TEE UIVERSTIY OF MICHIGAN INDUSTRY PROGRAM OF THEI COLIEGE OF ENGIEERING PRECISION-CAST ORDNANCE COMPONENTS R. A. Flinn G, A. Colligan June, 1960 IP-435

(- L'/-l, / ~d;

TABLE OF CONTENTS Page SUMMARY 1 GENERAL INTRODUCTION 4 I. DEVELOPMENT AND TESTING COMPLETED 5 Cast Drive Sprocket, Part No. 8671597 5 II. DEVELOPMENT COMPLETED; UNDER TEST 8 A. Shell-Cast Ductile Iron Followers, Front and Rear, Parts Nos. 739510 and 7360356 8 B. Precision-Cast Final Drive Gear, No. 7364141 11 III. ENGINEERING AND DESIGN COMPLETED; READY FOR PRODUCTION OF TEST CASTINGS 14 A. Precision-Cast Crankshaft for AV-1790 Engine, No. 8717036 14 B. Crankcase for AV-1790 Engine 16 IV. UNDER ENGINEERING STUDY 18 A. Suspension and Spindle Arm, No. DTA 15910-15 18 B. Outer Race, No. 7384006 18 C. Cupola, No. C8671475 19 APPENDIX A 21 Detailed Discussion of Statistical Study, Precision-Cast Final Drive Gear, No. 7364141 22 APPENDIX B 25 Supplementary Experimental Work 26 (1) Effects of Mechanical Clamping of Shell Molds on Improving Dimensional Tolerances 26 (2) Surface Quality and the Role of Mold-Metal Interface Reactions 28 APPENDIX C 35 Detailed Process Description 36 (1) Cast Drive Sprocket, Part No. 8671597 36 (2) Precision-Cast Final Drive Gear, No. 7364141 37 ii

SUMMARY The development and use of precision-casting methods for large ordnance components results in great potential savings in cost, critical materials, and lead time, in greater flexibility of design, and in an increased number of sources. Requirements for complex and expensive machining facilities are reduced drastically. The results of the first phase of a five-year program illustrate how vital components such as large sprockets, crankshafts, drive gears, and suspension and spindle arms made by other more expensive methods may be replaced by accurate castings by using new processes. Standard practices for these advanced applications of precision-casting techniques have been developed, and commerical production sources have been stimulated by integrating the program with industrial research groups. At the same time much needed supporting work of a general nature dealing with improved dimensional tolerances and surface finish has been conducted at The University of Michigan. The first phase of a projected five-year program is reviewed here, and some of the components selected for investigation are in more advanced states of development than others. Furthermore, the initiation of the more recent projects has required the knowledge gained from the earlier castings. To summarize briefly some of the principal findings, the major components which have received attention will be discussed beginning with the most advanced developments as follows: (I) development and testing completed (drive sprocket, No. 8671597); (II) development completed; under test (front and rear followers, No. 7359510, No. 7360356; final drive gear, No. 7364141); (III) engineering and design completed, ready for production of test castings [crankshaft for AV-1790 engine, No. 8717036; crankcase (completed, no further work)]; (IV) under engineering study (suspension and spindle arm, No. DTA 15910-15; outer race ring, No. 7384006; cupola, No. C8671475). I. DEVELOPMENT AND TESTING COMPLETED Drive Sprocket, No. 8671597. —The drive sprocket is a vital part in every tank or track-propelled vehicle. In this section, the production of cast sprockets poured in graphite permanent molds is described. These components are used without any nachining of the teeth or of the bolt holes. The 28 —in, diameter is maintained to + 0.1%. A substantial reduction in cost would be realized by the use of drive sprockets, cast in graphite permanent molds. A 62% reduction in scrap material, far less lead time, great flexibility in chemistry, and improved service life are obtained with the cast sprocket compared with the fabricated type, 1

'T DEVELOPMENT COMPLETED; UNDER TEST A, Front and Rear Followers Nos. 7360356 and 7359510o —At present the follower rings are made in an 83% copper, aluminum bronze. A shell-molding technique has been developed for producing these in the very low strategic index material, ductile iron, While preliminary tests indicated some scoring, new castings with superior properties with proper break-in surface coatings are expected to yield satisfactory results, At the time this project was initiated, this was the heaviest shell-molded casting in production. B. Final Drive Gear, No, 7364141, —Following the satisfactory production of sprockets, it was natural to attempt other large components with accurately cast teethe Accordingly, final drive gears have been produced in permanent molds which require fewer machining operations than the fabricated type. Tests of carburized and induction-hardened types have been initiated by Ordnance Tank-Automotive Command in the near future, III, ENGINEERING AND DESIGN COMPLETED, READY FOR PRODUCTION OF TEST CASTINGS Ao Crankshaft for AV-1790 Tank Engine, No, 8717036. —The production of precision-cast steel crankshafts offers very great potential savings in machining costs and critical materials. Although the present fabricated shaft requires extensive machining, its design is still severly limited. Experimental stress analyses and preliminary castings indicate a potential saving in cost and substantial reduction of regions of stress concentration by using the cast design, Production of this part by combined permanent mold and ceramic core techniques is strongly recommended for future works B, Crankcase for AV-1790 Tank Engine, —Improvement of the performance of tank engines has led to the use of very heavy (4-in0-thick) sections of aluminum alloys in the crankcase, Because of the susceptibility of these heavy castings to porosity and shrinkage, only very low design stresses (3500 psi) can be used. A lighters stronger crankcase using ductile iron or steel has been designed, Since no new application of precision processes was found necessary, this project is considered complete and ready for adoption by the manufacturing agency, IV. UNDER ENGINEERING STUDY A, Suspension and Spindle Arm, No, DTA 15910-15, —This is an exceedingly complex and expensive fabrication. It is anticipated that it can be produced much more economically, while increasing its strength, by further development of shell-molding techniques, Future work is recommended, 2

B. Outer Race Ring, No. 7384006. —If this 93-3/4-in.-diameter ring can be produced as a precision casting by extension of the technique developed for the sprocket, very substantial savings in cost and material will be obtained. Preparation and testing of pilot castings are recommended. C. Cupola, No. C8671475. —The present production of this part as a sand casting entails difficulties with dimensions and finish. The use of precision semi-permanent mold techniques should improve the product and result in reduced overall cost. 3

GENERAL INTRODUCTION Before considering the individual developments in detail, it may be helpful to review the background of this project. The Ordnance Tank-Automotive Command of the Department of the Army, realizing the rapid improvements taking place in castings, placed this contract to investigate the use of precision-casting methods for ordnance components. This work has been conducted at The University of Michigan to avoid overemphasis of any particular method. At the same time industrial sources have been developed for all the principal parts concerned so that, at the successful conclusion of a division of the work, production facilities would be immediately available to make the part under study. Only in one case (the follower castings) has it been necessary to do the work at The University of Michigan; commercial companies were unwilling to attempt this casting because of its large size. Even in this case the development has been conducted using a commercial shell-molding machine so that, upon completion, the pattern equipment could be transferred immediately to production. The selection of parts for development and the engineering studies have been conducted for the most part at The University of Michigan with the cooperation of personnel of the Ordnance Tank-Automotive Commando Several thousand parts received preliminary investigation before the final selection of parts was actually made. The careful detailed engineering work of the sprocket and final drive gear by personnel of the Griffin Wheel Company, and of the crankshaft by personnel of the Continental Aviation and Engineering Company, is gratefully acknowledged. As the investigation has progressed to include components of greater weight and complexity, certain supporting work has been necessary, such as the investigation of the deflections of shell molds during and after pouring and the reduction of harmful surface reactions between the liquid metal and the mold. Most of this work has been done with little cost to the Department of the Army as doctoral research work, it is available as background material when difficult casting problems arise. The specific parts of the investigation may now be considered in the same order as outlined in the Summary, from the completed investigations to those in the intermediate stages, For clarity and ease of comparison the same order of discussion will be used for each project insofar as data are available: (1) Purpose of the study (i.e., why was the casting selected for study); (2) The progress developed; (3) Properties of castings made by the new method (dimensions, stress analysis, etc.o) (4) Service performance; and (5) Advantages of the new method. 4

I. DEVELOPMENT AND TESTING COMPLETED Cast Drive Sprocket, Part No. 8671597 1. PURPOSE OF THE STUDY A review of the fabrication methods and the service performance of fabricated sprockets indicated great potential savings if a method could be developed for casting this part to avoid machining of the teeth and counter-bored bolt holes. The authors had had experience with semi-permanent graphite molds for railroad car wheels, and the Griffin Wheel Company was accordingly asked to consider producing sprockets in graphite molds. 2. THE PROCESS DEVELOPED Mold Design To emphasize the reliability of the new process, a related method, in successful commercial production for railroad car wheels, is illustrated in Fig. 1. In this technique a semi-permanent graphite mold is clamped above a pouring tube which dips into a covered ladle of liquid steel. Air pressure is applied at a controlled rate to the ladle and the metal is forced up the pouring tube and into the graphite mold at a smooth, predetermined rate. The resulting castings are free from slag and mechanical defects which are caused by irregular metal flow. Hundreds of thousands of these wheels are in highly stressed service with as-cast treads of this type. Several interesting developments were necessary to obtain a satisfactory mold for sprockets by application of this method. To form the sprocket teeth, mold inserts were machined. Originally these were made from segments of graphite but finally a one-piece gray-iron insert was found to be completely satisfactory. In this way simple and economical maintenance of the critical tooth surfaces was possible. The details of the mold design are indicated in Fig. 2. The bolt holes were produced using zircon-sand shell cores bonded with 4% liquid phenolic resin. The gating system is superior to that necessary for the car wheel. In place of the consumable stopper rod, used to prevent the liquid metal from running back into the ladle when the pressure is released, a restraining dam is employed. In the case of the sprocket, the metal rises in the center and flows over the dam. The pouring rate is 13 lb/sec. When the casting is filled, the pressure is released and only the excess metal drains back into the ladle. 5

As a result of these techniques, excellent mold life was obtained. No replacement of any parts was required and the mold was in excellent condition after pouring fifty castings. The final appearance of the gray-iron insert is illustrated in Fig. 3. Melting Analysis Control Melting was conducted in a 1000-lb high-frequency induction furnace for all the sprockets supplied for test. To show the general applicability to production, however, metal from a 20-ton arc furnace was used with equal success. In addition to the AISI 4150 analysis, sprockets were also poured successfully in plain carbon steel (0.70% C) and in cast armor analysis- This demonstrates rather wide flexibility in analysis. The surface obtained in cast armor is shown in Fig. 4. The analyses of the castings which were submitted for test are given in Table I and are all within the specification for AISI 4150 indicated for this part. 3. PROPERTIES OF CASTINGS MADE BY THE NEW METHOD Dimensional Variations of Castings: X-Ray Inspection Measurements made of three critical dimensions (the width of the base of each tooth, the radius to tooth root, and the radius to tooth tip) are listed in Tables II, III, and IV, respectively, for all castings shipped. To obtain a measure of the variation of these dimensions caused by the casting process, and to eliminate those which could be corrected by remachining of the gray-iron tooth chiller, the data of Figs. 5, 6, and 7 were developed. These graphs indicate that the variations of a given tooth, within 95% confidence limits, are as follows: Tooth-base width: ~ 0.005 in. Tooth-root radius: + 0.011 in. Tooth-tip radius: + 0.011 in. It should be emphasized that these are total variations in relatively large dimensions, not in in./in. These values compare very favorably, therefore, with other precision-casting processes. Sprocket Heat Treatment Pretreatment.-After casting, all test sprockets were heated to 15500F, held two hours and furnace-cooled. This resulted in stress relief and a homogeneous structure for flame hardening. 6

Flame-Hardening. —All sprockets were flame-hardened by the Detroit Flame Hardening Company. A typical hardness survey of a new tooth is shown in Table V. The data are well within the specification for this part. 4. SERVICE PERFORMANCE Cast sprockets were placed in comparative service tests with fabricated parts on the same tank at Milford, Michigan; Aberdeen, Maryland; Fort Stewart, Georgia; Fort Churchill, Canada; and Yuma, Arizona. The results of the tests to date are given in Table VI. It is evident that the cast sprockets provided better service life in all cases shown. In another instance, Fort Stewart, Georgia, an unexpected wear pattern was encountered as illustrated in Fig. 8. The peculiar clay and angular quartz soil in the Georgia installation resulted in the unusual pattern. The highly abrasive quartz-clay mixture packs beneath the track and remains on the sprocket. This condition was easily remedied by a small change in the flamehardening treatment and castings prepared in this manner are being tested. In general, the superior performance of the cast material is ascribed to the following causes: (a) fine grain structure and lack of the banding which is encountered in the fabricated sprockets (Fig. 9); and (b) greater hardenability. Deep hardenability is not desired by the fabricator of flame-cut sprockets because of the possibility of cracking during torch-cutting. The cast sprockets can be produced without trouble with any desired hardenability. In the cast sprocket both greater hardenability and absolute hardness can be obtained by increased carbon content, thereby saving critical alloys and reducing cost. Castings of this type are now being tested by Chrysler in another project which was initiated as a result of this work. 5. ADVANTAGES OF THE NEW METHOD (a) Material conservation: The data of Table VII indicate a 48% yield in the case of the cast sprocket compared to a 26% yield for the fabricated sprocket. This is a very conservative calculation since it does not take into account the substantial material losses in passing from ingot to rolled plate. (b) Lead time and production steps: The production steps for the cast sprocket are compared with the fabricated product in Fig. 10. 7

