WELDING RESEARCH

JANUARY 2004-S32

ABSTRACT. Furnace brazing is a joiningprocess used in the aerospace and otherindustries to produce strong permanentand hermetic structural joints. As in anyjoining process, brazed joints have variousimperfections and defects. At the presenttime, our understanding of the influenceof the internal defects on the strength ofthe brazed joints is not adequate. The goalof this three-part investigation is to betterunderstand the properties and failuremechanisms of the brazed joints contain-ing defects. This study focuses on the be-havior of brazed lap shear joints becauseof their importance in manufacturingaerospace structures. In Part 1, an averageshear strength capability and failuremodes of the single lap joints are explored.Stainless steel specimens brazed with puresilver are tested in accordance with theAWS C3.2 standard. Comparison of thefailure loads and the ultimate shearstrength with the finite element analysis(FEA) of the same specimens as a func-tion of the overlap widths shows excellentcorrelation between the experimental andcalculated values for the defect-free lapjoints. A damage zone criterion is shownto work quite well in understanding thefailure of the braze joints. In Part 2, thefindings of Part 1 will be verified on thelarger test specimens. Also, various flawswill be introduced in the test specimens tosimulate incomplete braze coverage in thelap joints. Mechanical testing and FEAwill be performed on these joints to verifythat behavior of the flawed ductile lapjoints is similar to joints with a reducedbraze area. Finally, in Part 3, the resultsobtained in Parts 1 and 2 will be applied tothe actual brazed structure to evaluate theload-carrying capability of a structural lapjoint containing discontinuities. In addi-tion, a simplified engineering procedurewill be offered for the laboratory testing ofthe lap shear specimens.

Introduction

In manufacturing of critical braze-ments, one of the main variables control-ling the strength of the lap shear brazejoints is the width of the braze overlap. Byallowing sufficient overlap width, thestrength of the braze joint (i.e., the load-carrying capability) can be made equal toor even greater than the strength of thebase metal. At the same time it is now wellestablished that the average maximumshear stress of the lap joint at the failureload decreases with the increase of theoverlap width (Ref. 1), as shown in Fig. 1.The presence of the defects, however, cancompromise the integrity of the brazejoints. Modern nondestructive inspectionmethods help to identify and weed out thedefective joints. In some cases, however, astructure or a pressure vessel containingdefective braze joints is too expensive tobe scrapped. Therefore, it is beneficial tounderstand how the defects affect the per-formance of the braze joints so the bestengineering decisions could be made toascertain the acceptability of the brazedstructure. An attempt by the authors tofind any references in open literature thatexamine the load-carrying capability ofbraze joints in the presence of internal dis-continuities was not very successful. A lackof research publications studying the ef-fects of discontinuities on the propertiesof the brazed joints is somewhat puzzling.It is particularly so considering that the is-sues of stress distribution, load-carryingcapability, damage tolerance, and failuremechanisms in structural bonded joints

had received a great deal of attention inthe research community (Refs. 2–4). Forexample, a theoretical analysis of stressesin adhesive lap joints had been presentedas early as 1938 (Ref. 5). Just the fact thatthere are at least 13 ASTM standards de-veloped to perform shear tests on variousadhesive-bonded test specimens indicatesa much higher level of effort devoted tounderstanding the properties of struc-tural-bonded joints.

A number of industrial and govern-ment quality control standards address theacceptable limits for internal discontinu-ities in brazed joints (Refs. 6–12). The ac-ceptance criteria are based either on themaximum aggregate braze area reductiondue to the voids, inclusions, or incompletebraze (up to 15–20% of the total area ofthe joint) or reduction of the leakage bar-rier width. Some documents allow up to15% reduction of the leakage barrier (Ref.7), but other ones permit the width of thelargest void or unbonded region to be upto 60% of the total joint width (Refs.8–10). Military specification MIL-B-007883C provided a more extensive treat-ment of the internal discontinuities, but ithas been cancelled. In addition, the guide-lines provided by this specification for ac-cepting or rejecting internal defects werenot very clear.

