Current and FutureApplications of MechanicalFasteners for Light-FrameWood StructuresProceedings of Mechanical FastenersPlenary Session at the Forest ProductsResearch Society Annual Meeting

Compiled by Leslie H. Groom

Presented byUSDA Forest Service andForest Products Research SocietyJune 24,199lNew Orleans, LA

CONTENTS

Introduction to Mechanical Fasteners Plenary SessionLeslieH.Groom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Wood Connection Design: 1991 National Design Specifications and BeyondT. E. McLain and D. G. Pollock, Jr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Eccentricities in Metal-Plate Connected JointsMichaelH.Triche................................:..............5

Determining Failure Mechanism in Bolted Joints Using Acoustic EmissionsMarcia Patton-Mallory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Strength and Stiffness of Shear Transfer Plate ConnectionsRon Wolfe and Dave Bohnhoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Mechanical Connections: The Weak Link in Light-Frame Wood ConstructionGaryJ.Klein..................................................13

..*1 1 1

Introduction to Mechanical Fasteners Plenary SessionLeslie H. Groom

The title of this publication-“Current and FutureApplications of Mechanical Fasteners for Light-Frame Wood Structures”-is the theme of this plen-ary session from the 1991 annual meeting of theForest Products Research Society (FPRS). This themewas chosen to address current issues in the forestproducts industry: changing design codes, fastenerperformance, and legal responsibilities. It is hopedthat this session helped provide a forum for forestproducts industries influenced by mechanical fas-teners and also for research agencies, resulting in anincreased exchange of current knowledge and con-cerns for the future.

Mechanical fastener research has blossomed inredent years as a result of the changing wood base andof liability issues and foreign competition. Liabilityhas become a significant problem because of the in-creasing frequency and severity of litigation, resultingin the financial collapse of some businesses and in-creased insurance premiums for all. A key defenseagainst litigation is proper assessment of product per-formance. This information is needed not only bymanufacturers of mechanical fasteners but also byother potential lawsuit defendants, such as designagencies and trade associations. Economical and in-novative designs of engineered wood structures(whose properties are largely governed by mechanicalfasteners) must be coupled with product performanceto be competitive in world markets.

The presentations given at this plenary sessionwere chosen because of their timeliness and impor-tance. Individual articles in this publication are basedon the presentations. The first article, by Tom Mc-Lain, addresses the current revamping of the Nation-al Design Specifications (NDS) for mechanical fas-teners, which are undergoing a fundamental change.The new design specifications will be based on theEuropean concept of yield theory and will be incor-

porated in the 1991 NDS manual. Tom discusses theNDS changes and how they will affect light-frameconstruction practices.

Product performance is addressed in the next threearticles, with topics progressing from theoreticalmodeling to mechanical behavior of mechanically fas-tened joints. The modeling article, by Mike Triche,focuses on the analysis and design of truss-platejoints. Marcia Patton-Mallory explores experimentaltechniques for determining the integrity of mechani-cal fastened joints in her article. She discusses the useof acoustic emissions to determine the failure modesof bolted joints. In the product performance article,Ron Wolfe addresses the mechanical behavior of sheartransfer plates. Traditionally, this mechanical fas-tener has been limited to laminated beam columns;however, shear transfer plates have exceptionalmechanical properties and’may prove useful in otherengineered systems such as timber bridge decks. Rondetails the load capacity of stress transfer plates andthe factors that affect their performance.

The final article of this publication is by Gary Klein,who discusses the consequences of a lack of productperformance understanding by examining the col-lapse of structures as a result of failure of themechanically fastened joints. Gary presents forensiccase histories and discusses the liability associatedwith the failures.

The contributors selected were chosen on the basisof their knowledge of the subject area and respectamong their colleagues. They represent industry,academia, and U.S. Government research facilities. Ishould note that there had not been an FPRS plenarysession on mechanical fasteners for several years, andthe enthusiasm of the contributors was very high. Ahigh level of interest was shown by both industry andresearchers. I hope the session was enjoyable andeducational and that discussions continue.

Leslie H. Groom is a research forest products technologist at Alexandria Forestry Center, U. S. Department of Agriculture, Forest Service,Southern Forest Experiment Station, Pineville, LA 71360.

