IC-NRLF 
 
 SB 27 073 
 
 pen Heart 
 Steel Castin 
 
 W.M.CABR 
 
LIBRARY 
 
 OF THE 
 
 UNIVERSITY OF CALIFORNIA. 
 
 Class 
 
OPEN HEARTH 
 STEEL CASTINGS 
 
 BY W. M. CARR 
 
 4 
 
 A complete exposition of the methods involved in the 
 manufacture of open-hearth steel castings by the basic and 
 acid processes. This work is compiled from a series of 
 articles by the author, written for and published by The 
 Iron Trade Revieiv and The Foundry. 
 
 UNIVERSITY 
 
 OF 
 
 Cleveland, Ohio, U. S. A. 
 
 The Penton Publishing Company, Publishers. 
 
 ipo; 
 
:AL 
 TABLE OF CONTENTS 
 
 CHAPTER I. 
 
 Page 
 Melting Stock for Acid Practice ............................. 8 
 
 Fuels and Alloys ............................................. 10 
 
 Molding Materials .................. . ........................ 13 
 
 Materials for Basic Practice .................................. 14 
 
 CHAPTER II. 
 
 Open-Hearth Furnace Construction ........................... 20 
 
 CHAPTER III. 
 
 Fuels and Accessories ........... ............................. 33 
 
 CHAPTER IV. 
 
 Manipulation of Heats in Acid Practice ....................... 40 
 
 CHAPTER V. 
 
 Manipulation of Heats in Basic Practice ...................... 
 
 Order of Charging ........................................... 
 
 Melting ..................................................... 60 
 
 Charging Cold Stock ........................................ 61 
 
 CHAPTER VI. 
 
 Chemical Analyses and Physical Tests ........................ ?2 
 
 CHAPTER VII. 
 
 Relation Between Composition and Physical Properties ........ 81 
 
CHAPTER VIII. 
 
 Blow Holes in Steel Castings 94 
 
 CHAPTER IX. 
 
 Discussion of the Causes of Cracks in Steel Castings 98 
 
 CHAPTER X. 
 
 Heat Treatment and Annealing 101 
 
 CHAPTER XI. 
 
 Repair of Steel Castings with Thermit Ill 
 
 CHAPTER XII. 
 
 Cost of Equipment for Open-Hearth Steel Foundries.. . 115 
 
 1.61533 
 
ILLUSTRATIONS 
 
 Page 
 
 Plan view, stationary type open-hearth furnace 21 
 
 Sectional elevation, open-hearth furnace 22 
 
 Cross section, open-hearth furnace 23 
 
 Gas Producer 34 
 
 Oil Burner for open-hearth furnace 36 
 
 Open-hearth furnace, arranged for burning oil 37 
 
 Diagram of variations in composition of a normal acid open- 
 hearth heat 42 
 
 Diagram of variations in composition of a normal basic open- 
 hearth heat 66 
 
 Standard Test Bar 78 
 
 Test Bar for Works' test 79 
 
 Shrink hole 95 
 
 Blow Holes caused by gaseous steel imperfectly deoxidized.... 96 
 
 Blow Holes caused by damp sand 96 
 
 Diagram of structural changes 103 
 
 Record of Government Tests 106 
 
 Specimen of cast steel as cast 1 108 
 
 Specimen of cast steel heated to 1,200 Degrees Cent 108 
 
 Specimen of cast steel heated to 1,200 Degrees Cent, air 
 
 quenched 109 
 
 Specimen of cast steel heated to 800 Degrees Cent 109 
 
pB*tAj^ 
 
 OF THE ^ 
 
 UNIVERSITY 
 
 OF 
 
 CHAPTER I 
 
 MELTING STOCK FOR ACID AND BASIC PRACTICE REFRAC- 
 TORIES FUELS ALLOYS MOLDING MATERIALS 
 FLUXES 
 
 In view of the growing interest manifested by both pro- 
 ducers and consumers of cast sections in the production of 
 steel castings and their increasing utility, the salient points 
 of their manufacture by the acid basic and open-hearth 
 processes will be presented in this series of articles which 
 cover : 
 
 First The selection and representative composition of 
 melting stocks, alloys, refractories, fuels, melting materials 
 and fluxes. 
 
 Second Furnace construction and the melting and ma- 
 nipulation of heats. 
 
 Third The conditions of melting as effecting the physi- 
 cal properties of products. 
 
 Fourth The analyses and physical tests of different 
 grades of products. 
 
 Fifth The effect of the constituent materials and metal- 
 loids usually present in open-hearth steel castings. 
 
 Sixth Heat treatment or annealing, with notes on micro- 
 scopic examinations. 
 
 Seventh Discussion of the causes of blow holes and 
 shrinkage cracks. 
 
 Eighth The repair of defects by thermit welding. 
 
 Ninth Approximate cost of open-hearth installations. 
 
8 MELTING STOCK FOR ACID PRACTICE 
 
 MATERIALS FOR ACID PRACTICE 
 
 MELTING STOCK. The composition of materials for acid 
 castings comes within well defined limits, for the main rea- 
 son that the process is nearer a melting, rather than a re- 
 fining one, owing to the fact that the metalloids, sulphur 
 and phosphorus, are not removed during the conversion of 
 the charge. As their presence in the finished product must 
 be subject to specification, it follows that the melting stock 
 must be bought with a limit to the contents of the elements 
 named. Only in regard to quantities of sulphur and phos- 
 phorus exists the distinction between acid and basic melt- 
 ing stock. Pig iron for ordinary practice analyzes as fol- 
 lows : , 
 
 Total Carbon 2-3.5 per cent 
 
 Silicon 0.50-1.5 per cent 
 
 Sulphur 0.04 or less 
 
 Phosphorus 0.04 or less 
 
 Manganese 0.50-0.75 
 
 In addition to pig iron, steel scrap known as "basic scrap" 
 is also bought on analysis or chemical specifications. Being 
 the product of rolling mills, etc., following basic practice, 
 the contents of sulphur and phosphorus are usually low, 
 and they are the only elements of composition taken into ac- 
 count. A representative analysis of steel scrap for acid 
 practice is as follows: 
 
 Sulphur 0.015-0.03 per cent 
 
 Phosphorus 0.010-0.03 per cent 
 
 Since the physical character is usually represented by 
 billets, crop ends, blooms, plate-shearings, defective castings 
 and waste metal from steel foundries, the nature of them 
 necessarily predetermines the composition in regard to the 
 presence of carbon, silicon and manganese. With a charge 
 of pig iron and steel scrap it is comparatively easy to keep 
 
REFRACTORIES FOR ACID OPEN HEARTH FURNACES 
 
 the composition of finished product within acid specifica- 
 tions. 
 
 REFRACTORIES 
 
 Silica sand forming the hearth of the furnace, known 
 chemically as an acid, lends its classification to distinguish 
 between the two processes. It does not present a condition 
 wherein certain elements are removed during melting of 
 stock, namely, sulphur and phosphorus. The amounts of 
 those elements charged will be found in the finished product. 
 The function of silica sand is mainly a refractory one. 
 That is, it must have heat resisting qualities, but not to the 
 extent that it will not set or sinter slightly to satisfactorily 
 preserve the contour of the shallow dish-like formation of 
 the hearth. It must not be too fusible, otherwise there 
 would be excessive scorification or cutting of the hearth. It 
 must set or sinter sufficiently to resist abrasion due to the 
 charging of stock. It is difficult to lay down strict chemical 
 specifications for silica sand. There are certain conditions 
 of composition not expressed by an analysis of the total 
 constituents. The combinations of them with their neigh- 
 bors cannot be ascertained. However, a silica sand of the 
 following analysis gave excellent results in practice: 
 
 Water 0.24 per cent 
 
 Silica 97-25 per cent 
 
 Alumina and Iron Oxide 0.16 per cent 
 
 Lime 0.08 per cent 
 
 Magnesia 0.39 per cent 
 
 Alkalies 0.36 per cent 
 
 Loss on Ignition 0.36 per cent 
 
 There are several deposits of silica sand in the west and 
 middle west that are supplied to steel foundries and no 
 difficulty will be found in getting the best quality for the 
 purpose. As a general rule a silica sand for hearth lining 
 must be low in lime, manganese and the alkalies (potash 
 
IO FUELS AND ALLOYS 
 
 and soda) ; an excess of any tends to lower the fusion point 
 of the sand, destroying its required sintering or refractori- 
 ness. Usually a silica sand with less than 95 per cent silica 
 will not answer for a refractory. 
 
 FUEL 
 
 It is a question of local conditions as to whether the fuel 
 may be natural gas, oil, tar, or producer gas. Natural gas 
 is by far the most satisfactory, owing to its high calorific 
 value and its non-contamination of the bath or molten 
 charge. It is fed directly into the working body of the fur- 
 nace without any preheating or a passing through the re- 
 generator chambers. Oil, next in efficiency, may be crude 
 petroleum or a grade known as residuum ; a by-product of 
 petroleum distillation. The heating values are high, and 
 in certain grades the composition will answer for acid work. 
 Some grades are rather high in sulphur, which is absorbed 
 by the stock in melting. 
 