I

(c) Cost: The potential savings by use of the cast method are illustrated in Table VII o While these amounts cannot be expected in pilot production, they represent potential, savings in full-scale production quantities in this and other vehicles. (d) Material Specifications: In addition to inherent economies of precisioncast sprockets there is no limit with respect to steel specification. By casting, a steel of any desired hardenability can be produced as a longer wearing sprocket. To use higher hardenability steels for a fabricated sprocket would substantially increase cost, delay procurement and further reduce the yield. Procurement in particular would be a problem since 8600 and 4100 series steels are not normally produced in plate form. 7b

II, DEVELOPMENT COMPLETED; UNDER TEST A, Shell-Cast Ductile Iron Followers, Front and Rear, Parts Nos, 7359510 and 7360356 1, PURPOSE OF THE STUDY The follower castings illustrated in Fig. 11 are used in the recoil mechanism of the 90-mm rifle of the M-48 tank, The present material (Spec, QQ-B671.-B) is cast aluminum bronze of the following composition Cu: 83% Al: 10l1.1l 5 Fe: 2-5 0% Mn: 0,5% Ni: 2.5% This cast;ing was selected for study for two reasons: (1) to determine if a noncritical ferrous alloy such as ductile iron could be substituted for the more expensive aluminum bronze, and (2) to investigate the use of shell molding for a larger casting with heavy sections as well, At the time this project was initiated, the heaviest commercial casting was an automobile crankshaft (70 lb) with only 1-1./2 in, sections, This casting provides 3-4 in, sections, and the pouring weight is 170 l.b for the front follower N o 73603560 2, THE PROCESS DEVELOPED Ductile iron. castings were produced successfully in shell molds, using ithe technique illustrated in Figs, 1.1. 12^ and 13, Mold. Design The mold design developed, which employs only one riser, is shown in Fig, 12, Various riser designs were employed in the investigation, but the one illustrated provided consistently satisfactory castings, The molds are produced on a -ommerical model Shell Process Company machine, Fig, 13, The shellmolding method is described below, Sand-Resin Mix 94% sand 6% resin (by weight) Solvent - 30o of resin content, 3 parts Ethyl or methyl alcohol, 1 part H20 8

Procedure for Mixing Place sand in muller. Add resin. Mix for 30 seconds. Add solvent. Mull until lumps are broken down to almost the original screen fineness. Remove from muller and screen to aerate the mix. Remarks Over-mulling should be avoided, i.e., do not mull until the mix is thoroughly dry. The coating is complete in approximately 3 minutes in a Simpson muller. Further mulling serves to evaporate the solvent. Air blown into the mix to accelerate the vaporization of the solvent should not be introduced until the coating of the sand grains is complete. A light stream is preferable; a heavy stream may tend to dry out the solvent in certain areas too rapidly, causing the resin on the sand grains to break away during the subsequent abrading of the sand grains. The sand should be thoroughly dry before being placed in the investment chamber. Otherwise, it will tend to pack, reducing its flowability and causing a mold to be of poor quality. Production Cycle Preheat pattern to 450~F. Spray pattern with release agent (Union Carbide LS-46 silicone). Invest pattern for 13 sec. Remove to oven; cure for 60 sec at 7000F. Repeat above cycle, then eject shell. Shell thickness approximately 1/2 in. Melting Practice The melts were made in a 200-lb high-frequency induction furnace by a practice which provides casting properties duplicating those of standard commercial sources. Standard charge materials were employed and the melts were inoculated with 1-1/2% alloy addition of 9% magnesium cerium ferrosilicon following accepted commercial practice. The shells were embedded in green sand and poured at 2550'F in 9 sec. After cleaning, inspecting, and cutting off risers, the castings were machined commercially. 9

3. PROPERTIES OF CASTINGS MADE BY THE NEW METHOD The use of shell molds resulted in castings free from burnt-on sand and with good surface appearance. The experiments illustrate that ductile iron sections of 3-4 in. can be cast satisfactorily in shell molds. The structure of the castings was determined metallographically and consisted of a matrix of 5-15% pearlite, 85-95% ferrite with spheroidal graphite (Fig. 14). It was decided to use the as-cast structure for initial tests since it is the least expensive to obtain. Microscopic examination, chemical analysis, X-ray and cobalt-60 radiography were used to insure sound castings of the desired structure. 4. SERVICE PERFORMANCE Firing tests were conducted at Erie Ordnance Dept., Port Clinton, Ohio, with the following conclusions. Scoring of the follower castings by the gun tube took place. There was some pickup of metal by the gun tube; the scoring was confined to plastic flow of the ductile iron as illustrated in Fig. 14. The ferrite in the ductile iron was apparently too soft a structure for this application. To avoid this effect, three new sets of followers have been prepared with the following characteristics: Matrix - 90-98% pearlite, balance ferrite and spheroidal graphite (Fig- 15). Surface coating to assist in break-in and avoid scoring: (a) Teflon-coated; the Teflon is bonded with DJ-855 phenolic resin by baking at 300~F, 1 hr. (b) Tin-plated - Type I, Spec. Mil-T-10727. (c) No surface coating, normalized heat treatment. 5. ADVANTAGES OF THE NEW METHOD The results of these tests should be available within two months to permit a final evaluation of the new method of production. In any event, the experiments indicate a satisfactory shell-molding technique for heavy sections of ductile iron. 10

B. Precision-Cast Final Drive Gear, No. 7364141 1. PURPOSE OF THE STUDY The purpose of this work was to determine whether final drive gears (Fig. 16) could be cast accurately by the permanent mold technique and thereby eliminate the rough machining operations nqw required. If successful, substantial savings in material, expensive machining equipment, and lead time, as well as greater flexibility in analysis can be achieved. Present gears are machined from AISI 4817 forged blanks and carburized. In these experiments cast gears of this material were made, machined, and carburized. In addition, advantage was taken of the precision-casting method to produce other gears of AISI 1062 for finish-machining and rapid induction-hardening, thereby eliminating carburizing. 2. THE PROCESS DEVELOPED The method used for gear production is somewhat similar to that used for sprockets and therefore does not require detailed description. This work was done under subcontract by the Griffin Wheel Company. Graphite molds were used for the body of the mold, and gray-iron inserts were machined for the teeth. Full details of the mold assembly are given in Fig. 17. Following preliminary work and the recommendations of the Brad Foote Gear Company, the following stock was allowed on the teeth: (a) For carburized AISI 4817 gears, 0.060-0.070 in. (b) For induction-hardened AISI 1062 gears, 0.030-0.040 in. Ten gears of AISI 4817 and eighteen gears of AISI 1062 were produced for shipment. Table VIII contains a summary of all experimental castings produced during this investigation and their final disposition. Table IX contains the chemical analysis of all test castings. It should be noted that the rough cut prior to carburizing on castings Nos. 62, 66, 69, and 70 resulted in insufficient stock after carburizing. The special quench and temper pretreatment used for forging blanks was not specified for precision castings and was the cause of the loss of these four gears. Proper pretreatment or modified machining procedures will eliminate this problem. Figure 18 is a photograph of an as-cast AISI 1062 final drive gear prior to any machining operations. Figure 19 is a photograph of an AISI 4817 cast gear after maching and carburizing, together with its mating pinion. The excellent surface and detail are obvious. 11

Tables X, XI, and XII contain the summaries of all dimensional studies made on these cast gears. These data include pin-diameter measurements in Tables X and XI, and in Table XII; A-1 through A-4: plate thickness, J-1 and J-2: hub diameter, L-l and L-2: outside diameter, and P-l through P-4: pin thickness. Appendix A contains both the data from which these summary tables were derived and a detailed discussion of the statistical analysis performed by the Griffin Wheel Company. Table VIII, the heat and casting summary, shows a great number of castings which were considered unacceptable due to an apparent cold-metal wrinkle on one or more teeth. Although this condition was improved somewhat when the mold tilt angle was changed from 3-1/2" to 2~, the loss was still considerable. For accurate evaluation, a gear casting with a typical wrinkle was machined to determine the depth of the defect. Pin-diameter measurements indicated an average of.030in. stock on the tooth surface. Inspection of the finish-machined gear indicated that an additional.030 in. would have reduced the scrap loss to 5o. Limitation of the graphite-mold and pressure-pouring process was defined by statistical analysis of measurements taken from 37 gear castings. Rim, plate, and hub dimensions were all within practical requirements as designed. Stock on the surface of the teeth was found to be inadequate, as originally designed. Data showed that a minimum of.060-in. stock should be provided on the face of each tooth to compensate for surface blemishes. Differences in pin-diameter measurements due to process variations were defined as +.032 in. at the 95% confidence level. Use of these figures makes it possible to design a gear casting which might be produced on a practical production basis, to define the accuracy of the casting, and to specify the amount of stock provided for machining. Reaction of the gears to the hardening treatment has indicated that investigation of procedures for pre-heat-treatment of the casting before carburizing or induction-hardening would be worth while. 3. ADVANTAGES OF THE NEW METHOD This development, together with that of the sprocket, indicates that any desired steel analysis can be poured in graphite molds. There is no appreciable carburization of the steel by the graphite mold as determined by careful metallographic inspection at 1000 diameters, 12

The gear castings indicate further that sections up to 3 in. can be cast accurately and with excellent surface finish. The grain structure of the cast gears will be free from any banding or other forging defects and should provide equivalent or better service, The precise casting operation eliminates a lengthy and expensive rough-hobbing operation of approximately two hours. Furthermore, if the unalloyed induction-hardened AISI 1062 gears are successful, a marked saving in critical materials will be realized. 13

III. ENGINEERING AND DESIGN COMPLETED; READY FOR PRODUCTION OF TEST CASTINGS A. Precision-Cast Crankshaft for AV-1790 Engine, No. 8717036 1. PURPOSE OF THE STUDY One of the outstanding advances in design and cost reduction of components for automobile engines has been accomplished by the use of precision-cast crank-shafts. By the replacement of wrought parts, greater flexibility of design and reduction in machining has been possible. It was natural, therefor, to consider the higher-performance alloy-steel crankshaft of the AV-1790 tank engine, which is 350% heavier than the automotive shaft, as a proper subject for engineering study. The AV-1790 tank-engine crankshaft poses many more difficult problems than automobile crankshafts, which are cast in malleable or ductile iron in relatively light sections. Shell-molding techniques are available for this work. By contrast, the high-performance AV-1790 engine requires an alloysteel crankshaft cast in heavy sections. No techniques were available for economical precision castings of this type at the start of the project. It now appears, however, that the part may be precision-cast through new technique developed on other components and in supplementary research. In the meantime, the critical sections of the crankshaft (journal and bearings) have been cast, subjected to stress analysis and redesigned. The excellent cooperation of the Continental Aviation and Engineering Corporation in the design and stress analysis work is gratefully acknowledged, A new, very promising design for casting has been developed. 2. THE PROCESS DEVELOPED Since this phase of the project has just passed the design stage, no complete crankshafts have yet been made. Discussion will, therefore, be confined to the production and testing of critical sections. The present forged and machined design is illustrated in Fig. 20. The design changes to take advantage of the casting process are illustrated in Fig. 20 and in the model of Fig. 21. Castings of the model were made of AISI 4340 steel using merely a good dry-sand molding technique. These provided test sections for stress analysis and were not intented as precision castings. 14

Previous work: at Continental Aviat i.o -i. ie., i.i r-::n CorLorat ion nau proved that torsion was not critical and that iendin,-,tress was the important consideration. It was prossi'bl theirefiore, to work with sirngle-throw sections of tihe complete shaft, Followingi the accepted methods of scress lysis al the regionsK of hilr heost stress were fiirst located with stre:...., arid c ti then resistance-type stra.iri gaes were employed for lmore accurate O':t.... s The manner of loadiin the test section is llustrated in Fig. 22. This procedure was developed in previous CAC research, The stresscoat indications -omparin- the fore d vs, the cast, ls. si1gns are shown in Figs, 23 and 24. Thl ma:n.Lm sitr;ain in the rast desrign is reduced to 5735 in.l/i:n.;rmpai.e vdi t! 6-'L,.g c'.;' 4: forgcei desirn —a reduction of 14.2^ 4 t u..ri^-iri,...Oj'.e ul'.Ji.. A.....,'ii/n.L.:..: tp e )- Jiai e r eu.ctions arc obtained at other stresses in the e ciasL. d.t:. t r tia i.cL with SR-4 resistance strain s:agaes as honir iin.Lg, 2 Additional determinations were made to find out if the hole in the crank cheek could be enlarged to simplify casting procedures. The data from these tests indicattd a 7/8-in.-diameter hole can be used in place of the present: 1/2-. in, hole if desired. The CAC investigation poirnired out; that still further improvement couldl be obtained by::-:eiifi3-:ti on of'the re-entrant fillets. 3. ADVANTAGES OF THE TEW MIETTIHOD The engineering study just described indicates that full-si: ci ast crankshafts should -be mrade and teste, becas e of ^ o olo.wing advanta es (a)! l{ct;i o0s in cost anJd. irn _ s of1 cri:1.:a matchine -ools In Figr 26 the surfaces reruiri- i a.cniliig or the cast crankshaft are the important Learinig surfaces Vwiic ari re: ll-a1uivecly sicmple to finish lihe tedious and expensive dri.lli., oi holes ro _i.-,r l the journals alnd the machiriing of the couinterweights has beeni eliminated, Again it should be emphasized that the design principles developed, for this shaft may be readily applied to any other similar shaft (after investig-atinprincipal produ-ction Iet:hods ad -co'firmation by the full -scale tests pro jected for this engine ) 1.,

