# Effect of the number of screws and screw size on moment capacity of furniture corner joints in case construction.

Abstract

Tests were carried out in order to determine the bending moment capacity of L-type glued-screwed corner joints constructed with multiple screws. 6 sizes of screws were used for assembling the specimens. 2, 3, and 4 screws were used in the joints of the specimens that have the same width. Specimens were prepared of 18-mm-thick particleboard (PB) and medium density fiberboard (MDF) with resin surfacing. The specimens were constructed with both screws and polyurethane adhesive. Specimens were tested under static compression and tension loads. Results have shown that the maximum moment capacity is obtained with the MDF specimens when the number of screws in the joints is 4. In both compression and tension tests, MDF corner joints were stronger than PB corner joints. Results of the tests also indicated that a screw corner joint becomes stronger as either screw diameter or screw length is increased. Screw length was found to have a greater effect on moment capacity than diameter. Furthermore, the average moment capacities of glued-screwed corner joints evaluated in this study in compression could be predicted by means of the expressions:

[CM.sub.PB] = -81.59 + 16.09 X + 14.52 Y + 1.47 Z

[CM.sub.MDF] = -173.9 + 29.07 X + 24.62 Y + 2.26 Z

and in tension by means of the expressions:

[TM.sub.PB] = -8.4 + 17.73 X + 44.55 Y + 1.32 Z

[TM.sub.MDF] = -136.1 + 32.42 X + 17.25 Y + 2.2 Z

where [CM.sub.PB], [CM.sub.MDF] = moment capacities of PB and MDF joints, respectively, under compression load (Nm); [TM.sub.PB]. [TM.sub.MDF] = moment capacities of PB and MDF joints, respectively, under tension load (Nm); X = number of screws; Y = screw diameter (mm); Z = screw penetration (mm).

**********

Case furniture is widely produced and extensively used in kitchens and offices for storage. Particleboard (PB) and medium density fiberboard (MDF) are the most common wood-based composite panel products used in constructing this type of furniture.

Screw-type fasteners are commonly used to construct corner joints with or without glue. The rational design of case-type furniture constructed with screws requires information on the moment capacities of these in PB and MDF. An important consideration in the product engineering of screw jointed cases is the specification of the number of screws that should be used in joining the sides to bottom and top of the case. In spite of their widespread use, limited information is available regarding moment capacities and stiffness of screw-type corner joints in case construction. Published information is mostly related to direct withdrawal resistances of screw-type joints (Eckelman 1974, 1975, 1978; Rajak and Eckelman 1993).

Ho (1991), Rajak (1989), and Zhang (1991) have helped to clarify the relationships between number of fasteners used and joint strength by defining what has been termed the "zone of influence" or "zone of failure" of the fastener. This term signifies that an individual fastener is supported for a finite distance by the material on either side of it. As a result, when a fastener connection such as a screw in the side of a case fails, it causes a portion of the side wall on either side of it to fracture. In his work on the bending moment capacity of case joints constructed with large screws, Rajak (1989) obtained "zones of failure" that extended from 40 to 50 mm on either side of the fastener. Similarly, in his work on the performance testing of the cases Ho (1991) obtained a value of a little more than 40 mm on either side. In the case of dowels in similar material, Zhang (1991) found that the "zone of failure" extended about 40 mm on either side of the dowel. These studies define the minimum fastener spacing that will yield maximum fastener strengths. Liu and Eckelman (1998) constructed 457-mm-wide joints with up to 36 fasteners. Screwed and doweled joints were tested in compression in order to determine the bending moment capacity. Results showed that the moment capacities increase rapidly until the "zones of influence" of the fasteners overlap. No increase in moment capacity was obtained beyond that point. Ho and Eckelman (1994) conducted cyclic performance tests on furniture cases constructed with sheet metal screws for determining joint strength, stiffness, and durability as a function of the number of screws, screw diameter, screw length, and screw position. Test results demonstrated that screw length significantly affected case strength and durability. Results also showed that maximum fastener strength was obtained when fasteners were placed no closer than 89 mm from one another.

Research on corner joints using other types of fasteners in case construction applications such as dowels and screw--nuts, were reviewed as reference in conducting moment capacity evaluations of screw-type corner joints in MDF and PB. In compression and tension tests of single-dowel corner joints with PB, Zhang and Eckelman (1993a) determined that the moment capacity in tension loads was higher than in compression loads. As dowel diameter and length increase, strength increases proportionally. In evaluating the moment resistances of multi-dowel joints in PB, Zhang and Eckelman (1993b) conclude that a space of 75 mm between two dowels yields the highest moment capacity per dowel. Zhang et al. (2005) investigated the effects of screw diameter and length, loading, board material, board surface condition, and gluing on moment resistances of three-screw L-type corner joints under tension and compression loads. Results showed that surfacing PB with synthetic resin and assembling joints with glue applied to the contact surfaces of the face and butt members significantly improve moment resistances of joints constructed of PB. According to results of the tests, 5-mm diameter by 50- or 60-mm-long screws were recommended for connecting corner joints of case-type furniture. A.N. Tankut and N. Tankut (2004) investigated the effects of some factors on moment capacity of corner joints constructed with wood biscuits. Results indicated that joint strength comes mainly from the edge gluing of the face and butt member and not from the glued biscuits. In another study, the bending moment capacity of corner joints in 32-mm case construction prepared from laminated PB and MDF were researched by testing under tension and compression loads. The dowel spacing effects were investigated on bending moment capacity of the corner joints. Test results showed that maximum moment is obtained in joints when the spacing between the dowels is at least 96 mm. In the tests, MDF corner joints were stronger than PB corner joints (Tankut 2005). Park et al. (2006) investigated the stability of several, mostly screw-based, fasteners for connecting 16-mm-thick furniture grade PB. Results indicated that screws on their own performed markedly better than either dowels or screws with PVC anchors in all three tests. The use of thicker screws or screws with anchors was not recommended for butt-jointed shelving applications using PB. Screws with specialized thread configurations such as beveled edges or alternating thread heights performed well, but were not significantly different from plain-threaded screws of the same dimensions in any of the loading configurations. The effects of screw and back panel on the corner joints in case furniture produced from MDF and PB were determined (Atar and Ozcifci 2007). Static compression and tension loads were applied to the L-type corner joint specimens. In conclusion, it was deduced that the edges of panels should be rebated and covered with solid wood or massive wood for case furniture.

Although L-type, screw connected corner joints are commonly employed in construction of cabinet furniture, limited information is available on effects such as number of screws, screw sizes (diameter and length) loading type, and board material type on moment capacity of the joints. Accordingly, the primary purpose of this study was to obtain practical information concerning the moment capacity of screwed corner joints that the furniture engineers could use in the design of case furniture. The following objectives were tested:

* How ultimate bending moment capacity of joints is affected by the use of different panel materials, namely PB and MDF,

* How number of screws affects the moment capacity of L-type corner joints,

* How screw sizes (diameter and length) affect the moment capacity of L-type corner joints.

* Could the average bending moment capacity of screwed corner joints evaluated in this study be estimated by developed predictive expressions?

Experimental design

Altogether, 36 sets of specimens consisting of 5 replications each, or, a total of 180 specimens were prepared for compression tests; in the same manner, 180 specimens were prepared for tension tests. Full linear models (models [1] and [2]) for the four-way factorial experiments were considered to evaluate the influence of material type (PB and MDF), number of screws (2, 3, and 4), screw diameter (4 and 5 mm) and screw length (40, 50, and 60 mm) on moment capacity of L-type corner joints under compression and tension loads. The form models are:

[C.sub.ijklm] = [[mu].sub.1] +[A.sub.i] + [B.sub.j] + [C.sub.k] + [D.sub.l] + [(AB).sub.ij] + [(AC).sub.ik] + [(BC).sub.jk] + [(AD).sub.il] + [(BD).sub.jl] + [(CD).sub.kl] + [(ABC).sub.ijk] + [(ABD).sub.ijl] + [(ACD).sub.ikl] + [(BCD).sub.jkl] + [(ABCD).sub.ijkl] + [[epsilon].sub.ijklm] [1]

[T.sub.ijklm] = [[mu].sub.2] + [A.sub.i] + [B.sub.j] + [C.sub.k] + [D.sub.l] + [(AB).sub.ij] + [(AC).sub.ik] + [(BC).sub.jk] + [(AD).sub.il] + [(BD).sub.jl] + [(CD).sub.kl] + [(ABC).sub.ijk] + [(ABD).sub.ijl + [(ACD).sub.ikl] + [(BCD).sub.jkl] + [(ABCD).sub.ijkl] + [[epsilon].sub.ijklm] [2]

where [C.sub.ijklm] moment capacity under compression load (Nm); [T.sub.ijklm] = moment capacity under tension load (Nm); [[mu].sub.1] = population mean moment capacity under compression load for all material type-number of screws-screw diameter-screw length combinations (Nm); [[mu].sub.2] = population mean moment capacity under tension load for all material type-number of screws-screw diameter-screw length combinations (Nm); A discrete variable representing effect of material type; B = discrete variable representing effect of number of screws; C = xdiscrete variable representing effect of screw diameter; D discrete variable representing effect of screw length; (AB), (AC), (BC), (AD), (BD), (CD) = effect of the two-way interactions among the four variables; (ABC), (ABD), (ACD), (BCD) = effect of the three-way interactions among the four variables; (ABCD) = effect of the four-way interactions among the four variables; [epsilon] = random error term; i = index for material type, 1.2; j = index for number of screws, 1.3; k = index for screw diameter, 1.2; l = index for screw length, 1.3; m = index for the replication, 1.5.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

General configuration and construction of the specimens

The configuration of the L-type corner joint specimens in this study is shown in Figure 1. The specimen consists of a face member and a butt member of the same type material, joined together by 2, 3, or 4 screws. Cross sections of the joint area for the specimens connected with 2, 3, or 4 screws are given in Figures 2(a), (b), and (c).

