United StatesDepartment ofAgriculture

Forest Service

CaSQEffect of Moisture Cycling On MechanicalResponse of Metal-Plate Connector JointsWith and Without an Adhesisve Interface

Southern ForestExperiment Station

New Orleans,Louisiana

L. H. Groom

Research Paoerso-291 -September 1995

5,000

DlSPLACEMENi ( INCH)

Wood trusses are frequently located in light-frame structures where theyare subjected to significant shifts in moisture conditions. However, little is

’ known about the effects of moisture cycling of the wood members on the mechan-ical behavior of metal-plate connector (MPC) joints. Thus, the primary objec-tive of this study was to quantify the effect of wood moisture content fluctuationson MPC joint behavior. A secondary objective was to investigate maintaininginterfacial integrity between the metal-plate teeth and wood by applying anepoxy adhesive at the interface and noting the changes in mechanical perfor-mance of MPC joints.

Results of this study indicate that moisture cycling significantly affects themechanical performance of MPC joints. Some of the decline in mechanical per-formance could be attributable to pIate “backout.” It was shown that in addi-tion to a decrease in mechanical performance, moisture-cycled specimensprobably have an altered stress transfer mechanism as exhibited by load-sliptraces. Initial stiffness of the load-slip traces is most drastically affected bymoisture cycling, but ultimate load and load-at-0.015~inch-slip were also af-fected. Addition of an adhesive interface between the teeth and wood appearsto increase the mechanical performance of MPC joints. However, the apparentincrease in mechanical performance could not be quantified due to the lack ofidentically matched specimens. The degrade ofultimate load and load-at-O.OlS-inch-slip of MPC joints with an adhesive interface was similar to that of stan-dard MPC joints. Initial stiffness of MPC joints with an adhesive interfacedegraded more rapidly than standard MPC joints, probably due to weakeningof the adhesion between the teeth and wood. Addition of an adhesive to thetooth-wood interface resulted in less variability of mechanical properties whencompared to standard MPC joints, with this variability remaining constantregardless of number and severity of moisture cycles.

.

.

N

Effect of Moisture Cycling on Mechanical Response of Metal-PlateConnector Joints With and Without an Adhesive Interface

L.H. Groom

INTRODUCTION

The fundamental purpose of a truss system in con-struction is to transfer applied loads to the structureframework and eventually the foundation, with someof the load being absorbed in the truss system throughinternal friction and damping. The trend in truss de-sign in the last 15 years has been to reduce the amountof material necessary to resist the design loads. Thisincrease in efficiency has led to higher stresses in thetruss members. However, restrictions placed on old-growth timber stands and an increase in plantation-grown timber will result in a timber base withmarkedly lower mechanical properties. This paradoxof increased demand for superior lumber with a de-creased availability of a quality-product lumber sup-ply will lead to research aimed at changing currentdesign values and practices (Galligan 1990).

In most cases, truss performance is a reflection ofthe properties of the joints (Wilkinson 1984). Failureof truss systems occurs primarily at the joints due tostress concentrations, whereas large deflections intruss systems are generally the result of small jointdisplacements (Reardon 1971). Thus, evaluation oftruss performance must begin by understanding themechanical behavior of metal-plate connector (MPC)joints.

Predictive models have been developed for the staticstructural behavior of MPC joints by various research-ers (Crovella and Gebremedhin 1990, Foschi 1977,Groom 1988, Triche and Suddarth 1988). However,these models concentrate on the fundamental mecha-nisms of stress transfer and ignore secondary vari-ables such as moisture cycling. Moisture cycling iscritical for truss systems because trusses are oftenfound in the most environmentally extreme areas:attics and basements. Joist and rafter moisture con-tent (MC) in these environments has been shown tofluctuate seasonally by anywhere from 10 percent(Cunningham 1984, Duff 1978) to as much as 18 per-cent (Cleary 1984, Cunningham 1983, Sherwood 1985).This type of moisture cycling can affect truss perfor-mance. Unfortunately, the effect of moisture cyclinghas not been quantified. As a result, design values have

been adopted that are too conservative and make forinefficient use of lumber whose mechanical propertiesare ever decreasing.

There are several studies that have evaluated theeffect of changing MC under a constant load for tim-ber connectors (Keenan and others 1983, Kunesh andJohnson 1968, Mayo 1980, Noren 1968, Wilkinson1971). The results of these studies share a commontheme: The combined effect of changing environmentwith loading significantly decreased truss stiffness.In most cases, strength was also significantly reduced.However, the research neither quantified the effect ofmoisture cycling on stiffness or strength nor definedthe level at which moisture cycling proved significant.

McAlister (1990) found that MPC joints containinglaminated veneer lumber and pine and subjected toan outdoor environment for a l-year period showed asubstantial decrease in stiffness and strength. Thedecrease in mechanical properties was due in part to“plate backout,” a phenomenon in which the metalplate backs out of the wood members due to shrinkingand swelling stresses. Plate backout significantlyweakens the interfacial contact between themetal-plate teeth and wood.

Recommended design values for truss members aredictated primarily by the materials comprising thetruss, the type of truss, and the expected loading dur-ing the life of the truss (Truss Plate Institute 1985).The design plan laid out by architects and, to somedegree, engineers, for a particular structure, gener-ally dictates the type of trusses that will form theframework of the roof and floor systems. The requiredmechanical properties of the trusses are determinedby the loadings as prescribed by local building codes.The truss manufacturer engineers the trusses basedon design values that are concerned primarily withthe lumber mechanical properties and the metal plateconnectors (Truss Plate Institute 1985).