(b) Reduction in lead time, complex machining, flexibility of analysis, and saving in critical material In many cases substantial delays are encountered because of shortages of quality steel billet stock such as AISI 4340. These can be circumvented by passing directly from raw materials to castings. In addition, other alloys may be easily substituted when desired in this process. As also shown by Table XIII a 135 increase in yield of critical material will result from the casting process. The elimination of complex machining, such as the special equipment needed for drilling of the crank pin holes in line for 5 ft, is also an important production advantage. (c) Stress reduction The substantial reduction in operating stresses in critical locations can be used either for lighter design or for greater safety under severe conditions such as fatigue i:n view of these pronounced advantages, it is strongly recommended that provision in future work be made for manufacture and testing of precision-cast crankshafts B. Crankcase for AV-1790 Engine 1. PURIPOSE OF THE STUTDY Tehe requirements for greater load-carrying ability in the crankcase sections have led to very heavy (4-in.-thick) designs. Because of the porosity and shrinkage encountered in sections of this size in the alloy used, only very low design stresses (3500 psi) are possible. Lt is believed that a lighter, stronger crankcase can be made in either ductile iron or steel, urtilizing the much higher, reliable design stresses for these materials, particularly in lighter sections. With this end in mind the crankcase was redesigned. 2. THE DES IGN DEVELOPED Since this study reached only the design stage, no actual experimental castings were produced. The present cast aluminum crankcase is shown in Fig. 27a and b. Figure 28 shows the cross section of the present crankcase and the new design for ferrous alloys' The lighter design of the new crankcase is evident from the drawings. A pattern of a section of the new design has been constructed. 16

It is recommended that this section be cast in steel and subjected to stress analysis before proceeding to full-size castings. An additional advantage of the new design is that repetitive segments could be cast individually and welded togethero This would simplify casting production and inspection. It was decided to leave further crankcase development to the manufacturer since precision castings were not required. The Superior Steel Casting Company can produce and weld the cast sections. 17

IV. UNDER ENGINEERING STUDY As the earlier studies have progressed, the new techniques which have been developed have led to the reconsideration of several parts previously considered too complex for immediate action. These new parts include: (A) Suspension and spindle arm, No. DTA 15910-15; (B) Outer race ring, No. 7384006; and (C) Cupola, No. C 8671475. A. Suspension and Spindle Arm, No. DTA 15910-15 This complex fabricated component is illustrated in Fig. 29. Since it must be made at present as a large forging and weldment, the production is quite complex and cumbersome. With the cooperation of the Research Laboratory of the American Steel Foundries, and at no expense to the Government, a simple redesign has been evolved. The very heavy 4-by-4-in. solid section in the main arm can be converted to a lighter box section by coring of the casting. It is anticipated that, by the use of advanced shell-molding technique and the development of the proper shell-mold assembly, this part can be produced as a single steel casting. This will eliminate the problem of pressing the spindle as well as the subsequent alignment problems in service. The savings in lead time, cost, and weight of critical materials should be very substantial. B. Outer Race, No. 7384006 This component (Fig. 30) is now made by extensive machining of a forged ring of AISI 4150 steel. The internal gear is then hardened by flame- or inductionhardening techniques. It is proposed to develop in place of this complex forging a precision casting using combined graphite-mold and ceramic-core techniques. The problem is more advanced than that of the sprocket because of the greater size (8-ft diameter) and because the teeth are internal. The successful solution of this problem would then permit the use of the graphite-permanent-mold technique for practically any gear or sprocket shape. The advantages should be at least as great as those demonstrated for the sprocket. 18

An alternate method, employing separately cast segments, is also under study at the present time. C. Cupola, No. C8671475 This part is now made as a steel casting. Because of the heavy sections, the part is subject to such defects as burnt-on sand related to the mold-metal interface reaction discussed later in this report, Furthermore in large castings of this type, considerable finishing is required to obtain the required dimensions. It is proposed to obtain better mechanical properties and closer dimensional tolerances by employing permanent-mold techniques. The rapid freezing coupled with the relatively inert mold surface should be very effective in accomplishing this result. Design modifications, such as elimination of projections which lead to hot tears, are also contemplated. 19

CONCLUSIONS: RECOMMENDATIONS FOR FUTURE WORK Based upon the data of the individual studies which have been summarized in this report, the following conclusions are justified: (1) The use of precision castings in an appreciable number of ordnance components results in substantial savings in cost, critical materials, and lead time, and to reduced use of critical machining and fabricating equipment. (2) Greater flexibility in design and better stress distribution and load-carrying ability can be achieved in cast designs than are possible with fabricated assemblies. (3) The development of alternate specifications and production methods greatly augments the number of commercial sources available in time of emergency. It is recommended, therefore, that this development of precision castings for ordnance be continued along the lines indicated in this report to include particularly the crankshaft, the suspension and spindle arm, and other vital components. It should be evident from the extensive work described in the text that these studies are not at all routine but require the fullest development effort of the group of industrial and university personnel which has been assembled. In other words, the original thought of Ordnance Tank-Automotive Command-namely, that the proper application of precision-casting processes for ordnance poses many problems requiring special development effort-seems adequately confirmed. It is confidently anticipated that future effort along those lines will be even more rewarding. 20

APPENDIX A

APPENDIX A DETAILED DISCUSSION OF STATISTICAL STUDY PRECISION-CAST FINAL DRIVE GEAR, NO. 7364141 EXTRACTED FROM GRIFFIN WHEEL CO. FINAL REPORT It is believed that a meaningful definition of the variation which might be expected in any lot of gears produced by pressure-pouring in graphite molds might be had by analysis of measurements taken from 37 castings produced on this project. Forty-five specific measurements for each casting are shown on the data sheet accompanying this report. Analysis of pin diameters is considered to be the most pertinent due to the close dimensional control and freedom from defect specified on the gear teeth. Rim thickness, plate thickness, hub diameter, and outside diameter will also be discussed. The 37 gear castings from which measurements were taken fall into two groups. Castings Nos. 17 through 54 were produced in two molds designed to produce castings having pin diameters large enough to provide approximately.030-in. finish on the surface of each tooth. Castings Nos. 58 through 70 were made in molds which had been enlarged to provide extra finish on the SAE 4817 gear castings. All dimensions other than pin diameter were the same for both lots. The 33 pin diameters for each costing in the group Nos. 17 through 54 were measured and recorded on the data sheet. The mean and standard deviation of the 33 diameters of each casting were then calculated and are presented in Table X. The mean pin diameters of each casting were then averaged and the standard deviation was calculated. These values also appear in Table X. The same procedure was applied to the measurements taken from the second group, castings Nos. 58 through 70, and recorded in Table XI. The standard deviation of the individual casting mean diameters and the average of the standard deviation values for the individual casting were found to be identical for both groups. The similarity of these figures is an indication of the accuracy which might be expected in relation to different products of generally similar size. The data presented in Tables X and XI also allow a statistical evaluation of the uniformity of the graphite-mold and pressure-pouring process in relation to the final drive gear. Using a confidence level of + 2 standard deviations from the mean, it can be stated that, if similar samples of gear casting were studied, 95% of the mean pin diameters of individual castings would not vary more than +.012 in. Going further, 95% of the individual mean pin diameters would not vary from the arithmetic average of the individual means by more than +.020 in. If the variation of the average of individual mean pin diameters is then corrected by the amount of tooth-to-tooth variation of the individual casting, the total 22

error which might be expected at the 95% confidence level becomes ~ 0032 in. or a range of.064 in, If this value is added to a theoretical pin diameter calculated to provide o060-in. finish on the surface of each tooth, it becomes possible to design a casting which has sufficient stock to compensate for process variations and surface defects. Recommended pin diameter for future production of Ordnance part No. 7364141, using pressure-pouring in graphite molds, is 20.034 in and is calculated from the data presented as follows. recommended Finish pin diameter + base stock per tooth + process variation = design 19.730 + 4( o060 in.) + o064 ino = 200034 in, Also calculated and shown in Tables X and XI is the average amount of stock on the tooth surfaces of the individual castings produced on this project. This figure is obtained by dividing the difference of the mean pin diameter and theoretical pin diameter by 4. It will be noted that for the entire lot of 25 castings the average stock is.027 in. The chiller ring used to produce this group was designed to produce o030-in, stock, indicating a design error of only.003 in. Since time was short, the same gear-cutting tool was used to enlarge the molds which were used to produce castings Nos. 58 through 70. Although the outside diameter of the casting was increased over,100 ino., Table XI shows an average increase of tooth-surface stock from.021 to o048 in., thus falling.012 in. short of the desired figureo The dimensions for rim thickness and plate thickness have specified tolerances of ~ 1/32 or.031 in. Table XII lists the arithmetic mean of the measurements taken at these locations from the sample of 37 castings. The mean plate thickness was found to be.870 in. with a standard deviation of.009 in. Comparing these values to the specified o875 in. ~.031 ino, it is determined that under similar conditions there is a chance of less than three castings in a thousand (+ 3 standard deviations) falling outside the specified range of from.844 to.906 in. Mean rim thickness determined from the 37 castings measured was 3.249 in. with. a standard deviation of.007 in. Adding o010-in stock for correction of parallelism errors, these values indicate that rim-thickness variation in similar samples would still be within the 2.219- to 2 281-in. range specified. Variation of rim thickness can be caused by foreign particles at the mold parting line, whereas uniformity of plate thickness can. be affected by inaccuracies in machining and subsequent remachining of the graphite mold, Both car. be affected by variation in thickness of the refractory coating put on the mold, Hub-diameter variation, given in Table XII, is of the same magnitude as reported for plate and rim thickness; and is well within the requirements for this area 23

Outside diameter was measured at only two locations on each casting. This measurement was difficult to obtain because the small amount of surface on the tooth tip did not afford accurate positioning points for the measuring instruments. Because of this it is believed the data obtained are not a true measure of the process. It is submitted that the foregoing information, documented by the data presented in the accompanying charts and illustrations, defines the limitations of the process in relation to casting design and product uniformity and proves the feasibility of producing the final drive gear by pressure-pouring in graphite molds. 24

APPENDIX B

APPENDIX B SUPPLEMENTARY EXPERIMENTAL WORK In addition to the development of technique for specific parts, supplementary investigations whose results have wide application were performed. This effect may be summarized now under two principal studies. (1) Effects of Mechanical Clamping of Shell Molds on Improving Dimensional Tolerances PURPOSE OF THE STUDY The problem of dimensional accuracy in shell molds is not merely a function of proper design of the pattern and of molding technique. Bulging, particularly of flat surfaces can be encountered during solidification and can lead to inaccurate dimensions. One of the first castings in which this was noted was the follower, discussed in Section IIA, Representative measurements showing distortion of as great as 0.033 in. are indicated in Table XIV. Conventional methods of gluing the shells, weighting with metal shot, or ramming in place with green sand did not improve this condition. Although this problem seemed most important with graphitic materials, it was decided to investigate cast steel as well as gray and ductile iron in a quantitative manner. A step test bar which had been developed for dimensional studies in previous research was selected (Fig, 31) Castings were poured from induction-melted heats of the following aim analysis: % C % Mn % P % S % Si % Mg Gray iron 3.60 0.30 0.05 0o05 2.20 Ductile iron 3.60 0.30 0.05 0.01 2.20 0.05 Cast steel 0.25 0,30 0.05 0.05 0.30 Melting practice followed procedures already discussed. The shell molds were subjected to different backing procedures. In addition, some molds were made by the "CO2 process" in which a very hard mold is produced by a silica bond. Dead-weight types of backing included metal shot (Fig. 32), green sand, iron weights, and sand, plus weights-all conventional foundry procedures. 26

A new method was developed using dynamic backup by employing calibrated spring clamps as illustrated in Fig. 33. The shell was ground to fit the steel bars and the end springs were set to the desired load. The three different alloys were first cast in the conventionally backed molds with the results shown in Tables XV, XVI, and XVII. In these experiments the variations in the thickness of the large step may be summarized: - deviation (in.) to + deviation (ino) Gray iron - 0.002 + 0,004 Ductile iron + 0.015 + 0.053 Steel - 0.007 + 0,023 Since the greatest problem and most consistent bulging encountered with ductile iron, further work was confined to this material, The effects of conventional weighting techniques and the new spring clamping may now be discussed. Changes in conventional weighting techniques or the use of very hard C02 molds had little effect upon bulging. In contrast, the data of Table II indicate reliable and consistent improvement by using the dynamic clamping method. The graph of Fig. 34, based on the data of Table XVII, indicates the decrease in bulge with applied spring pressure, At a total force of 240 lb, no bulging was encountered. It is also evident that the dimension "A" at the end of the step was less than for the conventionally weighted castings. In other words, ordinarily the entire casting bulges due to graphite precipitation in the final phases of solidification, but the reaction is more pronounced at the center due to the hot spot. CORRELATION OF RISERING WITH BULGING In all the castings discussed so far, the risering was adequate to prevent shrinkage defects. However, a question arose from the preceding data concerning the effect of clamping upon feed-metal requirements: if a casting is clamped to prevent bulging, would not the feed-metal requirements be reduced? If so, higher casting yields as well as greater precision could be expected from rigidly clamped molds, To test this theory, the riser sections of test molds were stopped off and a narrow runner was provided from the dowpsprue, A series of riserless ductile iron castings was then poured with different spring loading. The density of the castings was then measured and plotted vs. the amount of bulge (Fig. 34), Extrapolating the graph to zero bulge, the line approaches the density of a completely sound, risered casting, A20-3. 27

