The Iron Age 1920-03-11: Vol 105 Iss 11

NS et Oe ee eS = IKE IRON ACE Established 1855 New York BUNNY YNLEOLLEARLAA LLY LAALOAODDASLASOONOUULLLAQASOOUGAOOLQAOLUAEOOONEDLOLDLLASOGREDAMENSAALDUUUULGANUO0UUC4U4440000000000000000 0044000800 UUULOU0 NNO UUEO LONE DUO OAQEREUUUU GANGA UUAU DONNA UD AANAONRRRUUOEN AOD OAH OASA 0040000 ENAMOAMfNM TABLE OF CONTENTS - - - 771 Buyers’ Index Section........... 499 Contract Work Section Wanted Section aL UUUUNONUANLAAAAAARSTONASEONEATVOUAUUUAQeNAGnONNgnESQARONOLE}OOUUUOQdoUodnd4eu4dsdgsegengteuuuduunuuvdnsssengseeseaeersaneavenuantgeestaneengeeeeeertevpOUUUNAAUUUAAOANN ECONO UU LH TUAAAeAe ANAS WY, From the time. of ubel Cain— e Qua ity of a product has depen prin- cipally upon the standards set up and held to by its makers. ‘Red Cut Superior The Nationally known—First Quality HIGH SPEED STEEL contains the highest grade raw materials, which before being used are subjected to exhaustive scrutiny by our metallurgists. Every bar of the finished product is a composite expression of skilled Tool Steel makers to whom from father to son through many ae the word Quality has been a living ideal. Red Cut Superior The Best for All Machine Work. Are your tools made of Red Cut VANADIUM-ALLOYS STEEL Co. GENERAL. SALES OFFICES: PITTSBURGH, PA. WORKS: LATROBE, PA. BRANCHES: Buffale Boston Cleveland Chicago Cincinnati Detroit New Vork ( Warehouse) ( Warehouse) Philadelphia Pittsburgh Toronto ee Si. ne , MAKING IRON IN MAKING IRON a 2 ANCIENT INDIA BY ETRUSCANS fae ‘i ADVERTISING INDEX - - - 526 489 Business Opportunities. . 5g hae Pere Tr eee 478 Help and Situations Wanted..... 485 Clearing House Section.......... 425 March 11, 1920 THE IRON AGE March 11, 1920 = ea - ree nen a _ . — A A ATS NRA en SRS Nase gS SEU EE 10 ane The Paying Purchase | Steel for foundation and first floor of building shipped im- _ : mediately from warehouse. Construction finished one month ahead of time. | Saved one month’s income—over $1,000. The same principle applies to all modern construction and | manufacture. | Don’t wait for uncertain mill deliveries. Buy from the | Ryerson warehouse nearest your job or shop. Write for our Monthly Stock List. Orrices; ESTABLISHED 1642 INCORPORATED 1688 Orrices PHILADELPHIA SAN FRANCISCO sweet JOSEPH T. RyERSON & SON sss: TOLEDO BOSTON NEWARK CLYDE M. CARR, PRESIDENT JOSEPH T. RYERSON, VICE-PREsS. HOUSTON PITTSBURGH MINNEAPOLIS IRON STEEL MACHINERY PLANTS; CHICAGO NEW YORK BUFFALO ST. LOUIS DETROIT rt THE IRON AGE New York, March 11, 1920 ESTABLISHED 1855 VOL. 105: No. II Position of Tensile Tests in the Foundry Checking Metal, Melting and Molding Practices and Questions of Design with Special Reference to Aluminum Work BY W. A. GIBSON* this paper is mainly concerned. The diameter on a to the value of tension tests in a foundry. This doubt has arisen because of the fact that the same metal composition may give widely varying properties in a casting and also because of the fact that the properties of many test bars are widely different from the properties obtained in the castings made from the same heats. Doubt as to the value of a tensile test because of these apparent inconsistencies is caused by ignorance of the purposes [ the past considerable doubt has been expressed as machined specimen is made 0.505 in. for the reason that this gives an area of almost exactly 0.2 sq. in., thus facilitating computations in the formula: Stress equals load divided by area. The 2-in. length was arrived at because of the varia- tions in ductility which occur as the ratio of length to diameter is varied. This is more important in ductile materials than in brittle. For example, the curve re- produced shows the way in for which such tests are 50) which the ductility may be made. made to vary in a mild steel These purposes may be 45) as the ratio of length to subdivided into three main 40| diameter is changed. All divisions as follows: these tests were taken from 1—To check up metal and = 33) a the same bar of steel so that melting practice. = 30} = the material was as near 2—To check up molding VY | uniform as possible. The practice. . ae INS = curve shows that the ductil- —_ give assistance In K 20) | tee! ity may be made to vary z iad. | from 42 per cent down to A large number of people WW 15) a i = less than 12 per cent by believe that the main pur- « io) changing the’ above ratio. pose of a tensile test is for q | This brought about a stand- the third item given above, 5} ardization of 4 for the ratio namely, to give assistance of of length divided by diam- V in design, when by far its greater value is for the first and second purposes and more especially for the first. The first and _ second purposes might be grouped into one, but the variables in the molding are so many that usually they are kept separate and the tensile test is depended upon mainly to check up the metal and melting practice. In order to check up metal and melting practice, it is therefore necessary to standardize very carefully the methods of testing and of casting in order that results obtained in one foundry may be comparative with re- sults obtained in another, or that tests made one day may be compared with those made at another time. Variations caused by methods of testing may occur due to choice of section or to methods of operating the testing machine. The American Society of Testing Materials has done much to standardize the size of section. Errors Due to Choice of Section The standard test bar as specified by the American Society of Testing Materials on page 248 of their 1918 standards, consists of a 2-in. gage length % in. in diameter with ends to fit the testing machine. A radius of not less than % in. joins the gage length to the ends which are usually % in. in diameter. This is the stand- ard steel specimen for automobile work and has been adopted as the standard in cast aluminum, with which *Engineer of tests, Lynite laboratories of the Aluminum Manufacturers, Inc., Cleveland. [2s 4567T89 ON 2 BE LENGTH DivipeD By DIAMETER Relation of Ductility to the Ratio of Length to Diameter in Tension Tests of Structural Steel eter of test section. It also frequently led to the use of the per cent reduction of area in specifying the abil- ity