Part 6—Practical Applications Understanding Sheet Metal Formability You can reduce die tryout time and minimize breakage of stampings in production by using simple techniques described in this article. The result: better stampings faster. Fonningot sheet metas tundergoing a transition from art to science. As explained in preceding in- stallments of this article, guesswork hhas been replaced by engineering analysis The key to successful forming is to obtain as uniform a strain distribution as possible under the limit established by the eritical strain level of the mate- rial. If this limit is exceeded, the stamping will break during forming. ‘The direction and magnitude of ‘maximum strain are easily determined by electrochemically etching a pattern of small-diameter circles on the sur- face of the blank. During forming, the circles are deformed into ellipses. ‘The direction of maximum strain corres- ponds to the major axes of the ellipses and the magnitude of that strain is de- termined by measuring the increase in length of the ellipses over the diam- eters of the circles. Through the analysis of these visible strain patterns, die tryout and die mod- iffeation can be simplified, the selection THE AUTHOR. Stuart P. Keeler is Supervisor, Flat Rolled Products Ap- plications, Research and Development, National ‘Steel Corporation, Ecorse, Mich. He earned his Doctor of Science Gegtee at Massachusetts Institute of ‘Technology in 1961, Hie doctoral dis- sertation was on the subject of plas: Uc instability and fracture in. sheets stretched over rigid punches. Most of hie work im industry has also been on the formability of sheet metals. AGS VINa ft —__ of optimum materials and lubricants can be made easier, and breakage of stampings in production ean be mi mized, Successful forming requires the right combination of material, lubriea- tion and die design. Small changes in lubrication and die design are often more effective in reducing breakage of stampings than changing to materi- als with different properties. By ana- lyzing strain patterns, the changes needed to convert an “unformable” stamping into a formable stamping can be determined and the effects of those changes ean be quickly evaluat- ed. Ideally, this is done during die try- ‘out, where dies are given the necessary modifications for satisfactory produc- tion runs. The fact that a stamping is successfully formed during die tryout does not necessarily mean that it can be successfully formed in production During die tryout, press speeds are usually slower than those used in pro- duction. Good materials are used, blanking and trimming punches are sharp and welladjusted, gaging is carefully set and other press variables ‘correctly adjusted, lubricants are hand-applied. In. short, all variables are optimized in order to produce ac- ceptable stampings. When the dies are installed in pro- duction presses, conditions may no longer be optimum. A stamping that was successfully formed during die tryout may break when it is formed under production conditions. This can a lead to costly delays and the danger of missing deadlines for the startup of production ‘One of the major benefits of using the etched-cirele system during die tryout is that it makes it possible to find critical stampingsstampings that are likely to break when formed under production conditions—before produc tion trials are made. For many tool tnd die men, the visual display of strain pattems provides sulicient in- formation to suggest changes that will make the stamping formable in pro- duction, even when forming conditions are far from optimum. When a critical stamping is identified, the magnitude, directly and distribution of strain ean be altered to provide a margin of safe- ty in production forming Another advantage of the etched- circle system is that it provides a quantitative record of progress made during die tryout. At intervals, a pre pared blank can be formed. The peak strain or, more important, the safety factor, ean be compared with a pre- vious trial and the progress recorded. ‘A mathematical measurement of the current state of the dies replaces opinion. : Use of blanks with etched patterns of circles is also helpful in the selec- tion of materials, particularly when stampings are critical. When a blank of the material tentatively selected for the stamping is formed, its safety factor is determined by comparing its peak strain with the critical strain level. If a safety factor of, say, 10 MACHINERY 4 4 épercent Is present, the mechanical properties of that blank become the required properties and dictate the material grade, If the safety factor is less than that desired, materials of increasingly bet- ter quality will have to be tried until a satisfactory safety factor is attained The altemative is to modify the die design or try different lubricants until the peak strain is reduced sufficiently to provide an adequate