sign in sign up
<<
Home , Hardness, Strength of materials

mechanical.qxp 3/4/2005 9:15 AM Page 1 MECHANICAL UNDERSTANDING THE BASICS TESTING Load cell echanical testing deals F The results of tests Crosshead with the response of commonly used to to applied . M Considered in this article Screw measure the properties of are the common testing techniques heat treated parts provide related to the mechanical failure of AF=σA metals. These include tension (ten- 1 information about how Specimen sile and transverse rupture), torsion, well design requirements , , and - are met. Test data also can rupture, and impact tests. and statistical analysis of me- help ID production chanical property data are also F Base briefly covered. problems before they Fig. 1 — Typical tensile tester: a screw- driven electromechanical system. (Ref. 1) become critical. Tension Testing Tensile, transverse rupture, and Strain to Daniel H. Herring* other tension tests are widely used Uniform strain s Herring Group Inc. to provide basic design information Necking Fracture Elmhurst, Illinois on the strength of and as begins acceptance tests for the specification Offset strength Tensile of materials. The basic parameters strength used to determine the stress-strain Fracture stress behavior of a are its tensile Engineering stress, strength (ultimate tensile strength), yield strength (tensile yield strength) Engineering strain, e and yield point, elongation (or per- Fig. 2 — Engineering stress-strain curve. cent elongation), and reduction in Intersection of the dashed line with the curve determines the offset yield strength. (Ref. 2) area. The first three are strength properties; the other two, measures this portion of the test, the specimen of . will return to its original length. This Tensile tests are designed to “pull” is called elastic . The pro- a specimen to failure (Fig. 1). The de- portionality between the stress and structive test measures the and strain is the (mod- stretch of a during testing. ulus of or Young’s mod- The test plots “stress” versus “strain” ulus). The modulus of elasticity is de- (Fig. 2). Stress is the (applied) load termined by the binding forces of the divided by the cross-sectional area atoms and since these forces cannot of the test specimen at its center. be changed without changing the Strain is the change in a dimension basic nature of the material, it follows divided by the original dimension, that this is one of the most structure and is measured over the central por- insensitive mechanical properties. tion of the specimen length (where With continued loading and the cross section is constant). stretching, the tensile specimen per- As the specimen is stretched, the manently deforms, exhibiting load required to induce each level of deformation. The yield strength is strain is measured. When the load is the stress at which the specimen first applied the tensile specimen shifts from elastic (recoverable) stretches in proportion to the applied stretching to plastic (permanent) de- * Member, ASM Society load. If the load is removed during formation. By standard convention, HEAT TREATING PROGRESS • MARCH/APRIL 2005 31 mechanical.qxp 3/4/2005 9:16 AM Page 2