of a material to with- stand deformation. The error is greatest in a material which is subject to a large and localized reduction of area. Since cast aluminum alloys are but very little subject to this “necking,” this ratio is not as important in aluminum as in steel, but is maintained in order to follow the general practice in obtaining ductility meas- urements. Under ideal conditions all bars should be tested using threaded ends turned in the lathe and with self- aligning adapters. In this case the centering should be done upon the test section and not upon the threaded ends of the bar. In other words the thread must be concentric with the test section. In many cases it is not practicable to thread bars and the use of jaws must be resorted to. If a testing machine is in perfect align- ment, the use of jaws introduces but very little error. The trouble is that it is never certain just when a ma- chine is out of alignment and the error. caused by the use of jaws is therefore a variable one, varying from zero in some cases to as high as 100 per cent. In the majority of cases it will be close to 5 per cent. Fre- quently such an error is not objectionable, but any person using the data should always have this element of uncertainty in mind. Where a close comparison upon which much may depend is desired, threaded ends must be used if slight differences are to be detected. Flat test bars are even more objectionable than the 725 726 round from the point of accuracy. It is an exception when a jaw can be made to grip a flat bar evenly across the entire width of the piece. The amount of error that may be introduced in a test of a brittle or semi-brittle material by this eccentricity is almost unbelievable. In one example, a flat test bar pulled under the best con- ditions possible, an eccentricity of one hundredth of an inch introduced an error of 3000 Ib. per sq. in. into a total strength of 18,000 lb. per sq. in., thus causing the apparent strength to be 15,000 lb. persq.in. This was under the best conditions. The possible, and even the probable error, is much greater. Errors Liable in the Testing Machine The errors due to the testing machine itself are: 1.—-Machine out of balance. 2.—Wrong ratio of lever arms. 3.—Lack of alignment. 4.—Method of balancing beam. The first error is one which can easily be corrected by the operator. The second will probably be negligible with any standard type of machine and can only be checked up by means of a calibration which may be done by any of the well-known methods. The third, lack of alignment, is liable to occur at any time due to shifting of the table on the knife edges, sticking of the jaws, etc. As long as jaws are used it is never certain just how much of this error is intro- « is" \ ge Pw es 25...» + « SECTION A-A “6” fisersts Dia le 3 -« RS Method of Gating Individual Test Bar duced. The only way to eliminate it is to use spherical adapters and threaded ends. Under non-alignment would come errors in machining. If threaded ends are used the centering must be done upon the test section and not upon the ends. The fourth error is one which is entirely due to the operator. The amount of error which may be intro- duced by a dishonest operator, especially with a ma- chine whose capacity is far in excess of that necessary, is almost beyond belief. All of these errors in testing can be avoided easily if proper care is taken. The Effect of Variations in Pouring Variations in the method of casting are much harder to take care of and it is very seldom that test bar results from one company are exactly comparative with those from another. In fact, frequently those from one plant of the same company are not entirely comparative with those from another plant. An at- tempt has been made by the Aluminum Manufacturers, Inc., to eliminate this variation by the use of a standard test bar pattern. This pattern as developed by the Lynite laboratories of the company is shown in the accompanying line drawing. This method of casting was adopted for two reasons: 1.—It gives as near uniform conditions as possible, as it is very nearly independent of the speed of pouring. 2.—It gives a good quality of casting. It accomplishes this in two ways. (a)—-The metal enters the test section from opposite ends at the same time after which the risers are filled. This gives quiet metal in the test section at the time of solidification and also causes the risers to be filled with hotter metal than the test sections themselves, thus making certain that the test section will be solid long before the riser has solidified. THE IRON AGE March 11, 1920 (b)—The path of the metal from the sprue to the casting is long, giving a good chance for the dross to float toward the top of the runner and into the riser. For the sake of speed this pattern is usually put on a match plate and a large number of molds may be made by one molder. Since the gating is included in the pattern itself all variations due to the gating are eliminated. In order to eliminate variations due to ramming of the mold, the use of a squeezer is advisable. This method of casting has given extremely uniform results and is considered by the writer as the best pos- sible one to use in checking metal and melting condi- tions in aluminum. In addition to standardizing the method of molding, it is vitally necessary to standardize the temperature MaeenecennnnnnnnnnnncisensecnensertiiientpeneecaeneeanenenNeedeneenNRNONAAEANEEL iNELUONNEDOOEL ONAL OAAORONELUUUALITONEUDONEDOOOELOCAEDOCELENOELLUOOTOGECLOECO CLEC OU CULO HOON OO Table 1—Effect of Pouring Temperatures Sample PerCent Number Break Elongation 1827-M 19,883 1. 