The lack of quantifying data regarding environmen-tal and loading effects presents a difficulty in trussdesign. Design specifications for duration of loadingbased on the Madison Curve are available from theTruss Plate Institute located in Madison, WS. How-ever, the effect of moisture cycling is compensated for

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

1

by dividing the average load of (a) five moisture-con-trolled specimens (MC = 15 percent plus or minus 4percent) by 3.0, or (b) five moisture-response speci-mens (MC = 15 percent assembly, then 7 percent, thengreater than 10 percent) by 2.5. An adequate under-standing of environmental variables on MPC jointbehavior must be attained to knowledgeably assigndesign values for truss manufacture.

These environmental variables clearly need to bequantified for sensible design values to be assignedfor MPC joints. Thus, the primary objective of thisstudy was to quantify the effect of wood MC fluctua-tions on MPC joint behavior. Specifically, this studycompared the effect of severity and number of woodmoisture cycles with metal-plate backout, joint stiff-ness, critical slip, and strength. A secondary objectivewas to investigate maintaining inter-facial integritybetween metal-plate teeth and wood by applying anepoxy adhesive at the interface and noting the changesin MPC joint mechanical behavior.

MATERLALS AND PROCEDURES

Experimental Design

A randomized block design, summarized in table 1,was used to minimize the high variability customar-ily seen in lumber mechanical properties. Two sepa-rate blocks were constructed to evaluate the effect ofwood MC ranges and number of moisture cycles withvarious physical and mechanical properties of MPCjoints. The two separate blocks were as follows: (1)standard MPC joints and (2) MPC joints with an ad-hesive applied at the tooth-wood interface.

Each block consisted of two factors: the severity ofMC levels to which MPC joint specimens were sub-jected and the number of MC cycles. Ten replicatesper cell were subjected to varying MC conditions (i.e.,[al MC held constant at 12 percent, [bl MC cycled be-tween 9 and 15 percent, and [cl MC cycled between 5and 19 percent and a varying number of MC cyclesfrom zero to eight-table 1). These replicates yieldeda total of 150 MPC joint specimens per block (3 MCranges times 5 MC cycles times 10 replicates = 150specimens). Although the material used to constructthe replicates in the blocks containing standard MPCjoints and joints with an adhesive were from differentboards, the boards were chosen so as to have compa-rable specific gravity and stiffness values betweenblocks, thus allowing for a relative comparison be-tween blocks.

Lumber and Metal Plates

The MPC joints were constructed with No. 2 gradeKD15 southern pine lumber representing a narrow

2

Table l.- Summary of experimental design showing the numberof samples for each moisture cycle moisture range of noadhesive vs. adhesive metal-plate connector (MPC) jointsfor each number of moisture cycles

Number of moisture cyclesMoisture contentrange 0 1 2 4 8

--- Percent --- _______________ Number of samples ---__-_________No adhesive

1 2 1 0 1 0 1 0 1 0 1 09 to 15 1 0 1 0 1 0 1 0 1 05 to 19 1 0 1 0 1 0 1 0 1 0

Adhesive1 2 1 0 1 0 1 0 1 0 1 0

9ta 15 1 0 1 0 1 0 1 0 1 05 to 19 1 0 1 0 1 0 1 0 1 0

range of specific gravities and moduli of elasticity. Thelumber used to construct the MPC joints was nominal2- by 4-inch dimension and 1%ft length and wassampled from a local lumber retailer located in cen-tral Louisiana. Modulus of elasticity (MOE) of eachboard was estimated by applying a static center-pointload to each board oriented flatwise over a 16-ft spanand measuring centerpoint deflection. Specific grav-ity was based on volume at the time of the test, andMC was determined on an approximately 1.5- by 1.5-by 3.5~inch block of wood removed adjacent to themetal plate after testing.

The die-punched metal plates used in this studywere standard truss-plate type plates made of 20-gauge grade C sheet metal, 3 inches wide by 4 incheslong with an average tooth density of eight teeth persquare inch. Average tooth length of the metal plateswas 0.360 inch with teeth being pointed, slightlycrimped, and having no surface coating or visiblemachining oil on the metal surface. The metal plateswere not washed to remove any residual machiningoil before joint assembly.

Joint Construction and Testing

Figure 1 shows a standard MPC joint used in thisstudy. In accordance with Truss Plate Institute (1985),all teeth were removed within a lumber-end distanceof 0.5 inch. The teeth were removed with a milling bitat the plate surface. Thirty-six pairs of teeth werepresent in the upper member, and 24 pairs werepresent in the lower member. This arrangement en-sured failure in the lower member, which was equippedwith linear variable differential transformers (LVDT).

An epoxy resin with a cure time of 15 minutes wasapplied with a sponge to the tips of the metal-plateteeth immediately before pressing into the woodenmembers preconditioned to an MC of 12 percent. An

average of 0.0031 g of epoxy was applied to each toothwith the resin spread along the tooth length duringpressing. Metal plates were pressed into the woodenmembers so as to make intimate contact, but not somuch that the plate surfaces became embedded in thewood. The joints were placed in a hygrothermally con-trolled room at 79 “F and 66-percent relative humid-ity with an MC of approximately 12 percent afterassembly for about 4 weeks to allow for relaxation ofstresses induced by pressing. After conditioning, thejoints were moisture cycled by placement in one ofthree hygrothermally controlled units set to cycle be-tween the following MC ranges: 5 to 19 percent, 9 to15 percent, and 12 percent (constant). All moisturewas atmospheric in nature, and at no time were thejoints exposed to direct surface water. Moisture con-tent was monitored during cycling with a hand-heldsurface moisture meter. All joints were reconditionedto an MC of 12 percent for 4 weeks prior to testing.

Loading was applied in tension by a 30,000-lb ca-pacity, screw-driven crosshead testing machine (fig. 1).A constant displacement rate of 0.015 inch/min wasapplied so as to produce failure in 5 to 10 minutes.

