research report

Strength of Screw Connections Subject to Shear Force

RESEARCH REPORT RP04-2 NOVEMBER 2004 REVISION 2006

Committee on Specifications for the Design of Cold-Formed Steel Structural Members

American Iron and Steel Institute

The material contained herein has been developed by researchers based on their research findings. The material has also been reviewed by the American Iron and Steel Institute Committee on Specifications for the Design of Cold-Formed Steel Structural Members. The Committee acknowledges and is grateful for the contributions of such researchers. The material herein is for general information only. The information in it should not be used without first securing competent advice with respect to its suitability for any given application. The publication of the information is not intended as a representation or warranty on the part of the American Iron and Steel Institute, or of any other person named herein, that the information is suitable for any general or particular use or of freedom from infringement of any patent or patents. Anyone making use of the information assumes all liability arising from such use.

Copyright 2004 American Iron and Steel Institute Revised Edition Copyright 2006 American Iron and Steel Institute

Civil Engineering Study 04-1 Cold-Formed Steel Series

Final Report

STRENGTH OF SCREW CONNECTIONS SUBJECT TO SHEAR FORCE

Michael Rotimi Babalola Research Assistant Roger A. LaBoube Project Director

A Research Project Sponsored by the American Iron and Steel Institute

November, 2004

Department of Civil Engineering Wei-Wen Yu Center for Cold-Formed Steel Structures University of Missouri-Rolla Rolla, Missouri

PREFACE

The report summarizes a study related to the design of cold-formed steel screw connections. The study included a review of available literature and a compilation of the available test data pertaining to the strength of a screw connection subject to a shear force. Currently, there are equations for predicting the nominal shear strength of a screw connection given in the American Iron and Steel Institutes North American Specification for the Design of Cold-Formed Steel Structural Members. In an effort to increase the scope of application of the nominal shear strength equations, studies by Rogers and Hancock at the Department of Civil Engineering at the University of Sydney in Australia led to the development of a varied form of screw strength equations. A significant aspect of the University of Sydney study was the testing of screw connections using low ductility steels. The research reported herein analyzed screw connection test data from six different research programs. The test data was evaluated by comparison to both the AISI equations and the equations developed at the University of Sydney. The intent of this research was to determine the applicability and accuracy of the equations. Both normal and low ductility screw connections were included in the analysis. For applications in which t2/t1> 1.0, normal ductility steel, and connections with less than seven screws, the Rogers and Hancock equation provides a slightly more accurate prediction of the connection strength. For more than seven screws in a connection, the Rogers and Hancock equation was found to over estimate the tested connection capacity and thus a reduction factor of

0.85 has been proposed in order to provide satisfactory prediction of the connection strength. This report is based on a thesis submitted to the Faculty of the Graduate School of the University of Missouri-Rolla in partial fulfillment of the requirements for the degree of Masters of Science in Civil Engineering. Technical guidance for this study was provided by the American Iron and Steel Institutes Subcommittee on Connections (A. Harrold, Chairperson). The Subcommittees guidance is gratefully acknowledged. Thanks are also extended to H.H. Chen, AISI staff for her assistance.

1 1. INTRODUCTION

1.1

GENERAL Screws can provide a rapid and effective means to fasten steel metal siding and

roofing to framing members. Screws can also be used for connections in steel framing systems and roof trusses. Tapping screws are externally threaded fasteners with the ability to tap their own internal mating threads when driven into metallic materials. Cold-formed steel construction utilizes several types of tapping screws. The self-drilling screws are externally threaded fasteners with the ability to drill their own hole and form, or tap, their own internal threads without deforming their own thread. These screws are high-strength, one-piece installation fasteners. Self-piercing screws are high-strength, one-piece oneside installation fasteners with sharp point angles of 20 to 26 degrees and are used to attach rigid materials to 33mils (one thickness) or thinner. The self-piercing screws are externally threaded fasteners with the ability to self-pierce metallic material, form a sleeve by extruding metallic material and tap their own mating threads when driven. When choosing the proper fastener for cold-formed steel construction two fundamental questions must be answered: What materials are being joined? and what is the total thickness of the material in the connection? When the application has been defined, it is then possible to choose fasteners with the appropriate point design, body diameter, length, head style, drive, thread type and plating. Point types include self-piercing or self-drilling. Several types of tapping screws are available, including thread cutting, thread rolling and thread forming which all require

2 a pre-drilled hole. The body diameter is specified by the nominal screw size. The length of the fastener is measured from the bearing surface of the fastener to the end of the point. The length of self-drilling screws may require special consideration since some designs have an unthreaded pilot section or reamer wings between the threads and the drill point. Common head styles include flat, oval, wafer, truss, modified truss, hex washer head, pan, round washer and pancake.