1827-N 20,120 ‘ 1827-—O 19,831 : 1827—P 17,404 1827-Q 17,020 SOLENEEDOEENUDONEEONONEONNOHODEEDDONOEEAONOROAOEEOAOODDORNDDOOEEOLOORONOONED Pouring Temperature TO) of pouring. This effect is well shown in Table 1 which gives the effect of various pouring temperatures upon the standard No. 12 alloy containing 8 per cent copper, balance aluminum. From these values it is evident that in order to compare the qualities of metal and the correctness of melting practice, it is necessary to keep this tempera- ture very constant. The values given are for No. 12 alloy which is much less sensitive than some of the other more special alloys, which are now beginning to be used. Variations of Properties Due to Molding It must be borne in mind, that by this method of making test bars no check whatever is made upon the molding variables. The standard test bar checks up the history of the metal up to the time it goes into the mold. If the molding is improperly done, the only way to determine this is to take the casting itself, cut it up into parts and take tests from various points throughout its section. Several examples in which this has been done are given in Tables 2, 3 and 4. Te cee Table 2—Variation in Physical Properties Obtained at Various Points Throughout an Aluminum Crank Case Used in a Well Known Airplane Engine Per Cent Break Elongation 28,550 22,600 17,400 Location Thrust bearing near propeller end... Web supporting crankshaft bearing. Lugs for attachment to frame Wall at transmission end of case.... 22,600 Crankshaft bearing 19,900 (Each value is the average of five tests.) Table 3 a Bronze Worm mobile Axle bobs pore bo wNoo-S NoOoM- Variation in Properties Throughout the Section of Gear Made for a Widely Used Auto- Yield Tensile Per Cent No. of Point Strength Elongation Tests Outside edge of rim.. 25,806 44,546 19.! 12 Center of rim 24,505 34,810 7.9 12 Inside of rim, edge. 23,024 27,723 5.2% 1 Lug 23,2 26,394 Specimens 1-15 24,52 35,541 Specimens la .. 28,823 32,336 Specimens 1b—15b... .24,4! 34,136 tim opposite lug.... 23,95: 35,246 Rim between lugs... 25,095 36,101 Table 4 Location -Variations Throughout Various Sections of a Caet Aluminum Truck Wheel Per Cent Charpy Elongation Impact Outer edge of rim.. 10,337 18,697 5.9 2.123 Inner edge of rim.. 11,650 18, 2 Web 11,772 Outer edge of hub.. 13,225 Inner edge of hub.. 13,630 Webs at rim Locations Yield Break ot SO tote wo theo 57% o. 2.00 2.25 2.69 2.44 HUODEEUEOUNEOTONOEDONELOLONEDONEEEDOOEHOOONNEOOOEEHUNNCOONREROUNEOOUNRCOUONREDNONENONENOOND Coe AO 2 be tt he AID — These tables, especially Table 2, show that it is pos- sible to check design and molding practice by means of tensile tests. For example, in Table 2 the gates have been attached to the engine lugs, which should be the strongest part of the casting. This test showed March 11, 1920 that this method of gating was improper and a change was immediately made. Castings made with molten metal from a given heat may have better or worse properties than a test bar made from the same heat. Test bars cast from the same heat as the casting whose results are given in Table 1 tested 24,500 lb. per sq. in. No figures are available on separate test bars from the heat used in making the casting of Table 2. The individually cast bars from the same heat as that used in making the cast aluminum truck wheel of Table 4, gave a tensile strength of 20,000 lb. per sq. in. Unless the molding practice and design is perfect, a casting will vary from point to point throughout its sections as shown in the above tables. These variations are due to several reasons. 1.—Shrinkage. 2,—Gating. 3.—The personal element in foundry work. 4.—Poor design. Reasons for Variations in Castings Nearly everyone is familiar with the pipe in cast steel ingots. In some cases it is necessary to crop, that is, saw off and scrap, 15 per cent of the length of an ingot in order to avoid this pipe. This piping is caused by the fact that molten metal occupies more volume than the corresponding weight of solid metal. There are a few exceptions such as the so-called type metals containing antimony. This shrinkage from the molten to the solid state is the greatest cause of weak sections throughout castings. If one part of the casting is solid much before the other, the first part will then shrink and be fed up by the second, leaving insufficient metal to supply this second part. This will then be left spongy and weak. This can be eliminated in two ways: The gating may be so arranged that when the mold cavity is full the length of the path over which the metal has traveled will have been such that the heavy sections will be filled with colder metal than the thin sections. Due to mechanical difficulties this has a very limited application. It will be considered later under gating. © Risers may be used which are attached to the last part of the casting to solidify. These risers must be so located and of such a size that they will remain molten longer than the casting, otherwise they simply act the same as the first section mentioned and increase the sponginess of the casting at the point of attachment. The gate of a casting should be arranged so as to automatically skim the metal. This is more important in the iron and brass than in aluminum castings be- cause of the greater amount of dross with the first two metals. However, it is important in all. The problem of shrinkage may frequently be very much helped by gating. This can be done by so arranging the gates and the distance the metal has to pass through the mold before arriving at its final destination so that all parts will solidify at the same time. Proper gating of a casting is one of the fine arts of present day manufacturing processes. In gray cast iron the shrink- age is so small that this is not a serious problem, but with the high shrinkage metals such as brass, bronze, aluminum and magnesium, it requires a high degree of intelligence and thought. By personal element is meant care taken in melting and molding practice throughout the foundry. In other words, good or bad workmanship. It is the largest item affecting quality of product, but will not be con- sidered here as it is beyond the scope of this paper. Attention to Design In a low shrinkage metal like gray cast iron, it is frequently possible to cast a large section near a small one or to reduce a fillet to a minimum. However, this is bad practice even in that metal and always leads to bad results in brass or aluminum. As an example of the results of care in design, attention is called to the uniformity of physical properties maintained in the tests given in Table 4. Considering Tables 2, 3 and 4, from the above dis- cussion, Table. 