 Tar has been satisfactorily used when available as a by- 
 product in the manufacture of coke by Otto-Hoffman re- 
 tort ovens. The construction of burners permitted a sim- 
 ultaneous burning of the gas resulting from the coke retorts. 
 
 Producer gas is used extensively and when near a reliable 
 supply of coal, is considered cheaper than the aforemen- 
 tioned fuels. A description of gas producers and liquid 
 fuel burners will be given in subsequent chapters. Gas dis- 
 tilled in a producer does not have as high a heating value 
 as liquid fuels, nor is the efficiency so great, because many 
 of the total heat units are lost in the process of distillation 
 of the coal (soft or anthracite). With liquid fuels or natu- 
 ral gas the total thermal efficiency is available within the 
 working body of the furnace, there being no in- 
 termediate losses before delivery at the point of com- 
 bustion. Natural gas or liquid fuels are easy of control in 
 flame regulation ; furnace construction and repairs are sim- 
 plified and lessened ; regularity of product and longer cam- 
 
FUELS AND ALLOYS 1 1 
 
 paigns are assured. Producer gas is irregular, and owing 
 to heavy deposits of tarry and sooty matter, regular weekly 
 stoppages must be made to clean out mains and flues. 
 Liquid fuels or natural gas eliminate such losses of working 
 time. 
 
 HEATING VALUE OF FUELS 
 B. T. U. 
 
 Natural Gas 300-600 per cubic foot 
 
 Oil 14000-17000 per pound 
 
 Tar 15000 per pound 
 
 Producer Gas 100-150 per cubic foot 
 
 Bituminous Coal 10000-12500 per pound 
 
 One ton of bituminous coal yields in a modern producer, 
 160,000 cubic feet of gas with 65 per cent efficiency in 
 heating value of the coal. 
 
 ALLOYS 
 
 FERRO-MANGANESE. The standard quality contains 80 
 per cent manganese. A representative analysis will be as 
 follows : 
 
 Iron 12.14 per cent 
 
 Manganese 80.00 per cent 
 
 Carbon ; . 5.6 per cent 
 
 Silicon 0.5 -i.oo per cent 
 
 Sulphur 0.010-0.03 per cent 
 
 Phosphorus 0.100-0.75 per cent 
 
 FERRO- SILICON. The standard grade carries 13 per cent 
 silicon and is usually sold on a guarantee of n per cent of 
 that element. The following is a usual analysis: 
 
12 FUELS AND ALLOYS 
 
 Silicon 9 13 per cent 
 
 Carbon i 2 per cent 
 
 Sulphur 0.04-0.08 per cent 
 
 Phosphorus 0.10-0.50 per cent 
 
 In recent years there have been put on the market sev- 
 eral grades of electrolytic silicons that are very satisfactory. 
 The most economical grade is the one carrying 50 per cent 
 silicon, and considered on the basis of the unit cost of sili- 
 con, is cheaper than the commoner alloy. The following is 
 a typical composition of an electric furnace f erro-silicon : 
 
 Silicon 50 52 per cent 
 
 Iron 44 46 per cent 
 
 Carbon 0.15 -0.25 per cent 
 
 Sulphur 0.003-0.010 per cent 
 
 Phosphorus 0.04 -0.06 per cent 
 
 The purpose of the aforementioned alloys will be con- 
 sidered farther on. 
 
 IRON ORE. The purpose of iron ore is two-fold. One is 
 to increase the fluidity of plastic slags, the other as a car- 
 rier of oxygen to assist in the removal of the carbon from 
 the bath of molten metal. The total iron liberated in the in- 
 terchange between its oxygen and the carbon of the bath 
 adds to the yield of metal. The most satisfactory ores for 
 open-hearth practice are the magnetites or hard hematites. 
 The soft ores are apt to dissipate their combined usefulness 
 in the slag instead of oxidizing carbon. No particular 
 limits are placed on their compositions, excepting that they 
 be high in iron and moderately low in phosphorus. A fair 
 analysis is as follows: 
 
 Iron 60 68 per cent 
 
 Silica 15 per cent 
 
 Sulphur 0.05-0.100 per cent 
 
 Phosphorus 0.03-0.500 per cent 
 
MOLDING MATERIALS 13 
 
 MOLDING MATERIALS. Since all steel castings are poured 
 at higher ranges of temperature than gray iron or malleable 
 castings, it is essential that the sands and clays (binders) 
 be as refractory as possible. Pure silica is the most de- 
 sirable the purer the better. The following is an analysis 
 of a typical steel molding sand : 
 
 Silica 98. 5 per cent 
 
 Alumina 1.40 per cent 
 
 Iron Oxide 0.06 per cent 
 
 Lime 0.20 per cent 
 
 Magnesia 0.16 per cent 
 
 Combined Water 0.14 per cent 
 
 Alkalies 0.25 per cent 
 
 It must be of such a nature or structure physically that 
 the heated gases in the mold when displaced by liquid steel 
 will have a free passage outwardly. It is preferable that the 
 grains be sharp and irregular rather than rounded as would 
 be the case with sand subjected at some time to the action 
 of water. The color is often white or slightly tinged with 
 yellow. Its color is not necessarily a guide to its qualities, 
 but it is often an indication. 
 
 FIRE CLAY. Pure Silica sand having no binding proper- 
 ties, varying amounts of clay are mixed with it to give the 
 sand a needed bond and substantiality to the mold prepared 
 for the reception of the hot steel. The clay must also be 
 refractory and possess a maximum degree of plasticity. 
 Low grade sands and clays would fuse at the temperature 
 of liquid steel and cause the castings to be of an irregular 
 rough surface. An attempt to economize in the sand pile 
 is apt to spoil one's reputation for clean-looking castings. 
 
 It is not always reliable to have recourse to chemical tests 
 on refractories or molding materials, since actual practice 
 will affirm the desirable qualities in them. The following is 
 a typical composition of fire clay : 
 
14 MELTING STOCK FOR BASIC PRACTICE 
 
 Silica 60 66 per cent 
 
 Alumina 25 20 per cent 
 
 Iron Oxide nil- 2.00 per cent 
 
 Lime nil- i.oo per cent 
 
 Magnesia nil- i.oo per cent 
 
 Alakalies .- nil- 2.00 per cent 
 
 Combined Water 7.50-10.50 per cent 
 
 The value of a fire clay depends largely upon a low con- 
 tent of alkalies and a freedom from carbonates of lime. 
 Oxide of iron has a strong fluxing effect, but its presence 
 below 3 per cent is harmless. 
 
 CORE COMPOUNDS. Any reliable proprietary article will 
 answer and the list will include molasses water, rosin, flour, 
 linseed oil, etc., all of which are too well known to need 
 any description. 
 
 MATERIALS FOR BASIC PRACTICE 
 
 MELTING STOCK. Basic melting possesses a marred flex- 
 ibility in the selection of stock over acid melting. It is by 
 some considered a sort of metallurgical scavenger. While 
 it is true that there are greater latitudes in quality of pig 
 iron and scrap yet it must not be overlooked that the pro- 
 miscuous dumping of any kind of stock into a basic furnace 
 cannot yield a reliable product. If good castings are the 
 object sought, discretion must be observed in the selection 
 of materials entering into their manufacture. In regard 
 to quantity basic pig iron greatly exceeds acid pig so far as 
 availability is concerned. The ores of the southern and 
 south-western states are plentifully endowed by nature for 
 the yield of unlimited supplies of basic pig. As to scrap, 
 the situation is somewhat of an uncertainty, owing to the 
 inroads made by the larger interests engaged in the produc- 
 tion of ingots in basic bottoms. In consequence prices for 
 scrap have a tendency to gradually rise. The factors con- 
 trolling the choice between basic and acid practice for cast- 
 ings are ones of location and continguity to the sources of 
 
MELTING STOCK FOR BASIC PRACTICE 15 
 
 supply of raw materials. So far as the relative value of the 
 product of either process is concerned, it is true that basic 
 castings are fully as satisfactory, from the view point of 
 quality, as those made by the acid process. It must be re- 
 membered, one is a melting method and the other a refining 
 one. The basic process to get good results needs intelli- 
 gent handling and a higher development of melting skill. 
 The pig iron necessary is known as "standard basic" and 
 the following analysis represents the ulterior limits in com- 
 position : 
 
 Silicon i.oo per cent 
 
 Sulphur 0.05 per cent 
 
 Phosphorus i.oo per cent 
 
 "Off-basic" can carry as high as 1.50 silicon and again 
 as high as 0.07 sulphur. Shipments of these grades on 
 standard contracts can be accepted at a concession in price 
 and it is permissible to use a moderate amount of "off- 
 basic" in charges with no harm to follow. As was men- 
 tioned under "Acid Melting Stock" the sulphur and phos- 
 phorus charged in that process would equal that of the fin- 
 ished product. In basic melting it is possible to eliminate 
 50 to 75 per cent of the sulphur and 95 per cent of the 
 phosphorus, thanks to the character of the lining of the fur- 
 nace and slags formed by the liberal additions of 
 limestone with the charge. For castings it is de- 
 sirable to have on hand several brands of basic pig; 
 some with low phosphorus and some with high manganese. 
 Certain brands can be obtained with phosphorus as low as 
 O;2OO while standard in other particulars and at ruling 
 prices. Brands with high manganese ranging from 1.5 to 
 3.00 per cent command a higher price. It is not good prac- 
 tice to make the entire pig iron charge high phosphorus 
 stock. The reasons for mixing brands in regard to phos- 
 phorus and manganese will be considered under furnace 
 manipulation. 
 