1

Wood Connection Design: 1991 National DesignSpecifications and Beyond

T. E. McLain, D. G. Pollock, Jr.

Design provisions for engineered wood connectionsare changing substantially for the first time in over 50years. These changes can be found in the upcoming1991 edition of the National Design Specification(NDS) for Wood Construction and in a proposed loadand resistance factor or design (LRFD) Specificationfor Engineered Wood Structures. The purpose of thispresentation is to outline some of these changes.

BEHAVIORAL EQUATIONS

A major change in design basis is the adoption of awell-verified theoretical base for lateral yield strengthcriteria for connections using nails, wood screws,bolts, and lag screws. Yield strength is defined asshown in figure 1. The European Yield Models (EYM)are used to calculate design strength based on thebending resistance of the fastener, the crushingstrength of wood or member material, joint geometry,and a set of assumed mechanical relationships. The

5 % OF FASTENER DIAMETER

0DEFORMATION (INCHES)

Figure l.-Fastener yield strength as based on 5 percent of thenominal shank diameter. Note: PL = Proportional Limit.

EYM describe a set of possible yield modes for a singlefastener under lateral load (see fig. 2) for double-shearconnections. Results of EYM for common connectionsare calibrated to 1986 NDS criteria to minimizechange. However, because of the wholesale change indesign equations and calibration to only one point in

Yield Theory BehavioralFailure Mode Equation

Ii2 D 2f Fe,,,

3.2 (2+R,) Ke

2D23.2 Ke

Figure 2.-Yield modes and corresponding behavioral equations fordouble-shear connections with dowel-type fasteners.

T. E. McLain is professor, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061; D. G. Pollock, Jr. is director ofengineering, National Forest Products Association, Washington, DC 20036.

3

design space, there will be changes in future connec-tion design. For lateral loading of connections, theEYM require a dowel-bearing strength, Fe, and fas-tener yield Fy. The Fe is directly related to specificgravity, which is used as an index variable for specieseffects and replaces the traditional species connectiongroups.

Specific Gravity

Eight species were tested in the National In-GradeTest Program, and the average specific gravities (SC)obtained have been adopted. For all other species, themethods of ASTM D2555 with a lo-percent groupinglimit are proposed. These methods result in slightchanges from historic practice. The principal changeis not with the level of SG chosen for a species butelimination of species groupings in favor of a morecontinuous strength-SG relationship.

Dowel-Bearing Strength, Fe

Dowel-Bearing Strength, Fe, is determined througha simple compression test of a dowel into a predrilledhalf-hole. For solid wood, the USDA Forest ServiceProducts Laboratory (FPL) has obtained the relation-ship shown in table 1. For connections loaded at anangle to the grain, Fe is found with Hankinson’s for-mula. This is more convenient than applying Hankin-son’s formula when solving the connection yieldstrength, and it gives essentially the same results.

Fastener Yield Strength, Fy

The yield strength of the fastener, Fy, on a &percentnominal shank diameter (D) offset basis may be foundby standard tests or from published values. Histori-cally, with allowable stress design, 45 ksi (310 MPa)has been assumed as the F, of common steel bolts.This is not contraindicated by recent research fromFPL. Research at Virginia Polytechnic Institute and

elsewhere indicates that for common wire nails, Fy =130 ksi minus 2140 where D = nominal diameter. Fora 16-penny nail with D = 0.162 inch, Fy = 95 ksi.

MODIFICATION FACTORS

Most of the historic modification factors used todesign strength for end-use conditions remain un-changed. Exceptions in the NDS include minor adjust-ments to the penetration factor. The only substantivechange in the NDS is with the group action or multi-ple fastener factor, C,, which is now in an equationformat and also tabulated. This factor is changed fromMDS-86 to more accurately reflect the supporting re-search results, which show that C, is a function of con-nection stiffness. Consequently, C, for bolted connec-tions and for split ring/shear plate connections is nowdifferent.

CALIBRATION

Historically, engineered wood connections havebeen safe. The NDS-91 design values were derivedfrom yield equations and calibrated to the safetylevels found in NDS-86. Calibration was done by ayield mode that minimized change but did noteliminate it. The yield equations were directlycalibrated to steel side plate connections with a low-to-intermediate range of bolt L/D ratios. The NDS-91nail connection design values were generally in-creased over the NDS-86 values.