safety factor. If the safety factor is Tanger than required, it may be possible to use a lower-cost grade of matertal. The safe- ty factor of the new material must be checked, of course In some cases, the required mate- rial properties are in an area of over- lap between two grades. For example, the properties may fall in the upper range of a rimmed DQ steel and in the lower range of an aluminum-killed steel. In such cases, grade selection Gepends on other factors, including uniformity from shipment to shipment, fnventory turnover and aging of the rimmed steel. Ideally, the safety factor of the ma- terial selected is evaluated for the high, average and low portions of the specified grade range. From this, the sensitivity of the safety factor to vari tions in’ material properties can be Aetermined and appropriate safe guards imposed. In a few special cases, a harder steel (having higher yield stress) may have a higher safety factor than a softer steel. The harder steel appears to resist straining over the punch head and to pull more metal in from the flange, thus reducing the peak strain over the punch head and increasing the safety factor. This illustrates that it may be beneficial to evaluate steels that have forming properties above and below the tentative specification. The etched-cirele system can open up many possibilities in material se- lection, some of which ean be resolved only by a management decision. An example is shown in Figure 1. ‘This figure illustrates the strain dis- tribution in an automotive fender. ‘Two different steels are being tried out =a rimmed steel and an aluminum- Killed steel. Both steels ate fresh and have average mechanical properties The critical strain level for the fender is 47 percent. The peak strain for the rimmed steel is 45 percent, pro- viding a safety factor of only 2 percent. This safety factor is inadequate, since JULY 1908 FF crimica, STRAIN LEVEL ar percent Xo PERCENT PEAK STR InaMED Steet fy “XN XN yo" % Ns vo 7 perereers ONS eet 7 SRS ENGINEERING STRAIN (PERCENT) +1 be vainen LOCATION OF ORIGINAL GRID CENTERS 1. STRAIN DISTRIBUTION in an automotive fender. The critical strain level begins ot 47 percent. Peak strain for stampings made from the rimmod sioel is 45 percent, allowing a safety factor of only 2 percent. Slight variations in ‘material properties or press conditions can lead to breakage during forming. The aluminum-killed steel. with its 81 percent peak strain and 16 percent safety factor, will be more satisfactory for this stamping, Long storage times could be tolerated Fee ___ shmcat oan Gi Levets Bea 40 PRODUCTION STRAIN DISTRIBUTION -STANDARD STEEL 20 ENGINEERING STRAIN (PERCENT) REFERENCE STRAIN DISTRIBUTION LOCATION ON THE STAMPING 2, STRAIN DISTRIBUTIONS of a production stool and a standard (reference) steel are shown in this plot. Both strain distributions were obtained at the start of the production run. The reference strain distribution—even though different than the production strain distribution because of different properties —con be used as a norm to evaluate future dia conditions. Any difference between this norm and any future strain distribution obtained with the standard steel con indicate that the prass or or lubricants, have changed.SHEET METAL FORMABILITY any rimmed steel having borderline properties (resulting in a higher peak strain) would probably break during forming. Further, the properties of rimmed steels are changed by aging (higher yield stress and lower TS/YS ratio) so using rimmed steel that had been in storage for some time would probably result in breakage of stamp- ings in production, The aluminum-killed steel has a more+than-adequate safety factor of 16 percent, since parts formed from this steel exhibit peak strain of only 31 percent. Wide variations in proper ties and long storage times could be tolerated, If the decision were made to use rimmed steel, an attempt would have to be made to reduce the peak strain 10 oF 12 percent by modifying the dies or using a different lubricant. Also, it would be necessary to establish an {inventory control program to promote early usage of the rimmed steel. A quality control check on ineoming ma- terial would be desirable to make sure that properties were as specified. And any steel not purchased for immediate consumption should be aluminum- killed. The etched-circle system provides a reliable means of checking steel quality or evaluating trial steels. The advantage of this system is that— unlike traditional tests—it shows the formability of the material when used on a specific set of dies with a specific lubricant. It should be pointed out that no single, universal formability index ean be determined for a given lot of ma- terial. What counts is how suecessfully the material can be formed on a cer. tain press, using certain dies and a certain lubricant on a certain day. ‘The fact that a given steel cannot be