The face member measured 350 by 158 by 18 mm, whereas the butt member measured 350 by 140 by 18 mm. 6 different sizes of screws which included 2 diameters and 3 lengths (4 by 40, 4 by 50, 4 by 60, 5 by 40, 5 by 50, and 5 by 60) were utilized for constructing the specimens. All specimens were assembled with both screws and polyurethane based adhesive with 60 percent solid content. This procedure is a common application in construction of corner joints for furniture cases. Polyurethane adhesive was applied to joining surfaces area at a rate of 200[+ or -]10 g/[m.sup.2], also a liberal amount of adhesive was applied into the pilot holes.

18-mm PB and MDF were selected as board materials for this study. The panels were obtained from commercial suppliers. In preparation of the specimens, 1880- by 3660-mm full-size sheet of board were first cut into face and butt member strips. These strips were subsequently cut into the desired member lengths.

Steel Phillips-head flat wood screws with 40[+ or -]3 degree thread angle were chosen for this study. Root diameter, outside diameter, and thread per mm were 2.4[+ or -]0.25, 4.0[+ or -]0.3, 1.8 mm respectively for 4-mm-diameter screws; and 3.0[+ or -]0.3, 5.0[+ or -]0.35, 2.2 mm respectively for 5-mm-diameter screws. Screws were drilled to the center of the thickness of butt member. When attaching screws, pilot holes were drilled into the edge of the butt member. The diameters of the pilot holes were equal to approximately 80 percent of the root diameter of the screws, and depths of the pilot holes were equal to approximately 75 percent of the penetration of the screws (Eckelman 1991). Diameters and depths of the drilled pilot holes were presented in Table 1 with depth of the penetrations according to each screw sizes.

Before testing, all specimens were kept in a chamber which was set to produce equilibrium moisture content (EMC) of 8 percent.

Method of loading and testing

A total of 360 joint specimens were tested; 180 were tested in compression and the remaining 180 were tested in tension. Figure 3 shows loading diagrams in testing corner joint moment capacities to compression and tension, respectively.

All of the tests were carried out on a 30 kN capacity Seidner bending machine at a loading rate of 6 mm/ min. In the tension test set-up, the bottoms of each of the two legs of the joints were placed on rollers so that the two joint members were free to move outward and free of restraint as the joint was loaded. The loading was continued until a failure or full separation occurred in the specimens. Joint failure modes and ultimate load values were recorded. Ultimate applied load values, F, measured in N, were converted to corresponding bending moment values by means of the expressions M.sub.c] = 0. 08627 F and [M.sub.t] = 0.04949 F, for compression and tension, respectively. [M.sub.c] and [M.sub.t] measured in Nm are the ultimate bending moments for specimens subjected to compression and tension loads, respectively. Moment arms were calculated 0.08627 and 0.09988 m for compression and tension loading, respectively. Moment calculation equations were derived referencing to the study on the bending moment resistance of multi-dowel corner joints (Zhang and Eckelman 1993b).

Some physical and mechanical properties of the PB and MDF were tested in accordance with the procedures described in ASTM D 1037 (2001a) and ASTM D 4442 (2001b).

Results and discussion

Some physical and mechanical properties of test materials

Average moisture content (MC) values were 7.24 (3.16) and 6.27 (1.41) percent; panel density values were 0.57 (1.81) and 0.75 (0.81) g/[cm.sup.3], and internal bond strengths (IB) were 0.52 (8.37) and 0.70 N/[mm.sup.2] (9.35) for PB and MDF, respectively. The values in parentheses are coefficients of variation in percent.

Failure modes

In the joints, joint intersection areas were covered with polyurethane adhesive. The joint intersection consists of a face area and a butt (edge) area. Because of the weak bonding properties of butt area, in fact, the bonding that occurred in this joint intersection area was not primarily intended to influence the strength of the joint. The strength of the joint was significantly provided by the screws used. As a matter of fact, the failure patterns of adhesive and screw are contrary to each other, i.e., screws fail slowly while glue joints have an abrupt failure pattern. Thus, it is not recommended to use these two fasteners in one joint. It should be noted that usage of adhesive in such joints primarily aims to fill the pilot holes and minor cracks generated in driving screws, thus mending the constructional defects on the joint.

In general, all joints failed completely between 60 and 90 seconds. Joints opened up slowly, not suddenly. In both compression and tension loadings, axial stresses occurred at the glue line. It is assumed that, when axial stresses exceeded the maximum axial strength of the adhesive, the glue bond failed. Then, screws started to take the load. At this point, the withdrawal strength of screws from material became important.

[FIGURE 3 OMITTED]

This phenomenon was clearly observed in the tests. Typical mode of failure was the opening occurred under both compression and tension loads until the glue bond lost its strength. Opening started at the edge of the joint, then propagated toward the top of the joint with increasing load. After screws had taken the load, the failure pattern changed. Hereafter, for all joints, edge splitting occurred in transverse direction around the screws. Amount of edge splitting around screws in MDF specimens was significantly more than those of PB specimens. Furthermore, joints failed due to the fractures around screws along with edge splitting. Screws never broke or bent. In some PB specimens, screws withdrew with adhered particles, which clearly explained the presence and effect of adhesive in and around the pilot holes. The zone of influence for each screw was approximately 50 to 60 mm in diameter. None of the influence zones overlapped, but the zone diameters increased as the diameter of the screw was increased. During the tests, failures such as splits or fractures occurred on the outer face of butt member under compression loads, while on the inner face of the butt member under tension loads.

Moment capacities in general

Mean ultimate bending moment capacities and their coefficients of variation are summarized in Table 2 for compression and tension tests, respectively. Results indicate that, in general, moment capacities of the joints were significantly affected by the material types. Moment capacities of the joints constructed of M DF showed higher moment resistances than ones of PB. Tests results of the previous studies (A. Tankut 2005, Zhang et al. 2005, N. Tankut 2006, Kasal et al. 2006) agree with this study. These differences in moment capacities could be explained by differences in density and mechanical properties such as bending strength, IB strength, and screw holding capacity of the panel constructing the joints. It is a fact that the density, IB strength, and screw-holding capacity of the panels directly affect the strength properties of joints. Most strength properties of MDF are higher than those of PB.

It was considered that the other important factor for providing a strong adhesive bonding is the surface roughness. The specific adhesion between the smooth surface and the glue line can be stronger. It is expected that the edges of MDF give smoother surface than the PB after machining. Hence, the adhesion between the MDF and glue line is stronger than the adhesion between the PB and the glue line. It is accepted that the adhesion is lower on the rough surfaces.

According to Table 2, number of screws is very important factor on moment capacity under compression. In comparison of joints with 2, 3, and 4 screws, it would seem that the more screws used in the joint, the stronger the joint. This is true, but the load bearing capacity per screw decreases when the number of screws is increased on a given width. Results of the compression and tension tests clearly indicated that moment capacity of joints decreases as their "zones of influence" get closer to each other. Mean comparisons results point out that increasing either screw diameter or screw length tends to positively affect the moment capacities. Screw length was found to have a greater effect on moment capacity than diameter. When the screw length was increased from 40 to 50 mm the moment capacities of joints increased significantly.

Grouping data for joints subjected compression loading test yielded a mean moment capacity of 87.98 Nm while grouping data of joints tested in tension loading resulted in a mean ultimate moment capacity of 97.22 Nm. Therefore, in general, it can be deduced that the joints loaded in tension have greater moment capacities than those loaded in compression. According to Zhang and Eckelman (1993b), the reason for this phenomenon is that moment resistance of the joints loaded in compression is presumably related to the IB strength of the board, whereas moment resistance of the joints loaded in tension is presumably related to the surface tensile strength parallel to the plane of the board. The difference, however, is not great. Moment capacities for the joints loaded in tension averaged only 10 percent greater than those for the joints loaded in compression in this study.