1.2

STANDARD TEST There exists a standard for testing screw strength. The American Iron and Steel

Institutes Cold-Formed Steel Design Manual (AISI, 2002) gives test methods for determining the strength of a screw connection. In the AISI TS-4-02 and AISI TS-5-02, the standard test methods for determining the tensile and shear strength of screws connections and mechanically fastened cold-formed steel connections are defined.

1.3

CONNECTION STRENGTH Screw connection strength equations in the current American Iron and Steel

Institutes Cold-Formed Steel Design Manual (AISI 2001) are based on worldwide tests. Screw connection tests used to formulate the provisions included single fastener specimens as well as multiple fastener specimens. However, it is recommended that at least two screws should be used to connect individual elements.

4 2. REVIEW OF LITERATURE

2.1 GENERAL The following summarizes literature considered important for this study. 2.2 STANDARD TEST The following are the various sources that outline testing methods in use for determining the mechanical properties of screws and screw connections. 2.2.1. Society of Automotive Engineers J78 (SAE REV 1998). SAE J78 Self Drilling Tapping Screws (SAE, REV 1998) addresses mechanical requirements for selfdrilling screws, as well as dimensional, material, process, performance, selection and installation. The tests in SAE J78 specification focus on torsional strength, rather than the tensile or shear strengths of the screws. 2.2.2. American Society for Testing and Materials C1513-01. The standard specification for steel tapping screws for cold-formed steel framing connections covers steel self-drilling and self- piercing tapping. This standard also covers test methods for determining performance (hardness, ductility, torsional strength, drill drive, self-drilling tapping screw drill capacity) requirements and physical properties. The test standard does not cover tensile or shear strength. 2.2.3. American Iron and Steel Institute. The American Iron and Steel Institutes document, Test Method for Mechanically Fastened Cold-Formed Steel Connections (AISI, 1996b) outlines a lap-joint shear test. The shear test involves lapping two sheets together and connecting them with a self-drilling screw. The assembly is put

5 into a tension testing machine and a uniaxial tension force is applied. Tension tests are also specified for determining pull-over and pull-out of a screw. 2.2.4. Manufacturers Test Methods. The test procedure, results, and installation information was provided by several manufacturers. The previously mentioned documents SAE J78 (SAE, 1979), ASTM C1513-01 or the AISI Test Methods for Mechanically Fastened Cold-Formed Steel Connections were often cited as references by manufacturers. ITW Buildex,s standard is titled, Work Instruction QWI 10.6- Lab Instructions for Mechanical Properties Testing of Buildex Fasteners (ITW Buildex, 1995). Buildex specifies its fixtures and testing rate. The tests consist of pull-out, pull-over, torsion, tension and shear. Another manufacturers standard considered in this project is by Vicwest (Sommerstein, 1996). This test standard includes a fixture for testing pull-over, pull-out and shear strength of screw connections.

2.3 CONNECTION STRENGTH The references listed below present information on available data regarding the shear strength of a screw connection. The nominal strength of the screw Pns shall be determined by test according to section F1 (a) of 2001 edition of the AISI Specification. 2.3.1. Buildex Division Illinois (1979). The Buildex Division-Illinois Tool Works, inc. carried out a total of 141 tests on some of the more common types and sizes of screws and sheet materials.

6 In the shear test series, screw-fastened connections between two steel sheets in a single lap configuration were evaluated. The connections were subjected to forces parallel to the plane of interconnection. The ultimate shear value load of single lap connection was noted. Seven different types of screws were tested: Teks 1 to 5, mini-point (M-P) and Teks2-MBHT (Teks2-M) screws. The steel sheets had thicknesses ranging between Gauges 26 (0.018in) and 1/8 in, and had Fu/Fy ratios consistent with normal ductile steels. Shown in Appendix A are the sheet properties and the types of screws used. 2.3.2. Eastman (1976). DOFASCO in Hamilton, Canada sponsored a total of 160 screw connection tests to determine the ultimate shear load for the connection. Various types of screws ranging in sizes between No.8 and No.14 were used in the test program. The types of screws tested were screw Types A and AB, Teks 2F, Teks 1- Stitch and Teks 2-MBHT. The thickness of the steel sheets ranged from Gauge 24 (0.0239 in.) to Gauge 18 (0.0485 in.) and had Fu/Fy ratios appropriate for normal ductile steels. 2.3.3. Sokol (1999). Sokol, s work is summarized in Civil Engineering Study 98-3 (Cold-Formed Steel Series of UMR), titled, Determination of the tensile and shear strengths of screws and the effect of screw patterns on Cold-Formed steel connections. Sokol, s research established a standard test method for determining the screw strength. The study involved defining a test procedure and validating the test method concepts for practicality and reliability. The connection strength was also studied and involved the testing of 200 single lap connections of normal ductility steel sheets. Three sheet thicknesses (0.053 in., 0.040 in.,