The authors could not find any refer-ence in open literature addressing thestrength of the structural braze lap jointscontaining internal discontinuities. Obvi-ously, the criteria based on percentage re-duction of the aggregate braze area do notapply to long joints. One can be well belowthe maximum allowable area reduction,and, therefore, be within the “acceptable”limits but have all the flaws concentratedin a relatively short section of the seamcausing the entire lap joint to fail locally —Fig. 2.

This study focuses on lap shear jointsdue to their importance in the manufac-turing of large brazed pressure vessels.The following questions will be addressed:

1) What is the criterion of failure for alap shear braze joint?

KEY WORDS

AerospaceBrazementBrazingFurnace BrazingLap Joints

Flaw Tolerance in Lap Shear Brazed Joints — Part 1

Investigation brings a better understanding of the failure mechanism in brazed joints containing defects

BY Y. FLOM AND L. WANG

Y. FLOM is with NASA Goddard Space FlightCenter, Greenbelt, Md. L. WANG is withSWALES Aerospace, Beltsville, Md.

WELDING RESEARCH

-S33WELDING JOURNAL

2) Can the effect of internal disconti-nuities on the strength of ductile lap jointsbe viewed simply as a reduction of the ac-tual load-bearing area?

3) How do we determine load-carryingcapacity of braze lap joints containing in-ternal defects?

4) Is there a maximum flaw size thatcan cause a transition from ductile to brit-tle failure in a lap joint?

The objectives of the Part 1 effort wereas follows:

• Perform strength measurements ofthe lap shear brazed joints as a function ofoverlap width per AWS C3.2 (Ref. 12).

• Perform finite element analysis(FEA) of the test specimens.

•Develop failure criterion•Correlate FEA with the experimental

data.

Experimental Procedures

In this study, 347 stainless steel (seeTable 1 for composition) test specimenswere vacuum brazed using 99.9% pure sil-ver filler metal. These materials were se-lected for their important role in the aero-space industry. Using a single-phase fillermetal helps to eliminate possible influ-ence of the eutectic or multiphase micros-tuctures on the results of this investiga-tion.

A single lap shear test specimen (STS)configuration and fabrication was basedon the AWS C3.2 specification, with somedeviations due to material availability andfabrication cost. It is believed, however,that these differences did not compromise

the objectives set forth in the presentwork.

The “half of the dog bone” shapedblanks were electric discharged machinedfrom a 0.090-in. (2.3-mm) -thick stainlesssteel strip and plated with a 0.0005-in.(0.0127-mm) -thick layer of Ni prior tobrazing in order to facilitate wetting of thebase metal. The majority of the shear testspecimens (STS) were Ni plated usingelectroless process. Several of the STSwere electrolytically Ni plated to eliminatethe presence of low-melting Ni-P eutectic.All blanks were assembled for brazing inthe specially designed stainless steel fix-ture — Fig. 3. Silver filler metal in theform of the 0.0010-in. (0.254-mm) foil waspreplaced in the overlap. A pair of blanksform one complete “dog bone” shapedSTS. The top blank was allowed to float insuch a way that it could move down underits own gravity and along the longitudinalaxis of symmetry during brazing whilemaintaining an axial alignment with thebottom blank. Consequently, the brazejoint clearance was not rigidly controlledbut rather was allowed to form on its ownunder the combined action of the blankweight and the capillary forces of themolten filler. This freedom of movementassured that the braze joint clearance and

the alignment of the test specimens werenot affected by the differences in thermalproperties between the fixture and thespecimens. Each STS was brazed one at atime in the vacuum brazing furnace usingthe identical brazing cycle. The maximumbrazing temperature was 1020°C. Theoverlap width was adjusted manually foreach pair of blanks to cover the range0.05T to 5T, which corresponds to 0.045 in.(1.14 mm) and 0.450 in. (11.4 mm), whereT= 0.090 in. (2.3 mm) is the base metalthickness. Since the blanks were notclamped during brazing and the faying in-terfaces were allowed to move relative toeach other, the overlap widths on thebrazed specimen were slightly differentfrom the preset values due to variations inthe coefficients of thermal expansion ofthe specimens and various components of

Fig. 1 — A general representation of the maxi-mum average shear stress in the filler metal aswell as the tensile stress in the base metal as afunction of the overlap width in the braze joint.This graph is based on the hundreds of tests per-formed in the early 1960s (Ref. 1) and many sub-sequent experimental results reported in litera-ture.