2 shows especially the effect of poor molding practice and of poor design. The lugs attached to the frame, the third item given under Table 2, show THE IRON AGE 727 a tensile strength of 17,400 lb. per sq. in., which is at least 5,000 lower than the average and 11,000 lower than the maximum. This difference is explained when it is known that not only were the engine lugs much heavier than the parts of the casting to which they were attached, but also the gates were attached to these lugs. All the metal going into the mold therefore passed through this part of the cavity, warming it up and causing it to remain molten longer, thus causing shrinkage to occur in it as well as a very slow crystal- lization. After this series of tests the gating was moved to another part of the casting and much better results obtained. The value of 28,550 lb. per sq. in. taken from the thrust bearing near the propeller end was caused by the fact that this part of the casting was a long dis- tance from the sprues giving the metal a good skim- ming action as well as allowing it to be very cold when it reached this section. In addition this part of the mold filled up very quietly giving ideal conditions for high quality of material. Corresponding test bars were poured in individual molds from the same heat, the gating being similar to what would occur in the line drawing shown, with the center runner left off and one of the risers used as a sprue. These test bars gave a value of 24,500 lb. per sq. in. This comparison is interesting, in that it shows the harmful effect of a stream of metal being allowed to pass through a mold cavity, thus heating it up before the metal in it is allowed to solidify. The web supporting the crankshaft bearing and the wall at the transmission end of the case were of the same thickness and about the same relative posi- tion in the mold. This is well shown in their tensile results. The crankshaft bearing which has an average tensile strength of only 19,900 lb. per sq. in. was about twice as thick as the parts of the casting adjoining. This difference in section was partially compensated by the use of risers attached to the crankshaft bear- ing. However, as can be seen, these risers were not fully able to compensate for the weakness in the design. In Table 3 attention is called to the difference be- tween the tensile strength and the per cent elongation of the lug and of the outside edge of the rim together with the graduation from one to the other. A cast-iron circular chill was used at the outside edge of the rim causing the freezing to be progressive from the outside toward the lug. This is well shown in, physical prop- erties. Table 4 is introduced to show how nearly uniform results can be obtained when proper attention is paid to the design and to the gating. In the design of this casting particular attention was paid to the elimination of heavy sections. In the molding of it the gating was given very serious thought, and the arrangement of gates, chills and risers was the result of long ex- periment. This is well shown by the uniformity of physical properties. The weakest section, the web, has long been a difficult point in any wheel casting. The close approach of this web to the physical properties of the rest of the casting is a tribute to the care taken in molding, gating and pouring. Molding Variables Eliminated From the Test Bar It may be asked how it is possible to check up metal and melting practice without bringing in at least a part of these molding variations. In answer to this question the standard test bar is an almost ideal section in which it is very easy to obtain perfect feeding and skimming. In addition the shape is such that there are almost no cooling stresses left in it. No matter how carefully we attempt to standardize. the molding conditions there are a few variations from this source introduced and we can only keep them to as much of a minimum as possi- ble. As stated before, the method of gating shown in the line drawing almost entirely eliminates these varia- tions. In contrast to the test bar section, the average com- mercial casting is so complicated that perfect feeding and gating is mechanically impossible. The only thing which can be done is to approach as closely as possible perfect conditions. The greater the skill and intelli- gence of the man first laying out the job for production, the more nearly these perfect conditions will be approxi- 728 mated. The fact that the difficulties in his way are largely mechanical rather than scientific should not minimize the importance of his work or the high degree of intelligence required for it. Function of Tensile Tests in Design The third reason for carrying out physical tests would be to give assistance in design by furnishing data as to the mechanical strength of materials. There are very few cases where it is possible to compute the actual loads which a machine part has to stand. An example of this is the discussion which is constantly going on as to the proper method of computing stresses in the crankshafts of automobile and aeroplane engines. This doubt as to the amount of load acting on even a com- paratively simple machine part is usually much greater than the doubt of the quality of the material entering into it or of the formule used for computing the stresses after the load has once been determined. The only way to tell with certainty how any part will act in service is to actually put it in service and run it to destruction. Naturally, this is a long and expensive process and impossible in any but a very few cases. Therefore, we frequently are confronted with the so-called imitation tests in which a machine part is given an accelerated test under as nearly as possible the same conditions as those that would be obtained in practice. The question of imitation tests has recently been discussed in an article in Automotive Industries of Sept. 11, 1919, by Dr. Walter Rosenhain in which he says, “First, that in order to obtain definite and reliable results which have a meaning capable of definite quanti- tative interpretation, we should seek to make our me- chanical test as simple as possible, seeking to measure one physical property or constant at a time, and avoid- ing the snare and delusion of the ‘imitation’ test. Sec- ond, that we should seek to express our