Figure l.- fipical metal-plate connectorjoint specimen before ten-sile testing.

The specimen was attached to the testing machinewith universal joints to prevent potential momentsinduced by misalignment. The upper member slip wasrestrained by the attachment of two C-clamps. Foursimultaneous LVDT readings were taken every sec-ond, located in pairs at the two opposite sides of thewood-wood juncture (fig. 1). These LVDT values werethen averaged to provide overall slip between themetal plate and lower wood member. Load was moni-tored with a lO,OOO-lb capacity load cell. A computer-controlled data acquisition system acquired the loadand slip readings. Load-slip traces were standardizedto an MC of 12 percent by adjusting mechanical prop-erties by 5 percent per l-percent MC change (Bodigand Jayne l982).

Metal-plate backout is defined in this study as thedimensional difference between the wooden memberthickness and the wood-plus-plate thickness. Thismethod was chosen so as to minimize errors due towood dimensional changes related to stress relaxationor MC. The wood-plus-plate thickness was measuredpreceding and after MC conditioning; the dimensionwas based on the average of two readings located oneach side of the long axis center line of the metal plateand midway between the rows of teeth. Wood-mem-ber dimensions were recorded at the same time. Ini-tial stiffness was calculated as the slope of the load-slipcurve between 700 and 1,100 lb.

RESULTS AND DISCUSSION

The conditioning history of the moisture-cycledspecimens is shown in figure 2. The moisture historyfor the constant 12-percent MC specimens is not showndue to the lack of moisture fluctuations. The compres-sor motor needed maintenance during the sixth andseventh cycles in the mild-cycle conditioning cham-ber, whereas compressor problems were encounteredduring the fifth cycle in the more severe cycle condi-tioning chamber. The duration of each moisture cyclevaried with the operation of the conditioning cham-bers and was coordinated to be of about equal lengthfor both conditioning chambers.

Some of the physical properties of the MPC jointsat the time of the test are shown in table 2. The MCand specific gravity of standard MPC joints and thosejoints with an adhesive interface were very similar. Theaverage moduli of elasticity from static, flatwise bend-ing tests of the lumber comprising standard and adhe-sive MPC joints were 1.76 (lo? psi and 1.86 (lo? psi.

Plate Backout

Plate backout was significantly affected by the sever-ity and number of moisture cycles as seen in figure 3.

3

!A

Table 2.-Average value by moisture content mnge and number of cycles for selected physical and mechanical properties of metal-plate connector (MPC) joints tested in this study

Moisture contentr a n g e

Moisture contentNo. ofCycles No adh. Adhesive

9to15 0

1

5 to 19-

Percent

12 (constant) 0

1

-------- percent _-______ Percent tooth length _________ Lb _________ ____ Kips /inch* ---- __________ Lb _________

12.1 12.1 0.49 0.49 0 . 0 0 2 0.000 0 . 6 0 . 1

11.4 11.6 0.48 0.48 0 . 0 0 2 0.005 0 . 6 0 . 1

2 11.3 11.3 0 . 4 9 0.49 0 . 0 0 4 0.004 1.1 1.1

4 10.3 10.4 0 . 4 7 0.49 0 . 0 0 4 0 . 0 0 7 1.2 1.8

8 11.2 11.1 0.49 0 . 4 7 0 . 0 0 4 0 . 0 0 3 1.2 0 . 8

12.1 12.1 0.49 0.49 0 . 0 0 2 0.000 0 . 6 0 . 1

12.0 12.0 0.49 0 . 4 8 0 . 0 0 6 0.008 1.7 2 . 2

2 1 0 . 8 10.8 0.49 0 . 4 8 0 . 0 0 7 0.010 2 . 1 2 . 8

4 10.1 0.49 0 . 4 9 0 . 0 1 3 0.014 3 . 5 4 . 0

8 11.3 0.49 0.49 0 . 0 1 2 0.015 3 . 3 4 . 2

0 12.1

10.2

11.3

12.1

12.1

10.3

11.0

11.4

0.49 0 . 4 9 0 . 0 0 2 0.000 0.6 0 . 1

1 11.6 0 . 4 8 0 . 4 7 0 . 0 1 3 0 . 0 1 1 3 . 6 3 . 0

2 0.49 0 . 4 9 0 . 0 2 3 0.016 6 . 4 4 . 4

4

10.3

10.5

11.4

0 . 4 8 0 . 4 9 0 . 0 3 7

0 . 0 6 1

0 . 0 2 7 10.1

17.0

7 . 6

8 0 . 5 1 0 . 4 8 0 . 0 4 5 12.5

Load-at-Specific gravity Plate backout Plate backout Ult imate load StiEness 0.015~inch-slip

No adh. Adhesive No adh. Adhesive No adh. Adhesive No adh. Adhesive No adh. Adhesive No adh. Adhesive

4,080 4,431 729(11.2)+ (6.4 (14.6)

3 , 9 0 7 4 , 3 6 7 6 7 7(8.3) (7.4) (12.8)

3 , 7 7 8 4,201 6 1 3(7.0) (6.1) (9.5)

3,598 4,095 5 6 56.9) (7.3) (17.2)

3,490 3,979 5 2 2(9.4) (8.9) (14.2)

1,422(18.5)

1 , 3 1 6(13.2)

1,017(15.2)

8 5 3(20.9)

(1%

3,612 4 , 2 8 5(9.9) (7.4)

3 , 5 1 3 4 , 2 6 1(7.6) (8.8)3,462 4 , 0 6 20.2) (7.6)3,304 3 , 8 5 6(6.8) (9.6)

3,185 3 , 7 5 8(10.4) (9.2)

4 , 0 8 0 4 , 4 3 1(11.2) (6.4)3 , 8 2 1 4,035(8.4) 6.6)