Fig. 2 — Illustrates inconsistency in using maximum allowable area reduction approach. Consider twodifferent situations of brazing 10T long and 100T long pressure vessel. Assume, for example, that in bothcases the braze joint has a flaw (shaded area) and its size is 2T · 5T. In one case (A), the ratio of theflaw area to the total braze area is 10/40, i.e., 25%. In case B, the ratio is 10/400 or only 2.5%. If weapply 15% maximum allowable braze area reduction rule then the joint will be rejected in case A andaccepted in case B.

Fig. 3 — Shear test specimen blanks assembledfor brazing.

OVERLAP DISTANCE, RATION OF OVERLAP TO THICKNESS

AV

ER

AG

E U

NIT

ST

RE

SS A

T F

AIL

UR

E

WELDING RESEARCH

JANUARY 2004-S34

the brazing fixture. Fillets were removedfrom each end of the STS with the spe-cially ground end mill bit to maintain thesame fillet radius of 0.016 in. (0.4 mm) onall specimens.

A total of 37 lap shear specimens weretensile tested to failure. A 2-in. (50.8-mm)gauge length extensometer was used tomeasure elongation of each test specimen— Fig. 4. The extensometer arms used dif-ferent length knife edges to compensate

for the 1T offset in the geometry of the sin-gle lap shear specimen. As a result, the in-fluence of the test specimen rotation onthe elongation measurements was mini-mized at least during the initial portion ofthe tensile test. Each test was videotapedto ensure that the deformation behavior ofthe lap shear specimens, dynamics of frac-ture process, and other important featureswere captured.

Finite Element Analysis

Finite element analyses (FEA) wereperformed using COSMOS/M package.Plane strain nonlinear elasto-plastic ele-ments were used throughout the analysis.A typical model with parametric meshingis shown in Fig. 5. The total length of themodel is kept at 2 in. (50.8 mm). This isconsistent with the extensometer gaugelength including the lap width. A perfectbond is assumed between the filler metaland the base metal. The true stress straincurve of the 347 stainless steel was estab-lished by tensile testing of the material.The true stress strain curve of the fillermetal (annealed pure silver) was based ontensile testing of the material and the dataobtained from private communications(Ref. 13). The final form of the true stressstrain curves is shown in Fig. 6. Bothcurves are averages of at least three repet-itive test results. The loading process ofthe FEA model is accomplished by incre-mentally applying a uniform displacementat one end of the specimen while keepingthe other end fixed in the loading direc-tion. The overall tensile load applied tothe specimen at each loading step is the in-tegration of the tensile stress in the load-ing direction of the end elements. In aver-age, about one hundred loading stepswere used to complete a loading process.Instead of nodal, the elementalstresses/strains were used to describe thestress/strain level within the filler metalsince it is relatively thin and experienceslarge plastic deformation.

Experimental Results

Typical cross sections of the347/Ag/347 braze joints are shown in Fig.7. As expected, the microstructure of thefiller metal is a single phase consisting ofpure silver. The interface regions vary,however, depending on whether the elec-trolytic or electroless process was used forNi plating. In the case of electroless Ni,the brazed interfaces contain small inclu-sions of Ni3P compound since the brazingtemperature exceeded the nickel-phos-phorus eutectic point of 880°C.

Typical load vs. elongation records forthe tested specimens are shown on oneplot in Fig. 8. As was mentioned earlier,the actual overlap lengths in the lap shearSTS varied, slightly, from the exact multi-ples of 0.5T. Consequently, it was moreconvenient to group the individual recordsinto the 0.5T intervals, as shown in Fig. 8.All test specimens failed in the filler metal.During the initial part of the tension test,the braze joints rotated to reduce the load-ing axial offset between the top and thebottom ends of the test specimens. Afterthe alignment was achieved, no further ro-tation was observed. The specimens con-tinued to deform by uniform stretching ofthe base metal away from the joint. Whenthe failure load was reached, the brazejoint just slipped apart to cause the ap-plied load to drop. No apparent peeling ofthe joint edges at the moment of failurewas observed.

