
distribution were used in the Specification for flexure, beam
strength would be based on Sx. It is not. It is based on Zx. The
lower-bound ratio Zx/Sx for rolled wide-flange beams is about
1.11, though it varies somewhat among the shapes.
Section J4.2 addresses shear in connecting elements, which
are often rectangular sections, and bases the strength on the
gross area—the average stress. For example, when we check a
double coped beam, leaving what is essentially a rectangular
“narrow beam,” we base the strength on the average stress,
not because we feel that 1.5 is close enough to 1.0 but rather
because we are recognizing the inelastic redistribution of stress.
Larry Muir, PE
The complete collection of Steel Interchange questions and answers is available online.
Find questions and answers related to just about any topic by using our full-text search
capability. Visit Steel Interchange online at www.modernsteel.com.
Larry Muir is director of technical assistance and Carlo Lini is staff engineer—technical
assistance, both with AISC. Susan Burmeister and Bo Dowswell are consultants to AISC.
Steel Interchange is a forum to exchange useful and practical professional ideas and
information on all phases of steel building and bridge construction. Opinions and
suggestions are welcome on any subject covered in this magazine.
The opinions expressed in Steel Interchange do not necessarily represent an official
position of the American Institute of Steel Construction and have not been reviewed. It is
recognized that the design of structures is within the scope and expertise of a competent
licensed structural engineer, architect or other licensed professional for the application of
principles to a particular structure.
If you have a question or problem that your fellow readers might help you solve, please
forward it to us. At the same time, feel free to respond to any of the questions that you
have read here. Contact Steel Interchange via AISC’s Steel Solutions Center:
866.ASK.AISC • solutions@aisc.org