Effect of finger length on fingerjoint strength in radiata pine

Effect of finger length on fingerjoint strength in radiata pine Walford, G Bryan 1 SUMMARY Fingerjoints were cut in ten matched batches of dry 90x35 mm radiata pine by different manufacturers, using a variety of profiles with lengths ranging from 3.5 to 16 mm. All batches easily met the bending strength criteria of the Japanese standard for fingerjointed dimension lumber. Two batches of joints showed faulty manufacture and one of these failed the soak/dry glueline durability test in JASS 701. Analysis of the other eight batches showed that in radiata pine: Shorter joints are slightly stronger than longer ones but require greater precision in manufacture. Joint strength increases as wood stiffness or density increases. Wood strength governs in low density wood while glueline strength governs in high density wood. A small tip gap improves joint strength. It is concluded that the minimum limit of 12 mm on fingerjoint length required in the Japanese standard for fingerjointed dimension lumber is reasonable in timbers with poor glueline bond strength and/or low shear strength but is irrelevant in properly manufactured joints in radiata pine. BACKGROUND A mission by the NZ Ministry of Forestry, the NZ Forest Research Institute and the NZ timber industry to Japan in November 1995 discussed the restriction on the length of fingerjoints in the Japanese standard for fingerjointed dimension lumber (JAS 701). This standard specifies a minimum finger length of 12mm. Few manufacturers in New Zealand can produce joints of this length. For glulam, finger length is not specified but JAS 701 is sometimes quoted in glulam specifications. There seems an inconsistency then in that JAS 701, which deals with fingerjointed framing, requires the 12 mm minimum but the glulam standard does not. There is no restriction for non-structural fingerjointing. Research comparing the properties of fingerjoints of different lengths would be welcomed by the Japanese. SURVEY OF FINGERJOINT PROFILES There are about 30 companies producing fingerjointed timber in New Zealand. Some operate more than one fingerjointing machine. For example they may have one machine producing timber intended for structural uses while another produces for non-structural uses. A survey of most of these companies showed that machines and/or cutters are available to make the following range of fingerjoint lengths: Fingerjoint length, mm 4 6 8 10 12 16 Number of machines 13 2 3 13 2 1 4 and 10 mm fingerjoints are the most common sizes. Joints with lengths up to 8 mm are generally used in non-structural applications, joints 12 mm and longer are generally used in structural applications while 10 mm joints are used in both applications. THEORY OF FINGERJOINT STRENGTH Referring to the diagram in Figure 1, if pitch = p, tip width = t and length = L, tip gap = g, Milner, et al (1989), derived formulae for joint strength based on two modes of fingerjoint fracture under tension loads: 1 Senior scientist, NZ Forest Research Institute Ltd., Rotorua, New Zealand

(a) Failure in the glueline: ( bd / sin )( 1 2t p) Pg = f g θ / [1] (b) Failure of the wood in tension: Pt = f tbd( 1 t / p) where = ( p / 2 t) /( L + g) tanθ f g = glueline strength in shear/tension f t = wood strength in tension. θ = slope of finger b,d = member cross section dimensions [2] length L pitch p It is noted that finger length enters into equation [1] as it affects the slope of the fingers. Therefore, if glueline strength governs joint strength then finger length is important. Length does not enter into equation [2]. Therefore, joints made in a timber such as radiata pine which has high shear strength, good glue bonding characteristics and low tension strength in comparison to, say, Douglas fir, will be governed by equation [2] and show no effect of finger length. tip width t tip gap g Figure 1. Dimensions of fingerjoints DESCRIPTION OF THE STUDY About 600 m of 90x35 dry machine graded radiata pine was obtained in a range of qualities to fairly represent this resource. This timber was cut into 450 mm lengths, taking care to have half of each length defect-free. Each piece was weighed and the transit time for a sonic pulse to travel the length of each piece was measured. From these measurements density and Modulus of Elasticity (MoE) were calculated where MoE = density x (speed) 2 and density = (weight at test)/(volume at test). The pieces were sorted randomly into 10 batches of 120 pieces. Within each batch the pieces were sorted according to ascending MoE. One batch at a time was taken to a selected fingerjoint manufacturer to have the profiles cut in the defect-free ends of the lengths of timber. The manufacturers selected are listed in Table 1. Two joint orientations were available: where the profile appears on the wide face (face-face cut), or where it appears on the narrow face (edge-edge cut). Table 1. Fingerjoint manufacturers participating in the study. Company Finger length, mm Joint reference City 4 8 10 12 16 use* A Rotorua edge N B Taupo edge N C Tokoroa edge N D Rotorua face N E Rotorua face S F Auckland edge S G Rotorua edge S H Rotorua face S I Hamilton face N J Rotorua face N Total 3 1 4 1 1 * S = primarily structural use, N = primarily non-structural use The pieces were fed through the fingerjointing machines in sequence. Each batch was returned to the laboratory, and the joints were assembled within 24 hours using resorcinol glue (ICI R15). By joining them in sequence, pieces of similar MoE were joined to each other. The same cramping pressure was used in each case. Each batch was divided randomly into three