Some connection design strengths increased, andsome decreased as a result of the combined effect of allchanges. Overall, the design strength of most connec-tions with bolts and lag screws is within plus or minus15 percent of the NDS-86 levels. Little or no change isseen for split ring/shear plate connections or in thedesign strength of connections relying on shankwithdrawal.

Table l.- Dowel-bearing strength equations for various fastener types and grain angles

Fastener type Grain angle Dowel-bearing strength

Nails, spikes, wood screws All grain angles F = 16600 SG1.s4*Bolts, lag screws, large dowel Parallel-to-grain F: = 11200 SGBolts, lag screws, large dowel Perpendicular-to-grain Fe = 6100 SG1.45D-0.5

*SG = specific gravity based on ovendry weight and volume; D = nominal shank diameter(inches); and F,, = dowel-bearing strength (psi)

4

Eccentricities in Metal-Plate Connected JointsMichael H. Triche

Design methods for metal-plate connected jointshave evolved rapidly over the past several years.Specific requirements are published by the TrussPlate Institute, Design Specifications for Metal PlateConnected Wood Trusses, TPI-85, (1). A new specifica-tion, TPI-92, (21, will be published in the near future.Although TPI-92 includes a number of advances inconnector plate designs and testing requirements, itdoes not cover specific methodology for handling ec-centricities in connections. Typically, joints aredesigned such that eccentricities are very small, andthe resulting moment in the connection is negligiblecompared with the direct shear. However, in certainodd situations, connections may have to be designedto resist moment.

The method presented here accounts for ec-centricities in connections. The proposed methoddetermines forces on individual teeth of the metalplate using the nonlinear load-displacement relation-ships of the teeth, which are determined by standardtest (2). Variations in stiffness because of varyingangles of grain and tooth orientation are considered.

Several linear models have been used in the past todesign plates for moment resistance. The most com-mon is the polar moment of inertia approach, whichwas used successfully for years in steel construction.This method, however, is very conservative whencompared to actual experimental behavior; theAmerican Institute of Steel Construction currentlypublishes tables for standard connections used insteel buildings (3) that are based on a nonlinear “in-stantaneous center of rotation” method, not the linearpolar moment of inertia approach. The nonlinearmethod predicts joint capacity with much greater ac-curacy, and the connections designed by this methodare considerably more economical. The key to themethod is the use of the nonlinear load-displacementcharacteristics of the fasteners, which allows for theultimate strength of the connection to be determinedand a factor of safety to be applied to the ultimatestrength of the entire connection rather than to in-dividual fasteners.

When a load is applied to metal-plate joints, theload is transferred from the wood into the steel plateby the lateral resistance of the teeth of the plate. Fora joint under axial load, each tooth of the plate is as-sumed to share the load equally. This assumption,while not completely accurate, is consistent with thedesign procedure used for nailed joints. Under puremoment, however, the teeth revolve around a commonpoint, and the teeth farthest from the center of rota-tion undergo much greater movement than those nearthe center.

The use of a single linear load-displacementrelationship for the teeth, which does not account forwood grain direction or tooth orientation, results intooth forces directly proportional to the distance of thetooth from the center of rotation, This is the basis ofthe polar moment of inertia approach, which has oftenbeen proposed. However, the results obtained by thismethod are inaccurate. In reality, the lateral resis-tance of the teeth of a plate is distinctively nonlinearand direction dependent. This nonlinearity is benefi-cial to joints under eccentric load because the force fora given displacement is lower than that calculated if alinear load-displacement relationship were assumed.

Connections must also be designed on the basis ofultimate strength rather than allowable stresses.Under the allowable stress approach, the designcapacity of the connection is based on the load atwhich the first fastener of the connection reaches itsallowable load. In this procedure, a factor of safety istaken on the individual fasteners and not on thecapacity of the entire joint. At these low load levels,the fasteners are still in the linear range, and thebenefit of the nonlinear load-displacement behavior isnot realized. This approach is unnecessarily conserva-tive. The preferred method is to determine the ul-timate strength of the connection and to apply the fac-tor of safety to the entire connection.