3 , 5 8 5 3,89903.0) (5.6)

3 , 4 2 1 3,694(8.7) 6.7)3 , 3 1 2 3,596(8.6) (8.1)

7 2 9(14.6)

4 7 2(14.4)

3 8 3(10.7)

(:5:)

3 3 5(14.1)

1,422 3,612 4 , 2 8 6(18.5) (9.9) (7.4)

6 1 9 3,389 3 , 8 5 0(8.4) (8.3) (7.3)

5 3 2 3,187 3 , 6 5 203.0) (7.4) (7.8)4 7 0 3,020 3 , 4 2 8

(16.1) (9.7) (7.7)

4 7 1 2,878 3 , 3 7 5(16.6) (7.7) (8.2)

4,080 4,431 7 2 9 1,422 3,612 4 , 2 8 5(11.2) (6.4) (14.6) (18.5) (9.9) (7.4)

3 , 5 4 5 3,752 3 6 6 4 6 6 3 , 1 2 7 3 , 5 0 3(11.8) (6.1) (13.7) (12.5) (10.1) (6.5)3 , 1 9 9 3,438 2 7 1 3 5 8 2 , 7 2 5 3,174(12.6) (9.9) (14.1) (11.1) (10.5) (9.4)

3 , 0 2 7 3 , 2 4 1 2 6 7 3 3 6 2 , 5 7 5 2,954(18.8) (9.4) (20.2) (13.7) (14.3) (10.1)

2,812 2,980 2 4 9 2 9 5 2 , 4 5 7 2 , 6 7 8(17.2) (12.8) (20.0) (14.2) (15.9) (11.6)

-*Kips = k i lopounds+&eff ic ient of variat ion in parenthesis

1 6

1 5

35 1 4:!a2

e!z 1 3

!s8 1 2

wi5 1 1

Gg 1 0

9

8

25

0

3 1 9 . 1 18.9 I 5 21 8 . 6 18.9

I

44 . 6 5 . 8 5 . 3 5 . 6 5 . 5 5.2 5 . 5 5 . 6

I I I 1

0 5 0 loo 150 2 0 0 250

TIME (days)

Figure 2.-Moisture history of metal-plate connector joints that were cycled between (a) 9- and 15.percent moisture content (MCI and (b)5- and 19-percent MC.

5

Numerical values of plate backout are also shown intable 2. The MPC joints held at a constant la-percentMC showed plate backout values of approximately 1percent of the total tooth length. The mildmoisture-cycled specimens seemed to stabilize at about4 percent of total tooth length after eight moisturecycles. In both the constant la-percent MC and mildmoisture-cycled specimens, addition of an adhesivehad no apparent effect on plate backout.

The more severe moisture-cycled standard and ad-hesive MPC joints backed out 17 percent and 12.5percent of the total tooth length after eight moisturecycles. Unlike the constant la-percent MC and mildmoisture-cycled specimens, the rate of plate backoutfor the more severe moisture-cycled specimens showedno signs of leveling off. Addition of an adhesive to thetooth-wood interface retarded plate backout, but theadhesive specimens did continue to back out after eightmoisture cycles. This suggests that plate backout con-tinues for an extended period of time for MPC jointssubjected to wide variations in MC. A similar study,but of an extended timeframe, may help to define theupper limit of this variable.

Mechanical Behavior

Load-slip traces for standard MPC joints subjectedto a constant la-percent MC, mild moisture cycles,and more severe moisture cycles are shown in figure4, with the adhesive tooth-wood interface counterpartssummarized in figure 5. Average load-slip traces (mea-sured in kips, or kilopounds) along with plus and mi-nus one standard deviation for adhesive and standardjoints after zero, one, two, four, and eight moisturecycles are shown in the appendix.

Figure 4 shows that successive moisture cycles doappear to change the nature of the load-slip traces.Metal-plate connector joints subjected to a constantMC exhibited a slight degradative behavior with time.This is to be expected to localize relaxation of the woodsurrounding the tooth. Although the MPC joints ex-hibited diminished load-slip traces with time, eachtrace exhibited nonlinear behavior throughout theentirety of the loading.

Moisture cycling not only accelerated the rheologi-cal degradation of the MPC joints but also changedthe shape of the load-slip trace (fig. 4b, 4~). Mild mois-ture cycling of the MPC joints reduced the amount ofnonlinearity in the initial portions of the load-sliptraces, a factor more evident with each subsequentmoisture cycle. This effect is even more apparent inthe more severe moisture-cycled MPC joints, with theload-slip traces remaining almost linear up until ap-proximately 50 percent of the ultimate load. This sug-gests that a different stress transfer mechanism isbeing demonstrated. This altered mechanism is most

likely a result, either singularly or in concert, of lesssurface area between the tooth and wood due to platebackout, accelerated localized relaxation of the woodsurrounding the teeth, and/or an additional rotationalload being applied to the teeth due to eccentric load-ing of the backed-out plate (Chou 1987, Groom 1988).

This same phenomenon was also present in the MPCjoints with an adhesive tooth-wood interface (fig. 5).The adhesive interface appeared to remain intact forconstant MC specimens. However, moisture cyclingdegraded the adhesive interface rather rapidly, withmost of the degrade occurring in the first moisturecycle.

The addition of an adhesive to the tooth-wood in-terface reduced the variation in load-slip traces of MPCjoints (appendix). Table 2 and figures Al, A3, A5, A7,and A9 show that the variability of standard MPCjoints increased with the severity and number of mois-ture cycles. The tooth-wood interface of adhesive MPCjoints maintained more intimate contact during mois-ture cycling, resulting in mechanical performance vari-ability that seems little affected by moisture cycling(table 2; figs. A2, A4, A6, A8, and AlO). This suggeststhat the most important attribute of the adhesive in-terface may not be the average increase in mechani-cal performance that degraded during subsequentmoisture cycling, but instead the reduced variabilityassociated with moisture cycling.