Four-factor analysis of variance (ANOVA) general linear model procedure was performed for individual data to analyze main effects and interactions on moment capacities of joints under compression and tension loads. Summary of ANOVA results for both compression and tension tests are provided in Table 3. For compression test data, ANOVA results indicated that four-factor interaction, all three-factor and two-factor interactions, and main factor effects were statistically significant ([alpha] = 0.05). On the contrary, for tension test data, ANOVA results indicated that main factor effects, two-factor interactions, and three-factor interactions except for the number of screws-screw diameter-screw length interaction were significant. The four-factor interaction was not significant. Therefore, in compression data mean comparisons for four-factor interaction were analyzed while significant three-factor interactions were analyzed in tension data.

The least significant difference (LSD) multiple comparisons procedure at 5 percent significance level was performed to determine the mean differences of moment capacities of corner joints tested under compression and tension loads considering the significant interactions in the ANOVA results mentioned above.

Moment capacities under compression

Table 4 gives ranked mean comparisons of moment capacities of tested joints under compression regarding the effect of material type-number of screws-screw diameter-screw length interaction. The single LSD critical value of 4.64 Nm for compression tests was calculated based on the error mean square of the full model. Results showed that the highest moment capacity values were obtained with 5 x 60 screws when the number of screws was 4, while the lowest moment capacity values were obtained with 4 x 40 screws when the number of screws was 2 in the joint for both PB and MDF specimens.

Generally, MDF specimens showed higher moment capacities than PB specimens. When the screw diameter was increased from 4 to 5 ram, moment capacities increased significantly with any number of screws and screw lengths for all specimens. Particularly, increment of screw length from 40 to 50 mm has significantly increased the moment capacities. Influence of the screw diameter or screw length on moment capacity is directly affected by the number of screws. Also, results of the tests showed that if relatively shorter screws will be used in the joints, the number of screws should be increased for sufficient strength.

Moment capacities under tension

Table 5 shows the ranked mean comparisons of moment capacities of joints tested under tension with respect to the material type-number of screws-screw diameter interaction. The LSD critical value of 2.117 Nm was calculated based on the error mean square of the full model. Results showed that there is no significant difference in moment capacities between 4- and 5-mm diameter screws for PB specimens with any number of screws while there was a 17 percent increase in strength for MDF specimens. Results also indicated that moment capacities of PB specimens connected with 4- or 5-mm diameter screws increased at each level by an average of 24 percent as number of screws increased from 2 to 4 in increments of 1. On the other hand, moment capacities of MDF specimens connected with 4- or 5-mm diameter screws increased by 26 percent as number of screws increased from 2 to 3, while by 47 percent as number of screws increased from 3 to 4.

Table 6 gives the ranked mean comparisons of moment capacities of joints tested under tension considering the material type-number of screws-screw length interaction. The single LSD critical value of 2.592 Nm was found based on the error mean square of the full model. According to means comparisons results, with 2 screws in the joint, moment capacities increased by an average of 18 percent as screw length increased from 40 to 50 mm, while by 7 percent as screw length increased from 50 to 60 mm for both PB and MDF specimens. In the case of 3 screws, moment capacities increased by average 11 percent as screw length increased from 50 to 60 mm for both PB and MDF specimens, while by 23 and 37 percent as screw length increased from 40 to 50 mm for PB and MDF specimens, respectively. When the number of screws was 4, moment capacities increased by average 15 percent as screw length increased from 50 to 60 mm for both PB and MDF specimens, while by 31 and 45 percent as screw length increased from 40 to 50 mm for PB and MDF specimens, respectively.

Table 6 also indicated that, for PB specimens, moment capacities increased by 16, 24, and 28 percent as number of screws was 2, 3, and 4 for 40-, 50-, and 60-mm screw lengths, respectively. For MDF specimens, moment capacities increased by 32, 49, and 54 percent as number of screws increased from 2 to 3 for 40-, 50-, and 60-ram screw lengths, respectively. Moment capacities increased by 19, 27, and 33 percent as number of screws increased from 3 to 4 for 40-, 50-, and 60-mm screw lengths, respectively. It was clearly seen that influence of screw length on moment capacity is directly affected by the number of screws.

Table 7 lists mean comparisons of moment capacities of joints tested under tension considering the material type--screw diameter-screw length interaction. The single LSD critical value of 2.117 Nm was calculated based on the error mean square of the full model. Mean comparisons results indicated that moment capacities increased by an average of 26 percent as screw length increased from 40 to 50 mm while 10 percent as screw length increased from 50 to 60 mm for both PB and MDF specimens with any screw diameter. In this case, it can be deduced that when the screw length increased from 40 to 50 mm, the moment capacities of joints increased significantly. Moment capacities was not significant as screw diameter increased from 4 to 5 mm with any screw length for PB specimens, while 20 percent of increase was obtained for MDF specimens.

Predictive expressions

To provide a means of comparing the results of the compression and tension tests as well as to obtain functional relationships between moment capacity and the various joint parameters for PB and MDF. Curves were fitted to the individual test data points by means of regression techniques. The curves had the following forms:

[CM.sub.PB] = -81.59 + 16.09 X + 14.52 Y + 1.47 Z [3]

[CM.sub.MDF] = -173.9 + 29.07 X + 24.62 Y + 2.26 Z [4]

[TM.PB] = -8.4 + 17.73 X + 44.55 Y + 1.32 Z [5]

[TM.sub.MDF] = -136.1 + 32.42 X + 17.25 Y + 2.2 Z [6]

where [CM.sub.PB], [CM.sub.MDF] = moment capacities of PB and MDF joints, respectively, under compression load (Nm); [TM.sub.PB] [TM.sub.MDF] = moment capacities of PB and MDF joints, respectively, under tension load (Nm); X = number of screws; Y = screw diameter (mm); Z = screw penetration (mm).

The coefficients of determination ([R.sup.2]) values were 0.899, 0.868, 0.913, and 0.907 for the expressions [3], [4], [5], and [6], respectively. To provide a practical evaluation of how well the values predicted by these expressions agreed with the observed moment capacity results, differences between the predicted and observed values were determined and the differences were expressed as a percentage of predicted values (Table 8). With the exception of few specimens, especially the specimens constructed of PB and MDF and connected with 4 x 40 screws and the number of screw was 2, predicted and observed values agree well under both compression and tension tests.

Conclusions

This study was mainly carried out to obtain background information concerning the moment capacity of glued-screwed corner joints in PB and MDF, and to provide background information needed to formulate expressions for predicting the moment capacity of this type of joints.

The other purpose of this study was to determine the effect of number of screws on moment capacity of screwed corner joints for cases. Comer joint specimens were constructed of two material types and six different sizes of screws in order to investigate the effect of these factors on moment capacity, too. Significant differences occurred in moment capacities with respect to type of material, number of screws, screw diameter, and screw length.

Results indicated that joints constructed of MDF yield higher moment capacities than those of PB. Moment capacity comparisons showed that the 4-screw connected joints evaluated in this study had significantly higher moment capacities than 3-screw connections and 2-screw connections under both compression and tension loads. Results of the tests also indicated that a screw comer joint becomes stronger as either screw diameter or screw length is increased. Screw length was found to have a greater effect on moment capacity than diameter. For case type furniture constructed of the type of PB and MDF evaluated in this study and fastened mainly with the screws, a screw size of 5 mm in diameter and 60-mm long and 4-screw connections are suggested to obtain maximum moment capacities.

The most important conclusion is that the comparisons of the predicted and observed results indicated that the average moment capacity of glued-screwed comer joints evaluated in this study could be estimated by developed predictive expressions.

This study provides case furniture manufacturers some information concerning the effects of joint construction factors such as number of screws, screw size, material type on moment capacities of glued-screwed comer joints. The information could give supportive insight to engineers in product engineering of case furniture.

Literature cited

American Soc. for Testing and Materials (ASTM). 2001a. Standard test methods for evaluating properties of wood-base fiber and particle panel materials. D 1037-99. ASTM, West Conshohocken, Pennsylvania.

--. 200lb. Standard test methods for direct moisture content measurement of wood and wood-base materials. D 4442-92. ASTM, West Conshohocken, Pennsylvania.

Atar, M. and A. Ozcifci. 2008. The effects of screw and back panels on the strength of corner joints in case furniture. Mater. Des. 29(2):519525.

Eckelman, C.A. 1974. Which screw holds best'? Furniture Design and Manufacturing Magazine.

--. 1975. Screw holding performance in hardwoods and particleboard. Forest Prod. J. 25(6):30-35.

--, 1978. Predicting withdrawal strength of sheet-metal-type screws in selected hardwoods. Forest Prod. J. 28(8):25-28.

--. 1991. Textbook of Product Engineering and Strength Design of Furniture. Purdue Univ., West Lafayette, Indiana.