 STEEL SCRAP. The character of this material is not con- 
 
1 6 REFRACTORIES FOR BASIC FURNACES 
 
 sidered chemically because the physical nature of it brings 
 it well within working limits as to composition. It is usu- 
 ally designated "heavy railroad melting scrap," but liber- 
 ties are sometimes taken and unless the consumer exer- 
 cises circumspection almost anything may be found in it 
 from shop sweepings to tomato cans. Heavy scrap is the 
 desideratum, and may consist of, as an illustration, steel 
 rails, knuckles, draw bars, wheel centers, car-springs, fish- 
 plates, defective castings (steel), ingots, etc. Gray iron 
 castings should be religiously excluded when sold as steel 
 scrap. It is allowable to use limited quantities of defective 
 malleable castings although draw bars of such materials are 
 sold as steel. Five per cent of the scrap charge in malleable 
 scrap will not upset the melter's calculations. The scrap 
 charge will be augmented by daily waste from the foundry. 
 
 REFRACTORIES 
 
 The hearth of basic furnaces in American practice is 
 made w r ith magnesite, a substance classified chemically as a 
 base and possessing the quality, in addition to resisting 
 high temperatures, of being but slightly affected by a slag 
 highly charged with lime which would be fatal to a hearth 
 lined with silica sand. To lengthen the life and efficiency 
 of a basic hearth the first consideration is to keep out of the 
 charge as much silicious matter as possible. Magnesite, 
 the carbonate of magnesia, is an importation from Austria, 
 where it is calcined converting it into magnesia, the oxide of 
 the metal magnesium, by the removal of the major portion 
 of carbon dioxide. It is still considered commercially as a 
 magnesite. Its composition ranges as follows: 
 
 RAW 
 
 Magnesium Carbonate 93-!9 P er cent 
 
 Calcium Carbonate 1-43 per cent 
 
 Iron Carbonate 2.61 per cent 
 
 Silica 2.75 per cent 
 
REFRACTORIES FOR BASIC FURNACES 17 
 
 CALCINED 
 
 Magnesia 90 95 per cent 
 
 Lime i 2 per cent 
 
 Iron Oxide 0.5-3.50 per cent 
 
 Silica 0.5-2.75 per cent 
 
 Volatile Matter 0.5-1.00 per cent 
 
 DOLOMITE. This material is extensively deposited in the 
 United States. It is a double carbonate of lime and mag- 
 nesia. It is used either calcined or raw. Principally it is 
 used for patching slag lines, where scorification of the 
 hearth is the heaviest. It is not recommended for points 
 below the slag line. A typical analysis is as follows : 
 
 RAW 
 
 Silica 0.5 2.00 per cen* 
 
 Iron Oxide 0.5 2.00 per cent 
 
 Alumina 0.5 2.00 per cent 
 
 Calcium Carbonate 50 55 per cent 
 
 Magnesium ' 40 44 per cent 
 
 CALCINED 
 
 Silica 0.5 2.00 per cent 
 
 Iron Oxide 0.5 2.00 per cent 
 
 Alumina 0.5 2.00 per cent 
 
 Lime 50 55 per cent 
 
 Magnesia 37 45 per cent 
 
 CHROME ORE. A substance highly refractory to heat 
 and neutral to the action of acid and basic slags. Unfortu- 
 nately it has no bond and for that reason its uses are some- 
 what limited. It is used for patching parts of the hearth 
 where cutting above the slag line is severe upon the brick 
 work, usually in gas ports and door jambs. (In European 
 practice it is stated that entire hearths are lined with lump 
 
1 8 REFRACTORIES FOR BASIC FURNACES 
 
 chrome ore.) Aside from patching it is used as a neutral 
 lining between magnesite and silica bricks. Its composi- 
 tion is as follows: 
 
 Chromic Oxide 40-60 per cent 
 
 Iron 15-18 per cent 
 
 Alumina 5-30 per cent 
 
 Silica 1-5 per cent 
 
 FLUXES 
 
 The most efficient in basic melting- is ordinary limestone. 
 Its function is to form a slag that will readily absorb the 
 sulphur and phosphorus of the charge and act as a vehicle 
 for the oxidizable silicon, iron and manganese. The purer 
 the grade the better, that is, a richness in carbonate of lime 
 and a freedom from silica. 
 
 A fair analysis is as follows: 
 
 Silica 0.25- i.oo per cent 
 
 Oxide of Iron and Alumina 0.50- 2.00 per cent 
 
 Carbonate of Lime 95. -99. per cent 
 
 Carbonate of Magnesia 0.5 - i.oo per cent 
 
 FLUORSPAR. The function of this material is to thin a 
 limey slag when in the judgment of the melter it seems 
 thick or sluggish. A moderate addition of fluorspar will 
 liven it and its action may be likened to certain fluxes used 
 in brazing metals the property of dissolving at higher 
 temperatures metallic oxides. It is plentifully deposited in 
 the United States. A good grade will analyze as follows: 
 
 Calcium Fluoride 90-98 per cent 
 
 Oxide of Iron and Alumina 0.5-1.00 per cent 
 
 Silica nil-i.oo per cent 
 
 IRON ORE. See "Acid Melting Materials." 
 
FLUXES FOR BASIC PRACTICE 
 
 ALLOYS. See "Acid Melting Materials." 
 
 FUELS. See "Acid Melting Materials." 
 
 As a guide to purchasing stock the following tabulations 
 will be of assistance in furnishing approximate amounts 
 for regular consumption. The figures are based on the 
 different kinds of stock necessary to produce I net ton of 
 castings : 
 
 Acid. Basic. 
 
 Pig Iron 620 pounds 1227 pounds 
 
 Steel Scrap 1880 pounds 1227 pounds 
 
 Ferro-Silicon 54 pounds 57 pounds 
 
 Ferro-Manganese 28 pounds 35 pounds 
 
 Iron Ore 26 pounds 30 pounds 
 
 Aluminum 3-10 pounds 3-10 pounds 
 
 Limestone 300 pounds 
 
 Magnesite 34 pounds 
 
 Silica Sand 1800 pounds 1600 pounds 
 
 Fire Clay 300 oounds 350 pounds 
 
 Gas Coal 950 pounds 1250 pounds 
 
 Fuel Oil 55 gallons 80 pounds 
 
 Boiler Coal (power) 900 pounds 900 pounds 
 
CHAPTER II 
 
 FURNACE CONSTRUCTION AREAS AND VOLUME DRAFT 
 REGENERATION 
 
 In American practice, when larger tonnages in output 
 are sought, the general capacity of an open-hearth furnace 
 working on castings is about 20 tons per heat. This ca- 
 pacity approaches the maximum that can be economically 
 handled in jobbing shops and will amply represent a gen- 
 eral rule. Larger tonnages may be occasionally needed if 
 the class of product is in the shape of heavy work requiring 
 but a few molds to receive a heat of steel and consuming 
 but a moderate interval of time to pour them, but in cases 
 of a heat of steel to be put into a large number of molds 
 the pouring time may be so extended that the metal will 
 lose its temperature. Therefore, charges exceeding 20 tons 
 for miscellaneous castings are apt to result in losses due to 
 cold steel. The illustrations, Figs, i, 2 and 3, given here- 
 with, show the usual lines of a modern stationary furnace 
 of 20 tons capacity. The lines are the ultimate of experi- 
 ence in various plants and embody the best that is obtain- 
 able for that type of furnace at the present time. As they 
 re only representative they may be subject to some minor 
 changes which may be dictated by necessities arising from 
 local conditions in erecting. The principles of construction 
 are the same in both acid and basic furnaces, the differences 
 occurring in the character of the materials forming the 
 
OF THt 
 
 UNIVERSITY 
 
 OF 
 ^ 
 
 OPEN-HEARTH FURNACE CONSTRUCTION 
 
 21 
 
 hearth linings. Generally the furnaces are of the stationary 
 type. In some plants will be found furnaces of the rolling 
 or tilting kinds, each having some good points in its favor. 
 From the viewpoint of cheapness of construction the sta- 
 tionary furnace holds the ground. Movable hearth types 
 call for costly mechanical installation not required in sta- 
 tionary units. A potent argument in favor of the movable 
 
 FIG. 1. PLAN VIEW, STATIONARY TYPE OPEN-HEARTH STEEL 
 MELTING FURNACE. 
 