UZthute Load.-The effect of moisture cycling onultimate load is summarized in figure 6 and table 2.The MPC joints with an adhesive interface had ulti-mate-load values about 9 percent higher than the stan-dard joints. This difference may be partially due tothe different lumber samples used for the adhesiveand standard joints. Although the specific-gravity val-ues of the two samples were identical, the MOE of theadhesive joints was about 6 percent higher than thatof the standard joints.

The decrease in ultimate load for subsequent mois-ture cycles was consistent in both adhesive and stan-dard MPC joints (table 3). The percentage decrease inultimate load after eight moisture cycles for constantla-percent MC, mild moisture-cycled, and more severemoisture-cycled specimens was approximately 10,20,and 30 percent, respectively. The rate of decrease alsoseemed to decline for all specimens with subsequentmoisture cycles. However, the number of moisturecycles in this study was inadequate for determining alower limit for each moisture range.

Figure 7 shows that an inverse relationship existedbetween plate backout and ultimate load. Kline (1993)reported a similar relationship for MPC joints notsubjected to moisture cycling. This study showed thatthe ultimate load of standard MPC joint specimensheld at a constant la-percent MC decreased at a muchfaster rate than did the severely cycled specimens fora given degree of plate backout, with the mild mois-

6

MC = moisture content (4

l�M%c - 15%

= 12% MC

0 I I I I I I I

0 1 2 3 4 5 6 7 8 9

MC = moisture content 04

l 9% - 15%M C

n 12%MC

I I , I

0 1 2 3 4 5 6 I 8 9

NUMBER OF MOISTURE CYCLES

Figure 3.-Effect of number and severity of moisture cycles on plate backout of metal-plate connector joints (10 replicates per data point)constructed (a) without an adhesive interface and (b) with an epoxy adhesive interface.

7

r (4

09

0.03 II.04

I(c)

cm

DISPLACEMENT (inch)

Figure 4.-Load-slip traces in kips (kilopounds) showing the effect of sequential moisture cycling for standard metal-plate connector jointspecimens (10 replicates per moisture-content cycle) subjected to (a) constant 12-percent moisture content (MC), (b) moisturecycling between 9- and lb-percent MC, and (c) moisture cycling between 5- and 19-percent MC.

8

‘“,, :,_ ,,a I,>/

Table 3.- Percentage decrease, based on zero moisture cycles, by moisture content range and number of cycles ofselected mechanical properties

Moisture content No. ofrange cycles

Ultimate load

No adh. Adhesive

Stiffness Load-at-O.OlB-(Kips* /inch) inch-slip

No adh. Adhesive No adh. Adhesive

. . ..~...............____________________......... percent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

12 (constant) 0 0.0 0.0 0.0 0.0 0.0 0.01 4.2 1.5 7.2 7.5 2.7 0.62 7.4 5.2 15.9 28.5 4.2 5.24 11.8 7.6 22.5 40.0 8.5 10.08 14.5 10.2 28.4 51.7 11.8 12.0

9to 15 0 0.0 0.0 0.0 0.0 0.0 0.01 6.4 8.9 35.2 56.4 6.2 10.12 12.1 12.0 47.4 62.6 11.8 14.84 16.2 16.6 50.1 66.9 16.4 20.08 18.8 18.8 54.0 66.9 20.3 21.2

6 to 19 0 0.0 0.0 0.0 0.0 0.0 0.01 13.1 16.3 49.8 67.2 13.4 18.32 21.6 22.4 62.8 74.8 24.6 25.94 25.8 26.8 63.4 76.4 28.7 31.18 31.1 32.8 65.9 79.3 32.0 37.5

*Kips = kilopounds.

ture-cycled specimens falling between the two mois-ture ranges (fig. 7a). Because MPC joints in serviceare subjected to varying moisture cycles as well asin-service loads, it is difficult to develop a simple pre-dictive equation for plate backout and ultimate loaddue to other compounding factors.

Initial Stifie88.-The initial stiffness of MPCjoints is the mechanical property most affected bymoisture hysteresis, with results shown in figure 8and tables 2 and 3. Metal-plate connector joints keptat a constant la-percent MC showed a steady reduc-tion in initial stiffness for the first several months,with a reduced rate of degradation in the followingmonths. These constant la-percent MC specimensshowed about a 30-percent reduction in initial stiff-ness over the approximate g-month timeframe, witha lower limit appearing to be about 35 percent of theoriginal initial stiffness.

Moisture cycling significantly increased the degra-dation of MPC joint initial stiffness, especially withregard to the first moisture cycle (fig. 8a). Mildmoisture-cycled specimens showed a drop in initialstiffness of 35 percent after the first moisture cycle.Subsequent moisture cycling resulted in a 54-percentreduction in initial stiffness, with a lower limit ap-pearing to be about 60 percent of the original initialstiffness. More severe moisture cycling resulted in a50-percent reduction in initial stiffness after only onemoisture cycle. Subsequent moisture cycles had a mini-mal effect, with a lower limit of approximately 70 per-cent of the original initial stiffness.

Metal-plate connector joints with an adhesive in-terface showed a heightened response of initial stiff-

ness to moisture cycling (fig. 8b). In the constantla-percent MC specimens and the mild and more se-vere moisture-cycled specimens, the initial stiffnessof the adhesive interface MPC joints approached thatof the standard MPC joints with each successive mois-ture cycle. The difference between the adhesive inter-face and standard MPC joints diminished with theseverity of the moisture cycling. Although the initialstiffness of the adhesive interface MPCjoints dimin-ished more rapidly than that of standard MPC joints,the adhesive interface joints were still stiffer thanstandard joints in all cases.