groups of 20 joints. Soak/dry glue bonding tests and bending on the flat strength tests were done as specified in JAS 701. The third group of 20 joints was tested in tension. RESULTS AND DISCUSSION Joint profiles The dimensions of each joint as measured in the assembled condition are given in Table 2 where the dimensions were defined in Figure 1. Table 2 shows that the joints have similar ratios of length/pitch while the structural joints tend to have smaller ratios of tip width/pitch than the non-structural joints. Tip gap is relevant to the use of the joint. For structural joints a gap is desirable to ensure that the faces of the fingers are brought into close contact. For non-structural uses a tip gap is undesirable because the gap will create a hole if it is not completely filled with glue. In Table 2 it can be seen that generally the structural joints have larger tip gaps but there are exceptions. For instance manufacturer G produces a structural joint with no tip gap. Table 2. Dimensions of the joint profiles as measured on the assembled joints. Manufacturer Joint use length L, mm pitch p, mm tip width t, mm tip gap g, mm ratio p/l ratio t/p A N 3.5 1.6 0.3 0.1 0.46 0.19 B N 12 3.9 0.8 0.3 0.33 0.21 C N 9.5 3.6 1.0 0.1 0.38 0.28 D N 4.0 1.6 0.3 0.1 0.4 0.19 E S 11.0 4.0 0.6 1.0 0.37 0.15 F S 16.0 6.0 1.0 0.7 0.38 0.17 G S 9.5 3.8 0.7 0 0.4 0.18 H S 9.0 3.8 0.7 0.8 0.42 0.18 I N 8.0 3.8 1.2 0 0.48 0.32 J N 3.5 1.6 0.5 0.5 0.46 0.31 Data summary. Mean values of the properties measured on the fingerjoint groups tested in tension, in bending and for durability are given in Table 3. Table 3. Mean values of properties measured on fingerjoints. Manufacturer A B C D E F G H I J (a) Tension group Density, kg/m 3 478 472 501 499 508 504 508 490 488 507 MoE, GPa 13.2 13.0 14.5 14.0 15.3 14.7 15.3 13.9 13.6 14.0 Tension strength, MPa 26.46 34.57 31.46 21.52 39.02 31.59 31.30 34.57 28.97 40.78 % wood failure 38 17 15 63 24 39 46 23 34 39 (b) Bending group Density, kg/m 3 505 498 480 496 488 495 472 495 501 488 MoE, GPa 14.6 13.6 12.9 14.3 13.7 14.8 12.5 14.0 14.7 13.4 Bending MoE, GPa 11.4 10.5 9.8 11.2 10.5 11.5 9.9 11.2 11.5 10.5 MoR, MPa 40.6 46.5 41.6 42.9 49.4 49.3 45.0 51.1 44.7 50.1 % wood failure 62 25 27 45 23 40 44 37 35 19 (c) Durability group Density, kg/m 3 489 503 502 494 490 490 472 487 472 505 MoE, GPa 13.6 14.7 14.2 14.1 13.3 13.8 12.2 13.3 12.0 14.5 % glueline failure 5.73 0.20 0.00 1.78 0.00 0.10 0.75 0.13 0.65 0.07