To illustrate the differences in the polar moment ofinertia approach and the method proposed here, fivemoment splices were considered. The specimens con-sisted of a pair of 2 by 4’s spliced together with metal

Michael H. Triche is assistant professor, Civil Engineering Department, University of Alabama, Tuscaloosa, AL 35487.

Figure I.-Diagram of a moment splice joint; the “+“shows the loca-tion taken for the center of rotation used in the analysisof the left connection area.

plates as shown in figure 1. Each connection wasanalyzed by the two methods; plate size, orientation,and the resulting moment capacities are shown intable 1. For both methods, the center of rotation wastaken to be at the location of the bottom corner due tothe wood-to-wood bearing that occurred at the bottomedge of the 2 by 4 ends.

The joints analyzed were previously tested inanother study (4); the results are shown in table 2.Also given in table 2 are the ultimate capacities aspredicted by the proposed method, along with percenterrors. Although the proposed method of analysis un-derestimated the actual capacity of the joint in all butone case, it was in general agreement with the actualcapacity. Referring back to table 1, note the linear,polar moment of inertia method predicts considerablyless moment capacity and therefore would result inextremely conservative designs.

In conclusion, methodology is proposed for analyz-ing metal-plate connections subjected to moment. Themethod is based on ultimate strength and requiresload-displacement information, which is availablefrom standard test procedures. The method predictsactual joint behavior much more accurately thanother commonly used linear methods. Also, thismethod can be used to aid in the development ofreliability-based design procedures for wood trusses.

Table l.- Comparison of proposed method with linear method formoment capacity of metal-plate connected joints withvaryingplate size and orientation

Plate sizePlate Proposed

orientation methodLinearmethod

Inches Degrees In-lb In-lb2.5 by 5.25 0 6,091 4,8713.5 by 3.0 9 0 5,511 4,7963.0 by 3.5 0 5,254 4,3955.0 by 3.0 9 0 8,211 7,1103.0 by 5.25 0 8,533 6,902

Table 2.-Comparison of experimental and proposed model resultsfor moment capacity of metal-plate connected joints withvarying plate size and orientation

Plate Experimental Predicted PercentPlate size orientation capacity capacity error

Inches Degrees In-lb In-lb2.5 by 5.25 0 7,689 6,091 -203.5 by 3.0 9 0 6,435 5,511 -143.0 by 3.5 0 6,798 5,254 -225.0 by 3.0 9 0 7,920 8,211 +43.0 by 5.25 0 9.454 8,533 -10

1 .

2 .

3 .

4 .

REFERENCES CITED

American Institute of Steel Construction. 1986.Manual of steel construction, load and resistancefactor design. 1st ed. Chicago, IL: American In-stitute of Steel Construction. 1192 p.Triche, M.H. 1984. Analysis and design of metalplate connections. West Lafayette, IN: PurdueUniversity. 119 p. M. S. thesis.Truss Plate Institute. 1985. Design specificationfor metal-plate connected wood trusses, TPI-85.Madison, WI: Truss Plate Institute. 47 p.Truss Plate Institute. 1991. National designspecification for metal plate connected wood trussconstruction. TPI-92; draft 3. Madison, WI: TrussPlate Institute. 103 p.

6

Determining Failure Mechanism in BoltedAcoustic Emissions

Joints Using

Marcia Patton-Mallory

Understanding the mechanisms by which failureoccurs in wood connections can lead to better criteriafor specifying their design load and design details.The research described considered failure mecha-nisms that occur in wood connections containing steelbolts. Failure processes were evaluated duringlaboratory testing of connections. The sounds gener-ated during the failure process-termed acousticemissions @+-were recorded and compared withobserved differences in failure modes.

Connections with the shorter end distance failed ina brittle splitting mode with only minor crushing ofthe wood beneath the bolt. This was the primaryfailure mode for connections that had a small enoughbearing length that the bolt did not yield during thetest. The pattern of AE during these tests was a sud-den increase in event rate, high amplitude events, andhigh energy content of events just prior to failurefig. 1.