 (rolling or tilting) furnace is the ability offered to com- 
 pletely drain the hearth at the end of a melt thus emptying 
 any pools that may form in the bottom due to excessive 
 scorification and the ease with which they can be readily 
 repaired with proper refractories. In stationary furnaces 
 much time is lost, with much discomfort to the workmen in 
 emptying pools or "puddles" by means of rabbles or scrap- 
 ers. They cannot be thoroughly drained by such means 
 and the subsequent patching may or may not be satisfac- 
 
22 
 
 OPEN-HEARTH FURNACE CONSTRUCTION 
 
 torily accomplished. The patching may become loosened 
 in a succeeding melt. Such difficulties are more liable to 
 happen in basic bottoms than on acid. Another argument in 
 favor of the movable furnaces is that the tapping hole 
 troubles are eliminated. In stationary furnaces difficulties 
 and annoying delays are encountered through "hard-taps" 
 as a result of the materials used to temporarily close the 
 tapping hole becoming fused or hardened and offering 
 great resistance to tools necessary to open it at the proper 
 time. In well ordered plants such difficulties are a rare oc- 
 currence, but still the risk exists. With a movable furnace 
 the tapping hole is never tamped or closed so that there is 
 
 FIG. 2. SECTIONAL, STATIONARY TYPE OPEN-HEARTH STEEL 
 MELTING FURNACE 
 
 always the assurance that the metal can be drawn off when 
 desired. Against these favorable considerations is the com- 
 paratively heavier .first cost of the movable furnaces over 
 the stationary, so that the question as to which type is to be 
 approved will remain a debatable one. However, the rela- 
 tive volumes in regard to hearth area, regenerator cham- 
 bers, etc., on the basis of the capacity of output per heat 
 will be the same in any style of open -hearth furnace. The 
 
OPEN-HEARTH FURNACE CONSTRUCTION 
 
 next in order will be some general rules as to points of con- 
 struction and volumes for a 2O-ton unit. 
 
 CONSTRUCTION. The metal work such as buckstays, tie- 
 rods, hearth pan, doors, etc., should be of rigid construc- 
 tion to withstand the heavy duty due to brick work ex- 
 pansion when the furnace is at full working heat. All 
 walls should be bound at the ends. Rolled shapes should 
 be used whenever possible. Skew -backs of all arches 
 
 FIG. 3. CROSS SECTION, STATIONARY TYPE OPEN-HEARTH STEEL 
 MELTING FURNACE. 
 
 should be braced by the binding, particularly those of the 
 roof. In the regenerator chambers the ends of all outside 
 or partition walls should be bound, as should the ports, be- 
 cause the brick expansion may cause leaks and permit igni- 
 tion of the gas before it reaches the working body of the 
 furnace. The conditions of the subsoil at the selected site 
 should be well studied before putting in the foundation. 
 Foundations should be of rigid and first-class masonry to 
 guard against irregular settling. They should be of hard 
 red brick laid in cement or concrete. No part of the fur- 
 
24 OPEN-HEARTH FURNACE CONSTRUCTION 
 
 nace structure should extend below the lowest point at 
 which water may be found. Water is an enemy to smooth 
 furnace operation, if it finds its way into flues or chambers. 
 All underground flues not protected by clay should have 
 an outside course of red brick. 
 
 The reversing valves should be of such construction that 
 leakages and loss of gas will not occur when operated. 
 There are two distinct types, one known as the "butterfly- 
 valves" and another as the "turtle-back." The latter is 
 water sealed, which is an advantage. The stack should 
 be of such construction as will induce a good draught, de- 
 pending upon the damper for regulation. 
 
 Whenever possible the doors, door frames and furnace 
 fronts should be water cooled, features which add to the 
 operator's comfort in watching his furnace. The cost of 
 installation and maintenance may be heavy but will result 
 in a longer life for the parts and a consequent lessening of 
 their repairs offsetting the first cost. This arrangement is 
 only possible on stationary furnaces. On movable ones the 
 piping connections, etc., would be too complicated. 
 
 All flues, excepting those leading from the uptakes to 
 the furnace body, should be roomy to prevent choking and 
 cutting by the deposit and heat of waste gases. Roominess 
 is an essential in flues or conduits connecting the gas pro- 
 ducers with the regulating and reversing valves and it is 
 good practice to have such conductors above ground the en- 
 tire distance to allow ready access for the purpose of 
 cleaning out unavoidable accumulations of soot and tar. 
 Gas flues or uptakes leading to the furnace body should be 
 built with fake arches in their back walls, so that they can 
 be readily repaired when badly cut without disturbing the 
 rest of the furnace. They should also slope towards the 
 hearth so that the incoming gas will be directed downward- 
 ly and impinge upon the stock or charge. The air port 
 will also have the same direction, and with the gas and air 
 inlets and outlets working properly the sheet of flame will 
 
OPEN-HEARTH FURNACE CONSTRUCTION 
 
 be kept away from the roof which will be guarded against 
 burning or cutting. 
 
 AREAS AND VOLUMES. The hearth length of an open- 
 hearth furnace should be as great as possible in order that 
 the greatest possible benefit be derived from the calorific 
 value of the fuel. An undue shortening would be wasteful 
 because the heat of combustion would be spent in the out- 
 going gas and at the sacrifice of fuel consumption and ex- 
 cessive cutting of outlets leading to the chambers, meaning 
 an increased cost in furnace repairs. 
 
 Practical experience has taught that a 2O-ton furnace 
 can be safely operated with a hearth length of 20-25 feet 
 and a width of 9-11 feet. The total area of hearth surface 
 will work out to very close to nine square feet per ton of 
 capacity. The width is limited to a maximum of 15 feet 
 because furnace operatives cannot throw a shovelful of re- 
 fractory material much over that distance to reach the 
 back-wall. Generally the length will be 2-2^2 times the 
 width. 
 
 The depth of the fully lined hearth will depend upon the 
 dimensions of length and width. A shallow bath will give 
 rapid working but at the sacrifice of much burnt metal or 
 a yield which is only a small percentage of the metal 
 charged. A deep bath will retard the melting and present 
 difficulties in maintaining desirable thermal conditions. 
 The medium will be learned by the individuality of the fur- 
 nace and conditions of practice, but to put the problem in 
 figures the ranges for the choice will be between 15 inches 
 to 20 inches of depth. 
 
 An important consideration in furnace practice is re- 
 generator chamber volumes. The fuel efficiency will be 
 controlled largely by the length of the furnace as mentioned 
 and also upon the proper construction of the chambers. 
 The purposes of them will be subsequently considered. At 
 present attention will be given to their volumes. In that 
 particular there will be conditions to take into account as 
 to how much room can be allowed by the space in the build- 
 
26 OPEN-HEARTH FURNACE CONSTRUCTION 
 
 ing where the furnace or furnaces may be located both 
 above and below the charging floor and the depth to which 
 the foundations and flues can be safely carried. The effi- 
 ciency of a regenerator chamber depends upon the number 
 of "checkers" it can carry and the direction of incoming 
 and outgoing gaseous bodies. Direction is meant by the 
 flow of gases whether they be nearly horizontal or nearly 
 vertical in travel. 
 
 In American practice the longest dimension of the cham- 
 bers is horizontal, while in European practice some are 
 built with greater depth than length. It would seem, in 
 view of the natural tendency of heated gases to rise, that 
 the latter plan is the better, but as stated how they shall 
 be built depends upon variable conditions; however, with 
 a given chamber volume the efficiency varies with the 
 depth. There is quite a range of figures as to volumes of 
 regenerators per ton of capacity with different plants vary- 
 ing from 65 cubic feet to 140. A good figure to work by is 
 90 cubic feet per ton, allowing 1-3 for gas and 2-3 for air 
 chambers. 
 
 If the fuel should be oil or liquid or natural gas the vol- 
 ume can be decreased materially because such varieties of 
 fuel are led directly in the body of the furnace instead of 
 passing through the regenerators, thus offering a possibil- 
 ity of dispensing with the space occupied by regenerators, 
 commonly used as ducts for producer gas, but it is better 
 that a furnace be built with an eye to suitability for pro- 
 ducer fuel because the supply of liquid fuel or natural gas 
 is subject to possibilities of irregular deliveries and a fur- 
 nace built only for the latter fuels would cause some annoy- 
 ances were they to be short. By the same token, producers 
 should always be installed as a safeguard no matter what 
 fuel may be regularly used. Thus there would be but lit- 
 tle delay to put the producers for gas into service were 
 liquid fuel or natural gas to fail. 
 
 The uses of regenerator chambers will next be consid- 
 ered. The purpose is to store in them heat carried over by 
 
OPEN-HEARTH FURNACE CONSTRUCTION 2? 
 
 waste gases produced by the fuel combustion in the furnace 
 body, the heat being absorbed by a large number of No. 
 i fire brick, piled in such a manner that the gases in their 
 travel from the body of the furnace to the stack will have 
 to pass through innumerable ducts or passages. Bricks 
 piled in such a manner are called "checkers." The plan is 
 to pile the bricks so that they will form rectangular pass- 
 ages of about 3 to 3^> inches in width. The passages will 
 run horizontally and also vertically. Sometimes they will 
 be in a direct line in both directions, the length and width 
 of the chambers or the bricks may be piled in such a way 
 that the passages ,are zig-zag or "staggered." Generally 
 they are staggered in a vertical direction with straight 
 passages horizontally. 
 