successfully formed to make a given stamping under a given set of condi tions simply means that its forming characteristics are not suitable for those circumstances. The same steel ‘may have excellent characteristics for other stampings, or even for the same stamping if die’ design or lubrication are modifed When evaluating steel quality or {tial steels, the tests should be made under the same conditions as those used when forming the steel with which @ comparison is being made. If variables are changed, no true com- 80 Parison of forming characteristics of several steels is possible, Suppose, for example, that a safety factor of 10 percent is required for a Particular stamping. ‘The production steel originally met the requirements; a trial steel, with a safety factor of 2 percent, does not Normally, the trial steel would be rejected as not having suffetent form- ability. Itis possible, however, that the safety factor of the production steel Js no longer 10 porcent. A recent change in pad pressure, or a change im any one of a number of other variables, could reduce’ the current safety factor of the production steel to.as litle as 1 percent. The tral steel, with its safety factor of 2 percent, would actually be superior to the pro- Guction steel. This illustrates the iim- portance of continued monitoring, 3 FAILURE LOCATION on this instrument panel mounting plate is indi by the solid white line, The 0.2-inch-dia eiched ito the blank before forming. The area enclosed by the shown in Figure 4, Tha circles indicate the strain patterns of Production monitoring is simple when the etched-circle system is used. Alter die tryout is completed, a blank of a standard or reference steel is, gridded and formed. The strain dis- tribution pattern of this steel would be different from that obtained with the production material (Figure 2), but could be used as a norm to char. acterize the die, press and lubr conditions, A library of these standards or reference steel blanks is stored. If the blanks are of a stabilized, killed steel they need not be protected against aging. If the blanks are of rimmed steel, they should be stored in deep freezer. If stampings start to break in produe- tion the fault may lie in the material, the lubricant, or the die and press vari- ables. Its mechanical properties may oter circles were electrachemically shed line is ls stomping. MACHINERY atedPHYSICAL PROPERTIES OF SAMPLES USED IN INSTRUMENT PANEL MOUNTING PLATE Percent Yield Tensile Ratio Elongation Rockwell Value Value of Strength Strength of in 2 B of of Breakage i) i) TS/YS Inches (9%) Hardness at # 0 25,200 ‘44,500 175 a4 4 023 1.80 26,800 44,800 1.67 43 44 0.23 50 26,500 42,000 1.58 a 43 0.22 192 28,500 45,000 1.58 a, 8 0.22 100 31,500 45,100 1.43 a a 0.22 1.50 34,600 45,000 1.33 40, 48 0.20 1. Coalcant of work hordening. be different. When stampings start to break in production, a blank of the standard or reference steel can be formed. IF the strain pattern is identi cal to the initial pattern or norm for that stamping, the die and lubricant hhave not changed so the material is at fault. If the standard steel shows a greatly changed strain pattem or fnereased maximum strain from the norm, the dies, lubricant or some other production variable have probably changed. The etched-circle system ean be the key factor in identifying the cause of breakage. In the same way, standard lots of Iubticant can be made up for use in distinguishing between troubles caused by tools and troubles caused by lubricants. Dies can be monitored, too—say once a month-with the aid of the etched-circle system. Tool wear can cause changes in strain distribution Comparison of production stampings with stampings made from standard blanks from the library can show whether the safety factor is increasing or decreasing. If the safety factor is gradually becoming smaller, breakage can be anticipated and the dies modi- fied during a normally scheduled maintenance period so that breakage will not interrupt a produetion run, Standard blanks are also. useful when resetting tools into a press, especially when evaluating alignment and pitman arm adjustments. 1f the strain pattems of stampings produced fon the reset tools are identical with the norm, then the die alignment and pitman arm adjustments are similar to the initial conditions. The etched-circle system is useful when evaluating lubricants. Lubri- cants greatly influence peak strains and safety factors. Since each lubri JULY 1908 2 Average velue of plawic ensorepy. (See Part 3, Apa fave) etched cizelos have been deformed into ellipses in thie area, The direction of cant performs differently in different dies, bench type testing of forming lubricants is only marginally satisfac tory. For rapid comparison of Tubri- cants under forming conditions, a series of prepared blanks is formed, each with a different lubricant. A safety factor for each lubricant