opinion of the relative importance of the various mechanical proper- ties or constants in terms of definite ‘figures of merit’ which can be tested by comparison with service results.” While not going as far as Dr. Rosenhain in con- demning the imitation test, the writer believes that a wide experience with materials and with machine parts as well as a thorough knowledge of mechanics of forces is necessary to either plan or interpret such tests. Test bar results most certainly should not be used blindly in designing. Frequently, however, the differ- ence between test bar results and the properties of the casting, such as are shown in Table 2 and in Table 4, are allowed for in a factor of safety. When a factor of safety is given to cover this difference, considerable errors are liable to arise on either the safe or the un- safe side. The use of a factor of safety on test bar results without a proper appreciation of the limitations of such data has done much to give the “factor of safety” the title “factor of ignorance.” In order to use test bar results intelligently in machine design it re- quires not merely a wide knowledge of the mechanics of the distribution of stresses, but also a considerable experience of the way in which the properties of materials vary in different sizes and shapes of castings. Cincinnati Metal Trades Meeting The annual meeting of the Cincinnati branch of the National Metal Trades Association was held at the Business Men’s Club in that city, on March 4. Officers for the ensuing year were elected as follows: Presi- dent, J. B. Doan, American Tool Works; vice-president, J. W. Carroll, Lodge & Shipley; treasurer, H. C. Hoefinghoff, Cincinnati Grinder Co.; secretary; D. C. Jones, Lunkenheimer Co.; executive committee, R. T. Hazleton, N. C. Lamont and A. B. Breeze. J. W. O’Leary, Chicago, president of the National associa- tion, and H. H. Merrick, of the Great Lakes Trust Co., Chicago, were the principal speakers, Mr. O’Leary giv- ing a detailed account of the proceedings of the first industrial conference at Washington, of which he was a member, and Mr. Merrick reviewed conditions as they exist in the industrial world to-day, paying par- ticular attention to what he considered the iniquitous excess profits tax. THE IRON AGE March 11, 1920 Defects in Steel Revealed by Acid Etching WASHINGTON, March 9.—Results of experiments of the United States Bureau of Standards to determine defects in steel are described in a report by Henry S. Rawdon and Samuel Epstein. An abstract of this re- port, which has just been completed, follows: The deep etching of steel by means of concentrated acids, that is, hot concentrated hydrochloric acid is a very impor- tant means of revealing the structural condition of such material. The features revealed may be of three types: Chemical, such as heterogeneity, due to segregation, to com- position changes accompanying welding, carburization and similar processes, to lack of solution and uniform distribu- tion of special addition elements in alloy steel, etc.; me- chanical, such as initial stresses, due to preliminary mechan- ical and thermal! treatment; physical, such as internal, frac- tures and the similar discontinuities within the metal. Steel, which is heterogeneous in its chemical composition, shows a characteristic roughened surface when etched, cor- responding to the variations in composition. This roughening is due to the greater solubility of certain portions, particu- larly the non-metallic inclusions, sulphide, oxide, etc., and to the deepening and widening of the pits resulting upon the removal of these inclosures. As specimens typical of highly stressed ened steel balls were used. The stresses, due to the mechan- ical and thermal treatment the balls had received, were often of sufficient magnitude to cause the balls to crack spontaneously when subjected to the action of concentrated acid, The behavior of steel in this respect is identical with the corrosion-cracking of brasses and bronze under certain conditions. Physical discontinuities, which may exist within the steel, usually occur as tiny cavities or pores, giving rise to spongy metal, which is quite readily detected. Certain discontinui- ties are of the nature of internal fractures, the metal being still in such intimate contact, however, that they are de- tected only with extreme difficulty. This type occurs rather abundantly in defective rails of the transverse fissure type. Similar defects have been found in forgings, such as steel axles and tires. Deep etching reveals these internal fissures as short cracks which appear to bear a definite relation in their arrangement to the direction of the stresses to which the piece had previously been subjected. The presence of these defects in steel, previous to etch- ing, has never been demonstrated by previous investigators, due to the lack of a suitable method. Hence various con- flicting opinions have been held as to their nature and the seriousness of the defect. In order to show the presence of such internal cracks, a special magnetic method was used. The specimen was polished as for microscopic examination and, after magnetizing, was bathed in a suspension of fine iron dust in kerosene. -Discontinuities in the metal were clearly shown by the arrangement of particles of iron dust along their course. Subsequent etching tests showed that each crack revealed by the etching is due to the widening and deepening of a pre-existing internal fracture in the steel. material, hard- These defects are essentially fractures circular in area, which occur across what is otherwise sound metal. By suitable means the metal may be opened up along the line of the defect showing that there is a definite discontinuity within the metal and no coherence at all. These areas are identical in appearance with the starting point of the very serious defects known as transverse fissures in rails. 