While this phase of the program represents only a preliminary study, the following facts are established: (a) Control of bulging will reduce risering requirements. (b) In riserless castings that are nearly sound, clamping may produce complete soundness (c) The problem of bulging can be serious in small ductile iron castings. It also seems reasonable to suggest that large shell castings of any type) even if the metal is skin-forming, will be subject to bulging. (d) Normal variations in pouring temperature have little or no effecto (e) A rigid type of mold backing does not supply sufficient protection. (f) The means of controlling bulging have been developedo Dynamic steel clamps employing springs to provide restraining force will give excellent results. The necessary number of clamps and forces will vary with the application. (2) Surface Quality and the Roleof Mold-Metal Interface Reactions PURPOSE OF THE STUDY Even though very accurate patterns and molds are produced, there is no assurance that accurate castings will result unless the reaction of the molten irmital with the mold surface is under control. To control and minimize the mold-metal interface reaction, it is first necessary to understand what is occurring. Iwo important investigations bearing on this problem will be reviewed: the ironsilica interface reaction, and reaction of silica with iron-carbon alloys. The first is of considerable importance in understanding and improving the surface of steel castings; the second involves cast iron and ductile iron. THE IRON-SILICA INTERFACE REACTION The quality of the surface as well as the dimensional accuracy of a casting depends on the extent of the reaction between the metal and the mold material. Some of the variables affecting the reaction are the compositions of metal and sand, pouring temperature and casting cooling rate, metal pressure, and the mold atmosphere. The system iron-silicon-oxygen-carbon was chosen for study because iron and 28

silica and a gaseous atmosphere are the principal reactants in the mold-metal interface reactions. Once the mechanism of a simple system is known, the effects of additional elements may be explored. Each specimen consisted of a mixture of reagent-grade iron powder and quartz grains (Ottawa silica sand, 40-60) contained in platinum envelopes. Heating was done in a horizontal globar furnace. The degree of oxidation was controlled by premixing CO and C02 gases in desired ratios. The partial pressure of oxygen may be calculated from the known equilibrium constant of the equation CO + 1/2 02 = C02. In addition to providing a convenient control of the oxidation level, both CO and C02 are present in mold atmospheres. Furthermore, other atmospheres containing hydrocarbons and oxygen may be expressed in terms of equivalent CO2/CO atmospheres. The gas mixtures were analyzed with a mass spectrometer. After exposing the specimens for the desired periods of time, 3 min to 1 hr, they were cooled by pushing them into the cold end of the furnace while maintaining the same atmosphere X-ray and metallographic samples were prepared. The latter were impregnated with bakelite resin (BR-0014) prior to mounting and polishing for metallographic examination. The various treatments are summarized in Table XVIII. It should be noted that two of the temperatures, 1225~C (22370F) and 15250C (2777~F), are below the melting point of pure iron, 1537~C (2802~F), However, 1525~C (2777~F) is at the melting temperature of oxygen-saturated iron. The data indicate that the mechanism of mold attack in this system is as follows. Iron is oxidized at the mold surface, forming a separate, oxide liquid which wets the silica sand in the mold. This liquid phase penetrates into the pores of the mold. Silica is soluble in this oxide liquid to about 50 weightpercent. Solution of the silica enlarges the pores in the sand. This enlargement permits the molten iron to penetrate the mold at low pressures although the iron does not wet the silica sand. The depth of penetration into the sand by the iron depends on the length of time at elevated temperatures and the severity of oxidation. If the C02/CO ratio is maintained at a low level, only quartz grains and iron are present. No iron silicate melt is observed and no penetration occurs at low pressures because the iron does not wet the silica. Thus surface deterioration continues after solidification. The more rapid the cooling or the more reducing the mold atmospheres, the better the surface 29

t hat can be expected,, Thi.s indicates the basic superiority of the graphite semi-permanent mold process with its rapid cooli._ng and tendency toward reducirng mo.d atmospheres o When a mold. surface of quartz or other sand is to be used, it is evident that surface oxidation. should be avoided by suitabl.e additions to metal or to mold materials REACTION OF SILICA WITH IRON-CARBON ALL:OYS The surfaces of large shel.l,-molded iror castings are more subject to imperfections than are the surfaces of smaller casti.ngs An importarnt potential difficul.ty arises when large irorn castin.gs are made in silica shell molds because some of the silica may be reduced to SiO by certain alloying elements in the molten iron. This surface defect is shown in Fiyg 3, The SiO which is formed is a gas at ferrous casting temperatures, As a result, its formation at the metal-mol.d'nt'erface of the casting provides a porous surface if the gas cannot escape from tJhe.mold.d As the temperature decreases, the SiO dissociates to Si02 according +o the reaction: 2 SiO > SiO + Si Therefore, the porous metal srfac:e generally contain.s ar. SiO deposi.t This report presents.information concerring the nature an.d extent of the Si02 reduction at' t+he mold-metal2 interface, Th.e roles of temperat.u-re, casting size, al.oys, and selected mold materlals were investiigat;ed to i..ter pret the mechani sm of the reactions. On the basis of the results, met hods c minimizing the effects are suggested. EXPERIMENTAL PROCEDURES Mold Design. —Three mold designs were used i:n this investigation. (. a small step mold, (2) a small gear-.blank mold, and (3) a large step mold. T'! step mold designs are shown in Figs 31 and 36 W.ith these designs, it was hoped to determine the effects of mold volume, section thickness, al7nd parasite section upon the interface reaction. The small step mol.d hInas a vol..ume of 1.9,.21 cu inr, with an additional. 3 ^9 cu in, in the riser, gates and sprue. The heaviest section is 5 inr by.1.r by 2 in, thick. The thinnest section is 2-1/2 in, by I. in, by 1/16 in. The gear-blank mold consisted of three gear blanks with a total volume of 422^ ) in The casting was poured through the riser, The heaviest section was a.ri r. 3 i.r, in diameter, in, wde, and I in thick The large step mol d had a vi.. ume of 300 cu in. with an. additional 225 cu in. in gates, riser, and sprue. hFe. 30

largest section was 8 in. square by 4 in. thick. The total parasite volume was 43 cu in. and the smallest section was 4 in. by 1 in. by 1/8 in. Shell molds were made over the patterns described in the preceding paragraph. The sands used were. (1) two kinds of quartz sands, Geauga sand, and New Jersey beach sand; (2) forsterite sand; (3) Australian zircon sand; and (4) magnorite, Both the forsterite and the magnorite sands contained an excess of fines as received. The excess fines were removed by elutriation. All molds were made with coated sands. Six percent of resin by weight was used with all sands except the zircon which had 3% by weight of resin. Metal Compositions.- The metal compositions used were essentially ironcarbon alloys (1-4% carbon) which contained less than 0.05% of silicon and manganese. Other heats were made containing 0.5% silicon and 0.5% manganese to determine what, if any, effect these elements would have upon the reactions. In the first group of heats, aluminum in the amount of approximately 0.1% was added in the furnace and in the ladle to keep the oxygen level low and to control the carbon boil. No aluminum or other deoxidizer was used in the second group of heats. Instead, the surface of the metal in the furnace and in the ladle was completely covered with a basic slag. The large castings were poured directly from the furnace into the mold to minimize oxidation and to facilitate the control of casting temperatures. The basic slag was successful in preventing oxidation of the heat. The heats were made in either a 60-lb-capacity or a 200-lb-capacity induction furnace with rammed magnorite linings. The stock was armco iron, and spectroscopic grade graphite electrode as the source of carbon. Carbon levels ranged from 4% to 1% in steps of 0.5%. Casting temperatures ranged from 3000~F to 2500~F in steps of 100~F. RESULTS The defects shown in Fig. 55 consisted of porosity in the metal at the surface of the casting. The gas pockets contained an accumulation of white powdery material which was identified spectroscopically and optically as Si02 and A1203. Optical and electron micrographs revealed that the former had an extremely fine-grained fibrous structure. It is of prime importance that X-ray and electron diffraction patterns indicated a significant fraction of the fibrous material to be quartz. Furthermore, quartz was the only crystalline silica phase which was observed. The importance of this observation is that quartz is stable, and will form only below 16000F. Had the Si02 crystallized at a higher temperature, the tridymite and cristobalite modifications of silica would have been observed since they do not transform to quartz in the short period of time that is encountered during normal cooling of the casting. 31

The defects sh.cwn..In':ig.g 35 were the most pronorlcned under conditions of (ii greater s.per-heat,. 2.) larger metal sections, (3) higher carbon contents, [i)] wvith the presence cf alimni.nnm, and (51) i.n molds w:th a large amlont of free s li.ca,:l:rnd.Tvld'. al cc-.ns. ideration of these variables will show thei.r relative effect. Effect of'Temperat'..re -..n a % carb:n. al;amir carstin.g, the defects whi.ch have been described were nb w t encorntered at temperatu.res below 255F,. As th.e pculring tempera-t..res were i..ncreased, the gas poclkets and the accompanyig.:rg SI.O3 depcsits were mcre pricnc!:ncedlo.,h.is same general Increase with temperatu.,re was noted f(r all metal ccsmposiA.t-.onso Effect c:f Sect:orc S'.ze -W:ith comparable carbon co.ntents and with sl.I.) Jca sand, the defe(cts fi.rst appeared:n the 19-n~ir;. cast:i:rgs with 1-i.nr-thick: sect.ions at temperatu.-res cf abor.t 20.-0OC hiLgher thar that irn thle 50Cc..-:in... casting with. -:.. i,-thl: sectionso Effect of Carbo:._. C~on:tent -0th.er factcrs be-t:ig equal, these defects were more p:rnounced wi th. high.er carbron con.ten.ts.:For example a 1% carbcn,, 0,1% alunminri.m casting do-es not prcduce s.,.r:face pcrosity- evenr- at 30X"~00 whe:reas the defect occu.:rs at a termperat._ure as low as 255C0:F wher 4% carbon Is present. Effect cf Alley Add.ti.ncs — Aluminu.m add.itior.s t: the molten metal have a mere pronounc.ed effect,.ponr the defect tha.n aarborn. Wi:1th (.1% alumin.rlm, the clwest temperati.re at w.hi.ch the defect was enr.coun ltered i.n a 4% carbon melt was 25 5G, ~:.. With no al;mmi.numi the defect was not encc(.(:urt ered at 30'0~?W Si.l.con and manganese additi.or.s -ip to Oe.5 did nct h.ave any ncticeable effect upcon the presence or absence of the gas porosity ar.d the Sp02 deposits at the surface of the casti.r:.gs. Effect cf Mold C"T (-.mpCosiit.ioLns. — T he described defect i.s most prn.c.u.;Anced:i.-) a silica-sand shell rmoldo:In. shell molds of forster'ite (Mg2S:.iO(, ) zircn (ZrS',iUsgu or peri.clase (MgO cor.taLn.:ng 10% SlOp, defects were prodlced only when other factors were adverse, with. hl-gh casting temperatures, large sectio.r sizes high carbon., and the presence c;f alum..ln.Lmol The effect of mold compsiti.on.s especially n.ctable in the smaller castings, When a ea:rbon iron. containing 0. 1% of alumln:n-um was cast at 29C00 F iln i the small gear-blank mold made of zircon sand, little porosity was produced When iron of the same composition was cast at the same temperature into the large step mold made of zircon sand, a condition similar to that shown in Fig" 35 was observed. With silica sand the defect was found in both large and small castings 32

DISCUSSION The surface defects must originate from reactions involving silicon and oxygen. This is supported by the facts that (1) silica shells provided the most severe effects, and (2) and SiO2 deposit is produced. At least three hypotheses may be considered to account for reactions between the silicon and oxygen. The first possible mechanism would involve the oxidation of silicon in the metal to produce SiO2 in the gas pockets. Silicon oxidation is observed in higher silicon steels. However, gas porosity does not result. The resulting Si02 forms a liquid at temperatures above 1600~F. Furthermore, a silicon addition to the metal in this case had no noticeable effect upon the presence or absence of the defects. The possibility of this mechanism must be discounted for these castings. A second possible reaction mechanism would involve the reduction of Si02 in the sand by heated resins in the shell to produce an initial gas containing SiO. Although well known in chemistry, SiO is relatively unknown in metallurgy. xSi2 + Cx 2y..... SiO + xCO + yI2 (2) This reaction would require a pressure accumulation at the surface of the metal during solidifioation. There is some evidence to suggest that the contact of the molten metal softens the bonding resin behind the mold surface and clogs the pores in the sand. This produces an impermeable sand which will permit such pressure to build up. Under this mechanism, the SiO would revert at lower temperatures to SiO2. Thermodynamic considerations favor this reversal. The plausibility of this second suggested mechanism is greater than that of the first. Higher temperatures would increase the reaction by producing more SiO and more gas pressure. Likewise, larger castings would accentuate the results by providing longer cooling times. However, it is difficult to account for the significant effect of carbon and aluminum upon the presence of the defects. The third possible reaction also involves the reduction of the Si02 in the sand, but by the carbon and aluminum in the metal. C + SiO2... CO + SiO (3) 2A1 + 3 SiO2... A1203 + 3 SiO (4) Aside from the fact that the reducing agents come from the metal rather than from the resins, this mechanism is similar to the previously suggested one, Both of the reactions occur more strongly at higher temperatures. Both produce SiO and require a retention of the gas by pressure at the mold-metal surface 33

wuntl dissociation occurs at lower temperatures to produce Si02.* 2 SiO. O. SiO2 + Si (1) Most probably the true mechanism requires both the reducing action of alloys within the metal and the protective reducing gases of the heated resins. Otherwise, the SiO would diffuse into the mold where it would oxidize with infiltrating airo The avoidance of these defects may be suggested directly from the results of the tests, This wo-ld involve a minimum of super-heat, and avoidance of aluminum (and probably titanium and zirconium) in the metal, and in large castings the'.se of sands withou.-t free sil.ica. CONCLUSIONS Although. the reduction of SiO9 to a gasecu.s Si.C) form by carbon and oth.er elements has been known. for some time, its importance ir the control of.srface quality of ferrouas castings has not been generally appreciated. When condi.tions involving high. carbon contents, the presence of aluminu.im or other strong oxide formers, high temperatures, or large mcld sizes are encountered, SiO2 will be reduced to SiO when iron is cast into molds containing quartz or slicate sandso The SiO, even. though it is in the form of a gas, does not always escape from the mold cavity and may produce a surface porosityo Lower temperatures permit the dissociation of SiO to SiOpo *Equation (1) is exothermic; therefore, it occurs more strongly to the right at lower temperatures o 31