The maximum load in every test wasdefined as the failure load for that specificoverlap. A plot of the failure load as afunction of the overlap width is shown inFig. 9A. Maximum average or apparentfailure shear strength for each overlapwidth was calculated as the failure load di-vided by the area (overlap width · width ofthe specimen). Despite the fact that thisvalue changes with the overlap width, it isbeneficial to call it shear ultimate strength(SUS) of the lap shear braze joint in orderto be consistent with the engineering ter-minology widely used within the structuraldesign community. Hence, from now on,the term SUS will be used throughout thisthree-part investigation. A plot of the SUSas a function of overlap width is shown inFig. 9B. A close examination of the dataplotted in Fig. 9 shows that the data pointsseem to follow two different paths ofstrength values when the overlaps exceed

Fig. 4 — Extensometer installed on the shearspecimen prior to test shown on the top. Pictureon the bottom shows the gauge length portion ofthe test specimen under the load.

Table 1 — Composition of 347 Stainless Steel (%)

C Mn P S Si Cr Ni Nb/Ta Fe

0.08 2 <0.05 <0.03 1 17–19 9–13 0.62 balance

Fig. 5 — COSMOS/M finite element analysis (FEA) 2-D model used in this investigation showing displacement boundary condition (left) and a typical FEAmesh (right).

WELDING RESEARCH

-S35WELDING JOURNAL

2T. The higher strength path correspondsto electrolytic Ni-plated specimens andthe lower strength path to the electrolessplated ones.

FEA Results

The FEA exhibits satisfactory results insimulation of the loading process. A typi-cal result of the deformed specimen modelis shown in Fig. 10. A step-by-step exami-nation of model deformation process alsorevealed that the lap joint rotates at thebeginning of the loading process, as de-scribed earlier, followed by uniformstretching of the base metal as the loadingaxial offset diminishes. A FEA animationof the loading process vividly reproducedthe tensile test as recorded on the tape.

Discussion

As the mechanical testing demon-strated, for overlaps over 2T, the elec-trolytically plated braze joints are strongerthan the electroless plated ones, as shownin Fig. 9. Consequently, for the purpose ofthis investigation, only high-strength val-ues, obtained on the electrolytically platedtest specimens, will be considered for cor-relation with the FEA results, since FEAdoes not account for the presence of Ni3Pinclusions at the braze interface.

It is well established that plastic defor-mation of the lap joints loaded in tensionis controlled by the shear and peel (i.e.,tensile load acting perpendicular to the in-terface) forces. Indeed, high magnifica-tion images of the fracture surfaces shownin Fig. 11 clearly indicate the existence ofat least two distinct deformation regionswithin the braze joint. Fracture initiationregion located at the lap edge is a combi-nation of shear and peel, whereas the restof the fracture area is dominated by pureshear. No evidence of peel was observedon the fracture surfaces of the specimenswith overlaps < 2T.

While analyzing the deformation of thebrazed joints, it is instructive to discuss ourunderstanding of what constitutes a fail-ure of the flaw-free lap shear braze joint.In this study we are assuming that theyielding of the braze joint does not consti-tute its failure. This assumption is basedon the fact that some brazed structuressuch as rocket engines, for example, aredesigned to perform beyond their yieldpoint. Thus, in the context of our study wedefine the failure as an event of fracturewhen the joint loses its load-carrying ca-pability due to physical separationthrough the filler metal, base metal, orcombination of the two.

Since most of the stress-strain proper-ties of metals and alloys were obtained intensile tests, the von Mises stress and yield

criterion was used in the finite elementanalysis, even though the filler metal in thelap shear joint is almost in the pure shearcondition. Tresca yield criterion may givea better prediction of the yielding of thefiller metal. But due to the fact that thefiller metal experiences a large amount ofplastic deformation before failure, theonset of the yielding is less important inthe current investigation. In addition, thevon Mises stress, that is the effective stressin plasticity, is also the choice of the COS-MOS/M package to handle plasticity. Thevon Mises stress is defined as (Ref. 14)

(1)

where s1, s2, and s3 are principal tensilestresses. In the current investigation, all

ss s s s

s svm = ·

-( ) + -( )+ -( )

1

2

1 22

2 32

1 32

Fig. 6 — True stress-strain curves for 347 stainless steel base metal and pure silver filler metal used inthis work.