Tension tests (a) Effect of finger length. Figure 2 shows the weak trend found between tension strength and finger length. The relationship is highly non-significant with an R 2 value of 0.014. However, the joints from manufacturers A and D were not properly cut, as is discussed later. If these joints are deleted from the comparison in Figure 3 is obtained, with an R 2 value of 0.029. The correlation is still highly non-significant. It is concluded that finger length has no effect on joint strength, at least within the range of profiles covered in this study. 60 50 Joint tension strength, MPa 40 30 20 10 0 2 4 6 8 10 12 14 16 18 Finger length, mm Figure 2. Fingerjoint tension strength vs finger length - all data included 60 50 Joint tension strength, MPa 40 30 20 10 0 2 4 6 8 10 12 14 16 18 Finger length, mm Figure 3. Fingerjoint tension strength vs finger length excluding those from manufacturers A and D (b) Effect of density and MoE. Linear regression of tension strength against MoE is shown in Figure 4 and regression statistics for both MoE and density are given in Table 4. Correlations that are significant at the 1% level are shown bold, those significant at the 5% level are shown in italics. In all but two cases (manufacturers D and I), MoE gives a stronger

correlation than does density (i.e. higher R 2 value). It is concluded from this that increasing wood MoE and/or density gives increasing fingerjoint strength provided the joints are manufactured properly. 60 A 50 B C MTS, MPa 40 30 20 D E F G H 10 I J 0 5 10 15 20 25 MoE, GPa Figure 4. Effect of MoE on fingerjoint tension strength. Manufact urer Table 4. Regression statistics for tension strength vs density and MoE Density (kg/m 3 ) MoE (GPa) R 2 coefficient constant R 2 coefficient constant A 0.005 0.009 21.9 0.080 0.58 18.8 B 0.741 0.191-55.7 0.838 2.70-0.57 C 0.637 0.147-42.1 0.677 2.13 0.73 D 0.076-0.033 38.0 0.014-0.20 24.36 E 0.401 0.171-48.6 0.459 0.83 2.49 F 0.518 0.146-41.8 0.528 2.13 0.54 G 0.390 0.078-8.2 0.587 1.32 11.19 H 0.438 0.136-32.2 0.527 2.18 4.11 I 0.680 0.145-42.0 0.672 2.00 1.91 J 0.537 0.148-34.0 0.659 2.33 8.18 (c) Failure mode. Figure 5 shows the broken ends of the joints. They have been arranged in order of increasing stiffness from left to right, and in order of decreasing average tension strength from top to bottom. From top to bottom (strongest to weakest) the manufacturers of the joints were: J, E, B, H, C, G, F, I, A, and D. It can be seen in Figure 5 that there was a strong tendency for the low stiffness/low density timber to break away from the joint. In other words the joint was stronger than the timber being joined and 100% wood failure was obtained. In the stiffer/denser timber on the right of Figure 5 it can be seen by the darker colour that failure generally took place in the glueline of the joint. Thus the timber was stronger than the joint and wood failure percentages less that 100% were obtained. In assessing glued joints it is usual to seek high percentages of wood failure and consider low wood failure percentage as indicative of a glue strength problem. In this study low percentage wood failure is associated with the stronger joints. It is the author s contention that low joint strength in combination with low percentage wood failure indicates poor glueline quality. Thus wood failure percentage cannot be

taken as a criterion of glueline quality on its own, but must be considered in conjunction with the measured strength of the joint. (d) Effect of tip gap. Figure 6 shows the average tensile strength of the batches of joints plotted against tip gap where this has been expressed as a ratio of tip gap to finger length. Those joints intended for structural purposes are shown as + symbols. It is seen that a tip gap improves joint strength. Bending tests (a) Effect of finger length, density, MoE, and failure mode. Similar relationships to those for tension strength were observed for bending strength but with generally poorer correlations. This is a qualitative indication that tension is a more searching test than is bending. (b) Comparison with JAS 701. JAS 701, (JETRO 1992), requires that the bending strength of the joints tested shall equal or exceed the values given in Table 5. In other Japanese standards, (e.g. JAS 600), radiata pine is considered to belong to the Spruce-Pine-Fir species group. The observed 5 percentile and minimum values are given in Table 6. All but one of the joints, (manufacturer F), exceeded the requirements for the highest category (D Fir- L) and all easily met the requirements for the Spruce-Pine- Fir group. Figure 5. Fractured ends of the tension specimens 45 40 E, 11 J, 3.5 Average tensile strength, MPa 35 30 G, 9.5 I, 8 C, 9.5 B, 12 A, 3.5 F, 16 H, 9 25 D, 4 20 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 Ratio of tip gap to finger length Figure 6. Effect of tip gap on fingerjoint tension strength.