Ho, C. 1991. The use of performance tests in evaluating joint and fastener strength ni case type furniture. MS thesis. Purdue Univ., West Lafayette, Indiana.

and C.A. Eckelman. 1994. The use of performance tests in evaluating joint and fastener strength in case furniture. Forest Prod. J. 44(9):47-53.

Kasal. A., S. Sener, C.M. Belgin, and H. Efe. 2006. Bending strength of screwed corner joints with different materials. Gazi Univ. J. of Science. Gazi Universitesi Fen Bilimleri Dergisi 19(3): 155-161.

Liu, W.-Q. and C.A. Eckelman. 1998. Effect of number of fasteners oil the strength of corner joints for cases. Forest Prod. J. 48(I):93-95.

Park, H.J., K. Semple, and G.D. Smith. 2006. Screw thread shape and fastener type effects on load capacities of screw-based particleboard joints in case construction. Forest Prod. J. 56(4):48-55.

Rajak, Z. 1989. Efficient use of screws in the construction of corner joints for case goods. M.S. thesis. Purdue Univ., West Lafayette, Indiana. 124 pp.

-- and C.A. Eckelman. 1993. Edge and face withdrawal strength of large screws in particleboard and medium density fiberboard. Forest Prod. J. 43(4):25-30.

Tankut, A.N. and N. Tankut. 2004. Effect of some factors on the strength of furniture corner joints constructed with wood biscuits. Turkish J. of Agriculture and Forestry. TOBiTAK Turk Tarlm ve Ormancilik Dergisi. 28:301-309.

--. 2005. Optimum dowel spacing for corner joints in 32-mm cabinet construction. Forest Prod. J. 55(12): 100-104.

Tankut, N. 2006. Moment resistance of corner joints connected with different RTA fasteners in cabinet construction. Forest Prod. J. 56(4):35-40.

Zhang, J. 1991. Rational design of dowel joints in case construction. M.S. thesis. Purdue Univ., West Lafayette, Indiana.

-- and C.A. Eckelman. 1993a. The bending moment resistance of single-dowel corner joints in case construction. Forest Prod. J. 43(6): 19-24.

-- and --. 993b. Rational design of multi-dowel comer joints in case construction. Forest Prod. J. 43(11/12):52-58. --, H. Ere, Y.Z. Erdil, A. Kasal, and N. Han. 2005. Moment resistance of multiscrew L-type corner joints. Forest Prod. J. 55(10): 56-63.

The author is Assistant Professor, Dept. of Wood Sci. and Furniture Design, Mugla Univ., Mugla, Turkey (alikasal@mu.edu.tr). This paper was received for publication in May 2007. Article No. 10356.

Tests were carried out in order to determine the bending moment capacity of L-type glued-screwed corner joints constructed with multiple screws. 6 sizes of screws were used for assembling the specimens. 2, 3, and 4 screws were used in the joints of the specimens that have the same width. Specimens were prepared of 18-mm-thick particleboard (PB) and medium density fiberboard (MDF) with resin surfacing. The specimens were constructed with both screws and polyurethane adhesive. Specimens were tested under static compression and tension loads. Results have shown that the maximum moment capacity is obtained with the MDF specimens when the number of screws in the joints is 4. In both compression and tension tests, MDF corner joints were stronger than PB corner joints. Results of the tests also indicated that a screw corner joint becomes stronger as either screw diameter or screw length is increased. Screw length was found to have a greater effect on moment capacity than diameter. Furthermore, the average moment capacities of glued-screwed corner joints evaluated in this study in compression could be predicted by means of the expressions:

[CM.sub.PB] = -81.59 + 16.09 X + 14.52 Y + 1.47 Z

[CM.sub.MDF] = -173.9 + 29.07 X + 24.62 Y + 2.26 Z

and in tension by means of the expressions:

[TM.sub.PB] = -8.4 + 17.73 X + 44.55 Y + 1.32 Z

[TM.sub.MDF] = -136.1 + 32.42 X + 17.25 Y + 2.2 Z

where [CM.sub.PB], [CM.sub.MDF] = moment capacities of PB and MDF joints, respectively, under compression load (Nm); [TM.sub.PB]. [TM.sub.MDF] = moment capacities of PB and MDF joints, respectively, under tension load (Nm); X = number of screws; Y = screw diameter (mm); Z = screw penetration (mm).

**********

Case furniture is widely produced and extensively used in kitchens and offices for storage. Particleboard (PB) and medium density fiberboard (MDF) are the most common wood-based composite panel products used in constructing this type of furniture.

Screw-type fasteners are commonly used to construct corner joints with or without glue. The rational design of case-type furniture constructed with screws requires information on the moment capacities of these in PB and MDF. An important consideration in the product engineering of screw jointed cases is the specification of the number of screws that should be used in joining the sides to bottom and top of the case. In spite of their widespread use, limited information is available regarding moment capacities and stiffness of screw-type corner joints in case construction. Published information is mostly related to direct withdrawal resistances of screw-type joints (Eckelman 1974, 1975, 1978; Rajak and Eckelman 1993).

Ho (1991), Rajak (1989), and Zhang (1991) have helped to clarify the relationships between number of fasteners used and joint strength by defining what has been termed the "zone of influence" or "zone of failure" of the fastener. This term signifies that an individual fastener is supported for a finite distance by the material on either side of it. As a result, when a fastener connection such as a screw in the side of a case fails, it causes a portion of the side wall on either side of it to fracture. In his work on the bending moment capacity of case joints constructed with large screws, Rajak (1989) obtained "zones of failure" that extended from 40 to 50 mm on either side of the fastener. Similarly, in his work on the performance testing of the cases Ho (1991) obtained a value of a little more than 40 mm on either side. In the case of dowels in similar material, Zhang (1991) found that the "zone of failure" extended about 40 mm on either side of the dowel. These studies define the minimum fastener spacing that will yield maximum fastener strengths. Liu and Eckelman (1998) constructed 457-mm-wide joints with up to 36 fasteners. Screwed and doweled joints were tested in compression in order to determine the bending moment capacity. Results showed that the moment capacities increase rapidly until the "zones of influence" of the fasteners overlap. No increase in moment capacity was obtained beyond that point. Ho and Eckelman (1994) conducted cyclic performance tests on furniture cases constructed with sheet metal screws for determining joint strength, stiffness, and durability as a function of the number of screws, screw diameter, screw length, and screw position. Test results demonstrated that screw length significantly affected case strength and durability. Results also showed that maximum fastener strength was obtained when fasteners were placed no closer than 89 mm from one another.

Research on corner joints using other types of fasteners in case construction applications such as dowels and screw--nuts, were reviewed as reference in conducting moment capacity evaluations of screw-type corner joints in MDF and PB. In compression and tension tests of single-dowel corner joints with PB, Zhang and Eckelman (1993a) determined that the moment capacity in tension loads was higher than in compression loads. As dowel diameter and length increase, strength increases proportionally. In evaluating the moment resistances of multi-dowel joints in PB, Zhang and Eckelman (1993b) conclude that a space of 75 mm between two dowels yields the highest moment capacity per dowel. Zhang et al. (2005) investigated the effects of screw diameter and length, loading, board material, board surface condition, and gluing on moment resistances of three-screw L-type corner joints under tension and compression loads. Results showed that surfacing PB with synthetic resin and assembling joints with glue applied to the contact surfaces of the face and butt members significantly improve moment resistances of joints constructed of PB. According to results of the tests, 5-mm diameter by 50- or 60-mm-long screws were recommended for connecting corner joints of case-type furniture. A.N. Tankut and N. Tankut (2004) investigated the effects of some factors on moment capacity of corner joints constructed with wood biscuits. Results indicated that joint strength comes mainly from the edge gluing of the face and butt member and not from the glued biscuits. In another study, the bending moment capacity of corner joints in 32-mm case construction prepared from laminated PB and MDF were researched by testing under tension and compression loads. The dowel spacing effects were investigated on bending moment capacity of the corner joints. Test results showed that maximum moment is obtained in joints when the spacing between the dowels is at least 96 mm. In the tests, MDF corner joints were stronger than PB corner joints (Tankut 2005). Park et al. (2006) investigated the stability of several, mostly screw-based, fasteners for connecting 16-mm-thick furniture grade PB. Results indicated that screws on their own performed markedly better than either dowels or screws with PVC anchors in all three tests. The use of thicker screws or screws with anchors was not recommended for butt-jointed shelving applications using PB. Screws with specialized thread configurations such as beveled edges or alternating thread heights performed well, but were not significantly different from plain-threaded screws of the same dimensions in any of the loading configurations. The effects of screw and back panel on the corner joints in case furniture produced from MDF and PB were determined (Atar and Ozcifci 2007). Static compression and tension loads were applied to the L-type corner joint specimens. In conclusion, it was deduced that the edges of panels should be rebated and covered with solid wood or massive wood for case furniture.