Fig. 7 — Metallographic cross section through the braze joint. Image on the left captures the entire jointclearance width and shows basically single-phase silver filler metal. Image on the right is a higher mag-nification of the interface region showing a presence of small Ni3P inclusions (arrows). Also, a Ni layercan be seen separating the stainless steel base metal and silver filler metal.

10µm

4µm

347 s.s.

Nilayer

Base metal

Filler metal

Base metal

Fig. 8 — Typical load vs. elongation plots repre-senting overlaps ranging from 0.5T to 4.5T. Mostlikely, the three different slopes visible for all fam-ilies of curves correspond to elastic, plastic rota-tion, and plastic stretching portions of the sheartest specimen deformation during the pull test.

WELDING RESEARCH

JANUARY 2004-S36

the failures occurred in the filler metal.The elemental von Mises stress in the fillermetal is used as the indication of the stresslevel in the filler metal. The question is: atwhat stress level will the filler metal fail?This is the fundamental question of thefailure criterion of a brazed lap shearjoint, which has no satisfactory answers.

A literature review indicates that vonMises stress was successfully utilized (Ref.15) to demonstrate the decrease of the av-

erage shear stress in the lap joint with theincrease of the overlap width, as shown inFig. 1. Other studies (Ref. 16) consideredthe effects of the braze joint clearance,work hardening, and hydrostatic con-straint on shear and tensile stress distrib-ution, but did not discuss the braze jointfailure criteria.

Our preliminary attempt was to use theultimate tensile strength of the filler metalobtained from the tensile test as the fail-

ure criterion based on the single loadingcurve assumption (von Mises). The as-sumption is that the effective stress-straincurve coincides with the uniaxial tensilestress-strain curve, although the theorycannot predict the failure point. This at-tempt is apparently flawed, since the con-figuration of the filler metal in the brazedjoint is so different from tensile test sam-ples that the filler metal strength obtainedin the two configurations will always bevery different.

It is almost universally accepted thatthe ultimate failure strength of filler metalhas to be tested in situ. The lap shear sam-ple test is the most popular test. The lapshear test, however, determines thestrength of the joint rather than fillermetal strength — a property of the fillermetal. The main reason for that is anonuniform stress distribution in the fillermetal of the lap shear joint. The trend ofthe SUS reduction with the increase of thelap width is not due to the fundamentalproperty deterioration of the filler metal,but due to the joint configuration thatcauses the stress distribution to be less andless uniform.

The current analysis indicates that, forthe short overlaps (< 0.5T), von Misesstress is fairly uniformly distributed withinthe filler metal, even far past the yieldingpoint — Fig. 12. Thus, it is reasonable tothink of the maximum von Mises stress ob-served in the filler metal with the overlap≤ 0.5T as the shear strength of the fillermetal. As the overlap width increases, thevon Mises stress distribution becomes lessand less uniform. The middle portion ofthe overlap contributes less and less to theoverall load-carrying capacity of the joint,whereas the ends of the joint become“overloaded,” i.e., von Mises stress ex-ceeds maximum shear stress value of thefiller metal (we define it as critical). Themost striking feature shown in Fig. 12 isthat at the failure loads, all von Misesstress distribution curves converge in thesame point located approximately 5% ofthe overlap width away from the edge ofthe joint.

In other words, all lap shear test speci-mens failed when von Mises stress ex-ceeded a certain critical value over somelength that starts at the edge and extendsinto the joint. We can think of this lengthas a damage zone. The length of the dam-age zone is not constant but rather pro-portional to the overlap and equal to ap-proximately 10% (5% from each end) ofthe overlap width. Consequently, we candefine the failure of the ductile lap shearbraze joints as follows: the ductile lapshear braze joint fails through the fillermetal when the size of the damage zoneexceeds certain critical value. Based onour observations, the size of the damage

Fig. 9 — Maximum load (A) and maximum average or stress shear ultimate strength (B) plotted as afunction of the overlap size. Also shown are the FEA curves predicting similar values. As one can seeelectrolytically plated specimens show a better fit with the FEA curves.

Fig. 10 — Stress distribution in the lap braze joint at the failure load.

y

y

X

X

WELDING RESEARCH

-S37WELDING JOURNAL

zone for the stainless steel/silver combina-tion is about 10% of the overlap width.