Load-at-0.016-Inch-SZip.-Load-at-O.OlB-i,nch-slip was recorded instead of the traditional load-at-0.030~inch-slip as the top member of the joint wasrestrained and only the load-slip trace of the bottommember was recorded (fig. 1). Load-at-0.015~inch-slipfollowed a trend similar to the ultimate load of theMPC joints, with strength reductions of about 10 per-cent, 20 percent, and 30 percent for constant la-per-cent MC specimens, mild-cycled specimens, and moresevere cycled specimens, respectively (fig. 9). Adhe-sive interface specimens also showed an increase inload-at-0.015-inch-slip for all moisture-cycled speci-mens, along with a similar pattern of load-at-O.Olbinch-slip degradation. Metal-plate connector jointswith an adhesive interface did show an accelerateddegrade in load-at-0.015~inch-slip, with successivemoisture cycles when compared to standard MPCjoints, especially with regard to the first moisturecycle. This is most likely due to weakening of the ad-hesive interface caused by dimensional changes in thewood surrounding the teeth of the metal plate.

9

Figure 5.--Load-slip traces in kips (kilopounds) showing the effect of sequential moisture cycling for adhesive interface metal-plate connec-tor joint specimens (10 replicates per moisture-content cycle) subjected to (a) constant 12-percent moisture content (MC), (b)moisture cycling between 9- and 15-percent MC, and (c) moisture cycling between 5- and 19-percent MC.

10

r-

4,500(4

m 12% MC

+ 9%-15%M C

5%-19%M C

MC = Moisture contentI 1 I I

2,500 I I I I0 1 2 3 4 5 6 7 8 9

(‘3

MC = Moisture content1 ,

I, f I I

2,500 I0 1 2 3 4 5 6 I 8 9

NUMBER OF MOISTURE CYCLES

Figure 6.-Effect of number and severity of moisture cycles on ultimate load of metal-plate connector joints (10 replicates per data point)constructed (a) without an adhesive interface and (b) with an epoxy adhesive interface.

11

loo

Initial Stiffness9 0

MC = Moisture contentIII = 12 percent MC

0% 80=9to15prcentMC

!zA =5to19percentMC

d 7 0

8 60

5!n8 S O

ii 40

$2

3 0

2 00 5 10 15 2 0

PLATE BACKOUT (percent tooth length)

Figure 7.-Effect of plate backout on the normalized ultimate load and normalized initial stiffness of m&al-plate connector joints (10replicates per data point) constructed (a) without an adhesive interface and (b) with an epoxy adhesive intert?ux.

12

L

MC = Moisture content

1,400 -

800 -

n lZ%MC

*xi8 - 1 5 %

A - 1 9 %2 0 0 I I I I I I I 1 p$& .

0 1 2 3 4 5 6 7 8 9

MC = Moisture content 04

8 0 0

6 0 0

12% MC

200 L I I I I , I I I I0 1 2 3 4 5 6 7 8 9

NUMBER OF MOISTURE CYCLES

Figure 8.- Efict of number and severity of moisture cycles on initial stiffkss of metal-plate connector joints (10 replicates per data point)constructed (a) without an adhesive interface and (b) with an epoxy adhesive interfcrce.

1 3

29

3,000

(a)

-m 12%MC

+iti% - 15%

2,500 -M C

MC = Moisture contentI I 1 I I I I I2,000 -

0 1 2 3 4 5 6 7 8 9

4,500 r

04

m 12%MC

g

3,500

MC

2

2 3,000

MC I2,500

MC = Moisture content

2,000 -1I I J

0 1 2

NUMBER OF MOISTURE CYCLES

Figure 9.-Effect of number and severity of moisture cycles on load-at-0.015~inch-slip of metal-plate connector joints (10 replicatesper datapoint) constructed (a) without an adhesive interface and (b) with an epoxy adhesive interface.

1 4

_, I

CONCLUSIONS LITERATURE CITED

Moisture cycling significantly affects the mechani-cal behavior of MPC joints. Some of the decline inmechanical behavior could be attributable to platebackout, which was as high as 17 percent of the totaltooth length for more severe moisture-cycled speci-mens. Specimens subjected to a constant la-percentMC over the timeframe exhibited curvilinear load-sliptraces. However, more severe moisture cycling resultedin an almost linear load-slip trace up to about 50 per-cent of the ultimate load, followed by a curvilinearresponse to failure. Thus, it appears that there is achange in the stress transfer mechanism for moisture-cycled specimens. Although the average load-sliptraces of standard MPC joints are degraded by bothseverity and number of moisture cycles, the real gainin design values by addition of an adhesive interfacebetween the tooth and wood may be in variability ofthese traces, which is much less affected by moisturecycling.

Initial stiffness of the load-slip traces was most dras-tically affected by moisture cycling, with reductionsof approximately 30 percent, 55 percent, and 65 per-cent for constant la-percent MC, mild moisture-cycled,and more severe moisture-cycled specimens, respec-tively. Ultimate load was less severely affected, withconstant la-percent MC, mild moisture-cycled, andmore severe moisture-cycled specimens reduced ap-proximately 15 percent, 20 percent, and 30 percent,respectively. However, a study of longer duration isnecessary to determine the lower limit for each mois-ture range. Load-at-0.015-inch-slip showed reductionssimilar to ultimate load.

It appears that addition of an adhesive interfacebetween the teeth and wood increases the mechanicalproperties of MPC joints. However, the apparent in-crease in mechanical properties could not be quanti-fied due to the lack of identically matched specimens.The degrade of ultimate load and load-at-0.015-inch-slip of adhesive interface MPC joints was similar tothat of standard MPC joints. Initial stiffness of adhe-sive interface MPC joints degraded more rapidly thanstandard MPC joints, probably due to weakening of theadhesion between the teeth and wood. Increased adhe-sion between the metal plates and wood, either throughplate cleaning or improved adhesive system, would mostlikely further enhance structural performance.