Table 5. Minimum acceptable strength values from bending tests. Code for Code for species Bending strength (kgf/cm 2 ) Bending strength (MPa) species type group lower 5% ile minimum lower 5% ile minimum SI DFir-L 236 210 23.2 20.6 Hem-Tam 202 180 19.8 17.7 Hem-fir 202 180 19.8 17.7 SII Spruce-Pine-Fir 169 150 16.6 14.7 W Cedar 169 150 16.6 14.7 Table 6. Observed 5 percentile and minimum MoR values. Manufacturer A B C D E F G H I J finger length, mm 3.5 12 9.5 4 11 16 9.5 9 8 3.5 lower 5%ile, MPa 28.5 27.3 28.1 28.6 33.5 25.5 33.8 29.6 26.5 35.7 minimum, MPa 27.2 21.5 26.9 28.1 27.8 19.7 26.8 29.2 21.9 35.1 Durability tests. JAS 701 has a maximum acceptable average rate of delamination of 5% after one soak/dry cycle. The average observed rates are shown in Table 7, in descending order of severity. The joints from manufacturer A failed the test, while those from manufacturer D had noticeable delamination. Both of these joints were the short 4 mm type and were the two weakest of all the joints. It is possibly significant that these two joints also gave very low correlations between tension strength and MoE as seen in Table 4. The connection between the high delamination and low MoE-tension correlation in these two joints can be explained by the hypothesis that glueline strength is independent of wood strength (and stiffness or density). Thus, if glueline strength is high, wood properties will be critical and a high correlation with stiffness and/or density will be obtained. On the other hand if glueline strength governs then a poor correlation with stiffness or density will be obtained. Table 7. Average rates of delamination in fingerjoints. Manufacturer A D G I B F H J C E finger length, mm 3.5 4 9.5 8 12 16 9 3.5 9.5 11 Delamination % 5.7 1.8 0.7 0.6 0.2 0.1 0.1 0.1 0.0 0.0 Closer examination of the broken joints from manufacturers A and D revealed that these joints had been cut unevenly. This caused some fingers to fit tightly while others were loose. Figure 7 shows a close-up of the lower right corner of the photo in Figure 5. The lower, face-to-face cut joints from manufacturer A show light coloured horizontal lines where glue has been forced out of the joints. The lower ten fingers showed darker and partly shiny gluelines. This indicates that there was inadequate contact between the fingers over this portion of the joint. The edge-to-edge cut joints from manufacturer D show alternate dark and light coloured fingers. This indicates that one set of cutters in the bank of four cutters was misaligned, causing every second finger to be too thin. The delamination occurred along every fourth glueline, corresponding to inadequate glueline pressure on every second finger.

Figure 7. Close-up of the fractured joints from manufacturers A (upper) and D (lower). CONCLUSIONS This study shows that: The majority of fingerjoint manufacturers in New Zealand produce fingerjoint profiles with a length of 10 mm or less, including those intended for structural applications. Finger length has no effect on fingerjoint strength in radiata pine. Significant relationships between joint strength and both density and MoE were obtained for joints without manufacturing defects but no significant relationships were found for joints showing manufacturing defects. Joints in low density timber tended to fail in the timber rather than in the joint while those in high density timber showed a tendency to fail in the glueline rather than at the roots of the fingers. Tension is a better test than bending for joint examination in that tension strength shows a stronger relationship with wood variables than does bending strength and more of the joint surface is revealed in the fracture. REFERENCES JETRO, 1992; Finger jointed structural lumber for wood frame construction. Ministry of Agriculture, Forestry and Fisheries Notification No. 701, JETRO publication SIS-22, Japan. Milner, H R, Song, T, Yeoh, E H, 1989; Fingerjoint design and manufacture. Proc. of the Second Pacific Timber Engineering Conference, Auckland University, Vol 2 pp 159-164