Although L-type, screw connected corner joints are commonly employed in construction of cabinet furniture, limited information is available on effects such as number of screws, screw sizes (diameter and length) loading type, and board material type on moment capacity of the joints. Accordingly, the primary purpose of this study was to obtain practical information concerning the moment capacity of screwed corner joints that the furniture engineers could use in the design of case furniture. The following objectives were tested:

* How ultimate bending moment capacity of joints is affected by the use of different panel materials, namely PB and MDF,

* How number of screws affects the moment capacity of L-type corner joints,

* How screw sizes (diameter and length) affect the moment capacity of L-type corner joints.

* Could the average bending moment capacity of screwed corner joints evaluated in this study be estimated by developed predictive expressions?

Experimental design

Altogether, 36 sets of specimens consisting of 5 replications each, or, a total of 180 specimens were prepared for compression tests; in the same manner, 180 specimens were prepared for tension tests. Full linear models (models [1] and [2]) for the four-way factorial experiments were considered to evaluate the influence of material type (PB and MDF), number of screws (2, 3, and 4), screw diameter (4 and 5 mm) and screw length (40, 50, and 60 mm) on moment capacity of L-type corner joints under compression and tension loads. The form models are:

[C.sub.ijklm] = [[mu].sub.1] +[A.sub.i] + [B.sub.j] + [C.sub.k] + [D.sub.l] + [(AB).sub.ij] + [(AC).sub.ik] + [(BC).sub.jk] + [(AD).sub.il] + [(BD).sub.jl] + [(CD).sub.kl] + [(ABC).sub.ijk] + [(ABD).sub.ijl] + [(ACD).sub.ikl] + [(BCD).sub.jkl] + [(ABCD).sub.ijkl] + [[epsilon].sub.ijklm] [1]

[T.sub.ijklm] = [[mu].sub.2] + [A.sub.i] + [B.sub.j] + [C.sub.k] + [D.sub.l] + [(AB).sub.ij] + [(AC).sub.ik] + [(BC).sub.jk] + [(AD).sub.il] + [(BD).sub.jl] + [(CD).sub.kl] + [(ABC).sub.ijk] + [(ABD).sub.ijl + [(ACD).sub.ikl] + [(BCD).sub.jkl] + [(ABCD).sub.ijkl] + [[epsilon].sub.ijklm] [2]

where [C.sub.ijklm] moment capacity under compression load (Nm); [T.sub.ijklm] = moment capacity under tension load (Nm); [[mu].sub.1] = population mean moment capacity under compression load for all material type-number of screws-screw diameter-screw length combinations (Nm); [[mu].sub.2] = population mean moment capacity under tension load for all material type-number of screws-screw diameter-screw length combinations (Nm); A discrete variable representing effect of material type; B = discrete variable representing effect of number of screws; C = xdiscrete variable representing effect of screw diameter; D discrete variable representing effect of screw length; (AB), (AC), (BC), (AD), (BD), (CD) = effect of the two-way interactions among the four variables; (ABC), (ABD), (ACD), (BCD) = effect of the three-way interactions among the four variables; (ABCD) = effect of the four-way interactions among the four variables; [epsilon] = random error term; i = index for material type, 1.2; j = index for number of screws, 1.3; k = index for screw diameter, 1.2; l = index for screw length, 1.3; m = index for the replication, 1.5.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

General configuration and construction of the specimens

The configuration of the L-type corner joint specimens in this study is shown in Figure 1. The specimen consists of a face member and a butt member of the same type material, joined together by 2, 3, or 4 screws. Cross sections of the joint area for the specimens connected with 2, 3, or 4 screws are given in Figures 2(a), (b), and (c).

The face member measured 350 by 158 by 18 mm, whereas the butt member measured 350 by 140 by 18 mm. 6 different sizes of screws which included 2 diameters and 3 lengths (4 by 40, 4 by 50, 4 by 60, 5 by 40, 5 by 50, and 5 by 60) were utilized for constructing the specimens. All specimens were assembled with both screws and polyurethane based adhesive with 60 percent solid content. This procedure is a common application in construction of corner joints for furniture cases. Polyurethane adhesive was applied to joining surfaces area at a rate of 200[+ or -]10 g/[m.sup.2], also a liberal amount of adhesive was applied into the pilot holes.

18-mm PB and MDF were selected as board materials for this study. The panels were obtained from commercial suppliers. In preparation of the specimens, 1880- by 3660-mm full-size sheet of board were first cut into face and butt member strips. These strips were subsequently cut into the desired member lengths.

Steel Phillips-head flat wood screws with 40[+ or -]3 degree thread angle were chosen for this study. Root diameter, outside diameter, and thread per mm were 2.4[+ or -]0.25, 4.0[+ or -]0.3, 1.8 mm respectively for 4-mm-diameter screws; and 3.0[+ or -]0.3, 5.0[+ or -]0.35, 2.2 mm respectively for 5-mm-diameter screws. Screws were drilled to the center of the thickness of butt member. When attaching screws, pilot holes were drilled into the edge of the butt member. The diameters of the pilot holes were equal to approximately 80 percent of the root diameter of the screws, and depths of the pilot holes were equal to approximately 75 percent of the penetration of the screws (Eckelman 1991). Diameters and depths of the drilled pilot holes were presented in Table 1 with depth of the penetrations according to each screw sizes.

Before testing, all specimens were kept in a chamber which was set to produce equilibrium moisture content (EMC) of 8 percent.

Method of loading and testing

A total of 360 joint specimens were tested; 180 were tested in compression and the remaining 180 were tested in tension. Figure 3 shows loading diagrams in testing corner joint moment capacities to compression and tension, respectively.

All of the tests were carried out on a 30 kN capacity Seidner bending machine at a loading rate of 6 mm/ min. In the tension test set-up, the bottoms of each of the two legs of the joints were placed on rollers so that the two joint members were free to move outward and free of restraint as the joint was loaded. The loading was continued until a failure or full separation occurred in the specimens. Joint failure modes and ultimate load values were recorded. Ultimate applied load values, F, measured in N, were converted to corresponding bending moment values by means of the expressions M.sub.c] = 0. 08627 F and [M.sub.t] = 0.04949 F, for compression and tension, respectively. [M.sub.c] and [M.sub.t] measured in Nm are the ultimate bending moments for specimens subjected to compression and tension loads, respectively. Moment arms were calculated 0.08627 and 0.09988 m for compression and tension loading, respectively. Moment calculation equations were derived referencing to the study on the bending moment resistance of multi-dowel corner joints (Zhang and Eckelman 1993b).

Some physical and mechanical properties of the PB and MDF were tested in accordance with the procedures described in ASTM D 1037 (2001a) and ASTM D 4442 (2001b).

Results and discussion

Some physical and mechanical properties of test materials

Average moisture content (MC) values were 7.24 (3.16) and 6.27 (1.41) percent; panel density values were 0.57 (1.81) and 0.75 (0.81) g/[cm.sup.3], and internal bond strengths (IB) were 0.52 (8.37) and 0.70 N/[mm.sup.2] (9.35) for PB and MDF, respectively. The values in parentheses are coefficients of variation in percent.

Failure modes

In the joints, joint intersection areas were covered with polyurethane adhesive. The joint intersection consists of a face area and a butt (edge) area. Because of the weak bonding properties of butt area, in fact, the bonding that occurred in this joint intersection area was not primarily intended to influence the strength of the joint. The strength of the joint was significantly provided by the screws used. As a matter of fact, the failure patterns of adhesive and screw are contrary to each other, i.e., screws fail slowly while glue joints have an abrupt failure pattern. Thus, it is not recommended to use these two fasteners in one joint. It should be noted that usage of adhesive in such joints primarily aims to fill the pilot holes and minor cracks generated in driving screws, thus mending the constructional defects on the joint.

In general, all joints failed completely between 60 and 90 seconds. Joints opened up slowly, not suddenly. In both compression and tension loadings, axial stresses occurred at the glue line. It is assumed that, when axial stresses exceeded the maximum axial strength of the adhesive, the glue bond failed. Then, screws started to take the load. At this point, the withdrawal strength of screws from material became important.

[FIGURE 3 OMITTED]

This phenomenon was clearly observed in the tests. Typical mode of failure was the opening occurred under both compression and tension loads until the glue bond lost its strength. Opening started at the edge of the joint, then propagated toward the top of the joint with increasing load. After screws had taken the load, the failure pattern changed. Hereafter, for all joints, edge splitting occurred in transverse direction around the screws. Amount of edge splitting around screws in MDF specimens was significantly more than those of PB specimens. Furthermore, joints failed due to the fractures around screws along with edge splitting. Screws never broke or bent. In some PB specimens, screws withdrew with adhered particles, which clearly explained the presence and effect of adhesive in and around the pilot holes. The zone of influence for each screw was approximately 50 to 60 mm in diameter. None of the influence zones overlapped, but the zone diameters increased as the diameter of the screw was increased. During the tests, failures such as splits or fractures occurred on the outer face of butt member under compression loads, while on the inner face of the butt member under tension loads.