can be quickly obtained. The lowest-cost Iu- bricant that provides the required safety factor can then be selected, Causes of production dificulties ean be analyzed through the use of the etched-circle system. In one ease, au maximum stroin is along the major axes of the ellipses. Unexpectedly, this is ct an angle of 45 degrees to the failure direction shown by white line tomotive bumpers had a frosty ap- pearance after plating. The plating process could have been at fault, or a high strain in the base material might have been causing “orange peel. ‘There was no correlation between strain values and the degree of frosty ‘appearance, so attention was turned to the plating process. Incorrect place- ment of plating anodes turned out to be the cause of the difficulty. Slight differences in die construction can be found by analysis of strain distributions. This is particularly use aSHEET METAL FORMABILITY ful when symmetrical areas, and right- hhand and left-hand parts, are being ‘compared. In one case, troubles occurred in the forming of a home clothes dryer top. Splits occurred in one comer, a neck in another comer. ‘The third and fourth comers were formed without difficulty. According to the part print, all four comers were supposed to be identical, Strain readings, however, showed 65 percent peak strain in the torn comer, 55 percent peak strain in the necked comer, and peak strains of 45 percent and 35 percent in the other comers. Measurements of the radii of the four comers showed that strain levels were inversely proportional to the radii of the comers. When all adil were made the same as the larg- est radius, splits and necks disap- peared. The etched-oircle system can be used for many different types of form- ing operations, even for explosive forming. Strain distributions in ex- plosively formed domes proved to be very uniform in one case, enabling rather high domes to be obtained from 4 material having a very low ratio of tensile strength to yield strength. Strain distributions in tubing have been obtained by electrochemically INKED CIRCLES: marking a pattem of circles on a coil of steel entering a tube mill. The stee) was subsequently rolled into a round tube, welded, reformed into a rectane gular tube by a turk’s head and Ginally cut to length. The strain distribution around the perimeter of the tube was subsequently evaluated, The circle pattems are an excellent training aid. The metal flow ean be ob- served in successive forming steps. The effects of die changes are easily seen. Since the circle pattern can be applied electrochemically in 1 minute, tedious layout and hand seribing are avoided, and the educational process fs accelerated. Only limited trial-and- error experience is needed to make intelligent decisions when the etched- circle system is used to analyze strain distributions. How a typical stamping problem is solved, using prepared blanks and other techniques deseribed in this ar- ticle, can best be illustrated by a case history. ‘The problem stamping is an auto- motive instrument panel mounting plate. A section of the first draw is shown in Figure 3. When certain lots of electrogalvanized steel were formed, breakage was severe. Location of the break is indicated by the heavy solid white line, Looking at this photograph, one 5. SIMULATED FORMING of the stomping shown in Figures 3 and 4. The blank ‘completed stamping. The polyethylene foard di ‘As the dies are moved down- yard, the sheet is stretched over the punch. Inked circlos on the sheet indicate the direction and mognitude of maximum forming strain in this stomping. 2 might conclude that the metal is re. stricted in the” blank and/or in the @ upper and lower die radii, causing the direction of high strain to be vertical up the stamping wall. Maximum strain would probably be between 40 and 60 percent (from the empizical critical strain level presented in Part 4 of this article) and failure would be perpendicular to the ditection of max- imum strain, Three lots of steel had been used to make this stamping. ‘Their breakage rates were 100 percent, 50 percent and zero. To analyze the reasons for breakage, mechanical property tests were made for samples from each lot and circle patterns were eteched on blanks from each lot. ‘Tom stampings made from pre- pared blanks exhibited a condition Quite different from what was expect- ed. An enlarged view of the failure, showing the strain pattern, appears in Figure 4. The maximum strain diec- tion is along a diagonal of the wall of the stamping. Failure is at an angle of 45 degrees to the direction of maxi- mum strain. This is a failure condition often found in a tension test specimen after localized necking. ‘The failure strain was measured as 80 percent major strain (AC) and mi- nus 25 percent minor strain (BD). This strain is just on the bottom edge of the critical strain level curve. (See Figure 7 of Bart 4 of this article.) Tn