0. F. 8. Waste Material Dealers in Annual Meeting The activities of the newly-formed scrap iron divi- sion will be a feature of the annual meeting of the National Association of Waste Material Dealers, Inc., at the Hotel Astor, New York, March 15 to 17. This division will meet at the Hotel Astor at 3.30 p.m., March 16. First a chairman will be elected to serve a year. Consideration will be given to the advisability of asking the Interstate Commerce Commission for a formal ruling in reference to the shipment of machin- ery, boilers, etc., under the classification provided for scrap iron. A traffic attorney will be present to give advice. The metal division will meet at the same hotel at 2 p.m., March 15. The seventh annual banquet of the association will be held at the hotel at 7 p.m., March 17. The National Marine Exposition will be held at the Grand Central Palace, New York, April 12 to 17. The exposition will be opened by Secretary of Commerce Alexander and Secretary of the Treasury John Barton Payne. Pots and Boxes Used in Carbonizing™ Cast Steel as Compared with Cast Iron— Special Alloys as Materials—Cost per Heat Hour BY H. H. oped consistently for more than a quarter of a century and the heat-treating processes in their present scope of applications are practically an outgrowth of the automobile industry, no new process of carbonizing has been developed which eliminates the use of carbonizing boxes, with the exception of the gas carbonizing process for small parts and certain foreign developments in gas muffle carbonizing. Cyanide and lead hardening processes have changed very little and pots similar to those used 20 years ago are employed in the most modern practice. Three factors govern the service received from pots and boxes. They are: Design, methods of manufac- ture and material. \ ) ) HILE heat-treating furnaces have been devel- Variations in Design In certain large manufacturing plants there has been considerable improvement in the design of pots and boxes. Through the industry at large, however, a great majority of carbonizing boxes are of a design as well adapted for packing soap in as they are for carbonizing. The patterns are generally made by the handiest fence carpenter without a blueprint or a draftsman, unskilled in such matters, makes up a draw- ing of a carbonizing box, his only interest being the width, depth and length, without regard to how the box is used in the heat-treating process or the foundry methods required in making it. Frequently the dsign is left to the foundry, who knows as much about heat- treating as the average heat-treater knows about mold- ing practices and the blowing of steel. The foundry’s chief interest is to make the box heavy so that it will weigh more and can be manufactured with less care in molding. These practices insure poor design and consequent poor results. The life of a box is as much dependent upon design as upon analysis of material. Boxes can be designed to largely eliminate warping. Special design is necessary to facilitate handling the box, that is, taking it in and out of the furnace. Flanges, covers and methods of sealing are highly important and come under the head of design. A box must be designed to permit tight sealing with some refractory or other material, yet without the clay being permitted to enter the box and mix with the compound. Boxes must also be designed as regards their inside dimensions, with a view to the work which is to be carbonized, and a knowledge of how the materials within the box should be packed, the quantity of car- bonizing compound required, etc. The outside dimen- sions of the box must be made to conform with the size of the furnace, and be of such size as to permit the largest possible production of work from a given size of furnace. The thickness of the box must be as _ thin as possible to prevent warpage and still must be thick enough to permit proper flow of metal in casting. In this a knowledge of both heat-treating and foundry practice is necessary. Many foundries who know nothing about heat-treating practice and not too much about foundry practice are attempting to advise the heat-treating engineer as to the design of boxes. I do not know of a foundry to whom I would entrust the design of a carbonizing box. One of the alloy manufacturers specializes somewhat on design but is limited considerably because of the manufacturing diffi- culties with this material and their experience in de- *From a paper presented at the ‘January meeting of the New York Chapter of the American Steel Treater’s Society. The author is with the Quigley Furnace Specialties Co., New York. 729 HARRIS sign is, of course, confined to boxes and castings em- ploying the one metal only. Materials for pot and box manufacture may be briefly grouped under six classifications: Cast iron, cast steel, pressed and wrought steel, “trick” materials, alloys, and any of the first four calorized, or treated by the calorizing process. Cast Iron as a Material The cheapest first cost material obtainable has a fault common to most of the cheapest made—it is the poorest. Cast iron oxidizes very rapidly, is inclined to be very porous and allow leakage, and gives a very non-uniform service. Cast iron grows or expands under heat and does not contract. Carbonizing boxes fre- quently become so distorted that their tops do not fit and they leak gas and carbonizing material, becoming generally unsuitable for use. They also frequently scale on the inside if a small amount of oxygen is present when the box is sealed, the scale becoming mixed with the carbonizing compound giving generally poor results and sometimes spoiling the work entirely. Cyanide and chloride attack cast iron a great deal faster than they do steel or the alloys, and steel gives much better service under lead conditions than cast fron. Cast Steel Boxes Cast steel is generally much superior to cast iron. It costs about twice as much per pound but many times longer service can, as a rule, be expected. The prin- cipal objection to cast steel is that the grade of steel used for pots and boxes is generally of inferior qual- ity, sometimes even semi-steel being offered for steel. As most foundries sell pots and boxes at comparatively low prices, they run them from their lowest grade steel or the tail end of the ladle, making poor castings. Steel castings do not offer the ease of cast iron prac- tice in manufacture. Better molding is required and great care must be taken to exclude foreign matter, such as sand, slag, refractory material, etc., from