Moment capacities in general

Mean ultimate bending moment capacities and their coefficients of variation are summarized in Table 2 for compression and tension tests, respectively. Results indicate that, in general, moment capacities of the joints were significantly affected by the material types. Moment capacities of the joints constructed of M DF showed higher moment resistances than ones of PB. Tests results of the previous studies (A. Tankut 2005, Zhang et al. 2005, N. Tankut 2006, Kasal et al. 2006) agree with this study. These differences in moment capacities could be explained by differences in density and mechanical properties such as bending strength, IB strength, and screw holding capacity of the panel constructing the joints. It is a fact that the density, IB strength, and screw-holding capacity of the panels directly affect the strength properties of joints. Most strength properties of MDF are higher than those of PB.

It was considered that the other important factor for providing a strong adhesive bonding is the surface roughness. The specific adhesion between the smooth surface and the glue line can be stronger. It is expected that the edges of MDF give smoother surface than the PB after machining. Hence, the adhesion between the MDF and glue line is stronger than the adhesion between the PB and the glue line. It is accepted that the adhesion is lower on the rough surfaces.

According to Table 2, number of screws is very important factor on moment capacity under compression. In comparison of joints with 2, 3, and 4 screws, it would seem that the more screws used in the joint, the stronger the joint. This is true, but the load bearing capacity per screw decreases when the number of screws is increased on a given width. Results of the compression and tension tests clearly indicated that moment capacity of joints decreases as their "zones of influence" get closer to each other. Mean comparisons results point out that increasing either screw diameter or screw length tends to positively affect the moment capacities. Screw length was found to have a greater effect on moment capacity than diameter. When the screw length was increased from 40 to 50 mm the moment capacities of joints increased significantly.

Grouping data for joints subjected compression loading test yielded a mean moment capacity of 87.98 Nm while grouping data of joints tested in tension loading resulted in a mean ultimate moment capacity of 97.22 Nm. Therefore, in general, it can be deduced that the joints loaded in tension have greater moment capacities than those loaded in compression. According to Zhang and Eckelman (1993b), the reason for this phenomenon is that moment resistance of the joints loaded in compression is presumably related to the IB strength of the board, whereas moment resistance of the joints loaded in tension is presumably related to the surface tensile strength parallel to the plane of the board. The difference, however, is not great. Moment capacities for the joints loaded in tension averaged only 10 percent greater than those for the joints loaded in compression in this study.

Four-factor analysis of variance (ANOVA) general linear model procedure was performed for individual data to analyze main effects and interactions on moment capacities of joints under compression and tension loads. Summary of ANOVA results for both compression and tension tests are provided in Table 3. For compression test data, ANOVA results indicated that four-factor interaction, all three-factor and two-factor interactions, and main factor effects were statistically significant ([alpha] = 0.05). On the contrary, for tension test data, ANOVA results indicated that main factor effects, two-factor interactions, and three-factor interactions except for the number of screws-screw diameter-screw length interaction were significant. The four-factor interaction was not significant. Therefore, in compression data mean comparisons for four-factor interaction were analyzed while significant three-factor interactions were analyzed in tension data.

The least significant difference (LSD) multiple comparisons procedure at 5 percent significance level was performed to determine the mean differences of moment capacities of corner joints tested under compression and tension loads considering the significant interactions in the ANOVA results mentioned above.

Moment capacities under compression

Table 4 gives ranked mean comparisons of moment capacities of tested joints under compression regarding the effect of material type-number of screws-screw diameter-screw length interaction. The single LSD critical value of 4.64 Nm for compression tests was calculated based on the error mean square of the full model. Results showed that the highest moment capacity values were obtained with 5 x 60 screws when the number of screws was 4, while the lowest moment capacity values were obtained with 4 x 40 screws when the number of screws was 2 in the joint for both PB and MDF specimens.

Generally, MDF specimens showed higher moment capacities than PB specimens. When the screw diameter was increased from 4 to 5 ram, moment capacities increased significantly with any number of screws and screw lengths for all specimens. Particularly, increment of screw length from 40 to 50 mm has significantly increased the moment capacities. Influence of the screw diameter or screw length on moment capacity is directly affected by the number of screws. Also, results of the tests showed that if relatively shorter screws will be used in the joints, the number of screws should be increased for sufficient strength.

Moment capacities under tension

Table 5 shows the ranked mean comparisons of moment capacities of joints tested under tension with respect to the material type-number of screws-screw diameter interaction. The LSD critical value of 2.117 Nm was calculated based on the error mean square of the full model. Results showed that there is no significant difference in moment capacities between 4- and 5-mm diameter screws for PB specimens with any number of screws while there was a 17 percent increase in strength for MDF specimens. Results also indicated that moment capacities of PB specimens connected with 4- or 5-mm diameter screws increased at each level by an average of 24 percent as number of screws increased from 2 to 4 in increments of 1. On the other hand, moment capacities of MDF specimens connected with 4- or 5-mm diameter screws increased by 26 percent as number of screws increased from 2 to 3, while by 47 percent as number of screws increased from 3 to 4.

Table 6 gives the ranked mean comparisons of moment capacities of joints tested under tension considering the material type-number of screws-screw length interaction. The single LSD critical value of 2.592 Nm was found based on the error mean square of the full model. According to means comparisons results, with 2 screws in the joint, moment capacities increased by an average of 18 percent as screw length increased from 40 to 50 mm, while by 7 percent as screw length increased from 50 to 60 mm for both PB and MDF specimens. In the case of 3 screws, moment capacities increased by average 11 percent as screw length increased from 50 to 60 mm for both PB and MDF specimens, while by 23 and 37 percent as screw length increased from 40 to 50 mm for PB and MDF specimens, respectively. When the number of screws was 4, moment capacities increased by average 15 percent as screw length increased from 50 to 60 mm for both PB and MDF specimens, while by 31 and 45 percent as screw length increased from 40 to 50 mm for PB and MDF specimens, respectively.

Table 6 also indicated that, for PB specimens, moment capacities increased by 16, 24, and 28 percent as number of screws was 2, 3, and 4 for 40-, 50-, and 60-mm screw lengths, respectively. For MDF specimens, moment capacities increased by 32, 49, and 54 percent as number of screws increased from 2 to 3 for 40-, 50-, and 60-ram screw lengths, respectively. Moment capacities increased by 19, 27, and 33 percent as number of screws increased from 3 to 4 for 40-, 50-, and 60-mm screw lengths, respectively. It was clearly seen that influence of screw length on moment capacity is directly affected by the number of screws.

Table 7 lists mean comparisons of moment capacities of joints tested under tension considering the material type--screw diameter-screw length interaction. The single LSD critical value of 2.117 Nm was calculated based on the error mean square of the full model. Mean comparisons results indicated that moment capacities increased by an average of 26 percent as screw length increased from 40 to 50 mm while 10 percent as screw length increased from 50 to 60 mm for both PB and MDF specimens with any screw diameter. In this case, it can be deduced that when the screw length increased from 40 to 50 mm, the moment capacities of joints increased significantly. Moment capacities was not significant as screw diameter increased from 4 to 5 mm with any screw length for PB specimens, while 20 percent of increase was obtained for MDF specimens.

Predictive expressions

To provide a means of comparing the results of the compression and tension tests as well as to obtain functional relationships between moment capacity and the various joint parameters for PB and MDF. Curves were fitted to the individual test data points by means of regression techniques. The curves had the following forms:

[CM.sub.PB] = -81.59 + 16.09 X + 14.52 Y + 1.47 Z [3]

[CM.sub.MDF] = -173.9 + 29.07 X + 24.62 Y + 2.26 Z [4]

[TM.PB] = -8.4 + 17.73 X + 44.55 Y + 1.32 Z [5]

[TM.sub.MDF] = -136.1 + 32.42 X + 17.25 Y + 2.2 Z [6]

where [CM.sub.PB], [CM.sub.MDF] = moment capacities of PB and MDF joints, respectively, under compression load (Nm); [TM.sub.PB] [TM.sub.MDF] = moment capacities of PB and MDF joints, respectively, under tension load (Nm); X = number of screws; Y = screw diameter (mm); Z = screw penetration (mm).

The coefficients of determination ([R.sup.2]) values were 0.899, 0.868, 0.913, and 0.907 for the expressions [3], [4], [5], and [6], respectively. To provide a practical evaluation of how well the values predicted by these expressions agreed with the observed moment capacity results, differences between the predicted and observed values were determined and the differences were expressed as a percentage of predicted values (Table 8). With the exception of few specimens, especially the specimens constructed of PB and MDF and connected with 4 x 40 screws and the number of screw was 2, predicted and observed values agree well under both compression and tension tests.