7 0.030 in) were considered. Three self-drilling screw sizes, No. 8, No. 10 and No. 12 with the spacing of 2d and 3d (d is the diameter of the screw threads) were studied. 2.3.4. Daudet (1996). Daudet, s work is summarized in his Masters Thesis, titled, Self-Drilling Screw connections in Low Ductility Light Gage Steel. Daudet investigated double-lap and single-lap shear connections that used self-drilling screws. The steel used in the study included both normal and low ductility sheets with thickness of 0.029 in., 0.037 in., 0.040in., 0.043 in., 0.050 in. and 0.054 in. The studies include both single- screw and two- screw connections with screw sizes of No. 10, No. 12 and 0.25 in. screws. 2.3.5. Vicwest (1998). The fasteners considered by Vicwest included self-tapping and self-drilling screws with sizes between nominal 0.168 in. outside thread diameter (No. 8) and 0.348 in. The connection failures covered include fastener pull-out from base material, pull-over of fastened material over head of fastener, and shear failure. The connection tested in a shear test may fail in four possible ways including: bearing failure of material, material tearing due to tension failure of net section, shearing of the fastener and tilting of fastener. 2.3.6. Rogers and Hancock (1997). Rogers and Hancock carried out 88 different tests using six different types of screws. The types of sheets used were 042/042-G550, 060/060-G550, 042/060-G550, 0042/100-G550, 055/055-G300 and 055/080-G300. The screw diameters ranged from 0.165 in. to 0.252 in. and the sheet thickness was between 0.0161 in. to 0.0390 in. G550 steel sheet is a low ductility material where as G300 is normal ductility steel. Single-lap connections were investigated for the different

8 thicknesses of steel sheet stated above with two or four screw patterns. The failure modes investigated included bearing, tilting and bearing/tilting. Rogers and Hancock developed the following connection strength equations: For t2/t11.0 Pns = 4.2*(t23*d) 1/2*Fu2 Pns = C*t1*d*Fu1 Pns = C*t2*d*Fu2 For t2/t1 <2.5 Pns = C*t1*d*Fu1 Pns = C*t2*d*Fu2 Where C is d/t d/t< 6 6<d/t<13 d/t>13 Where: d = nominal screw diameter. t1 = thickness of member in contact with the screw head. t 2 = thickness of member not in contact with the screw head. Fu1 = tensile strength of member in contact with screw head. Fu2 = tensile strength of member not in contact with screw head. Pns = nominal shear strength per screw. t = the thickness of the smaller member. C 2.7 3.3-0.1d/t 2.0 Eq. 2.3.6-4 Eq. 2.3.6-5 Eq. 2.3.6-1 Eq. 2.3.6-2 Eq. 2.3.6-3

9 C = Varying coefficient determine by the value of d/t.

2.3.7. American Iron and Steel Institute (AISI 2001). Based on a study by Pekoz (1990), in the AISI Specification section E4.3 (2001) there are five equations to determine the nominal shear strength per screw, Pns: For t2/t1< 1.0 the smallest of the three equations controls. Pns = 4.2*(t23*d) 1/2*Fu2 Pns = 2.7*t1*d*Fu1 Pns = 2.7*t2*d*Fu2 Eq. 2.3.7-1 Eq. 2.3.7-2 Eq. 2.3.7-3

For t2/t1 > 2.5, Pns shall be taken as the smaller of the two equations controls. Pns = 2.7*t1*d*Fu1 Pns = 2.7*t2*d*Fu2 Eq. 2.3.7-4 Eq. 2.3.7-5

For 1.0<t2/t1 <2.5, Pns shall be determined by linear interpolation between the above two cases. Where: d = nominal screw diameter. t1 = thickness of member in contact with the screw head. t 2 = thickness of member not in contact with the screw head. Fu1 = tensile strength of member in contact with screw head. Fu2 = tensile strength of member not in contact with screw head. Pns = nominal shear strength (resistance) per screw. t = the thickness of the smaller member.