APPENDIX C

APPENDIX C DETAILED PROCESS DESCRIPTION (1) Pressure-Poured Precision-Cast Drive Sprocket, Part Noo 8671597 1. Machine the graphite mold according to drawings. 2. Complete required mold components, i.e., chiller ring, ingate sleeve, mold retainers, according to drawings. 3~ Machine pattern for mounting bolt hole shell cores according to drawings. 4. Make required quantity of shell core from zircon sand. 5. Heat mold cope and drag and spray with expendable refractory coating. 6. Set shell cores and assemble moldo 7. Melt a heat of the required analysis. 8. Tap heat at the required temperature (determined experimentally during project). 9, Place ladle in pressure pouring tank and record molten metal bath temperature using immersion thermocouple. 10. When proper metal temperature is reached, a properly preheated pouring tube is positioned over the pouring tank and clamped in place. 11. The assembled mold is positioned over the pouring tube and clamped in place. 12. Pouring is then started and the mold filled at the experimentally determined rate (details previously reported)~ 13. After completion of pour, the mold is released from the pouring station and the casting stripped from the mold after a suitable time (details previously reported). 14. The hub is torch cut to a diameter leaving approximately 1/4" for finish machining. 15. Casting is placed in heat treat furnace and annealed. 16. The mold is cleaned, reconditioned and reassembled for further use. 36

17. Casting is cooled to room temperature after annealing and shot blasted to remove scale. i8. The hub is then bored according to drawing requirements. 19. Casting is flame hardened according to specifications, inspected, and is then considered complete. (2) Precision-Cast Final Drive Gear, No. 7364141 1. Machine the graphite mold according to drawings. 2. Complete required mold components, i.e., chiller ring, ingate sleeve, mold retainers, according to drawings. 3. Heat mold cope and drag and spray with expendable refractory coating. 4. Set shell cores and assemble mold. 5. Melt a heat of the required analysis. 6. Tap heat at the required temperature (determined experimentally during project). 7. Place ladle in pressure pouring tank and record molten metal bath temperature using immersion thermocouple. 8. When proper metal temperature is reached, a properly preheated pouring tube is positioned over the pouring tank and clamped in place. 9, The assembled mold is positioned over the pouring tube and clamped in place. 10. Pouring is then started and the mold filled at the experimentally determined rate (details previously reported). 11. After completion of pour, the mold is released from the pouring station and the casting stripped from the mold after a suitable time (details previously reported). 12. The hub is torch cut to a diameter leaving approximately 1/4" for finish machining. 13. Casting is placed in heat treat furnace and annealed. 37

14o The mold is cleaned, reconditioned and reassembled for further use, 15. Casting is cooled to room temperature after annealing and shot blasted to remove scale. 16. The hub is then bored according to drawing requirements. 17^ Casting is processed for carburizing or induction hardening as per instructions of Ordnance Tank-Automative Command, 38

TABLE I CHEMICAL ANALYSIS - CASTINGS SUBMITTED FOR TEST PRECISION-CAST DRIVE SPROCKET NO. 8671597 Heat Sprocket Heat Sprocket C Si Mn P S Cr Mo No. No. R-179 35*.50 35 79 o 013.024 1.08.12 R-180 36 o.1.40.82.011.020 1.07.18 R-188 39.51.32.90 o06.028 1.08.23 R-189 40.50.35.84.014.021 1.03.23 R-191 41.53.39.94.013.028 1.08.21 R-192 42.52.38 o87.026.030 1.08.21 R-193 43.51.36.88.017.016 1.04.22 R-194 44 o1.35.82.016.026 1.05.22 R-196 4.53.33.81.010.030 1.08.22 R-197 46.51.34.90.013.024 1.07.22 R-198 47.53.33 75.007.026 1.05.22 R-201 48*.49.36 77.012.032 1.08.23 Shipped to Professor Flinn, The University of Michigan, All others shipped to Detroit Flame Hardening Company. 39

TABLE II PRECISION-CAST DRIVE SPROCKET, NO. 8671597 SPROCKET TOOTH-BASE WIDTH Cast- Tooth Numbers Heat No. No. 1 23 4 5 6 7 8 9 1011 R-179 35 4.258 4.254 4.253 4.257 4.255 4.251 4.243 4.241 4.245 4.240 4.248 R-180 36 4.260 4.250 4.247 4.250 4.248 4.247 4.236 4,235 4.242 4.44.2 R-181 37 4,262 4.250 4.250 4.250 4.247 4.248 4.234 4.235 4.28 4.241 4.260 R-188 39 4.261 4.250 4.249 4.252 4.252 4.250 4.245 4.241 4.247 4.249 4.256 R-189 40 4,262 4.257 40249 4.251 4.250 4.252 4.246 4.243 4.241 4.250 4.257 R-191 41 4.263 4.254 4.249 40249 4.247 4.248 4.24 424 4.244 4.247 4256 R-192 42 4.257 4.250 4,250 4.250 4.250 4.253 4.244 4.245 4.245 4.249 4.25 ^ ~R-193 43 4.264 4.250 40245 4.248 4.247 4.248 4.243 4_238 4,241 4.24.22 R-194 44 4.262 4.250 4.246 4.250 4.244 4.246 40237 40238 4.241 4.247 4254 R-196 45 4,263 4.255 4.248 4,252 4.250 4.253 4.243 4.242 4.244 4,250 4.255 R-197 46 4.263 40258 4.254 4.254 4.250 4.248 4.242 4.239 4.244 4.250 4.258 R-198 47 4.264 4.254 4.250 4.255 4.253 4.256 4.244 4.245 4.244 40249 4256 R-201 48 4.268 4.253 4.250 4.253 4.250 4.251 4.248 4.245 4.246 4.254 4.256 Mean 4,262 4.253 4.249 4.252 4.250 4,250 4.242 4.240 4.24 4.247 4.2 Std. Std..0027.0028.0034.0026.0028.0028.0039.0033.0024.0038.0038 De v. Mean Standard Deviation:.0030 in. Specified Dimension: 4.250 in.

TABLE III PRECISION-CAST DRIVE SPROCKET, NO. 8671597 SPROCKET TOOTH-ROOT RADIUS Heat Cast- Tooth Numbers Heat ing No. No. 1 2 3 4 5 6 7 8 9 10 11 R-179 35 9.873 9.877 9.878 9.882 9.876 9.873 9.871 9.875 9.878 9.880 9.872 R-180 36 9.870 9.884 9.877 9.873 9.868 9.867 9.868 o.876 o.874 0.865 o.863 R-181 37 9.872 9.873 9.872 9.877 9.876 9.876 9.874 9.877 9.876 9.885 9.887 R-188 39 9.868 9.872 9.874 9.879 9.875 9.869 9.879 9.889 9.885 9.873 9.862 R-189 40 9.878 9.882 9.873 9.877 9.881 9.886 9.884 9.879 9.884 9.877 9.872 R-191 41 9.875 9.881 9.879 9.880 9.873 9.875 9.879 9.880 9.874 9.871 9.867 R-192 42 9.885 9.885 9.878 9.878 9.875 9.875 9.872 9.878 9.883 9.879 9.882 Al: R-193 43 9.869 9.882 9.885 9.885 9.875 9.871 9.872 9.880 9.880 9.877 9.867 R-194 44 9.879 9.882 9.881 9.884 9.874 9.873 9.871 9.875 9.882 9.883 9.876 R-196 45 9.873 9.882 9.872 9.874 9.875 9.879 9.877 9.879 9.876 9.860 9.855 R-197 46 9.875 9.882 9.881 9.881 9.875 9.875 9.877 9 886 9.882 9.871 9.865 R-198 47 9.877 9.881 9.874 9.879 9.880 9.887 9.888 9.880 9.887 9.875 9.870 R-201 48 9.886 9.884 9.885 9.891 9.882 9.887 9.889 9.890 9.892 9.883 9.875 Mean 9.875 9.880 9.877 9.880 9.875 9.876 9.877 9.880 9.881 9.875 9.870 Std..0053 0.00.0043.0046.0.00.0064.0068.0048.0051.0070 0086 Dev. Mean Standard Deviation:.0054 in. Specified Dimension: 9.875 in.

TABLE IV PRECISION-CAST DRIVE SPROCKET, NO. 8671597 SPROCKET TOOTH-TIP RADIUS Heat Cast- Tooth Numbers ing No. No. 1 2 3 4 6 7 8 91011 R-179 35 13.85 13.856 13.855 13.856 13.855 13.848 13.853 13.857 13.858 13.862 13.857 R-180 36 13.847 13.856 13.863 13.86 13.851 13.843 13.853 13.852 13.854 13.853 13.844 R-181 37 13.853 135857 113858 13.858 13.858 13.856 13.861 13.863 13$864 13.861 13.850 R-188 39 13.849 13.835 15.858 13.861 13.861 13.845 13.861 13.867 15.868 135866 13.849 R-189 40 13.856 13.864 13.862 13.866 13.864 13.862 13.869 13.865 13.870 13.867 13.855 R-191 41 13.858 13.862 13.365 13.866 13.859 13.854 13.857 13.860 13.857 13o860 13.856 R-192 42 13.869 13.864 13.867 13.866 13.858 13.861 13.854 13.861 13.865 13.868 130868 R-193 43 13.849 13.863 13.866 1.874 13.864 13.857 13.862 13.866 13.869 13.864 13.853 R-194 44 13.864 13.864 13 866 130869 13.862 13.850 13.857 13.857 13.862 13.867 13.864 R-196 45 13.849 13.852 13.862 13.857 13.855 13.854 13.865 13.862 13.862 130864 13.842 R-197 46 13.858 13.867 13.871 13.870 130863 13.856 13.862 13.871 13.870 13.865 13.857 R-198 47 13.857 13.861 13.863 13.864 13.863 13.864 13.876 130864 13.874 13.869 13.857 R-201 48 13o864 13.868 13.870 13.870 13.867 13.862 13.871 13.879 13.881 130871 13.863 Mean 13.856 13861 8 13.864 1864 13.860 13.856 13.862 13 863 13.866 13.864 13.855 Std..0006.o0050.004.0053.0043.0064.0068.oo65.0071.0045.0072 Dev. Mean Standard Deviation:.0058 in, Specified Dimensionr 13.859 in,

VA F Location of hardness survey on sprocket teeth. TABLE V ROCKWELL "C" HARDINESS SURVEY OF SPROCKET TOOTH CAST AISI 4150-No. 195-AS-HARDENED Specified - 55-60 Rc at 1/8 in. Hardness - 50 Rc at 1/4 - 3/8 in. Distance, Hardness 1/32 in. A B C D E F 1 61.0 -- -- -- 2 61.3 61.5 -- 59.9 62.0 57.7 3 6161.61 618 59.1 61.7 59.8 4 61.5 61.5 62.0 58.6 61.7 61.1 5 61.2 61.5 60.0 58.0 61.5 61.0 6 61.4 61.5 61.8 58.0 60.8 61.1 7 59.5 61.6 60.9 55.8 61.2 60.3 8 57.8 61.5 60.3 55.4 61.0 59.7 9 53.4 61.2 59.2 54.5 60.5 58.2 10 52.3 61.0 8.5 53.8 60.5 55.7 11 46.3 61.1 57.5 51.2 58.7 55.0 12 42.8 59.5 56.6 51.2 58.5 50.3 13 21.0 58.5 54.6 50.0 56.8 45.3 14 -- 57.7 54.1 48.0 56.2 28.0 15 -- 56.3 52.2 46.1 53.0 16 -- 55.0 51.5 438 52.8 17 - 52.3 50.0 41.6 50.6 18 -- 51.4 49.6 38.2 50.2 -- 19 -- 49.5 46.9 28.0 46.2 -- 20 -- 48.1 46.8 25.5 46.8 -- 43

STIUMIARY OF S.PR'-OCKET' TEST RESTJL TS Averag-e Maximum m Wear Average Maxi:m.luil Wear of Pniace Mileaf e of Caasi' Sprocket Fabricated Sprocket (i.r. per 1000 miles) (in,, per 1000 miles) General Motors Mil.i iar-y P rong? GCround-! M!_ i f-0ri, Michi.gan 2410- 0.l 6 0 7. Aberdeecn Fr' Movi. ng Ground, Marylanrd 2359* 0<o.08 0o10o6 Forn (Ournh il Ontario. Iarada 1001 0 016 0023 YuL"ma Tlest Stat.ion* <'!.ma; Ari z onna under testTli. ckiness of caset sprockets 1i.-67.5 in,; of fabricated, 1!375 in, in this. est as only th:e tbinner parts were available at the time, In other tests, al. parts were- 1 675 in, thick. e',iCli-> "'he:cast sprockets lasted this distanrce, Another pair of fabricated:,r oc ke' Ca wac worn out in t;he first, 1.239 mi.les of -the test, "as" a. —nro -ke applied aft;er 514 mn.l.es T'iis is corrected for by using wear u-u"r 1.00. nJ(: ir les i n arove calcilati.Ions Wear is approximately li.nar /mile-.O.I.:F.-,FAR-OIS:'J OF O AS...AND F.ABLlFSRCAITE DRIVE S.PROCKETI N0, 8671597 Fa bric ated Sprocket- Cast, Sprocket -kr 0...n F...t. A!6S. 134 AISI 0150 WTeru -- tot~1-.~ sl 3563 195 lTei, f'' of * S-,:!,proc:ket b'. 9 -a:.'./... l 2 60 1 0 1.,~~~~~~~~~~r