the reported yield strength corre- elasticity in shear, torsional yield sponds to a plastic strain of 0.2% strength, and modulus of rupture. (where observable deformation has Any part subject to torsional loading Indent taken place) in service, such as shafts and axles, Diagonal or Tensile strength is the highest should be torsion tested. Test Indenter(s) diameter stress encountered in the tensile test. Torsion testers (Fig. 3) consist of a Brinell Ball indenter, 1–7 mm For many this corresponds to twisting head, with a chuck for grip- 10 mm (0.4 in.) (0.04–0.28 in.) the stress at fracture. For very duc- ping the specimen and for applying or 2.5 mm (0.1 in.) tile steels, the stress at fracture is the twisting moment, and a in diameter lower than the tensile strength. For weighing head, which grips the very brittle steels, the yield strength other end of the specimen and meas- is the same as the tensile strength ures the twisting moment, or torque. (and fracture strength). Deformation of the specimen is A common measure of ductility is monitored by a twist measuring de- Rockwell 120° diamond 0.1–1.5 mm the elongation (given as total stretch vice called a troptometer. cone, (0.004–0.06 in.) to failure divided by the initial spec- A torsion test specimen typically 1.6–13 mm (1/16–1/2 in.) imen-center test length, or gage has a circular cross section (the sim- length). Elongation is a dimension- plest geometry for calculation of Rockwell As for 0.1–0.7 mm Rockwell (0.004–0.03 in.) less number, expressed as a per- stress). Since the in the superficial centage. Another measure of ductility elastic range varies linearly from is reduction in area, also dimension- zero at the center of the specimen to Vickers 136° diamond Measure pyramid diagonal, less and expressed as a percentage. a maximum at the surface, it is fre- not diameter Reduction in area is a measure of the quently desirable to test a thin-wall change in cross sectional area at the tubular specimen. This results in a point of failure (change in area di- nearly uniform shear stress over the µ vided by the original area). cross section. Micro- 136° diamond 40 m The transverse rupture test is a Torsion test results can be used to hardness indenter or (0.16 micro-in.) strength test designed for low-duc- validate or expand on the informa- a Knoop indenter tility materials, including carbides tion gleaned from tensile testing. Tor- and powder (P/M) ma- sion testing provides a more funda- terials. This destructive test involves mental measure of the of a µ Ultrasonic 136° diamond 15–50 m bending rather than pulling of the material than does tension testing. It pyramid (0.06–0.2 specimen. Maximum load, specimen directly yields a shear stress-shear micro-in.) dimensions, and test time are used strain curve. In torsion, the critical to calculate the stress needed to shear stress for plastic flow is cause failure. A typical transverse reached before the critical normal rupture strength is 1.5 to 2 times the stress for fracture; while in tension, tensile strength the critical normal stress is reached before the critical shear stress. Torsion Testing ment, heat treaters probably are most Torsion testing is used to deter- Hardness Testing familiar with the hardness tester. Al- mine such properties as modulus of Among mechanical testing equip- most every shop has either a Rock- well-scale or Rockwell superficial- Hydraulic Servo-valve scale tester. The Rockwell scales are power Hydraulic motor Test specimen Rotary Torque cell those most often used for ferrous ma- transducer Radiant terials. Many shops also have Brinell furnace and microhardness or microinden- tation hardness testers. These and other tests are compared in Table 1. Saddle In general, hardness usually im- Lathe bed plies a resistance to deformation. For a metal, the property is a measure of its resistance to permanent or plastic deformation. Hardness measuring Vacuum Furnace pump controller methods can be sorted into three general types, depending upon the manner in which the tests are con- Fig. 3 — Typical torsion tester: a servo-controlled, hot torsion machine mounted on a lathe ducted: scratch hardness, indenta- bed. Shafts and axles are among the parts that should be torsion tested. (Ref. 3) tion hardness, and rebound, or dy- 32 HEAT TREATING PROGRESS • MARCH/APRIL 2005 mechanical.qxp 3/4/2005 9:16 AM Page 3

Table 1 — Comparison of indentation hardness tests (Ref. 4)

Method of Surface Tests per Depth1 Load(s) measurement preparation hour Applications Remarks Up to 0.3 mm 3000 kgf Measure Specially 50 with Large forged Damage to specimen (0.01 in.) and for ferrous diameter of ground area diameter and cast parts minimized by use 1 mm (0.04 in.), metals; indent under for measurements of lightly loaded respectively, down to microscope; measurements ball indenter; with 2.5 mm 100 kgf for read of diameter indent will then be (0.1 in.) and soft metals hardness less than a 10 mm (0.4 in.) from tables Rockwell indent in diameter balls 25–375 µm Major, Read hardness No 300 manually, Forgings, Measure depth of (0.1–1.48 60–150 kgf;directly from preparation 900 castings, penetration, not micro-in.) Minor, meter ornecessary on automatically roughly diameter 10 kgf digital display many surfaces machined parts 10–110 µm Major, As for Machined As for Critical A surface test of case (0.04–0.43 15–45 kgf; Rockwell surface, ground Rockwell surfaces of and micro-in.) Minor, 3 kgf finished parts 30–100 µm 1–120 kgf Measure Smooth clean Up to 180 Fine finished Small indent but high (0.12–0.4 indent with surface, surfaces, local stresses micro-in.) low-power symmetrical thin specimens microscope; if not flat read hardness from tables 1–4 µm 1 gf–1 kgf Measure Polished Up to 60 Surface layers, Laboratory test used on (0.004–0.016 indent with surface thin stock, brittle materials or micro-in.) low-power down to microstructural microscope; 200 µm constituents read hardness (0.008 in.) from tables 4–18 µm 800 gf Direct readout Surface better 1200 (limited Thin stock Calibration for Young’s (0.016–0.07 onto meter or than 1.2 µm by speed and finished modulus necessary; micro-in.) digital display (0.004 micro-in.) at which surfaces in 100% testing of finished for accurate operator any position parts; completely work; can read nondestructive otherwise, up display) to 3 µm (0.012 micro-in.) 1.The minimum material thickness for a test is usually taken to be 10 times the indent depth.