It should be noted that in-service MPC joints mostcommonly are subjected to MC somewhere betweenthe la-percent constant and mildly cycled specimens.Caution should be exercised when applying the resultsof this study to different metal-plate types and toin-service structures because of factors that have beenignored in this study, such as surface moisture andin-service loading of the truss structures.

Bodig, J.; Jayne, B.A. 1982. Mechanics of wood andwood composites. New York Van Nostrand ReinholdCompany, Inc. 712 p.

Chou, C. 1987. Modeling of nonlinear stiffness andnonviscous damping in nailed joints between woodand plywood. Corvallis, OR: Oregon State Univer-sity. 181 p. Ph.D. dissertation.

Cleary, P 1984. Humidity in attics-sources and con-trol methods. Lawrence Berkeley Lab. Rep. LVL-17590, July 1984. Berkeley, CA: University ofCalifornia. 15 p.

Crovella, P.L.; Gebremedhin, K.G. 1990. Analyses oflight frame wood truss tension joint stiffness. For-est Products Journal. 40(4): 41-47.

Cunningham, M.J. 1983. A new analytical approachto the long term behaviour of moisture concentra-tions in building cavities. 2: Condensing cavity.Building and Environment. 18(3): 117-124.

Cunningham, M.J. 1984. Further analytical studiesof building cavity moisture concentrations. Build-ing and Environment. 19( 1): 21-29.

Duff, J.E. 1978. Moisture content of a joist floor overa crawl space. Res. Pap. SE-189. Asheville, NC: U.S.Department of Agriculture, Forest Service, South-eastern Forest Experiment Station. 12 p.

Foschi, R.O. 1977. Analysis of wood diaphragms andtrusses. 2: Truss-plate connections. Canadian Jour-nal of Civil Engineering. 4: 353-362.

Galligan, W.L. 1990. The impact of raw material onlumber quality. In: Biographies and abstracts of the44th annual meeting of the Forest Products Re-search Society; 1990 June 24-27; Salt Lake City,UT. Madison, WI: Forest Products Research Soci-ety: 26.

Groom, L.H. 1988. Experimental verification and non-linear modeling of truss-plate joints by Runge-Kuttanumerical technique. Corvallis, OR: Oregon StateUniversity. 202 p. Ph.D. dissertation.

Keenan, F. J.; Suddarth, S.K.; Nelson, S.A. 1983. Woodtrusses and other manufactured structural compo-nents. In: Proceedings of the workshop on structuralwood research; 1983 October 5-6; Milwaukee, WI.New York: American Society of Civil Engineering:175-184.

Kline, G.J. 1993. Mechanical connections: the weaklink in light-frame wood construction. In: Groom,L.H., ed. Current and future applications of me-chanical fasteners for light-frame wood structures:Proceedings of mechanical fasteners plenary sessionat the Forest Products Research Society annualmeeting; 1991 June 23-26; New Orleans, LA. Gen.Tech. Rep. SO-92. New Orleans, LA: U.S. Depart-ment of Agriculture, Forest Service, Southern For-est Experiment Station: 13-17.

15

Kunesh, R.H.; Johnson, J.W. 1968. Strength of mul-tiple bolt joints: influence of spacing and other vari-ables. Rep. T-24. Corvallis, OR: Forest ResearchLaboratory, Oregon State University. 19 p.

Mayo, A.P. 1980. Long-term performance tests ontrussed rafters. Rep. CPlk30. Aylesburg, UK: Build-ing Research Establishment, Princes RisboroughLaboratory. 10 p.

McAlister, R.H. 1990. Tensile loading characteristicsof truss plate joints after weathering and acceler-ated aging. Forest Products Journal. 40(2): 9-15.

Noren, B. 1968. Nailed joints-their strength and ri-gidity under short-term and long-term loading. Rep.158B. Stockholm: The National Swedish Institutefor Building Research. 80 p.

Reardon, G.F. 1971. A structural analysis of frameswith semi-rigid joints. Tech. Pap. 69. Melbourne,Australiaz Division of Forest Products, CommonwealthScientific and Industrial Research Organization.

Sherwood, G.E. 1986. Condensation potential in highthermal performance walls-hot, humid summerclimate. Res. Pap. FPL-445. Madison, WI: U.S. De-partment of Agriculture, Forest Service, ForestProducts Laboratory, 29 p.

‘I&he, M.H.; Suddarth, SK 1988. Advanced designof metal plate connector joints. Forest Products Jour-nal. 38(g): 7-12.

Truss Plate Institute. 1985. Design specification formetal plate connected wood trusses. TPI-85. Madi-son, WI: Truss Plate Institute. 47 p.

Wilkinson, T.L. 1971. Theoretical lateral resistance ofnailed joints. Journal of the Structural Division,American Society of Civil Engineers. 97(STS): 1407-1422.

Wilkinson, T.L. 1984. Longtime performance of trussedrafters with different connection systems. Res. Pap.FPL-204. Madison, WI: U.S. Department of Agricul-ture, Forest Service, Forest Product Laboratory. 19 p.

16

,,,I,. I

A P P E N D I X

5

- - - - - - I-- l- - - -- - -

,0.01 0.02 0 .03 0 .04

DISPLACEMENT (inch)

Figure Al.-Average load-slip traces (solid lines), includingplus and minus one standard deviation (broken lines), for standard metal-plateconnector joints subjected to zero moisture cycles and (a) constant 12-percent moisture content (MC), (b) moisture cyclingbetween 9- and 15-percent MC, and Cc) moisture cycling between 5- and I9-percent MC.