Matched connections with longer end distancesfailed by wood crushing beneath the bolt. In contrastto the short end distance failure mode AE pattern,emissions were primarily of a lower amplitude, at afairly constant event rate, and with an increasingevent rate at the yield point (fig. 2). Higher amplitudeevents occurred only after considerable deformationin the joint and only when small shear cracks formedadjacent to the crushed wood.

The connection yield point was determined graphi-cally as the intersection of a line parallel to the majorlinear portion of the load-deformation curve but offsetby 5 percent of the bolt diameter. The yield pointdetermined in this matter corresponded closely to theincrease in AE events when bolted connections failedin a yielding (crushing) failure mode.

This study showed that AEl patterns correlatedclosely with observed brittle and ductile failure modesin bolted connections. However, care should be takento minimize small knots and drying cracks becausethese defects will cause unwanted AE during the test.

4,000I

1,000

(END D I S T A N C E = 4d, vd = 2)

IP

2; 3*ooo - 750 l

22

4 IrrkI

h 2 , 0 0 0 - - 500 T

2- - P L

ki:

4 1,000 - - 250 ‘,

%

2

0 I * 00 200 4 0 0 600

TIME (SECONDS)

Figure L-Brittle failure mode: acoustic emission increase below yield in bolted connection with ashort end distance. Note: d = bolt diameter; 1 = bolt bearing length; PL = proportionallimit; and A = acoustic emission event rate change.

Marcia Patton-Mallory is assistant director, U.S. Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experi-ment Station, Fort Collins, CO 80526.

7

4 , 0 0 0( E N D D I S T A N C E = 7d, I / d = 2)

;;; 3,DDD -

s- - Y I E L D

a - - P Ls 2 , 0 0 0 -

24 1,000 -

00 2 0 0 4 0 0 6 0 0

T I M E ( S E C O N D S )

Figure 2.-Ductile failure mode: acoustic emission increase near yield in bolted connection. Note:d = bolt diameter; 1 = bolt bearing length; PL = proportional limit; and A = acousticemission event rate change.

8

Strength and Stiffness of Shear Transfer Plate ConnectionsRon Wolfe and Dave Bohnhoff

A common problem faced by post-frame builders isthe difficulty of getting stress-rated solid sawn 8- by&inch timber posts. One solution to this problem is touse posts fabricated by laminating stress-graded 2- by&inch material. Glue-laminating may be prohibitive-ly expensive, but mechanically laminated posts can befabricated at a reasonable cost. Mechanically lami-nated columns have the additional advantage thatthey allow the builder to limit the use of treatedmaterial to the embedded portion of the post.

One approach to mechanically laminating woodposts is to use a double-sided connector plate called ashear transfer plate (STP) (fig. 1). Sandwiching STPsbetween adjacent layers provides needed post stiff-ness by limiting interlaminar shear displacement andlimits weakening effects of butt joints by transferringshear forces around them.

Although the concept of a double-sided nail plate forshear transfer between layers in a laminated assemb-ly is not new, its use has been limited to custom ap-plications. As a result, there have been few published

reports dealing with STP design properties or thestructural capacity of STP connected assemblies. Tomodel the behavior of mechanically laminated as-semblies, data are needed to characterize the load-slipbehavior of the connections.

The objective of this study was to add to the limitedavailable data base on strength and stiffness of STPconnections. Variables considered include fabricationmethods, plate size and type, joint confinement, loadorientation with respect to the grain, and woodspecies/specific gravity.

EXPERIMENTAL METHODS

The study was conducted in three phases, sum-marized in figure 2. The first phase concentrated onthe effects of plate size and fabrication method. Thesecond phase dealt with the effects of joint boundaryconditions. The third phase focused on effects of platetype, wood species, and load orientation with respectto the wood grain.

Figure L-Two sizes of shear transferplates were used in the initialphase of the study. One was a4- by 5-inch plate containing 10 plugs per side, and the other was a 2- by 5-inch platecontaining 5plugs per side.

Ron Wolfe is a research engineer, U.S. Department of Agriculture, Forest Service, Forest Products Laboratory, Madison, WI 53706; DaveBohnhoff is a professor, Department of’Agricultura1 Engineering, University of Wisconsin, Madison, WI 53706.