 DRAFT 
 
 The question of draft has to be considered and with as 
 many bricks as it may be possible to checker and with the 
 greatest possible depth of chamber the free working of 
 the furnace will be augmented by straight passages in both 
 directions. Indications as to heat absorption by checkers 
 can be gauged by the temperature of the waste gases enter- 
 ing the stack with the furnace at full working heat. 
 
 Pyrometrical observations by the writer show the nor- 
 mal working conditions of the stack gases to be an average 
 of 500 degrees Cent, with gases entering the outgoing 
 down-takes at 1,400 degrees Cent, and with air at atmos- 
 pheric temperatures entering chambers and passing 
 through them in the up-takes at 1,000 degrees Cent, will 
 suggest the heat absorption and radiation of the checkers. 
 
 The temperature of combustion is not sufficiently high to 
 maintain a continued liquation of a bath of molten metal as 
 its carbon decreases, because the air necessary to support 
 combustion, even with a forced draft, carries away or 
 absorbs the calorific energy of the flame playing upon the 
 
28 OPEN-HEARTH FURNACE CONSTRUCTION 
 
 bath of metal. In other words cold air lessens the full 
 heating value of combustion that should otherwise be spent 
 in work. If, then, the temperature of the necessary air for 
 complete combustion be raised, to that extent will the flame 
 efficiency be increased. On that rests the principle of re- 
 generation. 
 
 Let the course of the air be followed in the chambers of 
 an open-hearth furnace passing from left to right. The 
 reversing valves are in position to direct the inflow of 
 gas and air in their respective chambers on the left side of 
 the furnace. Passing through the checkers and innumer- 
 able ducts, they enter the up-takes. The gas upon reaching 
 the furnace body immediately ignites and draws upon the 
 accompanying air for complete combustion, the respective 
 volumes of each being under control by the operator. The 
 flame energy being dissipated in work, the waste gases are 
 now drawn by the draft and pushed along by a rear expan- 
 sion towards the stack but, before reaching it, nearly all 
 their heat units are absorbed by the checkers in the right- 
 hand set of chambers. After an interval of 15 to 20 minutes 
 the reversing valves are thrown and the gases are reversed 
 in direction. The air and gas now passing into the already 
 heated right-hand chambers carry back by radiation to the 
 furnace body some of the waste heat previously deposited 
 there to add to the heat units produced by combustion. 
 The efficiency of the flame will be greater from the right- 
 hand chambers work, assuming both sets to be of an equal 
 temperature at the beginning of the operation, and after 
 the second reversal the left-hand chamber will bring a still 
 greater increment of heat value than its neighbor. That is 
 to say, the heat units radiated to the incoming air and in- 
 creasing the flame value necessarily permits more heat for 
 the outgoing checkers to absorb. Thus it will be seen that 
 with the increase in the number of reversals there will also 
 be a gain in heat for work. 
 
 It would be possible to melt the best refractories by 
 reaching high ranges of temperature by the principle of re- 
 
OPEN-HEARTH FURNACE CONSTRUCTION 29 
 
 generation but by careful watching on the part of the ope- 
 rator, flame and air volumes are properly regulated to pre- 
 vent burning of the furnace. At the same time, enough 
 heat must be maintained during the progress of a melt to 
 preserve the fluidity and proper temperature of a bath of 
 steel at proper intervals. Without the system of regen- 
 eration it would not be possible to successfully handle 200 
 tons or less of liquid steel at a time in a single operation of 
 an open-hearth furnace. 
 
 The speed at which the gases travel through the furnace 
 when working is due to both draft and expansion. As 
 soon as the cold incoming air comes in contact with the 
 heated checkers it immediately expands and produces a 
 slight pressure which forces the body of air before it up- 
 wardly, and entering the furnace body it not only assists 
 the combustion of the fuel^ but washes and protects, so to 
 speak, the roof of the furnace with a film of air and at the 
 same time depresses the flame upon the bath of metal. 
 With free passages in the down-takes and checkers, the 
 stack will readily take care of the waste gases. Obstruc- 
 tions in either would make a slow working furnace and 
 disagreeable waste of flame out of the furnace doors. 
 
 Another important feature about regenerative chambers 
 is that they should never be under the furnace body or have 
 their up-takes directly below the ports. In the first in- 
 stance there would be danger of irregular settling causing 
 cracks in the partition walls between the air and gas cham- 
 bers, which would allow gas leakages and ignition of same 
 before entering the furnace. In the second place there is 
 always more or less dust and slag carried along by the 
 draft which would be deposited in the checkers with 
 chambers located as just mentioned, thus choking them and, 
 of course, crippling their life. Good practice requires that 
 the chambers be placed at the furnace ends and extend at 
 right angles to them under the charging floor. There 
 should also be spacious receptacles at the lowest point of 
 each down-take to retain accumulations of dust and slag 
 
30 OPEN-HEARTH FURNACE CONSTRUCTION 
 
 before they could reach the chambers. Such are known 
 as slag pockets and with proper construction they can read- 
 ily be cleaned out without disturbing the checkers. 
 
 ACID FURNACE BRICK WORK 
 
 The furnace body wherever subjected to uniform, high 
 temperature is lined by first grade silica bricks. Piers and 
 outside walls of the structure below the charging floor can 
 be red brick; linings of flues and chambers including 
 checkers are No. I fire brick. Silica brick will not answer 
 for checkers because they will crumble under the varying 
 ranges of heat. 
 
 The hearth pan is lined with fire brick to the depth of 
 nine inches or more, but above metal line of a fully lined 
 hearth, the sides, walls and roof are silica. With the brick 
 work complete the furnace is first dried moderately and 
 carefully with a wood or soft coal fire kept going for a few 
 days. 
 
 The gas or oil can then be turned on slightly at first and 
 then gradually raised to nearly full working temperature. 
 Layers of silica sand of the quality described in chapter I 
 are then spread over the bottom. They are put in in suc- 
 cession and between each interval the flame is allowed to 
 set or sinter the sand until hard. This operation is repeated 
 until the hearth lining will have reached a depth of 18 to 
 20 inches including the fire bricks. A hearth so lined with 
 a suitable refractory ought to last almost indefinitely under 
 favorable conditions. 
 
 There will be occasional patching of the slag line and 
 bottom with sand, at the end of each heat, the extent of 
 which will be controlled by the conditions and character 
 of stock used in melting. A hearth properly lined must be 
 set hard enough to resist attrition by the charging of melt- 
 ing stock. Under skillful handling an acid furnace ought 
 
OPEN-HEARTH FURNACE CONSTRUCTION 3! 
 
 to turn out normally 950 heats or more in a campaign at 
 the rate of at least 3 heats per working day. 
 
 BASIC FURNACE BRICK WORK 
 
 The designation basic is rather a misnomer. The nature 
 of the basic process requires a lining of such materials that 
 will resist the fluxing action of limey slags and vapors nec- 
 essary to purify and refine phosphoric melting stock. Un- 
 fortunately no materials are commercially available to com- 
 pletely line a furnace body, so recourse can only be had to 
 a hearth lined with basic materials, with roofs, sides and 
 walls of furnace body above the slag line consisting of 
 silica or acid bricks, the reasons being that bricks of basic 
 or neutral material, such as magnesite or chrome, while be- 
 ing refractory, do not give as good results as silica bricks, 
 owing to a liability to crumble if placed in the walls or roof. 
 Therefore a basic furnace is part acid and part basic lining. 
 
 With the exception of the furnace body in regard to brick 
 work, the construction is the same as an acid furnace. The 
 hearth pan is first lined with fire brick followed with mag- 
 nesite bricks. Usually the bottom is lined with ground 
 magnesite. It may be mixed with about 5 per cent of 
 anhydrous tar and rammed in to form the hearth and then 
 slowly and carefully brought to full temperature; or the 
 magnesite may be put in loosely in layers and gradually sin- 
 tered. A small percentage of ground basic slag is some- 
 times mixed with it to insure a partial fusing or sinter. 
 
 A magnesite hearth while costly gives the best results in 
 service and will sinter hard enough to withstand rough 
 usage by charging of the stock. Where the magnesite 
 bricks meet the silica bricks of the walls, a parting of 
 chrome ore is placed as a neutral separation of the two to 
 prevent a fluxing liable to ensue between them at full work- 
 ing temperature. 
 
 There will always be some scorification of the hearth at 
 
32 OPEN-HEARTH FURNACE CONSTRUCTION 
 
 the slag line and an occasional formation of holes in the 
 bottom, due to the action of silicious matter carried in with 
 the stock. The repairs to the hearth are made with raw 
 dolomite on the slag line, and with ground magnesite on 
 the bottom. Dolomite being so much cheaper it is fully as 
 effective as magnesite at the slag line. In the raw state it 
 is not recommended for bottom repairs, because at a high 
 temperature it is calcined, contracting greatly in bulk and 
 for that reason holes in the bottom cannot be satisfactorily 
 filled with it. Under the heat of fused stock it would 
 loosen, float upwards, and leave the condition as bad as be- 
 fore the patch. 
 