the metal. There is a large percentage of steel castings scrapped and a larger per cent patched up and sent out which should be scrapped. Sand and slag inclu- sions, blow holes, cold draws and cold shuts are com- mon defects in steel castings. Poorly made castings are inclined to be very porous, allowing uneven scaling and frequently leakage when used as containers for lead or cyanide. This can be largely overcome by extra care in manufacture and close inspection. Also to a considerable extent by close regulation of the compo- sition of the metal, without affecting the price to an appreciable extent. Steel pots and boxes now in use are mainly cast from ordinary steel or steel intended for miscellaneous castings. This class of steel will stand much improve- ment. A great many of the steel boxes used in the automobile industry are produced by.the foundries in the Detroit district, whose product in the past has been far from uniform, due to poor molding, bad in- spection, etc., brought about by labor troubles. The Eastern foundries as a whole turn out a more uniform quality of steel than foundries in the Detroit district and have better inspection. Their prices are a little higher, but I believe are warranted by the quality of the product. So far, judging from an average of various tests and conditions studied in many of the largest plants in the country, I do not believe any material has shown on the average a uniformly longer life per dollar investment than cast steel, with the exception of one 730 which has just been commercially introduced, although tests of it have been conducted for several years. This is a special alloy known as Q-Alloy, which will be dis- cussed later. Pressed and Wrought Steel Pressed and wrought steel cyanide, chloride and lead pots were quite generally used some time ago before the advance in price of this material. Its ad- vantage is its light weight and consequent small heat consumption. Its disadvantages are its high price and the comparatively few ‘shapes in which it can be secured. Several attempts have been made to produce satisfactory pressed steel carbonizing and annealing boxes, but have never been successful. The large majority of pressed steel pots are sold through specu- lative supply houses, who add a large profit, rather than sold direct from manufacturer to consumer, as cast material is sold. I know of a few companies manufacturing their own annealing boxes from wrought steel, riveting or welding them together. Some of these are giving satisfaction in extreme temperatures but not under general conditions. These boxes are not generally suitable for carbonizing. Very few are used as they do not hold gas and have a short life because of the minimum thickness necessitated by their construction. Wrought steel riveted or welded pots for cyanide, chloride or lead conditions have never been successful, owing to leakage. Trick Materials Under “trick” materials come the products of all firms who misrepresent their product or sell it under a trade name which does not correctly indicate its nature. For instance, a very poor grade of cast iron containing a small amount of silicon is sold as a “spe- cial heat-resisting metal” at a price several cents above its value. blanks with scrap in the cupola and selling the result- ing semi-steel as a “high heat-resisting steel.” Some foundries pour their regular run of steel used in making miscellaneous castings, from tractor wheels to window weights, into pots and boxes, yet advertise the pots and boxes as “made to resist high temperatures under furnace conditions” and sell them under a trade name implying that they were some special quality steel. Pots Made of Alloys Alloys for heat-resisting purposes are a compara- tively new development. Although there have been several very high-priced alloys manufactured for sev- eral years as platinum substitutes and the like, the first patent covering a heat-resisting metal to sell at less than $5 per lb. was issued in 1916. This was issued to John C. Henderson, Washington, D. C., and assigned to the Driver-Harris Co. A later patent issued to Mr. Henderson in 1918 covers the manufacture of carbon- izing boxes using the same alloy. These patents are not sufficiently broad to cover a fraction of the alloy field and are only the first step in the development of the carbonizing and annealing box alloys. Several steps have been taken since and from all indications the matter is still in its infancy. At this time there are more than 35 patents covering heat-resisting alloys. The analysis of several alloy materials manufac- tured under patents and otherwise is given in the table. A Nickel-Chromium Alloy The latest development in the alloy field is a special nickel-chromium alloy, the analysis of which I am not yet authorized to divulge. This material differs from nickel-chromium alloys on the market in that it retains many of the physical characteristics of the cold metal at a temperature of 2800 deg. Fahr., and rings like a bell when struck with a hammer at this temperature. It does not warp and is guaranteed against cracks and blow holes. As every heat of this material is carefully checked for uniform analysis, and as its manufacturing practice is highly standardized, any two boxes will show approximately the same number of heat hours in service under equal conditions. This is not generally true of other alloys on the market which, as THE IRON Some companies are melting down shell * AGE March 11, 1920 a rule, are far from uniform in service. This may be due to the fact that some of them are made from machine scrap containing copper and other elements not calculated to improve the quality of the metal. While far superior to ordinary alloys in many ways, this material is no higher in price than that generally asked by competitive manufacturers. This material can be also rolled and supplied in sheet form suitaLle for fabricating sheaths for the application of retorts, muffles, annealing pans, etc. This new alloy is known as Q-Alloy, Grade X. The manufacturers do not advocate it for the manufacture of cyanide or lead pots, as practice has found that nickel and chromium alloys do not as a rule compare favorably in service rendered per dollar