TABLE VIII PRECISION-CAST FINAL DRIVE GEAR, NO. 7364141 HEAT AND CASTING SUMMARY Mold Speed Type of Het Casting No. Date Tilt, of Pour, Risers, Disposition *________No.____ deg sec/lb ino diam R-33 1 10-28-57 2 1.75 None Scrap, mold did not fill R-335 2 10-30-57 2 1.75 None Scrap, mold did not fill R-338 3 11-4-57 2 1.75 1-2 Scrap, partial runback R-339 4 11-5-57 2 1.75 1-2 Scrap, partial runback R-341 5 11-7-57 2 1.75 1-2 Scrap, partial runback R-~48 6 11-19-57 2 1.75 3-2-1/2 Scrap, shrinkage between risers R-350 7 11-21-57 2 1.75 3-2-1/2 Scrap, shrinkage between risers R-350 8 12-9-57 3-1/2 1.75 3-2-1/2 Wheel steel analysis, visually sound R-50 9 12-17-57 3-1/2 1.75 3-2-1/2 Wheel steel analysis, visually sound R-363 10 12-20-57 2. 75 3-2-1/2 Scrap, poor surface R-367 11 12-26-57 2 1.75 3-2-1/2 Scrap, x-rayed, slight shrinkage R-367 12 1-11-58 3-1/2 1.75 3-2-1/2 Wheel steel analysis, experimental machining R-408 13 2-11-58 Mechanical failure, heatpigged R-410 14 2-13-58 3-1/2 1.2 3-2-1/2 Sectioned, visually sound R-414 15 2-18-58 3-1/2 1.2 3-2-1/2 Scrap, x-rayed, slight shrinkage R-417 Blank No. 1 2-25-58 3-1/2 1.2 4-2-1/2 Scrap, x-rayed, moderate shrinkage R-429 Blank No. 2 3-13-58 3-1/2 2.0 4-4 Sectioned, moderate shrinkage R-430 16 3-17-58 3-1/2 2.0 6-2-1/2 Sectioned, visually sound R-432 17 3-19-58 3-1/2 2.0 6-2-1/2 Scrap, off analysis R-434 18 3-21-58 3-1/2 1.75 6-2-1/2 Scrap, off analysis R-435 19 3-21-58 3-1/2 2.0 6-2-1/2 Scrap, off analysis R-436 20 3-24-58 3-1/2 2.0 6-2-1/2 Scrap, off analysis R-437 21 3-24-58 3-1/2 2.0 6-2-1/2 Scrap, off analysis R-438 3-25-58 Mechanical failure, heatpigged R-440 22 3-27-58 3-1/2 2.0 6-2-1/2 Scrap, off analysis R-444 23 3-31-58 3-1/2 2.0 6-2-1/2 SAE 1062, reserve stock, to Caterpillar 5

TABLE VIII (Continued) Mold Speed Type of Heat Casting No. Date Tilt of Pour, Risers, Disposition ~________No.____ deg sec/lb in. diam R-446 24 4-1-58 3-1/2 200 6-2-1/2 Scrapped, runback R-450 25 4-7-58 3-1/2 2.0 6-2-1/2 Scrapped, wrinkles on teeth R- 52 26 4-8-58 3-1/2 2o0 6-2-1/2 Scrapped, runback R-452 27 4-8-58 3-1/2 2o0 6-2-1/2 Scrapped, wrinkles cn teeth R-456 28 4-10-58 3-1/2 2o0 6-2-1/2 Wrinkle, 1 tooth R-456 29 4-10-58 3-1/2 2.0 6-2-1/2 SAE 1062, reserve stock, to Caterpillar R-h61 30 4-15-58 3-1/2 2.0 6-2-1/2 SAE 1062, stock, to Caterpillar R-461 31 4-15-58 3-1/2 2.0 6-2-1/2 SAE 1062, stock, to Caterpillar R-463 32 4-16-58 3-1/2 2o0 6-2-1/2 SAE 1062, stocR., tc: Caterpillar R-463 33* 4-16-58 3-1/2 2-0 6-2-1/2 SAE 1062, reserve stoca, to Caterpillar R-467 34 4-21-58 3-1/2 2.0 6-2-1/2 Scrapped, bad teeth R-467 35 4-21-58 3-1/2 2.0 6-2-1/2 Scrapped, large inclusicln in ibcre R481 36 4-29-58 3-1/2 2o0 6-2-1/2 Scrapped, wrinKles on teeth R-481 37 4-29-58 3-1/2 2~0 6-2-1/2 SAE 1062, reserve stock, hbld R-483 38 4-30-58 3-1/2 2~0 6-2-1/2 SAE 1062, stock, to Caterpillar R-483 39 4-30-58 3-1/2 2~0 6-2-1/2 SAE 1062, stock, to Caterpillar R-486 40* 5-5-58 2 2o0 6-2-1/2 SAE 1062, stock, to Caterpillar R-486 41 5-5-58 2 2o0 6-2-1/2 Scrapped, bad wrinkle o-re tooth R-487 42 5-6-58 2 1.75 6-2-1/2 SAE 1062, stock, to Caterpillar R-487 43 5-6-58 2 1.75 6-2-1/2 SAE 1062, stock, to Cater pillar R-489 44 5-12-58 2 1o75 6-2-1/2 Scrapped, bad wrin.le one tooth. R-489 45 5-12-58 2 1.75 6-2-1/2 SAE 1062, stock, to Caterpillar R-490 46 5-13-58 2 175 6-2-1/2 SAE 1062, stock, to Caterpillar R-490 47 5-13-58 2 175 6-2-1/2 SAE 1062, stock, to Caterpillar R-491 48 5-14-58 2 1075 6-2-1/2 Scrapped, bad wrinkles three teeth.

TABLE VIII (Concluded) Mold Speed Type of Casting No. Date Tilt, of Pour, Risers, Disposition deg sec/lb in. diam R-491 49 5-14-58 2 1.75 6-2-1/2 Scrapped, bad wrinkles one tooth R-493 50 5-16-58 2 1.75 6-2-1/2 SAE 1062, stock, to Caterpillar R-493 51* 5-16-58 2 1.75 6-2-1/2 SAE 1062, stock, to Caterpillar R-496 52 5-25-58 2 1.75 6-2-1/2 Scrapped, bad wrinkles one tooth R-496 53 5-25-58 2 1.75 6-2-1/2 Scrapped, bad wrinkles three teeth R-497 54 5-26-58 2 1.75 6-2-1/2 SAE 1062, stock, to Caterpillar P-497 55 5-26-58 2 1.75 6-2-1/2 Scrapped, wrinkles on teeth R-498 56 5-27-58 2 1.75 6-2-1/2 Scrapped, gas hole in tooth R-498 57 5-27-58 2 1.75 6-2-1/2 Scrapped, wrinkles on teeth R-499 58 5-28-58 2 1.75 6-2-1/2 SAE 1062, stock, to Caterpillar R-499 59 5-28-58 2 1.75 6-2-1/2 Scrapped, wrinkles one tooth R-501 61 6-2-58 2 1.75 6-2-1/2 SAE 4817, stock, to Brad Foote R-527 60 7-7-58 2 2.0 6-2-1/2 SAE 4817, gas holes, machined for test R-527 62** 7-7-58 2 2.0 6-2-1/2 SAE 4817, stock, to Brad Foote R-528 63 7-7-58 2 2.0 6-2-1/2 SAE 4817, stock, to Brad Foote R-528 64 7-7-58 2 2.0 6-2-1/2 SAE 4817, stock, to Brad Foote R-529 65 7-9-58 2 2.0 6-2-1/2 SAE 4817, stock, to Brad Foote R-529 66** 7-9-58 2 2.0 6-2-1/2 SAE 4817, stock, to Brad Foote R-530 67 7-9-58 2 2.0 6-2-1/2 SAE 4817, stock, to Brad Foote R-530 68 7-9-58 2 2.0 6-2-1/2 SAE 4817, stock, to Brad Foote R-531 69** 7-10-58 2 2.0 6-2-1/2 SAE 4817, stock, to Brad Foote R-531 70** 7-10-58 2 2.0 6-2-1/2 SAE 4817, stock, to Brad Foote Radiographically sound. Scrapped after carburizing due to decrease in size Scrapped after carburizing due to decrease in size.

TABLE IX PRECISION-CAST FINAL DRIVE GEAR, NO. 7364141 CHEMICAL ANALYSIS - CASTINGS SUBMITTED FOR TEST Heat Casting Chemical Analysis No. No. C Mn Si P S Ni Mo R-444 23.62.98.31.010 o020 R-456 29.65.90 33 010.016 R-461 30.63 98 30.012.018 R-461 31 63 98.30.012.018 R-463 32.67.91.28.011.026 R-463 33 67.91.28.011 o026 R-483 38.64.96.24 012 o020 R-483 39 o64.96.24 o012.020 R-486 4 0.60.95 35.011 024 R-487 42.60.93 26.007.014 R-487 43.60.93.26.007 o014 R-489 45 59.90.25 o008.018 R-490 46.64.91.25 o013 o016 R-490 47.64.91.25.013 o016 R-493 50.60.91.26.009.026 R-493 51.60.91 o26.009 026 R497 54.56.90.23 o 014.024 R-499 58*.58.94.29.019 o 022 R-501 61.21.66.28.019 o018 3.47 ~23 R-527 60.15.69.28.012.014 3~44 o25 R-527 62.15.69.28.012 o014 3~44 ~25 R-528 63.18.62.24.014.012 3046.25 R-528 64.18.62.24.014.012 3046 25 R-530 65.16.65.27.010.012 3.50,25 R-530 66.16.65.27.010 o012 3.50.25 R-531 67.22.65.26.009 o016 3.53.25 R-531 68.22.6.26.009 016 3 53 25 R-532 69.20 72.26.009.012 3051.25 R-532 70.20.72.26.009.012 3051.25 *Cast in larger mold. Pin diameter approximately.120 in. larger. lg

TABLE X PRECISION-CAST FINAL DRIVE GEAR, NO. 7364141 SUMMARY OF PIN-DIAMETER DATA* CASTINGS NOS. 17 THROUGH 54 Gear YEX X E x2X x - 19.730 Number 4 17 654.782 19.842 12,992.106.005.028 18 655.191 19.854 13,007.452.007.031 19 655.238 19.856 13,010.209.007.0270 20 654.704 19.839 12,990.440.007.0272 22 654.724 19.840 12,989.804.004.0275 23 654.310 19.825 12,973.382.005.0237 29 655.224 19.823 13,009.653.007.0232 30 654.903 19.843 12,996.908.005.0282 31 655.538 19.832 13,022.125.007.0280 32 654.991 19.846 13,000.402.007.0290 33 654.695 19.837 12,988.654.006.0267 34 654.419 19.829 12,977.705.006.0247 35 655.787 19.840 13,012.144.007.0275 38 655.287 19.825 13,012.153.003.0237 39 655.224 19.823 13,009.652.008.0232 40 654.875 19.843 12,995.796.004.0282 42 654.380 19.828 12,976.158.005.0245 43 654.138 19.820 12,966.562.0oo.0225 44 654.432 19.829 12,978.221.007.0247 45 654.723 19.838 12,989.765.006.0270 46 654.916 19.844 12,997.425.007.0285 47 654.981 19.846 13,000.242.oo4.0290 50 654.667 19.836 12,987.543.oo6.0265 51 654.671 19.837 12,987.701.004.0267 54 654.895 19.843 12,996.590.004.0282 E x 495.918.143.6644 x 19.837.oo6.027 EX2 9837.389 cr X.010 Range 19.856 to 19.820.0310 to.0225 *Q-1 through Q-33 individual measurements. LAQc,

TABLE XI PRECISION-CAST FINAL DRIVE GEAR, NO. 7364141 SUMMARY OF PIN-DIAMETER DATA* CASTINGS NOS. 58 THROUGH 70 Gear ZE XX2 19 730 Number 4 58 657.878 19.934 13,115.258 o007 o0510 60 658.096 19.910 13,123951 o006 o0450 61 658.648 19.927 13,145o976 ~004.0492 62 658.092 19910g 13,123o792 0007 o0450 63 658,720 19.929 139148.851 oo006.o49 64 659.018 19.908 13,160.750 o005 o0445 65 658.802 190931 13,152.125.007.0502 66 658.582 19.925 13,143o343.007.0487 67 658.891 19.934 13,155o678 o005 o0510 68 658~300 19.916 13,132.089 o006 o0465 69 659.151 19.912 13 166.064 0007.0455 70 658o505 19.922 13,140.268 o004.o480 E x 2390058.071?43 x 19.922 o006 o48 Z X2 4762.395 o X.010 Range 19.934 to 19.910 o0450 to.Oi! 0Q-1 through Q-33 individual measurements.