 Undue hearth scorification can be controlled by proper 
 care in the character of stock. Hence the consumption of 
 refractories for hearth patching can be kept at the mini- 
 mum figure. With proper case a basic hearth of magnesite 
 should last indefinitely, and the life of the brick work of 
 the roof and walls ought to yield 400 or more heats at 3 
 heats per day. 
 
CHAPTER III 
 
 FUELS AND ACCESSORIES DISCUSSION OF THE USES OF 
 COAL AND OIL 
 
 As the choice of fuel may rest between producer gas or 
 oil, a description of the operation of either will be briefly 
 considered. Referring to the illustration of a gas producer, 
 Fig. 4, a general idea will be formed of its construction. 
 The one shown is of the simpler kind and entirely hand fed 
 and poked. In some of the large rolling mills coal is fed 
 in continuously by a mechanical device, and the bed of the 
 fuel is poked by a mechanical contrivance. The principle 
 of operation is the same in either case so far as making 
 gas goes. 
 
 For a continuous supply of gas, air and steam are forced 
 through an incandescent bed of bituminous coal on top of 
 which is fed at regular intervals fresh coal. Frequently the 
 mass is poked with long bars to break up the decomposing 
 coal and to prevent holes or passages being formed which 
 might allow air to pass through them and dilute the gas. 
 In the vicinity of the grate, which is water sealed, the fuel 
 is completely burned, while near the top the fuel gives off 
 its volatile matter, forming copious volumes of smoke with 
 some tarry matter. As the fuel descends towards the grate 
 it is gradually burned to ash. 
 
 For proper working conditions the bed of coal should 
 be kept at a constant height, and vigorous poking should be 
 frequently and persistently followed. 
 
34 
 
 FUELS AND ACCESSORIES 
 
 The object of water-sealing the grate is to permit the 
 amount of air necessary to gasify the coal, to be under con- 
 trol at all times. 
 
 The use of steam lessens the temperature of combustion 
 
 lop Vleiv of Water Flan, 
 
 Section E- 
 
 FIG. 4. GAS PRODUCER. 
 
 at the grate and so lengthens the life of the grate bars. At 
 the same time the steam chemically combines with the fuel 
 to form water gas as will be shown. It also prevents the 
 
FUELS AND ACCESSORIES 35 
 
 formation of clinkers, making it easier to keep the fires 
 clean. 
 
 To make good gas the fuel must be hot, and close attention 
 must be given to the admixture of air and steam forced into 
 the producer. Too much steam tends to cool the fires and 
 pass into the flues, undecomposed, causing an extravagant 
 loss of fuel efficiency. A deep, hot bed of coal will yield 
 the richest gas. 
 
 It will not be amiss to consider some of the chemical 
 changes that take place in a producer. Roughly the bed 
 of fuel in it can be divided into two zones. The lower 
 one, nearest the grate, can be called the CO2 zone and the 
 upper one the CO zone. The air coming into union with the 
 fuel near the grate forms 
 
 C + 2 O = CO. 
 
 CO2 is of course, non-combustible, but as it passes up- 
 wards it combines with the glowing carbon of the CO zone 
 and absorbing some becomes 
 
 CO 2 + C = 2CO. 
 
 the latter constituent forming the larger volume and chief 
 calorific agent of producer gas. By the action of steam we 
 have 
 
 C + HsO = CO + 2H (water gas). 
 
 The calorific value of this last product is greater in equal 
 volume than the CO formed in second equation but at the 
 expense of the heat in the bed of fuel. The following is 
 an analysis of producer gas by volume : 
 
 C O 27.00 per cent 
 
 C O 2 5.00 per cent 
 
 H 10.00 per cent 
 
 C H 4 -f C 2 H 4 1.50 per cent 
 
 O + N by difference 56.50 per cent 
 
 100.00 per cent 
 The amount of oxygen in the gas will be about I per cent, 
 
FUELS AND ACCESSORIES 
 
 and represents the air that passes through the producer 
 uncombined. The index to the proper working of the pro- 
 ducer is the amount of COs present. Under the most ad- 
 vantageous conditions it will rarely fall below 2.5 per cent 
 and with bad conditions it will exceed the average of 5 
 per cent. The causes of an excess are due to insufficient 
 poking, a shallow fire and faulty brick work allowing air 
 leakage to ignite the gas before it can be delivered to the 
 furnace. 
 
 9 Threads to \'J ,, Thrcafl!i to r 
 
 Rubber Hose 
 Connects Here 
 
 FIG. 5. OIL BURNER FOR OPEN-HEARTH FURNACE 
 
 At the best, producer gas is unsatisfactory, and the steel 
 melter is always at the mercy of the vigilance or lack of 
 it of the gas man. 
 
 One ton of bituminous coal yields 160,000 to 170,000 
 cubic feet of gas with a calorific value, at the producer, of 
 about 137 B. T. U. per cubic foot. The gas in traveling 
 to the furnace loses heat units at a variable rate. The 
 actual amount of gas delivered to the furnace is hard to 
 determine, owing to leakage and the consumption in drying 
 ladles. 
 
FUELS AND ACCESSORIES 
 
 37 
 
 Liquid fuels, such as crude petroleum or residuum, pos- 
 sess a high calorific value, usually expressed at 14,000 to 
 17,000 B. T. U. per pound of oil. Because the oil being 
 delivered directly to the furnace (see oil burning device 
 and furnace construction for same, Figs. 5 and 6), and 
 igniting, when atomized by steam or compressed air, 
 yields its entire thermal efficiency to work with no interme- 
 diate losses as is the case with gas, the value of oil over 
 the latter is marked. 
 
 FIG. 6. FURNACE ARRANGED FOR BURNING OIL 
 
 It is difficult to make an actual comparison between oil 
 and coal for steel melting on the basis of the cost of a ton 
 of metal produced. The figures may be in favor of coal 
 in certain localities, and in favor of oil in others. Yet the 
 advantages of oil over coal in working results are so pro- 
 
38 FUELS AND ACCESSORIES 
 
 nounced that discrepancies in cost against oil are offset by 
 its usefulness. 
 
 Ignoring the relative costs, the principal points in favor 
 of oil against gas will be considered. 
 
 First, the higher thermal value: A cubic foot of gas 
 will yield 137 B. T. U. Taking 16,000 B. T. U. as an 
 average of one pound of oil and allowing a cubic foot of 
 oil at 57.11 pounds then 57.11 X 16,000 = 913,760 B. 
 T. U. a substantial gain in favor of oil against an equal 
 volume of gas. 
 
 Second, the simplicity of installation. One furnace will 
 require a storage tank with a capacity of about 17,000 
 gallons. From this the oil is pumped to the burner which 
 essentially is the producer in the sense that the arrange- 
 ment .of the burner permits a necessary atomization of the 
 oil by steam or compressed air before ignition. It is su- 
 perfluous to make a further comparison on this point, in 
 view of the crudity of the gas producer. 
 
 Third. The use of oil lessens furnace repair costs and 
 allows longer campaigns before shutting down for general 
 repairs. This point alone, is perhaps the strongest one in 
 favor of oil. Conditions of brick work in regenerator 
 chambers do not require the same attention with liquid 
 fuel as they would with gas. That is to say, should there 
 be leakages in the partition wall between air and gas 
 chambers, they can be ignored, using oil; but with gas 
 they would necessitate a shut-down of the furnace to re- 
 pair them. The same applies to ports and down-takes. 
 
 Fourth. The character of the oil not being subject to 
 the same latitudes of irregularity as the composition of the 
 gas, there results a decided gain in the certainty of the 
 furnace's work. The temperature of the bath is under 
 control, and regularity of output can be expected a feat- 
 ure not so dependable with gas. 
 
 Fifth. The labor cost is greatly lowered. One man per 
 working day attends the pumps. In the gas house there 
 will be a foreman and several laborers to feed and poke the 
 
FUELS AND ACCESSORIES 39 
 
 fires, and to wheel away the ashes. The labor in unloading 
 and stocking coal is also eliminated. With these features 
 can be mentioned the removal of the attendant dirt and 
 smoke with gas producers; the loss of time in cleaning 
 gas mains and other conditions that would occupy too much 
 space to mention. 
 
 Oil fuel will also be useful for drying ladles, firing an- 
 nealers, core ovens, etc. 
 
CHAPTER IV 
 
 MANIPULATION OF HEATS IN ACID PRACTICE COMPOSI- 
 TION OF CHARGES DETAILS OF RECARBONIZING 
 
 Given an acid lined hearth and stock for melting pur- 
 poses, the next step will be to consider some of the changes 
 that take place in the conversion of the materials charged 
 into steel. As has been mentioned the only elements, that 
 are confined within stated limits, are the sulphur and 
 phosphorus. Considerable latitude remains in making up 
 the charge in regard to the available silicon, carbon and 
 manganese carried in by the stock. Assuming the stock 
 to be made up of pig iron, billets, blooms, plate-clippings, 
 axle-butts, defective steel-castings, shop scrap or wasters 
 in varying proportions, a charge of 24,000 pounds will be 
 studied because the diagram herewith shown (Fig. 7) was 
 plotted on a heat of that size. The proportions and 
 changes would be relatively the same in a 2O-ton heat. 
 