10 3. DATA ANALYSIS

3.1 INTRODUCTION Tests were compiled from a variety of sources for the shear strength of screw single-lap connections. These test data were used to compare with AISI equations, Eq.2.3.7-1 to Eq.2.3.7-5 and the equation from Rogers and Hancock, Equations 2.3.6-1 to 2.3.6-5.

3.2

BUILDEX DATA COMPARED WITH AISI METHOD The Buildex Division-Illinois Tools Works, inc. carried out a total of 141 tests on

the more common types and sizes of screws and sheets materials. In comparing test data with the AISI equations, the governing principal parameter is t2/t1; the nominal shear strength per screw is the smallest of the five computed values (Eqs 2.3.7-1 to 2.3.7-5). The ratio of failure shear strength for test to computed results (Pt/Pns) was recorded for the 141 samples tested. The Mean (1.054), Standard Deviation (0.240) and Coefficient of Variation (0.228) for (Pt/Pns) of the 141 samples were also computed.

3.3 BUILDEX DATA COMPARED WITH ROGERS AND HANCOCK METHOD In comparing analysis between Buildex test results and that of the Rogers and Hancock method (Eqs 2.3.6-1 to 2.3.6-5), the governing principals parameters are t2/t1, d/t and the varying coefficient (C); the nominal shear strength per screw was the smallest of the computed values. The ratio of failure shear strength for test to computed results

11 (Pt/Pns) was recorded for the 141samples tested. The Mean (1.109), Standard Deviation (0.232) and Coefficient of Variation (0.209) for (Pt/Pns) of the 141 samples were also computed.

3.4

DOFASCO DATA COMPARED WITH AISI METHOD DOFASCO carried out a total of 160 tests. For comparison analysis with AISI

equations, the governing principal parameter is t2/t1; the nominal shear strength per screw is the smallest of the five computed values. The ratio of failure shear strength for the test to computed results (Pt/Pns) was recorded for the 160 samples tested. The Mean (0.984), Standard Deviation (0.182) and Coefficient of Variation (0.185) for ratio Pt/Pns of the 160 samples were also computed.

3.5

DOFASCO DATA COMPARED WITH ROGERS AND HANCOCK METHOD For the comparison analysis between the DOFASCO test results and the Rogers

and Hancock method, the governing principal parameters are t2/t1, d/t and the varying coefficient (C), the nominal shear strength per screw was the smallest of the computed values. The ratio of failure shear strength for test to computed results (Pt/Pns) was recorded for the 160 samples tested. The Mean (0.996), Standard Deviation (0.181) and Coefficient of Variation (0.182) for (Pt/Pns) of the 160 samples were also computed.

12 3.6 TEST DATA FROM ROGERS AND HANCOCK COMPARED WITH AISI METHOD Rogers and Hancock developed their data from 150 different tests using six different types of screws. For the comparison analysis between Rogers and Hancock and AISI equation, the governing principal parameter is t2/t1, the nominal shear strength per screw was the smallest of the computed values. The ratio of failure shear strength for test to computed results (Pt/Pns) was recorded for the 180 samples tested. The Mean (0.997), Standard Deviation (0.204) and Coefficient of Variation (0.206) for (Pt/Pns) of the 150 samples were also computed and recorded.

3.7 TEST DATA FROM ROGERS AND HANCOCK COMPARED WITH ROGERS AND HANCOCK METHOD In comparing the analysis between University of Sydney test and the equations developed by Rogers and Hancock, the governing principal parameters are t2/t1, d/t and the varying coefficient (C); the nominal shear strength per screw was the smallest of the computed values. The ratio of failure shear strength for test to computed results (Pt/Pns) was recorded for the 88 samples tested. The Mean (1.019), Standard Deviation (0.194) and Coefficient of Variation (0.190) for (Pt/Pns) of the 150 samples were computed.

3.8 VICWEST DATA COMPARED WITH AISI METHOD Vicwest carried out tests using two types of screws, the self-tapping and selfdrilling. A total of 520 tests were carried out on self- tapping screw and 680 tests on self drilling screw.

13 In the comparing analysis with AISI equations, the governing principal parameter is t2/t1; the nominal shear strength per screw was the smallest of the five computed values. The ratio of failure shear strength for test to computed results (Pt/Pns) was recorded for the 520 self- tapping samples and 680 self-drilling samples. The Mean for Self-tapping screw (1.111) and for Self-drilling screw (1.035), Standard Deviation for Self-tapping screw (0.261) and for Self-drilling screw (0.199) and Coefficient of Variation for Self-tapping screw (0.235) and Self-drilling (0.193) (Pt/Pns) of the total 1250 samples were computed.