Conclusions

This study was mainly carried out to obtain background information concerning the moment capacity of glued-screwed corner joints in PB and MDF, and to provide background information needed to formulate expressions for predicting the moment capacity of this type of joints.

The other purpose of this study was to determine the effect of number of screws on moment capacity of screwed corner joints for cases. Comer joint specimens were constructed of two material types and six different sizes of screws in order to investigate the effect of these factors on moment capacity, too. Significant differences occurred in moment capacities with respect to type of material, number of screws, screw diameter, and screw length.

Results indicated that joints constructed of MDF yield higher moment capacities than those of PB. Moment capacity comparisons showed that the 4-screw connected joints evaluated in this study had significantly higher moment capacities than 3-screw connections and 2-screw connections under both compression and tension loads. Results of the tests also indicated that a screw comer joint becomes stronger as either screw diameter or screw length is increased. Screw length was found to have a greater effect on moment capacity than diameter. For case type furniture constructed of the type of PB and MDF evaluated in this study and fastened mainly with the screws, a screw size of 5 mm in diameter and 60-mm long and 4-screw connections are suggested to obtain maximum moment capacities.

The most important conclusion is that the comparisons of the predicted and observed results indicated that the average moment capacity of glued-screwed comer joints evaluated in this study could be estimated by developed predictive expressions.

This study provides case furniture manufacturers some information concerning the effects of joint construction factors such as number of screws, screw size, material type on moment capacities of glued-screwed comer joints. The information could give supportive insight to engineers in product engineering of case furniture.

Literature cited

American Soc. for Testing and Materials (ASTM). 2001a. Standard test methods for evaluating properties of wood-base fiber and particle panel materials. D 1037-99. ASTM, West Conshohocken, Pennsylvania.

--. 200lb. Standard test methods for direct moisture content measurement of wood and wood-base materials. D 4442-92. ASTM, West Conshohocken, Pennsylvania.

Atar, M. and A. Ozcifci. 2008. The effects of screw and back panels on the strength of corner joints in case furniture. Mater. Des. 29(2):519525.

Eckelman, C.A. 1974. Which screw holds best'? Furniture Design and Manufacturing Magazine.

--. 1975. Screw holding performance in hardwoods and particleboard. Forest Prod. J. 25(6):30-35.

--, 1978. Predicting withdrawal strength of sheet-metal-type screws in selected hardwoods. Forest Prod. J. 28(8):25-28.

--. 1991. Textbook of Product Engineering and Strength Design of Furniture. Purdue Univ., West Lafayette, Indiana.

Ho, C. 1991. The use of performance tests in evaluating joint and fastener strength ni case type furniture. MS thesis. Purdue Univ., West Lafayette, Indiana.

and C.A. Eckelman. 1994. The use of performance tests in evaluating joint and fastener strength in case furniture. Forest Prod. J. 44(9):47-53.

Kasal. A., S. Sener, C.M. Belgin, and H. Efe. 2006. Bending strength of screwed corner joints with different materials. Gazi Univ. J. of Science. Gazi Universitesi Fen Bilimleri Dergisi 19(3): 155-161.

Liu, W.-Q. and C.A. Eckelman. 1998. Effect of number of fasteners oil the strength of corner joints for cases. Forest Prod. J. 48(I):93-95.

Park, H.J., K. Semple, and G.D. Smith. 2006. Screw thread shape and fastener type effects on load capacities of screw-based particleboard joints in case construction. Forest Prod. J. 56(4):48-55.

Rajak, Z. 1989. Efficient use of screws in the construction of corner joints for case goods. M.S. thesis. Purdue Univ., West Lafayette, Indiana. 124 pp.

-- and C.A. Eckelman. 1993. Edge and face withdrawal strength of large screws in particleboard and medium density fiberboard. Forest Prod. J. 43(4):25-30.

Tankut, A.N. and N. Tankut. 2004. Effect of some factors on the strength of furniture corner joints constructed with wood biscuits. Turkish J. of Agriculture and Forestry. TOBiTAK Turk Tarlm ve Ormancilik Dergisi. 28:301-309.

--. 2005. Optimum dowel spacing for corner joints in 32-mm cabinet construction. Forest Prod. J. 55(12): 100-104.

Tankut, N. 2006. Moment resistance of corner joints connected with different RTA fasteners in cabinet construction. Forest Prod. J. 56(4):35-40.

Zhang, J. 1991. Rational design of dowel joints in case construction. M.S. thesis. Purdue Univ., West Lafayette, Indiana.

-- and C.A. Eckelman. 1993a. The bending moment resistance of single-dowel corner joints in case construction. Forest Prod. J. 43(6): 19-24.

-- and --. 993b. Rational design of multi-dowel comer joints in case construction. Forest Prod. J. 43(11/12):52-58. --, H. Ere, Y.Z. Erdil, A. Kasal, and N. Han. 2005. Moment resistance of multiscrew L-type corner joints. Forest Prod. J. 55(10): 56-63.

The author is Assistant Professor, Dept. of Wood Sci. and Furniture Design, Mugla Univ., Mugla, Turkey (alikasal@mu.edu.tr). This paper was received for publication in May 2007. Article No. 10356.