investment with high grades of alloy steels for use under cyanide and lead conditions, particularly lead. Many alloy manufacturers have not always shown much discretion in recommending their product for use under conditions where it would not prove a saving for the user. Each material is best suited for certain specific applications and should be recommended only for use under conditions where it will give maximum results. Analysis of Alloy Materials for Pots and Bowes No. Chromium Manganese Iron (0, 2—Nickel, approximately Chromium, approximately Iron, approximately. i: CIEE 0S Wiha c. whe albus Sw ias @ oS eee 0.50 to 3 Chromium 1: Manganese Silicon Molybdenum ee Carbon 2.5 wets i ww- sos rer ro Oo ~Onfuro wr ote ANAIS n Chromium Silicon Aluminum Manganese Iron Aluminum Silicon TO Nickel or chromium alloys, for instance, are not suit- able for use as cyanide or lead pots, particularly the latter. The lead has an amalgamative action and tends to break down the structure of the alloy by uniting with certain elements. Steel, on the other hand, is highly unsuited for the construction of aluminum melting pots, as the steel not only contaminates the aluminum, but the aluminum makes short work of the steel pot. It is regrettable that so many companies manufacturing only one material recommend it to serve under all heat-treating conditions, as no one metal is a cure-all. Methods of Manufacture Granting that both design and material be right, if the .methods of molding, gating and pouring are not thoroughly understood, the product is a compromise. The mold must be prepared with a full knowledge of the conditions under which the casting is to serve. The metal must be poured at the proper temperature and all impurities removed by skimming. Ample risers must be provided so that every molecule of steel is under positive pressure at the moment of solidification. Density of metal is of utmost importance, To make good pots or boxes, the manufacturer should be thor- oughly familiar with heat-treating practice. Calorized Pots and Boxes Calorizing is a process for infusing aluminum into the outer portion of any metal article, forming a pro- tective coating of aluminum over aluminum oxide when exposed to high temperature. This process was in- vented by the General Electric Co. and is controlled by the Calorizing Corporation of America. Calorizing has failed completely in many instances as applied to pots and boxes, and I would advocate its use in the heat-treating processes only as applied to rotary retorts such as used in the American Gas Furnace Co.’s rotary carbonizing furnaces, pyrometer tubes and wrought pipe for special processes. The price is excessively March 11, 1920 high, being $1.50 to $2 per square foot of surface treated. The Maxwell Motor Car Co., in its Detroit labora- tories, conducted a test some time ago under the direction of E. W. Upham, chief metallurgist, to ascer- tain the value of calorizing as applied to carbonizing boxes. Four classes of boxes were entered in the test as follows: No. 1—Calorized welded sheet steel box No. 2—Calorized cast steel box No. 3—Calorized cast iron box No. 4—Uncalorized cast steel box. The results were as follows: Boxes of Class No. 1 broke down and burned out very rapidly owiug to the fact that calorizing over the welds gave way in a very short time, exposing the porous weld direct to oxidation. Class 2—Calorized cast steel boxes stood up fairly well but calorizing burned away in spdts, allowing oxidation to destroy portions of the box, thereby rendering it useless. Class 3—Calorized cast iron box warped and “grew,” breaking the calorizing, thereby exposing the cast iron be- neath, causing destruction of the box by oxidation. Class 4—Uncalorized cast steel box, oxidized uniformly, and was in fully as good condition after the completion of the test as any of the calorized boxes, but had lost more weight than the calorized cast steel box, owing to the pro- tection afforded by the calorizing which remained intact. This test proved calorizing of no value whatever applied to the Maxwell company’s boxes, and they gave up further tests. The Tool Steel Gear & Pinion Co., Cincinnati, Ohio, conducted a test of calorized riveted sheet steel boxes which proved calorizing of. little or no value under their conditions. The Brown Pyrometer Co. sold a great many calorized pyrometer protection tubes and are in a position to advise in regard to calorizing applied to this equipment. It is my under- standing that they found some calorized tubes gave Salisbury Iron Corporation Organized The Salisbury Iron Corporation has been incorpo- rated under the laws of Delaware with an authorized capitalization of $500,000 in 6 per cent preferred stock and 20,000 shares of common stock of no par value to succeed to the business and property of the Barnum- Richardson Co., Lime Rock, Conn. The new corpora- tion has authorized the issuance of $200,000 of secured serial 7 per cent notes, of which $150,000 will be pres- ently issued for the purpose of providing additional working capital. The business to which the Salisbury Iron Corpora- tion has succeeded was founded in 1734 and its princi- pal product, Salisbury charcoal iron, is known through- out the iron trade as the highest grade of pure char- coal iron. It was from this iron that a large part of the ordnance used by the American forces in the Revo- lutionary War was made. Contrary to the growth in all other classes of iron production, the past 20 years has seen a marked curtailment in the output of high grade iron in which only charcoal without the addition of coke is used as fuel. The demand for this class of iron, however, is large and the comparatively small tonnage now being produced in this country is chiefly attributable to the fact that the pure charcoal iron market cannot follow the trend of large tonnage pro- duction of other grades. The Salisbury Iron Corporation owns its own mines, foundries and furnaces. In addition to increasing its present capacity to approximately 10,000 tons of Salis- bury iron per annum, it will etxend;its foundry opera- tions in the manufacture of charcoal iron castings, chilled iron castings, gray iron castings and indus- +rial and railroad car