terms of true strains, the major true strain is 0.59 and the minor true strain is minus 0.29. The ratio of minor true strain to major true strain is minus 0.49. This is almost identical to the ratio of minus 0.5 found in uniaxial tension tests. From the strain ratio and failure mode, it could be concluded that the failure was similar to that in tension testa long axis AC. Ac cordingly, the formability of the steel for this stamping should be directly related to its stretchability, ag ev denced by a steep stress-strain curve, a high tensile-to-yield stress ratio and a high uniform elongation, Such a material would tend to dis. Uibute the strain more uniformly in the presence of a stress gradient, No correlation with the anisotropy ratio r ‘would be anticipated. The mechanical properties of the steels shown in the accompanying table confirm these re- lationships. Therefore a steel with a minimum longitudinal tensileto-yield stress ratio of 1.75 was specified. For successful forming, a good aluminum- MACHINERYkilled steel is undoubtedly requiced. Once a suitable steel was selected, the origin of the strain was determined to see if the maximum strain could be reduced. To do this, it was necessary to understand the metal flow. Since no partially formed stampings were available for study, metal flow was simulated as shown in Figure 5, using 1 sheet of polyethylene as the blank and a completely formed stamping as the punch. One edge of the sheet was taped to the “punch”; the other end was free. Two cardboard dies simu- Iated the die radii and the hold-down plate, The sheet was sandwiched be- tween these dies, which had profiled edges to fit the contour of the punch. During simolated forming, the poly- ethylene sheet was slowly pulled over the punch and was allowed to slip between the cardboard dies, Inked circles on the sheet helped to make the deformation (direction and magni- tude of strain) visible. ‘When the sheet was fully formed, it had stretched (elongated circles) along diagonal AC and compressed into wrinkles along diagonal BD. This duplicated, to some degree, the de- formation of the metal stamping. The next step in this analysis was development of the model shown in Figure 6. A section of the blank, A’BPO’, is merely folded down to form the wall ABPO. Strain within this area is practically zero. There is, ‘of course, a bending and unbending of the steel in this section as it passes from the blank over the die and then flattens out into the wall. The initial and final states can be visualized by the flap being folded down, however. A similar process takes place to form wall CDML. To maintain geo- metrical continuity of the blank, the middle section A’BCD’ therefore must deform to fill space ABCD of the final stamping, Further visualization shows that line A’C elongates to become diagonal ‘AC, From the geometry of the stamp- ing, Figure 7, a strain of 62. percent was required along the line AC. This agrees exceptionally well with an av- erage strain of 63 percent measured between A and C on the stamping. ‘The other diagonal, BD, had a calculated strain of minus 22 percent and a measured strain of minus 25 percent ‘The strains are generated only by the material being forced to conform to the geometrical shape of the punch, JULY 2068 6. DEFORMATION MODE in the instrument panel mounting plate shown in Figure 3,4 and 5. Tho shaded flaps are bent down, forcing material A’BCD’ to conform to area ABCD in the final part. Averoge strains ore calculated from this. PERCENT STRAIN=I00% Bf + 305-189 109 138 STRAIN = 62 PERCENT 7. MAXIMUM STRAIN along diagonal AC of the instrument panel mounting plate stamping can bo calculated from the geometry of the blank. This calcu- Toted strain corresponds with the actuel strain measured on the formed part. For this reason, changes in die radii, © blank size and lubrication would not sxeatly affect forming. Changes in Beometry of the stamping would have a great effect on the formability of the stamping but, since the geometry was fixed, the only way to avoid breakage ‘was to carefully control the properties of the material. Although the advent of strain mess- ‘urement with the aid of the etched- circle system represents an important forward step in the engineering evalu- ation of stampings, further research is needed to broaden our understanding of sheet metal formability, Work is being done to determine the effects of cold working, cleanliness, aging, yield point elongation and many other vari- ables. As research findings reach the stage of practical application, they will be reported in Machinery, aa Further Reading The author has prepared an extensive list of worthwhile technical articles, papers and books for use by readers who would like to broaden their understanding of sheet metal form- For a free copy of this bi liography, write to the editors of Machinery, 200 Madison Ave., New York, N. Y. 10016. 83