namic, hardness. Only indentation  hardness is of major engineering in- 160 (1103) terest for metals.  A part is usually hardness tested 120 Low-cycle 40,000 cycles at ~110 ksi (760 MPa) fatigue after heat treating, and the value ob- (827)  tained is an indication of the effec- 160,000 cycles at ~70 ksi (485 MPa) 80  (552 tiveness of the treatment. A test tip   (indenter) of defined configuration  40 

Tensile stress, ksi (MPa) Tensile (276) is pressed into the surface of the spec- Infinite life imen using a constant, predeter- mined force. The tip deforms the 104 105 106 107 108 such that a high hardness cor- No. of cycles responds to a small impression or in- Fig. 4 — Typical S-N (stress vs. load cy- dent. The hardness value is a func- edly without failure. A fatigue failure cles) curve. A linear reduction in applied tion of the size of the indent (Brinell is particularly catastrophic because stress results in an exponential improvement and Vickers) or the depth of penetra- it occurs without warning. Three in fatigue life. (Ref. 5) tion by the indenter (Rockwell). basic factors are necessary to cause a fatigue failure: a maximum tensile Fatigue Testing stress of sufficiently high value, a Fatigue is a measure of the stress large enough variation or fluctuation that a material can withstand repeat- in the applied stress, and a suffi- HEAT TREATING PROGRESS • MARCH/APRIL 2005 33 mechanical.qxp 3/4/2005 9:16 AM Page 4

ciently large number of cycles of the (N) depends on the maximum ap- applied stress. plied stress (S). This particular graph The basic method of presenting shows that a linear reduction in ap- engineering fatigue data is by means plied stress results in an exponential of the S-N curve (Fig. 4) that shows improvement in fatigue life. Thus, a how the life of the specimen, ex- 35% reduction in stress — from 110 pressed in number of cycles to failure ksi (760 MPa) to 70 ksi (485 MPa) — results in a 400% improvement in fa- Specimen tigue life — from 40,000 cycles to Bearings Specimen Bearings 160,000 cycles. Additional reductions in stress result in significantly more fatigue life enhancement. Fatigue testing equipment (Fig. 5) is usually designed to induce cyclic Load (a) (b) Load loading and unloading to a known (peak) stress and measure the Fig. 5 — Schematics of rotating- fatigue testing machines. (a) Four-point loading R.R. Moore testing machine. (b) Cantilever number of such cycles to failure of loading machine. (Ref. 6) the specimen. Variants of the test in- clude tensile, bending, and rotating. The average stress at which a steel Lever arm can withstand 10 million loading cy- cles without failure is reported as the Double fatigue strength (also called the en- knife-edge durance limit). As stress increases, alignment coupling the number of cycles to specimen failure decreases. Fatigue failures are a combination of crack initiation and crack growth (propagation). Sophisticated design Weight equations are available to predict train component life under cyclic loading

Tube conditions. Fatigue strength is al- furnace ways less than yield strength, and is Load frame often 30 to 50% of tensile strength. Specimen Fatigue strength is significantly re- holders duced by the introduction of a stress raiser (stress concentrator) such as a keyway, screw thread, fillet, press fit, Elapsed timer or hole, many of which are present in Extensometer most structural parts. One of the best ways to minimize fatigue failures is to reduce as much as possible the number of stress raisers via careful design and proper heat treatment. Double knife-edge alignment coupling Creep and Stress-Rupture Testing It’s often necessary to consider the effects of service on the life of a heat treated part. Since the

Weight pan mobility of atoms increases rapidly with temperature, diffusion-con- Weight trolled processes can have a very sig- elevator nificant effect on high-temperature mechanical properties. The effect of long-term exposure to these environ- ments is a critical design considera- tion. Thus, successful use of these parts requires knowledge of how their strength varies with time at Fig. 6 — Schematic of a test stand used for creep and stress-rupture testing. (Ref. 7) high . 34 HEAT TREATING PROGRESS • MARCH/APRIL 2005 mechanical.qxp 3/4/2005 9:17 AM Page 5