 The charge will be as follows: 
 
 Acid pig iron 3,600 
 
 Mixed scrap = 20,400 
 
 Lbs. 24,000 
 
 15 per cent 
 
 85 per cent 
 
 100 per cent 
 
 AVERAGE COMPOSITION 
 
 C 0.90 per cent 
 
 Mn 0.47 per cent 
 
 Si 0.40 per cent 
 
 S 0.024 per cent 
 
 P 0.028 per cent 
 
MANIPULATION OF HEATS IN ACID PRACTICE 4! 
 
 The order of charging will be as follows : 
 First. Two-thirds of the pig iron. 
 Second. Lightest sections of scrap. 
 Third. Heaviest sections of scrap. 
 Fourth. One-third or remainder of pig iron. 
 
 The object in charging the pig iron in two portions with 
 the larger amount on the bottom is to protect it from scori- 
 fication caused by the oxide of iron always formed in the 
 melting of the scrap which oxidizes at a faster rate than the 
 pig iron. The portion of pig iron on top is the first to melt 
 and in dripping over the scrap lowers the melting point of 
 the latter and in a measure protects it from undue burn- 
 ing or oxidation until the whole mass sinks below the slag 
 formed during the exposure of the stock to the flame action. 
 
 The length of time occupied in charging varies as to the 
 size of the pieces of scrap charged and the room it offers to 
 follow with the rest of the stock. 
 
 Occasionally there may be some little time elapse before 
 the bulky stock may have partly melted and subsided before 
 the charge can be completed. Usually the length of time 
 consumed in charging is about one to one and one-half 
 hours. 
 
 The charging of stock being completed the history of the 
 heat is divided into two stages : First, the melting ; second, 
 the complete fusion and conversion of the materials. Dur- 
 ing the first period little or no change takes place in the 
 composition, the main action being the transmission of the 
 solid stock to the liquid form and with it the formation of 
 the slag which is to play an important part in the subsequent 
 conversion of the stock to steel. 
 
 The length of time required to liquify the stock is normal- 
 ly from 2 to 2^2 hours, and during that time the tempera- 
 ture of the furnace is gradually increasing owing to the 
 method of regeneration already explained. 
 
 It is certain that some changes occur as soon as the stock 
 begins to melt and the slag begins to form. Perhaps the 
 most pronounced change takes place by flame action on the 
 
MANIPULATION OF HEATS IN ACID PRACTICE 
 
 Diagram Showing Variations in Composition 
 of a Formal A.ci<l 
 Open Hearth Heat 
 
 FIG. 7 SHOWS VARIATION IN COMPOSITION OF A NORMAL ACID 
 OPEN-HEARTH HEAT 
 
MANIPULATION OF HEATS IN ACID PRACTICE 43 
 
 exposed stock, it being strongly oxidizing ; but not until the 
 charge becomes entirely fused or liquid does the active part 
 of conversion begin. 
 
 With complete liquation of the charge the flame simply 
 becomes a vehicle of heat and, under the practiced eye of 
 the melter, the thermal conditions are so maintained that 
 the temperature of the bath is gradually increased as the 
 conversion progresses. It is necessary that the tempera- 
 ture of bath be gradually increased because of the influ- 
 ence of carbon. In general terms the fusing point of iron 
 or steel depends upon the amount of that element, the high- 
 er percentage fusing at a lower point thermometrically 
 than the lower amounts. 
 
 Glancing at the diagram the carbon will be seen as gradu- 
 ally increasing as the conversion progresses. Without a 
 corresponding increase in temperature, the bath of metal 
 would become pasty ; and with the bath in that condition 
 there would be losses. 
 
 In order to induce liquation at the first stage of conver- 
 sion it is necessary to introduce carbon in addition to that 
 being furnished by the pig iron of the charge. Fluidity 
 and proper thermal conditions succeeding are then easily 
 attained. 
 
 Assuming the stock to be melted and having passed from 
 direct flame action below a covering of slag the functions 
 of that will next be considered. The existence of the slag 
 is derived from sand carried in mechanically by the stock, 
 the oxidation of the silicon contained in the pig-iron and 
 scrap to silica, the oxidation of the manganese brought in 
 by them to manganous oxide, some scprification of the 
 hearth and the oxidation of iron to ferrous oxide. 
 
 Normally the slag is automatically formed, in regard to 
 composition, throughout the progress of the heat. 
 
 Should there be an excess of FeO due to low silicon 
 pig-iron, there would be an excessive cutting of the hearth 
 and it might be necessary to add sand to furnish the needed 
 SiO to form the adjustment of proper slag composition. 
 
/| /j MANIPULATION OF HEATS IN ACID PRACTICE 
 
 Approximately a normal slag in an acid heat consists of 
 nearly equal parts SiOa acid and FeO + MnO (bases) and 
 this composition will exist throughout the heat with a grad- 
 ual increase in volume resulting from the continued oxi- 
 dation of silicon and manganese and by the addition of 
 iron-ore. 
 
 The oxidation of the carbon causes a lively bubbling in 
 the bath by the liberation of carbon monoxide (CO) as a 
 result of the exchange between the FeO of the slag and the 
 oxygen furnished by iron-ore additions, which may be ex- 
 pressed as follows: 
 
 C + O^CO or C + FeO = Fe + CO. 
 
 The metallic iron reduced from the FeO of the slag and 
 that brought in by the iron ore (FezOs) is immediately ab- 
 sorbed by the bath. Were there no additions of ore the 
 slag would become thick and pasty owing to the decrease 
 of the base FeO. 
 
 Test samples should be taken regularly after melting, the 
 fractions of which indicate the amount of carbon in them, 
 giving guidance to the melter as to the necessary additions 
 of ore. 
 
 In acid practice, only the carbon is considered in pre- 
 liminary tests, but for detailed information they may be ex- 
 amined for the usual elements. When in the judgment of 
 the melter the bath needs no more iron ore and the decreas- 
 ing carbon has reached a predetermined point (which can 
 be accurately estimated by the practiced eye) preparations 
 are then made to finish the heat. 
 
 The thermal conditions being satisfactory and the slag 
 normal, deoxidizers and recarburizers in the shape of ferro- 
 silicon and ferro-manganese are then added and the heat of 
 finished steel is ready to tap or draw off into a hot ladle. 
 
 DETAILS OF RE-CARBONIZING 
 
 Assuming that the chemical composition of the metal 
 going into castings shall be as follows : 
 
MANIPULATION OF HEATS IN ACID PRACTICE 45 
 
 C 0.25 per cent 
 
 Si 0.300 per cent 
 
 S 0.040 per cent or less 
 
 P 0.040 per cent or less 
 
 Mn 0.75 per cent or less 
 
 which may be considered as representative of regular prac- 
 tice on medium hard cast-steel. Taking as a basis 24,000 
 pounds of metal charged and the weights of ferro-silicon 
 and ferro-manganese 792 and 305 pounds respectively, from 
 them there will be furnished silicon, manganese and carbon 
 plus those several elements contained in the bath at the time 
 of final additions. According to these analyses the available 
 elements will be first, silicon from the FeSi with 10 per cent 
 silicon, 
 
 792 X o.io = 79.2 pounds Si; 
 
 second, manganese from the FeMn with 80 per cent manga- 
 nese, 
 
 305 X 0.80 = 224 pounds Mn; 
 
 third, carbon from both the FeSi and FeMn. 
 
 792 .X 0.015 = 11.88 
 305 X 0.055 = 16.77 
 
 28.65 Ibs. C 
 
 Taking into account the residual silicon, manganese and 
 carbon of the bath, and adding to them those furnished by 
 the FeSi and FeMn there will be, 
 
 Si = o.oooi X 24,000 + 79.2 = 81.6 total available 
 Mn^ 0.0003 X 24,000 + 244.0 = 251.2 total available 
 C = 0.0012 X 24,000 + 28.7 = 57.5 total available 
 
 With the number of pounds of the several elements di- 
 vided by the weight of the charge there will be found the 
 approximate analysis of the finished product, 
 
46 MANIPULATION OF HEATS IN ACID PRACTICE 
 
 100 X 8l.6 
 
 ioo X 251.2 
 ioo X 57.5 
 
 24,000 = 0.34 per cent Si 
 24,000 = 1.05 per cent Mn 
 24,000 = 0.24 per cent C 
 
 or calculating another way the approximate composition can 
 be determined in finished product using the same quantities 
 of values in terms of available silicon, manganese and car- 
 bon: 
 
 ioo X 79.2 Residual 
 
 + o.oi = 0.34 per cent Si 
 
 24,000 
 
 ioo X 244 
 
 + 0.03 = 1.05 per cent Mn 
 
 24,000 
 
 ioo X 28.65 
 
 + 0.12 = 0.24 per cent C 
 
 24,000 
 
 In making the foregoing computations the amounts of 
 sulphur and phosphorus have been ignored. With a slag 
 highly charged with silica no absorption of these takes place, 
 and the amount of either carried in by the melting stock 
 nearly equals the finished product in regard to their con- 
 tent. Usually there are slight gains. The sulphur increases 
 because of flame contamination and a loss of metallic iron 
 by oxidation. Phosphorus also increases slightly through 
 the latter cause. Both may be slightly augmented by move- 
 ments from the deoxidizers. 
 