3.9 VICWEST DATA COMPARED WITH ROGERS AND HANCOCK METHOD Vicwest carried out tests using two types of screws, the self tapping and self drilling. A total of 520 tests were carried out on self tapping screw and 680 tests on self drilling screw. In the comparing analysis, the governing principal parameters are t2/t1, d/t and the varying coefficient (C); the nominal shear strength per screw was the smallest of the computed values. The ratio of failure shear strength for test to computed results Pt/Pns) was recorded for the 1250 samples tested. The Mean, Standard Deviation and Coefficient of Variation for (Pt/Pns) of the 1250 samples were computed as 1.035, 0.199, and 0.199.

3.10 SOKOL DATA COMPARED WITH AISI METHOD

14 Sokol carried out tests on self-drilling screws with spacing of two times and three times the screw diameter under different patterns of screw arrangements. In this comparison, the connections with three times the screw diameter are used. For the comparison analysis between the Sokol test results and the AISI Method, the governing principal parameter is t2/t1; the nominal shear strength per screw was the smallest of the five computed values. The ratios of failure shear strength for the test to computed results (Pt/Pns) were recorded for the samples. The Mean (0.855), Standard Deviation (0.126) and Coefficient of Variation (0.147) for the ratio Pt/Pns of all the samples were computed.

3.11 SOKOL DATA COMPARED WITH ROGERS AND HANCOCK METHOD For the comparison analysis between Sokols test results and the Rogers and Hancock Method, the governing principal parameters are t2/t1, d/t and the varying coefficient (C), the nominal shear strength per screw was the smallest of the computed values. The ratio of the failure shear strength for test to computed results (Pt/Pns) was recorded for all the samples. The Mean (0.854), Standard Deviation (0.126) and Coefficient of Variation (0.147) for (Pt/Pns) were recorded for all the samples.

3.12 DAUDET DATA COMPARED WITH AISI METHOD Daudets test results used in this study were from tests using low ductility steel. A total of 111 tests was performed.

15 For the comparison analysis between Daudets test results and the AISI Method, the governing parameter is t2/t1; the nominal shear strength per screw was the smallest of the five computed values. The ratio of failure shear strength for the test to computed results (Pt/Pns) was recorded for the 111 samples tested. The Mean (0.866), Standard Deviation (0.168) and Coefficient of Variation (0.193) for ratio Pt/Pns of the 114 samples were also computed.

3.13 DAUDET DATA COMPARED WITH ROGERS AND HANCOCK METHOD For the comparison analysis between Daudets test results and the Rogers and Hancock Method, the governing principal parameters are t2/t1, d/t and the varying coefficient (C); the nominal shear strength per screw was the smallest of the computed values. The ratio of failure shear strength for test to computed results (Pt/Pns) was recorded for the 114 samples tested. The Mean (0.866), Standard Deviation (0.168) and Coefficient of Variation (0.193) for (Pt/Pns) of the 114 samples were also computed.

16 4. EVALUATION OF DATA

4.1 INTRODUCTION Tests were compiled from a variety of sources for the shear strength of single-lap screw connections. Between the six different sets of data, there were a total of 1890 test data points considered in the analysis. To analyze each of the different equations, a spreadsheet was developed to evaluate the nominal shear strength. This value was then compared to the tested value of shear strength, forming a ratio of Ptest/Pns. The mean, standard deviation and coefficient of variation were determined for each set of data.

4.2

BUILDEX RESULTS WITH AISI AND ROGERS AND HANCOCK RESULTS In addition to all the test data from Buildex being evaluated together, the data was

divided into sub-groups according to screw-sizes and also evaluated. Table 4.1 summarizes the statistical data, showing the number of tests, the mean, standard deviation and coefficient of variation for the Buildex test data for each screw sizes. The statistical parameters in Table 4.1 show a smaller coefficient of variation and a higher ratio of Ptest/Pns when the Rogers and Hancock equations are used. This indicates that for data from Buildex the Rogers and Hancock equations are in fact more accurate at predicting the shear strength of the screw connection.