The concept of the damage zone is notnew. In fact, it has been used quite exten-sively to define failure of adhesivelybonded joints (Ref. 4). The authors, how-ever, did not find any references where thedamage zone concept was used to describethe failure of the braze joints. It is inter-esting to note that in the 1960s, whenstudying deformation of the lap joints withthe help of photoelastic methods, the re-searchers were surprised to see the jointedges experienced loads far exceeding theyield point of the filler metal, while theoverall load applied to the test specimenswas still relatively small (Ref. 17)

It would be constructive to see if thedamage zone concept can be applied tothe experimental results reported by otherresearchers. To stay within the same fam-ily of base metal/ filler metal combina-tions, we selected the classic work ofBredzs and Miller (Ref. 18) to see if theirtest results can be correlated with our fi-nite element analysis based on the 10%damage zone failure criterion. For thispurpose, AWS BAg13 filler metal (Ag, Cu,and Zn alloy) was obtained and mechani-cally tested to generate a true stress-straincurve (Fig. 13), which is necessary forFEA-procedure. Indeed, the correlationbetween the experimental results and theFEA-predicted shear strength for theBretz-Miller test specimens is quite goodas shown in Fig. 14. Limited resources pre-vented the authors from comparing thedamage zone based FEA results withother experimental data reported in theliterature.

Although the strength of the fillermetal in the lap shear braze joints is higherthan the strength of the filler metals testedalone, it is realized that the von Misesstress cannot be as high as one would as-sume by looking at the edge regions in Fig.12. Finite element analysis does not alwaysgive realistic values for the von Mises peak

stress at the cor-ners or edges. Thevon Mises curveusually cappedaround the discon-tinuity points usingcertain simplifiedvalues. For exam-ple, in the case ofadhesive joints, theshear stress withinthe joint calculatedusing FEA proce-dure is capped bythe shear strengthof the adhesive(Ref. 4). In ourcase, however, itwas decided toleave the von Misesstress plot as is,since it is not clear

Fig. 11 — Fracture surface of the electrolytically plated shear test specimen showing peel + shear (Re-gion 1 ) and shear -dominated (Region 2) fracture modes. Shape of the dimples revealed on higher mag-nification images (300· and 2000· , progressively) confirms this observation.

Fig. 12 — Calculated values of the von Mises stress in the filler metal of thebraze joints corresponding to the fracture of the shear test specimen. For con-venience, the distance within the lap is given as a normalized value.

Fig. 13 — True stress–strain curve generated for AWS BAg13 filler metal.Using these data, the damage zone based FEA was applied to compare ourprediction with the results of Bredzs and Miller (Ref. 18.)

Fig. 14 — Good agreement between the data by Bredzs and Miller (Ref. 18)for lap shear specimens brazed at two different temperatures and predictedSUS using a 10% damage zone.

WELDING RESEARCH

JANUARY 2004-S38

what value should be set as an upper limitfor the von Mises stress in the lap shearbraze joint damage zone. However, thisinvestigation has demonstrated that,within the damage zone, the filler metalcan be overloaded, but the joint does notfail until the size of the damage zonereaches a critical value.

Conclusions

• It appears that for small overlaps (≤ 0.5T) stress distribution within the lapshear braze joint is uniform.

•The maximum von Mises stress cal-culated at failure in the 0.5T lap joint canbe used as a critical or shear ultimatestrength of the filler metal for the lap joint.

• Within the joint clearance sizestested, 0.001–0.008 in. (25.4–203 mm),strength is not sensitive to clearance sizes.

•In the stainless steel lap shear jointsbrazed with silver metal, the failure occursin the filler metal for overlaps up to 5T.

•It appears that the following damagezone criterion can be used: the lap shearbraze joint fails if von Mises stress exceedsthe critical value over 10% of the overlaparea.

• The 10% damage zone observationchecks well against other stainlesssteel/silver filler metal systems.

The authors would encourage other re-searchers to apply the damage zone fail-ure criterion to their experimental data tosee if this observation holds true for otherbase metal/filler metal combinations.