Creep and stress-rupture tests are by the impact load at the opposite used to evaluate the performance of end. materials for elevated-temperature The most common test machine service. The creep test measures the for measuring , or impact Scale dimensional changes that occur in a resistance, involves a swinging pen- specimen during exposure to high dulum, with a test specimen posi- Starting position (all potential temperature, while the stress-rup- tioned at the bottom of the pen- Pointer energy) ture test measures the effect of tem- dulum swing (Fig. 7). A specimen of Hammer a tough material requires consider- Bottom perature on the specimen’s long-time position load-bearing characteristics. The test able energy for failure, consuming (all kinetic End of swing energy) stand shown in Fig. 6 can be used for much of the pendulum’s kinetic en- h both creep and stress-rupture testing. ergy. Consequently, the pendulum Specimen Other tests measure special proper- will not swing very high after h’ ties such as thermal shock resistance breaking the specimen. If the spec- Anvil and stress relaxation. imen has low toughness, then the Note that creep and stress-rupture pendulum will end its swing consid- data also play important roles in the erably higher after fracture. Fig. 7 — Standard impact testing appa- heat treating industry. For example, Toughness defined: The tough- ratus. (Ref. 8) when selecting high-temperature al- ness of a material is a measure of its loys for furnace interior components ability to absorb energy in the plastic High- spring steel or heat treating fixtures, a good range. The ability to withstand occa- sional stresses above the yield stress measure of the reliability of a candi- s A date , or as a comparison of the without fracturing is particularly de- B Structural steel expected performance of several dif- sirable in certain applications; for ex- Stress, ferent alloys, is “1% creep in 10,000 ample, freight car couplings, gears, hours” data at the intended service chains, and crane hooks. Toughness temperature. may be visualized as the total area Strain, e under the stress-strain curve. This Fig. 8 — Comparison of stress-strain Impact Testing area is an indication of the amount curves for high- and low-toughness steels. Brittle fracture is catastrophic, and of work per unit volume that can be Cross-hatched regions represent the modulus factors that contribute to it include: done to the material without causing of resilience (UR) of the two materials. UR is triaxial stress, such as exists at a it to rupture. determined by measuring the area under the stress-strain curve up to the elastic limit of notch or stress raiser; low tempera- An example is shown in Fig. 8, the material (point A for the spring steel; point ture; and a high strain rate or rapid stress-strain curves for high- and B, the structural steel). (Ref. 9) rate of loading. All three do not have low-toughness materials. The high- to be present at the same time. carbon spring steel has a higher yield cleavage change with temperature. Impact tests are designed to deter- strength and tensile strength than the Temperature has a strong effect on mine a toughness value, which indi- medium-carbon structural steel. basic flow and fracture properties. cates the susceptibility of a material However, the structural steel is more For metals, the yield stress (flow to brittle fracture (tougher materials ductile and has a greater total elon- stress) increases with decreasing are less susceptible). Notched-bar im- gation. The total area under the struc- temperature. Above the transition pact tests are used most often. These tural steel’s stress-strain curve also temperature, the flow stress is tests detect differences between ma- is greater, which means it is the reached before the fracture stress; terials that are not revealed by a tor- tougher of the two materials. There- below it, the fracture stress is reached sion test. fore, toughness encompasses both first. The relative values of these two Notched Charpy and Izod speci- strength and ductility. properties determine whether the mens are considered the standards Data scatter: Results of tests using fracture will be ductile or brittle. for impact testing. Charpy “bars” notched impact bars are subject to Most of the scatter in impact test have a square cross section and a “V” considerable scatter, particularly in data is due to local variations in the or “keyhole” notch at the center of the region of the ductile-to-brittle properties of the steel, while some is their length. The Charpy bar is struck transition temperature. (There actu- due to the difficulty of preparing per- at its center by the impact load to en- ally are two distinct temperatures in fectly repeatable notches. In general, sure failure at that point, which re- this region: the ductility transition the reported value of the transition sults in a lower toughness value (re- temperature, which is related to the temperature is based on an amount flecting a more conservative design material’s fracture initiation tenden- of transition energy absorbed, a parameter). Izod specimens have ei- cies, and the fracture transition tem- change in the appearance of the spec- ther circular or square cross sections perature, which is related to crack imen fracture surface, or a transition and contain a “V” notch near one propagation.) The transition temper- in ductility. end. The specimen is clamped at the ature owes its existence to the way Note that a slow-bend test is some- end with the notch and then struck in which resistance to shear and times used to determine the transi- HEAT TREATING PROGRESS • MARCH/APRIL 2005 35 mechanical.qxp 3/4/2005 9:17 AM Page 6