17 TABLE 4.1- Buildex test data comparison AISI METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.186in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.212in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.251in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.137in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.164in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.246in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns ROGERS & HANCOCK METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.186in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.212in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.251in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.137in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.164in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.246in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns

141 1.054 0.240 0.228 30 1.070 0.298 0.278 55 1.035 0.192 0.185 6 0.765 0.092 0.120 5 1.453 0.189 0.130 20 1.142 0.222 0.194 19 1.008 0.177 0.175

141 1.109 0.232 0.209 30 1.109 0.300 0.278 55 1.095 0.197 0.180 6 0.930 0.079 0.085 5 1.486 0.151 0.102 20 1.174 0.211 0.180 19 1.048 0.179 0.170

18 TABLE 4.1- Buildex test data comparison (cont.) 0.243in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.243in (Screw diameter) No Mean Standard Deviation Coefficient of Variation

6 0.953 0.204 0.214

6 1.132 0.156 0.138

4.3 DOFASCO RESULTS WITH AISI AND ROGERS AND HANCOCK RESULTS The data from DOFASCO indicated both self-drilling and self-tapping screws were used in the test program. The screw data was analyzed by dividing the data into subgroups according to screw-types and sizes also evaluated. Table 4.2 summarizes the statistical data, showing the number of tests, the mean, standard deviation and coefficient of variation for the DOFASCO test data for each screw sizes. The statistical parameters in Tables 4.2, 4.3 and 4.4 show a smaller coefficient of variation and a higher ratio of Ptest/Pns when the Rogers and Hancock equations are used. This indicates that for the data from DOFASCO, the University of Sydney is in fact more accurate at predicting the shear strength of the screw connection. DOFASCO data indicated that the screw type, self-tapping or self-drilling, had little influence on the strength of the connection.

19 TABLE 4.2- Dofasco test data comparison AISI METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation SCREW A No Mean Standard Deviation Coefficient of Variation SCREW AB No Mean Standard Deviation Coefficient of Variation TEKS/2F No Mean Standard Deviation Coefficient of Variation TEKS/1 STITCH No Mean Standard Deviation Coefficient of Variation TEKS/2 MBHT No Mean Standard Deviation Coefficient of Variation Pt/Pns ROGERS & HANCOCK METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation SCREW A No Mean Standard Deviation Coefficient of Variation SCREW AB No Mean Standard Deviation Coefficient of Variation TEKS/2F No Mean Standard Deviation Coefficient of Variation TEKS/1 STITCH No Mean Standard Deviation Coefficient of Variation TEKS/2 MBHT No Mean Standard Deviation Coefficient of Variation Pt/Pns

160 0.984 0.182 0.185 48 0.989 0.180 0.182 56 0.956 0.178 0.186 16 1.032 0.220 0.213 32 1.000 0.181 0.181 8 0.980 0.153 0.156

160 0.996 0.181 0.182 48 1.001 0.180 0.180 56 0.970 0.180 0.186 16 1.035 0.221 0.214 32 1.014 0.171 0.168 8 0.994 0.159 0.160

20 TABLE 4.3- Dofasco test data comparison (self-drilling screws) AISI METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.164in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.186in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.212in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.243in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns ROGERS & HANCOCK METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.164in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.186in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.212in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.243in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns

56 1.006 0.187 0.186 16 1.032 0.220 0.213 16 1.003 0.189 0.188 8 0.980 0.153 0.156 16 0.996 0.178 0.179

56 1.017 0.182 0.179 16 1.035 0.221 0.214 16 1.010 0.189 0.187 8 0.994 0.159 0.160 16 1.018 0.157 0.154

TABLE 4.4- Dofasco test data comparison (self-tapping screws) AISI METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.164in (Screw diameter) No Mean Standard Deviation Pt/Pns ROGERS & HANCOCK METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.164in (Screw diameter) No Mean Standard Deviation Pt/Pns

104 0.972 0.179 0.184 32 1.032 0.198

104 0.985 0.180 0.183 32 1.034 0.198

21 TABLE 4.4- Dofasco test data comparison (self-tapping screws) (cont.) Coefficient of Variation 0.186in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.243in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.192 Coefficient of Variation 0.186in (Screw diameter) 32 No 0.990 Mean 0.180 Standard Deviation 0.182 Coefficient of Variation 0.243in (Screw diameter) 40 No 0.909 Mean 0.146 Standard Deviation 0.156 Coefficient of Variation 0.191 32 0.996 0.178 0.179 40 0.935 0.157 0.167

4.4 ROGERS AND HANCOCK RESULTS WITH AISI AND ROGERS AND HANCOCK RESULTS Rogers and Hancock used low ductility steels in carrying out many of their test on the shear strength of a screw connection. Besides all the test data from University of Sydney being evaluated together, the data was divided into sub-groups according to screw-sizes and also evaluated. Table 4.5 summarizes all of this data, showing the number of tests, the mean, standard deviation and coefficient of variation for the University of Sydney test data for each screw sizes. The statistical analysis shows a higher mean value for the Rogers and Hancock method compared with AISI method but again has a lower standard deviation and coefficient of variation.