18

0.01 0.02 0.03 0.040 0.01 0.02 0.03 0.04

_--- - -I - - - - w

- - - - - -

0 0.01 0.02 0.03 0.040 0.01 0.02

DISPLACEMENT (inch)DISPLACEMENT (inch)

0.03 0.04

Figure AZ- Average load-slip traces (solid lines), including plus and minus one standard deviation (broken lines), for adhesive interfacemetal-plate connector joints subjected to zero moisture cycles and (a) constant 12.percent moisture content (MC), (b) moisturecycling between 9- and 15-percent MC, and (c) moisture cycling between 5- and 19-percent MC.

19

(4_---- -

--- ---

0.01 0.02 0.03 0.04

03- - - - -

- - - - - -- - -- - - -

0.02

DISPLACEMENT (inch)

Figure A3.- Average load-slip traces (solid lines), including plus and minus one standard deviation (broken lines), for standard metal-plateconnector joints subjected to one moisture cycle and (a) constant l&percent moisture content (MC), (b) moisture cycling between9- and 15-percent MC, and (c) moisture cycling between 5- and 19-percent MC.

20

5

4

?7g 3

29 2

1

0

5

4

72.n 3&

29 2

1

0

/.--- -- -- (4_--------- -_

0 0.01 0.02

L

0.01 0.02 0.03 0.04

0 0.01 0.02

DISPLACEMENT (inch)

Figure A4.- Average load-slip traces (solid lines), including plus and minus one standard deviation (broken lines), for adhesive interfacemetal-plate connector joints subjected to one moisture cycle and (a) constant 12-percent moisture cohtent (MC), (b) moisturecycling between 9- and 15-percent MC, and (c) moisture cycling between 5 - and Is-percent MC.

,

(4

0.02 0.03 0.04

/---- _------

----- -----

I

0

1

03

0 0 . 0 1 0.02

DISPLACEMENT (inch)

0.03

Figure AS.- Average load-slip traces (solid hes), includingplus and minus one standard deviation (broken lines), for standard metal-plateconnector joints subjected to two moisture cycles and (a) constant 12-percent moisture content (MC), (b) moisture cycling be-tween 9- and 15-percent MC, and (c) moisture cycling between 5- and Is-percent MC.

22

5

- - (4_---4 - - - - - -

0.02 0.03 0.04

(b)

0 0.02 0.03 0.04

503

DISPLACEMENT (inch)

Figure A6.- Average load-slip traces (solid lines), including plus and minus one standard deviation (broken lines), for adhesive interfacemetal-plate connector joints subjected to two moisture cycles and (a) constant 12-percent moisture content (MC), (b) moisturecycling between 9- and Z-percent MC, and (c) moisture cycling between 5 and IO-percent MC.

0 . 0 2 0 . 0 3 0 . 0 4

04

_---_---- - --

----------

0 . 0 2

DISPLACEMENT (inch)

0 . 0 3 0 . 0 4

Figure AT.- Average load-slip traces (solid lines), including plus and minus one standard deviation (broken lines), for standard metal-plate Iconnector joints subjected to four moisture cycles and (a) constant 12-percent moisture content (MC!, (b) moisture cycling Ibetween 9- and 15-percent MC, and (c) moisture cycling between 5- and 19-percent MC. I

24

5(4-

- - -- - - -

4 -- - c_---

0 0.01 0.02 0.03 0.04

5

4

1

00 0.01 0.02 0.03 0.04

0 0.01 0.02

DISPLACEMENT (inch)

0.03 0.04

Figure A&- Average load-slip traces (solid lines), including plus and minus one standard deviation (broken lines), for adhesive interfacemetal-plate connector joints subjected to four moisture cycles and (a) constant 12-percent moisture content (MC), (a) moisturecycling between 9- and IS-percent MC, and (c) moisture cycling between 5- and 19-percent MC.

25

5

4 -

(4

- - - - -/-----

- - - - -

0 0 . 0 1 0.02 0.03 0.04

(‘4

------ --_---/.--- -- ------

0 0 . 0 1 0.02 0.03 0.04

0.02

DISPLACEMENT (inch)

Figure A9.- Average load-slip traces (solid lines), including plus and minus one standard deviation (broken lines), for standard metal-plateconnector joints subjected to eight moisture cycles and (a) constant la-percent moisture content (MC), (b) moisture cyclingbetween 9- and 15-percent MC, and (c) moisture cycling between 5- and 19-percent MC.

26

(4- - -- -4 - /----/

- - - - - -

0 0.01 0 . 0 2 0 . 0 3 0.04

0 0.01 0.02 0.03 0.04

5w

4 -

3. a 3

_--- - - - --

25 _--- - - -/---

0.02 0.03

DISPLACEMENT (inch)

O.d4

Figure AlO.-Average load-slip traces (solid lines), including plus and minus one standard deviation (broken lines), for adhesive interfacemetal-plate connector joints subjected to eight moisture cycles and (a) constant 12-percent moisture content (MC), (b) moisturecycling between 9- and 15-percent MC, and (c) moisture cycling between 5- and 19-percent MC.

27

The United States Department of Agriculture (USDA) prohibits discrimina-tion in its programs on the basis of race, color, national origin, sex, religion,age, disability, political beliefs and marital or familial status. (Not all prohib-ited bases apply to all programs.) Persons with disabilities who require alter-native means of communication of program information (braille, large print,audiotape, etc.) should contact the USDA Office of Communications at (202)720-2791.

To file a complaint, write the Secretary of Agriculture, U.S. Department ofAgriculture, Washington, DC 20250, or call (202) 720-7327 (voice) or (202) 720-1127 (TDD). USDA is an equal employment opportunity employer.