9

Phase Species Plate Pressing Testa Load to Joint codegrain

Size Type

SYP

SYP

SYP

2 b y 5 ONE 25-lSTAGE

4by5 G A L V 1 45-1

2by5 TWO 2 5 - 2STAGE

4by5 4 5 - 2I

1 WOA

4by5 GALV T W O 2 A H 0

STAGE 3 A L 0

4 AHI

ss SYP-S-E

GALV I SYP-G-E

ss

GALV

ssIII SPF

TWO4by5 GALV STAGE

ss

GALV

ss

GALV

DFss

GALV

’ Confinement conditions:’ No test apparatus used.‘Apparatus used with no restraint.3 Apparatus used with 20 psi restraining pressure.4 Apparatus used with 40 psi restraining pressure.

SYP-S-A

IISYP-G-A

SPF-S-E

1I SPF-G-E

II SPF-S-A

II SPF-G-A

I DF-S-E

IDF-G-E

II DF-S-A

I IDF-G-A

Figure 2.-Experimental design: Species considered included southern yellow pine (SYP), spruce-pine-fir (SPF), and Douglas-fir (DF); plates include stainless steel (SS) and galvanizedsteel (GALV) 20-gauge shear transfer plates with one four-tooth roqette (plug) per squareinch. Twenty-two joints were tested for each configuration. Joint code: S=stainless steel,G=galvanized steel, E=perpendicular, A=parallel.

The control joints referenced in each phase werefabricated using the two-stage pressing method,southern pine lumber, and 20-gauge, 4- by &inch gal-vanized metal STPs. Each joint was composed of threeparallel laminations, connected using two STPs.Load, applied perpendicular to the grain of the middlemember, was transferred through the STPs to theouter members, which were supported on the bed ofthe test machine. In two-stage pressing, plates werecompletely pressed into both sides of the center mem-ber, using special steel pressing plates that fit overthe STP. The outside pieces of wood were then posi-

tioned over the STP and pressed into place. In one-stage pressing, STPs were positioned between thewood members, and the four wood-plate interfaceswere pressed simultaneously.

SUMMARY OF TEST RESULTS

The control joint had an average maximum loadcapacity of 455 pounds per plug with a coefficient ofvariation of only 8 percent. For design purposes, theaverage load at 0.015 inches slip was 390 pounds perplug with a coefficient of variation of 9 percent. Test

10

results suggest that these values are significantly af-fected by fabrication method, wood specific gravity,plate size, and joint boundary conditions. No sig-nificant effect was noted for load-to-grain orientationor for galvanized versus stainless steel STPs.

Joints fabricated using single-stage pressingmethods required greater total fabrication energy andhad less stiffness and strength than those fabricatedusing the two-stage process. For obtaining joints of

comparable appearance, fabrication pressure require-ments for single-stage pressing were on the order of25 percent greater for single-stage pressing than fortwo-stage pressing. The single-stage 4- by 5-inch platejoints were only 75 percent as stiff and 85 percent asstrong as the two-stage joints. For the 2- by 5-inchplates, this effect on joint stiffness was slightly less(89 percent); but strength effect was the same.

Joint codeStiffhess Strength

lb/in Rel lb/plug ReI

II-WOA 50790 1.05 449 0.99

Control III-HP-G-E 50641 1.05 470 1.08

146-2 42628 .90 447 .96

Total 46241 1.00 465 1.00

Species III DF-G-E 41022 .55 448 .98

III SPF-G-E 28582 .59 359 .79

Pressing I 46-l 35975 .75 381 .84

Orientation III SYP-G-A 34708 .72 472 1.04

Plate Type III SYP-G-E 31449 .55 511 1.12

Plate Size I 26-2 29776 .62 558 1.28

IIANO 25122 52 266 58Confinement IIALO 25364 53 440 .97

IIAMI 24601 51 484 1.06

Species/Plate Type III DF6-E 82667 58 452 .99

III SPF-8-E 18536 .59 360 .79

Species/Orientation III DFG-A 32241 .57 300 1.10

III SPF-G-A 13352 58 3so .88

Orientation/Plate. III SYP-G-A 25965 .54 488 1.07

Species-Orient-Plate III DF-G-A 25537 .53 501 1.10

III SPF-O-A 11467 .24 337 .74

Pressing and Plate Size I 26-1 26682 .55 402 1.08

Figure 3.-Relative values of average joint load capacity. The stiffness values given were derived asthe stiffness (K) parameter derived on a per-plug basis to fit the Foschi nail-plate model.Strength values represent the maximum load for the joint uniformly distributed to theshear transfer plate plugs.