Acknowledgments

This effort was funded exclusively bythe internal Goddard Space Flight Centerflight programs. Specifically, the authorswould like to thank Tom Venator andKevin Grady (EOS AQUA), Robert Lillyand Martin Davis (GOES), and JohnGagosian and Tony Comberiate (TDRSS)for their support and encouragement.Also greatly appreciated are numerousdiscussions on the subject matter withRobert Peaslee and his help in obtainingmany valuable resources used in this work.

References

1. American Welding Society. 1963. C3.1-63, Establishment of a standard test for brazedjoints, A committee report.

2. Johnson, W. S., ed. 1985. ASTM STP 876,Delamination and Debonding of Materials,American Society for Testing and Materials.

3. Johnson, W. S., ed. 1988. ASTM STP 981,Adhesively Bonded Joints: Testing, Analysis and De-sign, American Society for Testing and Materials.

4. Tong, L., and Steven, G. P. 1999. Analysisand Design of Structural Bonded Joints, KluwerAcademic Publishers.

5. Volkersen, O. 1938. Die niektraftverteil-ing in zugbeansprunchten mit konstantenlaschenquerscritten. Luftfahrforsch 15: 41–47.

6. Military Specification MIL-B-007883C,1986. Brazing of Steels, Copper, Copper Alloys,Nickel Alloys, Aluminum and Aluminum Alloys.Currently obsolete.

7. NASA. 1968. John F. Kennedy Space

Flight Center Specification KSC-SPEC-Z-0005. Specification for Brazing Steel, Copper,Aluminum, Nickel and Magnesium Alloys.

8. ANSI/AWS C3.6:1999, Specification forFurnace Brazing. Miami, Fla.: American Weld-ing Society.

9. ANSI/AWS C3.4:1999, Specification forTorch Brazing. Miami, Fla.: American WeldingSociety.

10. ANSI/AWS C3.5:1999, Specification forInduction Brazing. Miami, Fla.; AmericanWelding Society.

11. ANSI/AWS C3.7:1999, Specification forAluminum Brazing. Miami, Fla.: AmericanWelding Society.

12. ANSI/AWS C3.2:2001, Standard Methodfor Evaluating the Strength of Brazed Joints inShear. Miami, Fla.: American Welding Society.

13. Watkins, T., Jr. 1999. Private communi-cations.

14. Dieter, G. E. 1976. Mechanical Metal-lurgy. McGraw-Hill, Inc.

15. Lugscheider, E., Reimann, H., andKnotek, O. 1977. Calculation of strength of sin-gle-lap shear specimen. Welding Journal 56(6):189-s to192-s.

16. Shaw, C. W., Shepard, L. A., and Wulff,J. 1964. Plastic deformation of thin brazedjoints in shear. Transactions of ASM 57: 94–109.

17. Peaslee, R. L. 2002. Private communica-tions.

18. Bredzs, N., and Miller, F. 1968. Use ofthe AWS standard shear test method for evalu-ating brazing parameters. Welding Journal47(11): 481-s to 496-s.

PublisherJeff Weberjweber@aws.org, Extension 246General Management,Reprint Permission, Copyright Issues

Editor Andrew Cullison cullison@aws.org, Extension 249Article Submissions

Senior EditorMary Ruth Johnsenmjohnsen@aws.org, Extension 238Feature Articles

Associate Editor Susan Campbellcampbell@aws.org, Extension 244Society NewsPersonnel

Associate Editor Ross Hancockrhancock@aws.org, Extension 226New ProductsNew Literature

Production Editor Zaida Chavezzaida@aws.org, Extension 265Design and Production

Advertising Sales Director Rob Saltzstein salty@aws.org, Extension 243Advertising Sales

Advertising Production Coordinator Frank Wilsonfwilson@aws.org; Extension 465Advertising Production

Advertising Sales & Promotion Coordinator

Lea Garrigangarrigan@aws.org, Extension 220Production and Promotion

Peer Review Coordinator Doreen Kubishdoreen@aws.org, Extension 275Peer Review of Research Papers

Welding Journal Dept. 550 N.W. LeJeune Rd. Miami, FL 33126 (800) 443-9353FAX (305) 443-7404

CAN WE TALK?The Welding Journal staff encourages an exchange of ideas with you, our readers. If you’d like to ask a question, share an idea or

voice an opinion, you can call, write, e-mail or fax. Staff e-mail addresses are listed below, along with a guide to help you interact withthe right person.