TABLE XII PRECISION-CAST FINAL DRIVE GEAR, NO. 7364141 SUMMARY OF GEAR MEASUREMENT DATA CASTINGS NOS. 17 THROUGH 70 Zx X EX2 aX A-i 31.305.869 27.226.010 A-2 31.320.870 27.252.010 A-3 31-333.870 27.266.007 A-4 31.315.870 27.244.009 X.870.009 J-1 202.422 5.621 1,138.189.010 J-2 202.423 5.621 1,138.199.008 X 5.621.009 L-1 720.827 19.477 14,043.160.016 L-2 721.353 19.491 14,063.643.025 X 19.484.020 P-1 120.275 3.250 390.977.007 P-2 120.240 3.249 390.750.008 P-3 120.288 3.250 391.062.008 P-4 120.222 3.248 390.632.0oo6 X 3.249.007 A-1 through A-4 - Plate Thickness J-1 and J-2 - Hub Diameter L-1 and L-2 - Outside Diameter P-1 through P-4 - Rim Thickness

TABLE XIII COMPARISON OF FORGED VERSUS CAST CRANKSHAFT 14-v - 1790 Engine Forged 1790 Crankshaft Cast 1790 Crankshaft Material SAE 4340 SAE 4340 Weight of Stock, lb 418 315 Finished Weight, lb 250 232 Scrap Material, lb 168 83 Finished Cost $500.00 $250.00 Estimated savings per crankshaft: $250.00 Estimated production, 1958: 900 Total estimated savings: $225,000.00 TABLE XIV DIMENSIONAL VARIATIONS IN REAR FOLLOWER CASTINGS Edge Center Heat No. 676 Heat No. 682 Heat No. 671 Center Edge Center Edge Center Edge 2.605 2.572 2.578 2.560 2.615 2.590 2.588 2.576 2.571 2.557 2.611 2.598 2.576 2.567 2.591 2.578 2.610 2.594 Differences 0.033 0.018 0.025 0.012 0.014 0.013 0.009 0.013 0.016 Average Difference of Bulge 0.018 or 0.015 or 0.018 or 0.007 in./in. 0.006 in./in. 0.007 in./in. b.

TABLE XV DIMENSIONAL VARIATION IN STEEL STEP CASTINGS, CONVENTION SHELL-MOLDING WITH GREEN SAND BACKING Heat Dimensions Bulge No. "A" "C " in./in. 5 1.964 1.960 -.002 14 1.979 2.026 +.023 15 2.047 2.083 +.018 20 2.019 2.004 -.007 21 2.014 2.043 +.015 25-C 1.959 1.964 +.002 27-C 1.985 1.980 -.002 Range of Analysis C Mn Si.2-.3.2-.4.15-.5 TABLE XVI DIMENSIONAL VARIATION IN GREY CAST IRON STEP CASTINGS Heat Dimensions Bulge No. "A" "C " in./in. 691-1 1.964 1.969 0.002 691-3 1.982* 1.989 0.004 691-2 1.979* 1.988 0.003 A581 1.981 1.977 0.002 A582 1.984 1.985 0.000 A584 1.967 1.968 0.001 A585 1.965 1.965 0.000 A586 1.962 1.960 0.000 Analysis... as charged C Si Mn S P 3.60 2.20.10.05.05 *Cast with green sand backing. 53

TABLE XVII DIMENSIONAL VARIATION IN SHELL STEP MOLD CAST IN DUCTILE IRON Bulge Total Force Bulge Dimension "A' in./in. on Mold, lb Edge of Step (in 2-in. section) Edge of Step Static Backing Metal shot number A20-2 40 0.029 2.038 685-8 80 0.055 2.019 682 24 0.105 2.072 Sand, mold number 662-1 -- 0.017 2.012 A20-3 45 0.032 2.002 664-1 7 0.054 2.064 Sand + weight A20-4 45 0.015 1.989 685-7 44 0.055 2.012 A55-1 27 0.104 2.114 C02 process mold A28-3 -- 0.033 2.021 A28-2 -- 0.039 2.027 Dynamic Backup - Spring Clamps Mold number A54-1 18 0.017 2.006 A55-2 60 0.017 1.992 A54-2 60 0.015 1.990 671-1 60 0.007 1.986 A54-3 60 0.007 1.998 683 90 o.006 1.979 A55-4 120 0.009 1.986 A54-4 120 0.002 1.968 673-1 120 0.005 1.995 676-3 150 0.006 1.974 A54-5 180 0.003 1.975 A55-5 180 0.006 1.976 680 180 0.001 1.996 A28-1 180 0.008 1.981 A20-1 180 0.002 1.984 A54-6 240 0.000 1.978 A55-6 240 0.000 1.960 5

TABLE XVIII SUMMARY OF EXPERIMENTAL DATA IRON-SILICA INTERFACE REACTION Specimen Gas Mixture Temperature, 0C Time, min Phases Present No. CO2 % co 1 50 50 1225 3 2 50 50 1225 5 { Metal, silica, fayalite 3 50 50 1225 15 4 50 50 1225 60 5 50 50 1525 3 6 50 50 1525 5 Metal, silica, fayalite 7 50 50 1525 15 8 50 50 1525 60 9 50 50 1565 60 Silica, fayalite, magnetite 10 10 90 1225 60 Silica, metal 11 10 90 1525 3 12 10 90 1525 5 ~~~~~~~~13 10 90 1525 15 qMetal, silica, fayalite 13 o10 90 1525 15 14 10 90 1525 60 15 2.7 97.3 1525 60 Metal, silica 16 2.7 97.3 1565 60 Metal, silica

^ PNEUMATIC CLAMPS PLUNGER Showing arrangement of graphite mold, ladle, and tube during pres. sure-pouring operation. C IR-TIGHT.AMB CHAMBER bOg00 Mold risers not shown.' b' o:; Pouring speed is close- -*'oj Qoe ly regulated to prevent _.. __- ". mold erosion. Operation o'- z is pushbutton controlled a....E -~.:I, - METAL.-._ I/ I~~AIR' ___PRESSURE Fig. 1. Production of railroad car wheels in pressure-poured graphite mold..o: 6:.

^~C —-------- ~ 55 COPE > XXx3:)IX DRAWING NO DE2-Mf>402 -'^^ ^ ((NOT DRAWN) ___ SPROCKET TOOTH x \X ^^ X^ SEE NG NO. E 1/2 BO WKK~~x\ ~s~i ~L"2" 12 41 ~~~~~~~~~~., ~~~~~~~~~~~~~~~~~~~a,, ~ -. ~~~~~1 a P1'o I8A. 7 -AI. FAM~ CUT X <'~ ~~~~~~~~~~~~~33 DIA/,~~~~~~~~~~~~~~f I NGATE-'/ ING NQ1402 —- /:Rl, INOT MANN)~C DRAG /'<' ~,,~~~~~~~~~~~RWING NQE-402.44 ~, x~ (NOT DRAWN).!, - n~K ~~ oi 6~~~~~~ FOR DETAILS OF C.I WAFERS SEE DRG. 5200 I-5-E Fig. 2. Permanent mold design for precision-cast drive sprocket( Part No. 8671597.

X.: qiiii::'iii~':i liii:::::: S ~ iiiv:i: mm,::::i: ~::::_:::::ii- Ji:::::::-I -:-:::::: 1 4::.:1-~I:-:: E~~91 ~2j:::::::::j::::::::H:::::::::: -:::::::::::::::::::::::::::__:::i......... ~~iiiliiii~~~~-:iii~~~~::iiiifiiii-i~~~~~~~i -..........:-iii:- ~si:-...,~~~~~....... ~~~~~~~~A ~~~~~~~~~~~~~~~~~~~~AKAM.: ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~:::::::::::~.: _::::::::~~~~~~~~~~~~~W 00 P.7-1~~~~~~~~~~~~~~~~~~~~~~_.:::::-.: —.i-'':,_~~ —''"iii~ii ii.-. i I-i -;::I- -i... Al~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~:::-::::::,::~-iiii::ii:iii~ —ii~:ii:..... jk WIN~~ ~~~~i:i~..,,,,,,~~iBB8~~:'i j.. -::-ji-jj:iiiii:i ~~~::ii~ig. Surfce of gay iron nsert fo sprockt imold fter proucing 2:'%: i':'DE'i'i'-~~:l:::-i~;i-iiiii —i'::: —:i i r~i Casting s.~i~i

l -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l l" i''-:r *WW:: F2 * sto Ijo a I ~i ~' ~O __.4 Diesprce cas in aro nl _ l _~~T

inches allowed tolerance* mean dimension sample range specified dimension --- --- 95% probability limit 4.290 4.270 4.250 42.50 4.210 ^-~~~~ —-. L L 0.0 1 2 3 4 6 7 8 9io 11 Tooth Numbers *Specified tolerance equal to this dimension. Fig. 5. Tooth base width showing sample range and 95% probability limits, precision-cast drive sprocket, Part No. 8671597.

inches allowed tolerance* mean dimension sample range specified dimension - 9, probability limit 9.900 9.880 o-N 9.860 9.840 0.0 o.o L 1 l I1 h 1. 11 11 1i1 i 1 1 2 3 4 5 6 7 8 9 10 11 Tooth Numbers *Allowed tolerance modified to be equi-distant from specified dimension. Fig. 6. Tooth root radii showing sample range and 959 probability limits, precision-cast drive sprocket, Part No. 8671597.

inches allowed tolerance* mean dimension sample range specified dimension --- 9 probability limit 13.880 - - - - 13.86o.. ON 13.84o _ _ 13.820 0.0 1 2 3 6 7 8 9 10 11 Tooth Numbers *Allowed tolerance modified to be equi-distant from specified dimension. Fig. 7. Tooth tip radii showing sample range and 95% probability limits, precision-cast drive sprocket, Part No. 8671597.

-...-..- NO. 197 2416 SERVCE' MILES- MILFORD, MICHIGAN ~- - NO. 194 347 SERVICE MILES-FT STEWART, GEORGIA I; r WO i X// 7\ / Fig. 8. Comparative wear patterns, precision-cast drive sprockets, Part No. 8671597. I - ~~~97 0

X'.'... "' ~^ -, ~.!?.'x ^^c::'x W^'~ -'. -'-r: ~'^~.-1 —'^~~'^,~ 5 - t' -skP^^ ~^ ^.K^ ~~~~~~~?..... 1~^"^ ^"1^^ ^ ^v - ^ ^'.c ~..?ra~nsii~on Zone ii00X 2~0 Nitai Fig. 9a. Cas-t AISI ~:i90 sprocket, No. 197. *e f it'- ~'"^ "'~^ a-' -t " C I.. ~ ~.' ~: rans -i~ ion Zone,00OX 27~ lNitai Figl. 9b. Fabricated AISI 1549 sprocket A. 64 a'4~~~~~~~~~~~~ ~. rans t i n oe -Ox2%Nia F ic.9a. Cs-G' ISI -150 prockt, No 197 ~~~~~~,e ~ ~ ~ ~ ~~i7 " Wj"I.: ~~ 1. kl lrans-, ption Zone -ssl-a 2 D Ti t a 1 F i l)-b Fbriate ASI 545sprcke A

Charge material 12 Arc furnace Pressure poured 2) 3) Machine bore I.D. Flame harden 4) 5) a Precision-Cast A/IS 4150 Charge__ moaterial Open hearth Ingot 1) 2) 3) l _ __J I \ Forged to billet Rolled to plate Flame cut 4) 5) 6) Machine bore and lame harden teeth 7) b. Fabricoted A/SI 1345 Fig. 10. Manufacturing operations for production of cast and fabri-cated drive sprockets, Part No. 8671597. 65

99 )azTs 9/T 0% peonpatH'~~ol'~i~~~;65X L'o~~~~~~~~;.L::!G~~~~~~~~~~~ ML.;?"a~~~~~~ ~, ~a s?ses-i~-.'-'- -T q iI::~: Kuwaiti: ~~s q/T 0 panpa 960{~'oH: l' ~e xOTIO, %uos,1ses:ileMS.elI.}$S; i::: f), T:':'/.............::::; 0'Sf:73:0::~::f:,'Dj0:::~~~~~~~~~~~~~~~~~~~~~~~~~' [:'::[':...'9-[ | | | l:: |: ~ |,':, 0'::;::.: "..S:2::f: f::::i::::f::: iE:E:i::;EEi..:2:: -:: 7:. SS$ h5:: E-: E-t:E# zfi//............'ON i4-Tnd::&i 00....... cTl,/-OTTO:;;:'...... $;S-D Q.T-9T4 t — eT'2 i i:.Ejtt "', Y+,SE,'';,,,iE,: St004A:X5it:fEf:t2,X,:ii............................E,,,,, j ZZS...t.::iy E1 ~TS

Fig. 12a. Shell mold for front follower, Part No. 7360356. Fig. 12b. Shell-mold for rear follower, Part No. 7359510. 67

FIRMr::ii:f ~~~:::~:~~:i:Ok:~I Ei-:-lb_:a -`1~~~~~~~~~~~~~~~Ow "V1 i:~~~~~~IL' G"\~~~0 rr? ii 4, 9 rvs,~~~~~~~~~~~~~~~~~7 Fig 13 Shll rocssComanymolingmacin with front follower pattern, Part No. iii 736o356.