11

3 6 0 0

HIGH LATERAL6 -s 500 PRESSURE (dtil)

WITHOUT TESTINGB3 APPARATUS (WOA)

400-

PRESSURE (ANO)

0 . 0 2 0 .03

SLIP ( INCHES)

Figure 4.-Confinement effect on joint performance. Joints tested inthe confinement apparatus with no lateral pressure hadno external forces acting to resist lateral movement.

Results of tests conducted to evaluate plate size ef-fect are inconclusive. The 2- by &inch plate size ap-peared to give a per-plug load capacity 25 percentgreater than that of the 4- by &inch plate. This dif-ference was significant at the go-percent confidencelevel. Differing modes of failure, however, suggestthat this difference in strength may have been par-tially influenced by the size of the joint wood membersrelative to the size of the STPs used. For the 4- by 5inch plate joints, wood deformation caused by bearingstresses influenced joint strength.

Joint strength and stiffness are directly related tospecific gravity. Relative values of average joint loadcapacity (fig. 3) measured at incremental displace-ments were fairly constant over the range from 0.005to 0.05 inch and roughly equal to average relativespecific gravity values. Differences between the

southern pine and Douglas-fir joint strengths werenot significant at the go-percent confidence level, butthe southern pine joints were significantly strongerthan the spruce-pine-fir joints.

Test boundary conditions have a major effect onfailure mode and joint strength (fig.4). Joints testedwith no lateral restraint had a failure load 55 percentless than those tested with some form of restraint.Joint strength was not improved significantly by sub-stituting lateral pressure for bearing friction or by in-creasing lateral pressure beyond 20 psi.

CONCLUSIONS

Our research on strength and stiffness of STP con-nections led to the following conclusions:

l The two-stage method of joint fabrication ispreferable to single-stage pressing.

l Further study is needed to verify/quantify a platesize effect. Additional studies should considerwood member size relative to STP size as a vari-able.

l Results affirm a strong and direct relationshipbetween joint strength and member specificgravity.

l The effect of load orientation with respect towood grain direction appears to be species de-pendent. Although it was not significant at thego-percent confidence level, the effect of loadorientation was most noticeable for Douglas-firjoints.

l Further study is needed to relate test boundaryconditions to in-use conditions.

12

Mechanical Connections: The Weak Link in Light-FrameWood Construction

Gary J. Klein

Metal plate-connected trusses are one of the mostwidely used structural systems in residential andlight commercial construction. When properly de-signed and built, plate-connected trusses can be ex-pected to perform very well under most conditions.

In spite of advances in quality control, poorly in-stalled connector plates often result in trusses thatare much weaker than intended. Among the mostcommon problems are mislocation and partial embed-ment of connector plates. These and other problemswith truss connections are described in this paper.Methods for repairing deficient joints are alsopresented.

CONNECTOR PLATE MISLOCATION

Connector plates are sometimes mislocated duringfabrication, resulting in a significant reduction in thecapacity of the connection. Often the contact area of

the connector plate on the diagonal web member isless than required to develop the computed diagonalforce at service loads (fig. 1).

Sometimes individual plates are mislocated. Thisproblem is particularly likely where very small platesare used. On the other hand, plate mislocation can besystematic. Generally, this condition is not apparentuntil the shear stresses between the connector platesand wood members are computed.

Related deficiencies include connector plates thatare missing, improperly oriented (fig. 21, or under-sized. Also, wane or knots in the joint region can sig-nificantly diminish the capacity of the connection(figs. 3,4X These problems can result in joint failuresand sometimes collapse, although it has been my ex-perience that wood truss systems have a remarkableability to hang together in spite of numerous connec-tion failures.

Figure L-Pull-out of overstressed connectorplate.