Table 1.--Screw sizes, screw penetration depths, and diameters and depths of the drilled pilot holes for each screw type. Diameter of Depth of Depth of Screw Screw pilot pilot penetration diameter length holes holes of screw (mm) 40 2.5 15 22 4 50 2.5 24 32 60 2.5 32 42 40 3.0 15 22 5 50 3.0 24 32 60 3.0 32 42 Table 2.--Average moment capacity values of the corner joints under compression and tensions loads with their coefficients of variation. Moment capacity (Nm) Under Number compression of Screw Screw screws diameter length PB MDF (mm) 2 4 40 51.22 (2.48) * 53.41 (3.44) 2 4 50 61.04 (3.00) 63.75 (4.24) 2 4 60 64.42 (2.54) 66.79 (2.84) 2 5 40 55.64 (4.33) 57.49 (1.42) 2 5 50 69.15 (4.01) 72.38 (3.45) 2 5 60 81.85 (2.25) 83.69 (6.13) 3 4 40 59.26 (6.09) 65.92 (4.79) 3 4 50 73.05 (4.39) 89.61 (2.96) 3 4 60 77.60 (5.05) 97.07 (3.45) 3 5 40 66.77 (5.72) 83.50 (3.30) 3 5 50 86.23 (3.48) 113.27 (3.80) 3 5 60 105.66 (5.55) 145.76 (2.99) 4 4 40 69.82 (3.04) 81.00 (4.43) 4 4 50 90.11 (4.85) 113.27 (3.80) 4 4 60 102.14 (8.09) 127.83 (4.51) 4 5 40 82.33 (2.58) 94.69 (3.33) 4 5 50 100.93 (3.05) 142.88 (2.95) 4 5 60 131.21 (3.06) 186.61 (3.25) Moment capacity (Nm) Under Number tension of screws PB MDF 2 59.95 (4.18) 60.92 (4.42) 2 68.97 (2.02) 72.27 (3.35) 2 72.75 (5.27) 76.63 (3.69) 2 60.03 (3.05) 69.26 (2.97) 2 69.46 (3.53) 84.38 (2.93) 2 74.69 (2.91) 91.67 (1.86) 3 71.78 (3.97) 80.52 (3.07) 3 89.73 (4.27) 110.30 (2.67) 3 96.52 (3.26) 120.29 (1.02) 3 68.87 (4.26) 90.70 (2.87) 3 83.91 (3.11) 129.51 (2.27) 3 95.07 (3.61) 138.73 (2.66) 4 80.52 (3.07) 92.74 (1.57) 4 107.19 (3.13) 136.21 (2.03) 4 119.80 (2.44) 157.14 (3.88) 4 81.96 (1.71) 111.56 (3.30) 4 106.41 (1.63) 160.56 (2.19) 4 122.71 (2.12) 185.97 (2.49) * Values in parenthesis are coefficients of variation. Table 3.--Summary of the ANOVA results for compression and tension tests. Degrees of Sum of Mean Source freedom squares squares ANOVA for compression Material type (A) 1 13368.862 13368.862 Number of screws (B) 2 61253.255 30626.628 A x B 2 5370.781 2685.390 Screw diameter (C) 1 17256.675 17256.675 A x C 1 1145.977 1145.977 B x C 2 2151.951 1075.976 A x B x C 2 600.369 300.185 Screw length (D) 2 42312.793 21156.396 A x D 2 1927.456 963.728 B x D 4 6217.038 1554.259 A x B x D 4 1118.711 279.678 C x D 2 4398.567 2199.283 A x C x D 2 312.458 156.229 B x C x D 4 495.247 123.812 A x B x C x D 4 295.530 73.882 Error 144 1983.421 13.774 Total 179 160209.089 ANOVA for tension Material type 1 26769.975 26769.975 Number of screws 2 75496.367 37748.184 A x B 2 6673.011 3336.506 Screw diarneter 1 3174.779 3174.779 A x C 1 3530.125 3530.125 B x C 2 395.732 197.866 A x B x C 2 276.728 138.364 Screw length 2 39015.429 19507.714 A x D 2 2635.904 1317.952 B x D 4 7593.830 1898.457 A x B x D 4 815.001 203.750 C x D 2 185.048 92.524 A x C x D 2 131.641 65.820 B x C x D 4 8.508 2.127 A x B x C x D 4 29.182 7.296 Error 144 1238.481 8.601 Total 179 167969.481 Source F value Prob. (sig) ANOVA for compression Material type (A) 970.6040 0.0000 Number of screws (B) 2223.5496 0.0000 A x B 194.9643 0.0000 Screw diameter (C) 1252.8664 0.0000 A x C 83.0000 0.0000 B x C 78.1178 0.0000 A x B x C 21.7940 0.0000 Screw length (D) 1535.9934 0.0000 A x D 69.9684 0.0000 B x D 112.8421 0.0000 A x B x D 20.3051 0.0000 C x D 159.6720 0.0000 A x C x D 11.3425 0.0000 B x C x D 8.9890 0.0000 A x B x C x D 5.3640 0.0005 Error Total ANOVA for tension Material type 3112.5843 0.0000 Number of screws 4389.0367 0.0000 A x B 387.9404 0.0000 Screw diarneter 369.1363 0.0000 A x C 410.4528 0.0000 B x C 23.0062 0.0000 A x B x C 16.0878 0.0000 Screw length 2268.1906 0.0000 A x D 153.2402 0.0000 B x D 220.7364 0.0000 A x B x D 23.6904 0.0000 C x D 10.7579 0.0000 A x C x D 7.6530 0.0007 B x C x D 0.2473 NS A x B x C x D 0.8483 NS Error Total NS: Not significant. Table 4.--Results of mean comparisons for material type- number of screws-screw diameter-screw length interactions on moment capacities under compression. * Material Number Screw Screw Moment type of diameter length capacity screws (mm) (Nm) MDF 4 5 60 186.7 A MDF 3 5 60 145.8 B MDF 4 5 50 142.9 B PB 4 5 60 131.2 C MDF 4 4 60 127.8 C MDF 4 4 50 113.3 D MDF 3 5 50 113.3 D PB 3 5 60 105.7 E PB 4 4 60 102.1 EF PB 4 5 50 100.9 FG MDF 3 4 60 97.07 GH MDF 4 5 40 94.69 HI PB 4 4 50 90.12 IJ MDF 3 4 50 89.62 J PB 3 5 50 86.24 JK MDF 2 5 60 83.70 KL MDF 3 5 40 83.51 KL PB 4 5 40 82.34 KL PB 2 5 60 81.85 KLM MDF 4 4 40 81.01 LM PB 3 4 60 77.61 MN PB 3 4 50 73.06 NO MDF 2 5 50 72.38 O PB 4 4 40 69.83 OP PB 2 5 50 69.15 OP MDF 2 4 60 66.79 PQ PB 3 5 40 66.77 PQ MDF 3 4 40 65.93 PQ PB 2 4 60 64.43 QR MDF 2 4 50 63.75 QRS PB 2 4 50 61.04 RST PB 3 4 40 59.53 STU MDF 2 5 40 57.49 TUV PB 2 5 40 55.64 UVW MDF 2 4 40 53.42 VW PB 2 4 40 51.23 W LSD value: 4.640 Nm * Values followed by the same upper case letter are not significantly different. Table 5.--Results of mean comparisons for material type-number of screws-screw diameter interactions on moment capacities under tension. Material Number Screw Moment type of screws diameter capacity (mm) (Nm) MDF 4 5 152.7 A MDF 4 4 128.7 B MDF 3 5 119.7 C MDF 3 4 103.7 D PB 4 5 103.7 D PB 4 4 102.5 D PB 3 4 86.02 E PB 3 5 82.62 F MDF 2 5 81.78 F MDF 2 4 69.95 G PB 2 5 68.07 GH PB 2 4 67.23 H LSD value: 2.117 Nm Table 6.--Results of mean comparisons for material type-number of screws-screw length interactions on moment capacities under tension. Material Number Screw Moment type of screws length capacity (mm) (Nm) MDF 4 60 171.6 A MDF 4 50 148.4 B MDF 3 60 129.5 C PB 4 60 121.3 D MDF 3 50 116.9 D PB 4 50 106.8 E MDF 4 40 102.2 F PB 3 60 95.80 G PB 3 50 86.82 H MDF 3 40 85.62 HI MDF 2 60 84.16 I PB 4 40 81.25 J MDF 2 50 78.33 K PB 2 60 73.73 L PB 3 40 70.33 M PB 2 50 69.22 M MDF 2 40 65.10 N PB 2 40 60.00 O LSD value: 2.592 Nm Table 7--Results of mean comparisons for material type-screw diameter-screw length interactions on moment capacities under tension. Material Screw Screw Moment type diameter length capacity (min) (Nm) MDF 5 60 138.8 A MDF 5 50 124.8 B MDF 4 60 118.0 C MDF 4 50 106.3 D PB 5 60 97.50 E PB 4 60 96.36 E MDF 5 40 96.51 F PB 4 50 88.64 FG PB 5 50 86.60 G MDF 4 40 78.07 H PB 4 40 70.76 I PB 5 40 70.29 I LSD value: 2.117 Nm Table 8.--Comparison of observed test results obtained in this study with values obtained with predictive expressions. Material Number of Screw Screw type screws diameter length (mm) PB 2 4 40 2 4 50 2 4 60 2 5 40 2 5 50 2 5 60 3 4 40 3 4 50 3 4 60 3 5 40 3 5 50 3 5 60 4 4 40 4 4 50 4 4 60 4 5 40 4 5 50 4 5 60 MDF 2 4 40 2 4 50 2 4 60 2 5 40 2 5 50 2 5 60 3 4 40 3 4 50 3 4 60 3 5 40 3 5 50 3 5 60 4 4 40 4 4 50 4 4 60 4 5 40 4 5 50 4 5 60 Moment capacity (Nm) Under compression Material Predicted Observed Difference type (Nm) (percent) PB 41.24 51.22 -19.4 56.03 61.04 -8.2 70.81 64.42 9.9 55.76 55.64 0.2 70.55 69.15 2.0 85.34 81.85 4.2 57.33 59.26 -3.2 72.11 73.05 -1.2 86.90 77.60 11.9 71.85 66.77 7.6 86.64 86.23 0.4 101.43 105.66 -4.0 73.42 69.82 5.1 88.21 90.11 -2.1 102.99 102.14 0.8 87.95 82.33 6.8 102.73 100.93 1.7 117.53 131.21 -10.4 MDF 32.58 53.41 -39.0 55.23 63.75 -13.3 77.87 66.79 16.5 57.21 57.49 -0.4 79.86 72.38 10.3 102.50 83.69 22.4 61.65 65.92 -6.4 84.30 89.61 -5.9 106.94 97.07 10.1 86.28 83.50 3.3 108.93 113.27 -3.8 131.58 145.76 -9.7 90.71 81.00 11.9 113.36 113.27 0.0 136.01 127.83 6.3 115.34 94.69 21.8 137.99 142.88 -3.4 160.65 186.61 -13.9 Under tension Material Predicted Observed Difference type (Nm) (percent) PB 54.31 59.95 -9.4 67.51 68.97 -2.1 80.69 72.75 10.9 53.87 60.03 -10.3 67.06 69.46 -3.4 80.25 74.69 7.4 72.04 71.78 0.3 85.24 89.73 -5.0 98.43 96.52 1.9 71.60 68.87 3.9 84.79 83.91 1.0 97.98 95.07 3.0 89.78 80.52 11.5 102.98 107.19 -3.9 116.17 119.80 -3.0 89.33 81.96 8.9 102.53 106.41 -3.6 115.72 122.71 -5.7 MDF 46.31 60.92 -24.0 68.37 72.27 -5.4 90.42 76.63 18.0 63.56 69.26 -8.2 85.62 84.38 1.4 107.68 91.67 17.4 78.73 80.52 -2.2 100.79 110.30 -8.6 122.84 120.29 2.1 95.98 90.70 5.8 118.04 129.51 -8.8 140.10 138.73 0.9 111.14 92.74 19.8 133.20 136.21 -2.2 155.27 157.14 -1.1 128.40 111.56 15.1 150.46 160.56 -6.2 172.52 185.97 -7.2

Printer friendly Cite/link Email Feedback | |

Author: | Kasal, Ali |
---|---|

Publication: | Forest Products Journal |

Geographic Code: | 1USA |

Date: | Jun 1, 2008 |

Words: | 7796 |

Previous Article: | Prevention of drying defects and drying degrade in hardwood lumber by predrying treatment. |

Next Article: | Use of fire-impacted trees for oriented strandboards. |

Topics: |