W mm,''*'.w9, 5 4sr,r.i.^ t i.A Fig. 14a. As-cast ductile iron front follower, Part No. 7360356. 4,AW~~~~~~~~~~~~~~~. Part 4o.. Scored Edge lOOX 2o Nital Fig. 14b. As-cast ductile iron rear follower, Part No. 7 559510. 69

is — ^,- ^. ^t ^ Con: t S i l OOi2 Ni t ai::: t: 5 MMa Fig. 1 mlid- dutie io for rpaceet folowrcsings.'*O, 70 a::.. - ^ ^. ^ ^: ^ ^,. ^,' vI::: sf 4,,'p: -~~ * h d i:::' y *. *~ I: ^\~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-,, -. jCI~~~~~~~~~~~~~~~~~~~~~~~i.. -II, Control Specimen 100X 2g Nital Fig. 15. Normalized ductile iron for replacement follower castings. 70

INVOLUTE MUST BLEND WITH ROOT RADIUS AT OR BELOW 18.470 DIA. WITHOUT ANY STEP.004 UNDERCUT WILL BE FINISH BEFORE PERMITTED.8976 SHOT PEENING 25.44897_.004.306MIN.269.296.219 CHAMFER TIPS OF -1 TEETH (BOTH SIDES) 18.470 DIA. 134 MA RAI NE INVOLUTE MUST EXTEND TO OF STRAIGHT LINE 9.38 DIA DETAIL B BASIC RACK (CUTTER) DETAIL A SHOW PIECEMARK, MANUFACTURER'S SEE AIL IDENTIFICATION AND SERIAL NUMBER THESE SURFACES MUST BE SQUARE SEE DETAIL B / AND THE WORD SET TOGETHER WITH WITH THIS DIAMETER WITHIN.003 A SET NUMBER. TOTAL INDICATOR READING =/^^^^^~ /\^ ^^^^^ 13/16- 13/4 1 R 1/ / i ~ _ \ ADMI NOM IEQUALLY SPACED =91_~-t / 775025 4IRTO. R. V98 >6 /V. V \\ 16 ISPLINES T PITCH D TCHD.OF 5.505-00.~~ — 5.5131 ROLLIN WITH THID. V19.0 Ir ~ ic. 16. F Ia' i\143 19.00 I''~'^^ /81 >< ///0 ~s ol RF3 —OPT -- ^^^^^^ ^^^ -3iDIA. 3 HOLES 1/16 R -^^^ ^ _____ ^HOLES MUST FREELY ^^^^^ iIADMIT NOMINALLY OF LOCATED GAGE PINS _______,.010 UNDER MIN. HOLE SIZE 1/, ~-6'R.5760 DIA. /45 OPT5 O- THIS DIAMETER WITHIN.002 TOTAL INDICATOR READING ORDNANCE PART NO. 7364141 Fig. i6. Final drive gear, Part No. 7364141.

- -3/4" RISER i PLUG RISER SAND~__., I~~~~~~~ II: _________.37 "DIAM~: > 13. 625"DIAM FOR 3-1/2"DIAM HOLES N:xx~ 6-SPACED EQUALLY x ll" A X>^:. —: ~..^V~~~~y^^^XX'-V^ ^^ <.\^\^^ U STANDARD STEEL WHEEL v\ ^^<^:;~'_.~ ^'^ ^ ^V\^ /~~~~~~~~~~~~~~HUB RISER, ALTERED V\ )<V\;::~' ~ ~"'~ X X~XCK-^XXXJ ///'/;?'"/''- -' TO SUIT CUTTER V x'~''f/~'"'~' //"'"''/< ON DWG 5200.1-6G DIA^^'~ i ^525^j^MM'";>^X5/16"ALLOWANCE FOR FIN. I >y<I~X~ RII ~'2 LLEaTAPPEDHOLES y$$^'~'~'~~'~'~'^^KK'^\^X^ I 5~~~~~~~~5600' DIAM ^^\ ^ ^ ^\ FOR I" EYE BOLT-SPACED (IO X^^X'Z\'.X~,K)^~..~.~.~' ~~.,^^X.X"X^~-\~ -. ~-~REN VG SP:R:<NX "'"X <, ^^<><'~'~~ ~'~~~^^\X.X^ I I ^^V^^\\^\\X^SPURGEAR CHILLER RING) Ix 1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'1 -1/ 2'-.'L- \ F ^^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Il I I'x! 162" DDAM __ II T ^^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ""'1^^''I R'" SPRUE^XX^ ~~-6" DIAM IT _SAME ASCNGATE ON140%215-.\\\V ^^^XV\~ ^X^X^ ^\X DWG E- 503.3-1 EXCEPT DIM. \ \ Fig7. 17. Permanent moid. design for precision-cast final drive gear, Part No. 7564141.

!;iii,,~!11i:~~: ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i"~:: ~~~~~~~~~~~~~~~~~~:: at i: ~ TA:0~::-:!:: E E ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~::-:;:f: 0 w;i_::::~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*i~~ i.. I~~: i- is ii F-i~~~~~~~~~~~~~~~~~~~~~~~~~~~ii_:~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~i~ ~~ ~ ~~:i:-::~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ii Fig,. 1&. Precision cast AIS! 1062 final drive,r a ~.7'>11l:,I ~ ~ ~ ~ ~ ~ ger Part: _o }g4!

Fig. 19. Precision-cast AISI 4817, Pa t No. 756411i, with pinion ready for test. -Cz- 1~~~~~~~~~~~~~IV1 -'4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. Figure 19. Precision-Cast AISI 4817, Part No. 7564141, with Pinion Ready for Test.

m4 I \\ I I^'''I I \ J Present Forged Design x.^'AV - 1790 Engine 1L\ \_ \.] New Cast Design Figure 20. Cast and Forged Crankshaft, AV-1790 Engine. 75

Present Forged Design -\ - AV- 1790 t^^^^^^ Engine >1r - Figure 20. Cast and. Forged Crankshaft, AV-1790 Engine. 75

,~::I/. Figure 23. Stresscoat Pattern on Present Design, Forged Crankshaft, P/N 529083 - S.L. 258. 77

...........ii ti~~~~~~~~~~~~~~~~~~C?: gi a~ ~::07: I:ji iiL —-: —-"" s *;, 3~~~~~~~~~....:s........:::~ Figure 22. Test Set-Up for Experimental Cast Steel i~~~~~~~~~~~~~~~~i::: P~~~~~~~~~~~:::::11~~ 1~~~~~~~: Crankshaf~~~~at Deig.L 28.I: 7 ~~~~~~~~~~~~~~~~~~~~~~~~~:- i -.:.-.',..:,- i~ii:ii,::: -.................... - s F~~~~~~~igur l2 T i!es0t — Set-U fo Ex 0 0 0 t04 iw -st > - --.. shat w! sig S.Lf'y.f"'-0W-0-f~E ~ES0000fffDA t l0004 S005 *:;0'.......';0::'S...'l;T';002ll700V'l:Q000||glt7; 0;\,|0 0i 98 xi,'4'

470 f~~~~~~~~~~~i:!'i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~iii ~ ~ ~ ~ ~ ~ ~ 3iiiii5 7,b ~ ~'/ ~' ii'~''~~:iii:......... 4,11'/"::;tr a in VaI.lue, 3as du I sc on 1 0, 000 i n LI. -1 L di r. },t,n t iii::: ~: ~:.:..:~ ii... 3'8h a -~~~~~~~~~~~~~~~~j" acdet. Figure 24. Stresscoat Pattern on Experimental Cast Steel Crankshaft (Modlified F iet) SIL. 238. 79

0- 0 Cost steel-modified with 250 radius in cheek.-* Forging- present design, P/N 529083 700 -3 2 3 - 600 Strain gage Uw location 2'- 2 -2 C-),400 zt 300 200 0 __00 200_____ _____ L 0 100 200 BENDING LOAD, POUNDS C( 50.5 IN. Fig. 25. Strain vs. bending load for cast and forged design crankshaft, AV-1790 engine. 80

16~ ~ All sufcsmc,. e,c~j-sufce m lle FOREDC cas+ 4 requ~~~~~~~~~~~~~~~~~~

(a) (b) Fig. 27. Sand-cast aluminum-alloy crankcase for AV-1790 engine. 82 r~~~~~~~~~~~~ ( b )~I Fig. 27. Sand-cast aluminum-alloy crankcase for AV-1790 engine.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:~:: 82i:~iiii~::~:i~::

' SBONU'BIu XOITI-snoJx3jJ S'o aau (q) pusR'G(UUS o6LI-AV zOJ asoesyrjBso XoITI-amuTmInB;sos-pues;uasald (a) Jo uoTOI S'85'T2, iU._ i ~ i l T~ s T- t~:__~ ~r__ I -________ (e) X r ~ X I: 11

33/8 OMT ONE TOOTH ON THS e AS SHOWN \5 5 \~ ~ 0'"2~2 3M35 WEAK COR~NER 118R R/' Z~/53L'"~ ~ 1RI SCCTION ~ ~ DI MAX -_ 4__4_ t cm 0~~ -I 31I I Ilia 3f4R~~~~~~~ Figr. 29. Suspension and spindle arm. PartNo.TA191 —-— 0-1. 3 318 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~9 y4R~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 314 R 3 112 SECTIONAFi.29 usesinan p de r,,PrtN. T 190I5

NO. OF TEETH 300 DIAMETRAL PITCH 3.5 PRESS ANGLE 25" PITCH D. 857143 INTERNAL D. 85304 BASE CIRCLE D 776835 CIRCULAR PITCH 8976 3/64 MJX (OPT) A DDENDUIM.2051 WHOLE DEPTH.6529 13/16 48HOLES EQUALLY SPCED~.010 (NOT ACCUMULATIVE) CHORDAL AD0EN0UM.2048 rLEAVE 5/8 MAX SOFT SPOT 64MAX. CIRCULAR THICKNESS 360) -3643 CHORDAL R D. OF MEASURING PIN 432 V4 2J MEASUREMENT BETWEEN PINS 85.4034 - 854104 MAX TOOTH -TOOTH ERROR 001 MAX PARALLELISM ERROR WITH DTAI AXIS/IM LENGTH.001 DTAIL MAKXINVOLUTE PROFILE ERROR.0003 FART NO. OF MATING GEAR 7354162 CENTER DISTANCE 4L000 / BACKLASH WITH MATING GEAR.010-0165 0.0. OF MATING GEAR 4.446 74-100 090 R A 1-f% 0; 205 -i- A 460'.39 REE 3_6 _4 i V 48 TAMP FRPONT 60' 2 ^/ R 1/8 R ^^' ^ ^1111^^'^^ 03= BASIC RACK DETAIL A)OPT) 92J25D.aC I/B — ~ c 89 125.+0300. OPT. STEEL HOT ROLLED FS8650 FS8750 OR FS4150 3 )L - 1116~~~~~~~~~~~~~~~~~~~~~~~~~~~~111 2R C~~ i ti~t- i~ uax~~b R~p~N~t~~6 2yV C MATERIALL: FORGING STEEL -FS $650 -FS SM OR FS4150 1O /16XIS'CHAM WER SPEC-5 1 -105-1 E ~_ OPT STEEL HOT ROLLED FS OM0 FS 850 OR FS 4150 SPEC QQ-S624 SECTION B- - -~ ~ OPT. STEEL TS-4150 - 3 HEAT TREAT- 235-311 BRINELL HARDNESS BEFORE FINISH MACHNNG 89 ~3 / PI MAGNAFLUX AT FINAL INSPECTION~ g 3, __ _ __ _ _ ~ 875-.oo________ "^ 1^~ -^90.8~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Q75-M 03 SECTIONA-A Fig. 30. Outer race, Part No. 7384006.

.....?""''^it t i.... - ~.~ i.'"^ ^ ~:i- ^iiiii~i: ~^"'w;: "~:. ~';.~'^^~.~ - -' ".~ ~ _'. ^ ^~~ "~~;~~ ^'~ ii ^ ^ i ~ f-''"" ^~-.:~~~~~~~~~~~~~~[~ 2~-^^... ii-:''/;ai::~.ii?: i:iii? {{i[[!i~ ii~ll~~i~i C2."'".' ~. ^ —:-^^^".'-. ii? Fig. 31. Mating surfaces of small shell step mold. 86!!!Pii i!?ii~~~~~~~~~~~~~~~~~~~~-ii~i iiii-iii_:, iii:i.l_-':.......:!%~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-iiiiii:-:i-~iii,:i-iae~~ F i g.: 3 1. -: Mat:ing sufae of:: s mll she;ll step mold.:;~;~~~

:S - *) C) C 5,, ^:.\^:, GREEN SAND Figure 32. Cross Section of Assembled Shell Step Mold. S~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,' 0 5,, ~ 5, Figur 52. Cros Secion f Asembld ShllSep old

i iiiii~iiii i~~11~!J ~'~~...... d: HA i~~id:S..... i..i ~ I:iiE0: _:i~iti:i:~::::~:it::::51. a:S:B::::ii::::::::~ ~~~~70:;:0::~:;::;f::::::0:;:: 0:::* Fig. 33. Calibrated spring clamps. i;:fii fi"':: ff?"':iisc: ~:i:: t~; i i: A; t:/~ y t iiiiii:;iiiiii;i:: i X' 000V t -lf40Stl00dt::;000 00 0:,0-0,S! SI;. tt Tt ttftt*-::.........:.s -: ttE..... -..:: -;i Fig. 33. Calibrated spring clamps.

7.30 A20-3 7.20 7.10 7.00 6.90 6.80 6.70 6,60...010.020.030.040.050.o60 Bulge - Inches/Inch Fig. 34. Casting density vs bulge of 2-inch step for unrisered castings. 89

'.::' ^ ~ - -,^"".' ^,t:,. J.. ~ -...' i ~.. ~ iii l' l'>........... Fig. 35. Surface defect caused by SiO2 reduction. maM Fig. 36. Mating surfaces of large shell step mold. 90

UNIVERSITY OF MICHIGAN 3 9015 02826 I 3393 3 9015 02826 3393