Gary J. Klein is a consultant at Wiss, Janney, Elstner Associates, Inc., Northbrook, IL 60062.

13

Figure 2.-Pull-out of misoriented plate.

Figure 3.-Wane in joint region.

14

Figure 4.-Diagonal failure at knott below connector plate. Note that plate does not cover the entirejoint, concentrating the diagonal force at the knot.

PARTIAL PLATE EMBEDMENT

Often connector plates are not properly pressedwhen trusses are fabricated, or the plates tend to“walk out” as wood dries during service (fig. 5). Partialembedment can also occur because of a mismatch inmember thickness or width (fig. 6). The latter problemis particularly common on 2- by 4-inch parallel chordtrusses.

Partial embedment is characterized by a uniformgap between the connector plate and the wood mem-ber. Joints with partially embedded plates are muchweaker than joints with fully embedded plates. Thegap causes weak axis bending of the teeth. The prob-lem is particularly severe for 20-gauge plates loadedalong the length of the plate. A limited research pro-gram has shown that a Vl6-inch gap below a 20-gaugeplate reduced the joint strength by roughly 50 percent(fig. 7).

REPAIRS

Joint deficiencies can be repaired with plywood gus-sets attached to one or both sides of the truss usingthru-bolts or lag bolts. A limited testing program wasconducted to verify the effectiveness of %-inchplywood gussets (APA B-B Plyform, class 1) connectedto southern pine 2 by 4’s using %-inch bolts.

Recommended design values are as follows:

SuggestedConnection Type design value

%-inch plywood each side (double shear) 900 lband %-inch diameter thru-bolts per bolt

%-inch plywood one side (single shear) 350 lband !&inch diameter lag bolts per bolt

These values are based on a proportional limit loadand allow for time effects and a safety factor. A reduc-tion per the National Design Specification (ND@ formultiple fasteners in a row should be considered, al-though the test results did not show an apparentstrength reduction for up to three thru-bolts and sixlag bolts in a row.

A hydraulic C-clamp can also be used for a varietyof repairs. The most frequently used C-clamp repair isto repress partially embedded connector plates (fig. 8).The C-clamp can also be used to install new connectorplates at truss joints following shoring (if required)and removal of the existing plates.

15

Figure L-Partially embedded connector plate.

Figure &-Pull-out ofpartially embedded connectorplate due to mismatch of 2 x 4 width.

1 6

\

\

\0

0

20 GA. PLATE316 IN. TEETH

0 - “A” SPECIMENS0 - “8” SPECIMENS

sy 6 0 -

k\

\0l

\

0 0.02 0.04 0.06 0.06 0.10 0.12 0.14P L A T E G A P ( I N C H E S )

Figure ‘7.-Relative joint strength versusplategap for a 20-gauge plate with s/s-inch teeth, where 0and l represent specimen data from separate sample sets.

Figure C-Field repressing ofpartially embedded connector plates.

17

CLOSING REMARKS

The kinds of problems described are becoming lessfrequent. In recent years, the plate-connected trussindustry has made very significant advances in estab-lishing and enforcing quality. It is hoped that theseadvances continue and that ongoing research and de-velopment will strengthen the reliability of mechani-cal connections in light-frame wood structures.

18

Groom, Leslie H. 1993. Current and future applications of mechani-cal fasteners for light-frame wood structures; Proceedings ofmechanical fasteners plenary session at the Forest Products Re-search Society annual meeting. Gen. Tech. Rep. SO-92. New Or-leans, LA. U.S. Department of Agriculture, Forest Service,Southern Forest Experiment Station. 18 p.

Articles based on the five presentations at the mechanical fas-teners plenary session of the 1991 Forest Products ResearchSociety annual meeting are given. The information presented isaimed at designers of light-frame wood structures and researchersof mechanical properties of fasteners.

Persons of any race, color, national origin, sex, age, religion, or with anyhandicapping condition are welcome to use and enjoy all facilities, programs,and services of the USDA. Discrimination in any form is strictly against agencypolicy, and should be reported to the Secretary of Agriculture, Washington, DC20250.

"U.S. GOVERNMENT PRINTING OFFICE: 1993-766-017/80009