22 TABLE 4.5- Rogers and Hancock test data comparison AISI METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.165in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.192in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.214in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.252in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns ROGERS & HANCOCK METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.165in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.192in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.214in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.252in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns

150 0.997 0.204 0.206 41 1.003 0.203 0.203 67 1.068 0.222 0.208 34 0.869 0.097 0.112 8 0.926 0.093 0.101

150 1.019 0.194 0.190 41 1.021 0.203 0.199 67 1.092 0.200 0.183 34 0.896 0.097 0.108 8 0.926 0.093 0.101

4.5 VICWEST RESULTS WITH AISI AND ROGERS AND HANCOCK RESULTS The data from Vicwest indicated both self-drilling and self-tapping screws were used in the test program. The screw connection data was divided into sub-groups according to screw-types and sizes and also was evaluated.

23 Tables 4.6 and 4.7 summarizes all of this data, showing the number of tests, the mean, standard deviation and coefficient of variation for Vicwest test data for each screw sizes. The statistical analysis shows a higher mean value of the Rogers and Hancock method compared with AISI method and again a lower standard deviation and coefficient of variation.

TABLE 4.6- Vicwest test data comparison (self-drilling screws) AISI METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.189in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.215in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.246in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns ROGERS & HANCOCK METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.189in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.215in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.246in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns

680 1.035 0.199 0.193 90 1.036 0.164 0.158 340 1.082 0.227 0.210 250 0.970 0.146 0.151

680 1.043 0.207 0.198 90 1.048 0.172 0.164 340 1.087 0.235 0.216 250 0.982 0.156 0.159

24 TABLE 4.7- Vicwest test data comparison (self-tapping screws) AISI METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.246in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.254in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.290in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns ROGERS & HANCOCK METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.246in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.254in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.290in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns

520 1.111 0.261 0.235 310 1.159 0.244 0.211 150 1.108 0.286 0.258 60 0.870 0.098 0.113

520 1.123 0.252 0.225 310 1.171 0.232 0.198 150 1.128 0.273 0.242 60 0.865 0.101 0.116

4.6 SOKOL RESULTS WITH AISI AND ROGERS AND HANCOCK RESULTS Normal ductility steel was used in the Civil Engineering Study 98-3 and the screw sizes were No.8 (0.165 in.), No.10 (0.186 in.) and No.12 (0.215 in.) with spacing of 2d and 3d (d is the diameter of the screw threads). In this study, only the 3d spacing test data were evaluated against the AISI and Australian equations. Besides all the test data from Sokol being evaluated together, the data was divided into sub-groups according to screw-sizes and also evaluated. Table 4.8 summarizes all of this data, showing the number of tests, the mean, standard deviation and coefficient of variation for Sokol test data for each screw sizes.

25 The statistical analysis shows less than 1% difference in the mean, standard deviation and coefficient of variation values for all the 128 tests data using the two methods in question. In analyzing the screws by there sizes, it shows a less than 0.1% difference in any of the two methods consider. The statistical parameters listed in Table 4.8 shows no difference in the two different methods.

TABLE 4.8- University of Missouri-Rolla (Sokol) test data comparison AISI METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.165in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.186in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.215in (Screw diameter) No Mean Standard Deviation Coefficient of Variation ROGERS & HANCOCK METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.165in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.186in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.215in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns

128 0.855 0.126 0.147 42 0.833 0.134 0.161 36 0.856 0.139 0.162 50 0.873 0.108 0.124

128 0.855 0.126 0.147 42 0.834 0.134 0.161 36 0.856 0.139 0.162 50 0.873 0.108 0.124

26 4.7 DAUDET RESULTS WITH AISI AND ROGERS AND HANCOCK RESULTS Daudet, s work is summarized in his Masters Thesis, titled, Self-Drilling Screw connections in Low Ductility Light Gage Steel. Daudet investigated double-lap and single-lap shear connections that used self-drilling screws. The steel used in the study included both normal and low ductility sheets with thickness of 0.029 in., 0.037 in., 0.04 in., 0.043 in., 0.050 in. and 0.054 in. The studies include both single- screw and two- screw connections with screw sizes of No. 10, No. 12 and 0.25 in. screws. In this study, the low ductility steel sheet tests data were evaluated with AISI and Australian equations. Besides all the tests data from Daudet being evaluated together, the data was divided into sub-groups according to screw-sizes and also evaluated. Table 4.9 summarizes all of this data, showing the number of tests, the mean, standard deviation and coefficient of variation for the Daudet tests data for each screw sizes. The statistical analysis shows the same Mean, Standard deviation and Coefficient of variation values for all the 111 tests data using the two methods in question. In analyzing the screws by there sizes, it shows also the same values for the two methods in question.