tion temperature. A biaxial stress bility of heat treated parts. They also useful everyday tools, while design state is produced during bending of can influence how a material reacts of experiments and frequency distri- an unnotched beam (having a width to externally applied stresses. bution techniques are invalu- much greater than its thickness). This able to the design engineer. test represents a severity interme- Statistical Methods diate between that of the tension test There are at least three reasons References and the notched impact test. why a working knowledge of statis- 1. Metals Handbook Desk Edition, Howard tics is needed in mechanical testing. E. Boyer and Timothy L. Gall (Eds.): ASM Residual Stress • Mechanical properties are struc- International, Materials Park, Ohio, 1985, Heat treaters deal with issues of ture-sensitive, so they frequently ex- p. 34·3. 2. ASM Handbook, Vol. 8, Mechanical residual stress every day. These are hibit considerable variability or Testing and Evaluation, Howard Kuhn and the stresses that exist in a material scatter. This makes statistical tech- Dana Medlin (Eds.): ASM International, when it is free from external forces. niques useful, and often necessary, Materials Park, Ohio, 2000, p. 99. (Orig- They also are sometimes referred to for determining the precision of the inal source: Mechanical Metallurgy, 3rd as internal or locked-in stresses. measurements and enabling valid Ed., by George E. Dieter: McGraw-Hill, Residual stresses are produced conclusions to be drawn from test New York, N.Y., 1986.) whenever a material undergoes a data. 3. ASM Handbook, Vol. 8, Mechanical nonuniform plastic deformation. In • Statistical methods can assist in Testing and Evaluation, Howard Kuhn and quenching (hardening), for example, designing experiments to provide Dana Medlin (Eds.): ASM International, a temperature differential between the maximum amount of informa- Materials Park, Ohio, 2000, p. 190. (Orig- the rapidly cooling surface and the tion at minimum cost. inal source: S. Fulop et al.: Journal of Testing and Evaluation,. Vol. 5, No. 6, 1977, slower cooling core produces a mis- • Statistical methods (which are p. 419.) match of strain. The residual stress based on probability theory) can be 4. ASM Handbook, Vol. 8, Mechanical pattern that results is a combination used to help explain certain prob- Testing and Evaluation, Howard Kuhn and of the thermal volume changes and lems or phenomena such as the size Dana Medlin (Eds.): ASM International, those resulting from the transforma- effect in brittle fracture and fatigue. Materials Park, Ohio, 2000, p. 261. tion of austenite to martensite. Heat treaters use statistical meth- 5. “Selecting the Best Residual stresses are responsible ods daily. For example, control charts Method for the Heat Treatment of for warping and dimensional insta- and linear regression analysis are Gears,” by Daniel H. Herring, David J. Breuer, and Gerald D. Lindell: AGMA 2002 Fall Technical Meeting, ISBN: 1- 55589-807-6, 13 p. 6. ASM Handbook, Vol. 8, Mechanical Testing and Evaluation, Howard Kuhn and Dana Medlin (Eds.): ASM International, Materials Park, Ohio, 2000, p. 689. 7. ASM Handbook, Vol. 8, Mechanical Testing and Evaluation, Howard Kuhn and Dana Medlin (Eds.): ASM International, Materials Park, Ohio, 2000, p. 374. 8. Metallurgy for the Non-Metallurgist, by Harry Chandler: ASM International, Ma- terials Park, Ohio, 1998, p. 206. 9. ASM Handbook, Vol. 8, Mechanical Testing and Evaluation, Howard Kuhn and Dana Medlin (Eds.): ASM International, Materials Park, Ohio, 2000, p. 102. (Orig- inal source: Mechanical Metallurgy, 3rd Ed., by George E. Dieter: McGraw-Hill, New York, N.Y., 1986.)

For more information: Dan Herring is president, Herring Group Inc., P.O. Box 884, Elmhurst, IL 60126-0884; tel: 630/ 834-3017; fax: 630/834-3117; e-mail: [email protected]; Web: www.heat-treat-doctor.com. 36 Circle 11 or visit www.adinfo.cc HEAT TREATING PROGRESS • MARCH/APRIL 2005