27

TABLE 4.9- University of Pittsburgh (Daudet) test data comparison AISI METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.188in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.190in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.210in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.212in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.240in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.243in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns ROGERS & HANCOCK METHOD ALL DATA No Mean Standard Deviation Coefficient of Variation 0.188in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.190in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.210in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.212in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.240in (Screw diameter) No Mean Standard Deviation Coefficient of Variation 0.243in (Screw diameter) No Mean Standard Deviation Coefficient of Variation Pt/Pns

111 0.866 0.168 0.193 24 0.847 0.169 0.199 18 0.854 0.169 0.198 12 0.848 0.150 0.177 30 0.916 0.172 0.188 12 0.926 0.126 0.136 15 0.780 0.173 0.222

111 0.866 0.168 0.193 24 0.847 0.169 0.199 18 0.854 0.169 0.198 12 0.848 0.150 0.177 30 0.916 0.172 0.188 12 0.926 0.126 0.136 15 0.780 0.173 0.222

28

5. CONCLUSIONS

A total of 1890 test data from six different sources (235 in low ductility and 1655 in normal ductility steels) were analyzed using both the AISI equations and the equations from Rogers and Hancock at the University of Sydney. Based on the data analysis, the following design recommendations were deduced: 1. For connections within two to seven screws in low or normal ductility steels, Rogers and Hancock equations provide a marginally more accurate prediction of the connection strength. 2. For connection with more than seven screws in low and normal ductility steels, Rogers and Hancock equations should be multiplied by a reduction of 0.85. The 0.85 reduction factor is based on tests by Sokol in which it was determined that as the number of screws increased the connection capacity was not proportional to the number of screws in the connection. Although Rogers and Hancock tested connections with four or fewer screws, based on engineering judgment the 0.85 reduction is recommended to be applied to the Rogers and Hancock equations. 3. For single screw connections with normal ductility steels, Rogers and Hancock equations are marginally more accurate prediction of the connection strength. 4. For connections with a single screw in low ductility steel, Rogers and Hancock equations should be multiplied by a reduction of 0.85. 5. The equations are valid for self-drilling and self-tapping screws.

29 BIBLIOGRAPHY

American Iron and Steel Institute (2001), North American Specification for the Design of Cold-Formed Steel Structural Members, Washington, D.C. Daudet, R.L. (1996), Self-Drilling Screw Connections in Low Ductility Light Gage Steel, submitted to the Graduate Faculty at the University of Pittsburgh in partial fulfillment for the degree Master of Science in Civil Engineering. Eastman, R.W. (1976), Report on Screw Fastened Sheet Steel Connections, Two Volumes, Report for Canadian Steel Industries Construction Council, DOFASCO, Hamilton, Canada. Janusz, M., Sledz, M. and Moravek, S. (1979), Teks Fasteners, Pullout and Shear Characteristics In Various Thicknesses of Steels, Second Edition, Buildex Division-Illinois Tools Works, Inc. Pekoz, T. (1990), Design of Cold-Formed Steel Screw Connections, Proceedings of the Tenth International Specialty Conference on Cold-Formed Steel Structures, Department of Civil Engineering, University of Missouri-Rolla Rogers, C.A.and Hancock, G.J.(1997) Screwed Connection tests of Thin G550 and G300 Sheet Steels, Centre for Advanced Structural Engineering, Department of Civil Engineering, University of Sydney, Australia, December. SAE J78 Self Drilling Tapping Screws (SAE, REV 1998) Society of Automotive Engineers, Warrendale, Pennsylvania. Sokol, A.M (1998)., Determination of the Tensile and Shear Strengths of Screws And The Effect of Screw Patterns on Cold-Formed Steel Connections, Civil Engineering Study 98-3, University of Missouri-Rolla. Sommerstein, M. and Mandelzys, B. (1996). Vicwest Fasteners Manual, Vicwest- A Division of Jannock Steel Fabricating Company. Test Method for Mechanically Fastened Cold-Formed Steel Connections (2002), American Iron and Steel Institute, Washington, D.C.

American Iron and Steel Institute

1140 Connecticut Avenue, NW Suite 705 Washington, DC 20036

www.steel.org

Research Report RP-04-2