 The chemical composition as shown in the foregoing fig- 
 ures will differ from that shown by the ultimate analysis 
 of a sample taken when the steel is going into the molds. 
 Comparing the approximate and ultimate figures we find: 
 
 Approximate Ultimate 
 
 Analysis. Analysis. 
 
 C 0.24 per cent 0.22 per cent 
 
 Mn 1.05 per cent 0.71 per cent 
 
 Si 0.34 per cent 0.31 per cent 
 
MANIPULATION OF HEATS IN ACID PRACTICE 
 
 The main difference is between the manganese added and 
 that found, the cause of which will be understood in the ex- 
 planation of the purpose in using the deoxidizing agents 
 FeSi and FeMn. During the second stage of the conver- 
 sion, the bath of molten metal carries variable quantities of 
 dissolved ferrous oxide (FeO) and with that substance 
 there is a vigorous chemical action between the carbon and 
 the oxygen of the FeO producing throughout the bath co- 
 pious bubbles of the gas CO. So long as the bath remains 
 liquid this chemical action will go on, charging it with that 
 gaseous body. Were the metal in that condition to be 
 poured into castings they would be found to be unsound or 
 honey-combed with blow holes. 
 
 Since solidity of product is one of the objects sought in 
 the physical properties, recourse must be had to some agent 
 or agents that will remove the cause of the gas-forming 
 action in the bath of metal before it can be drawn off. The 
 agents must possess greater affinity for the oxygen of the 
 dissolved FeO than the carbon in intimate association with 
 it. Practical experience has shown that manganese and sili- 
 con accomplish the purpose and these elements are com- 
 mercially available in the alloys, FeMn and FeSi. Hence 
 their designation as deoxidizers. The functions may be ex- 
 pressed as follows : 
 
 (1) FeO + Mn = Fe + MnO 
 
 (2) F 2 FeO + Si Fe 2 + SiO 2 
 
 In the interchange between the C and the FeO we have 
 a gas impregnating the bath. In the interchange, as shown 
 in the above equations, we have a solid or an easily fusible 
 slag formed by the SiO2 and the MnO, which being lighter 
 than the molten metal quickly floats to the surface of the 
 bath. Thus with normal conditions the metal will cease bub- 
 bling and pour quiet or "dead" and tend to make solid cast- 
 ings. 
 
 The amounts of silicon and manganese in the finished 
 metal will in the main depend upon the condition of the bath 
 
48 MANIPULATION OF HEATS IN ACID PRACTICE 
 
 at the time the deoxidizers are added, and the difference in 
 analyses between the approximate and ultimate figures may 
 be taken to represent the amount consumed in "washing" 
 the bath of metal. 
 
 The usual practice in adding the deoxidizers is as follows : 
 The carbon having dropped to, say, o.io to 0.12 per cent, 
 the dose of FeSi, broken into short pigs, is placed on the 
 breast of the furnace in order to heat it up before pushing 
 it into the bath. After a lapse of about eight minutes the 
 whole is pushed into the bath, allowing a little time to pass 
 during which the FeSi is melting and dissolving. To in- 
 sure a complete distribution the bath is agitated with bars 
 of iron. Ten minutes after the FeSi is added the metal is 
 ready to tap, but before doing so, the dose of FeMn is 
 thrown in and then the tapping may take place. It is some- 
 times the practice to put part of the FeMn into the fur- 
 nace and part into the ladle, the latter being done while the 
 steel is flowing into it. 
 
 In an acid heat the losses of silicon in the act of deoxidiz- 
 ing are not great, but of the manganese used for the same 
 purpose, considerable is consumed, being greater when en- 
 tirely added to the furnace than when divided between the 
 furnace and ladle. For this reason it is necessary to make 
 allowances for such losses when calculating the necessary 
 dose. Further, the losses are greater when using oil for 
 fuel than when gas is used, the latter flame being "soft" and 
 the former sharp. An oil flame may be likened to a blow- 
 pipe and its effect always strongly oxidizing. 
 
 As already shown, silicon will reduce the FeO, but the 
 reduction is not always complete, and what may escape the 
 silicon may further unite with the manganese. Hence the 
 reason for simultaneously using two powerful and active 
 reducing agents. The influence of the two agents remaining 
 in the finished steel after their reducing function will be 
 considered later. 
 
 The yield of metal after conversion against the weight 
 charged is a variable one and depends upon the character of 
 the stock and manipulation. If the stock, other than the pig 
 
MANIPULATION OF HEATS IN ACID PRACTICE 49 
 
 iron, should be of light, thin sections, there will be a heavy 
 melting loss due to excessive burning or oxidizing. If the 
 flame should be very sharp during the melting period, the 
 same condition will arise. The melting losses are also 
 heavier on oil fuel than on gas for the reason stated. Not 
 only will the conditions (when abnormal) result in heavy 
 melting losses, but the effect will be seen in certain physical 
 properties of the product. That is to say "over-oxidation" 
 from any cause is bad practice; yet even with a greater 
 tendency towards "over-oxidation" following the use of 
 oil, the condition is still subject to control. To express the 
 losses in figures to represent the difference between metal 
 charged and that yielded is not so easy, but it may be found 
 to be about 5 per cent. 
 
 The principal points about the manipulation of an acid 
 heat may be given as follows : 
 
 First To charge enough pig iron so that there will be a 
 high enough initial carbon to ensure an easily maintained 
 fluidity of bath. A low percentage of pig iron will mean a 
 rapid heat, but at the sacrifice of quality and a very high 
 flame temperature. The less carbon charged the greater 
 the chances of over-oxidation. 
 
 Second Charge as heavy sections of scrap as possible 
 and protect it as far as possible from burning with a cov- 
 ering of pig iron. 
 
 Third Watch the flame conditions, to guard against un- 
 due burning of stock. This point calls for a high degree of 
 skill to maintain flame conditions and yet reach the thermal 
 ranges necessary to melt the stock. 
 
 Fourth Be judicious in oreing heats. Too much ore is 
 harmful. 
 
 Fifth Aim for uniformity of product in a given class of 
 work. 
 
 Sixth Over-anxiety for tonnage will make the scrap pile 
 grow. 
 
CHAPTER V 
 
 BASIC PRACTICE SLAG SILICON SULPHUR PHOS- 
 PHORUS MANGANESE CARBON SLAG COM POSI- 
 TION ORDER OF CHARGING MELT- 
 ING CHARGING COLD STEEL 
 
 The problem of charging into a basic open-hearth fur- 
 nace, phosphoritic and otherwise impure materials, convert- 
 ing the same into good castings, offers many interesting 
 features. To get satisfactory and uniform results, certain 
 conditions require close attention and niceties of adjustment- 
 When with a suitably lined basic hearth and the proper 
 melting stock it is possible to begin with a high initial phos- 
 phorus and sulphur content and end with an almost com- 
 plete removal of them, it is clear that the chemical history of 
 the manipulations in conversion is more varied than that of 
 an acid heat. As already stated, 95 per cent of the phos- 
 phorus and 60 per cent to 75 per cent of the sulphur can be 
 eliminated in a basic bottom. Concerning the other con- 
 stituents the conditions in regard to their removal during 
 conversion and their presence in the finished product are 
 practically the same as in acid castings. 
 
 SLAG 
 
 Given, then, a basic bottom and melting stock, the changes 
 in a heat will be considered. The function of the basic bot- 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 51 
 
 torn is only a refractory one and the material entering into 
 it plays no part in the direct purification and conversion of 
 the stock. The calcerous slag is the active agent and the 
 bottom must be of such material that the action of such a 
 slag will have the least possible cutting effect. To mini- 
 mize that effect, the best results will be obtained with a mag- 
 nesite lining prepared as previously described. Limestone 
 or calcium carbonate forming the slag and classified as a 
 "base" will not flux with magnesite, also a "base." That is, 
 an admixture of two or more bases will successfully resist 
 high temperature or those common to steel melting; but a 
 combination of an acid (silica, etc.) and a base (lime or 
 magnesite, etc.) will at similar ranges of temperature read- 
 ily fuse. Hence the metallurgical necessity of the hearth 
 lining being of a similar character chemically to that of the 
 slag which may be formed in open-hearth melting. 
 
 Silica being greedy for any base and always ready to 
 fuse in the presence of a base, the life of a basic hearth 
 will be subject to its influence, therefore it is important that 
 the substance be allowed to enter the charge only in the 
 smallest amounts possible. Through the charge it will be 
 carried in as silicon (subsequently changed to silica) and 
 sand adhering to the stock. With pig iron cast in chills, 
 the amount of sand from that source will be practically nil. 
 Defective castings charged directly from the molding floor 
 may have more or less sand on them and unless they are 
 carefully cleaned to free them from burnt cores, etc., there 
 will be danger of undue hearth scorification both on the 
 slag line and bottom, causing an increased use of refrac- 
 tories for patching. 
 
 SILICON 
 
 Theoretically silica or silicon should be absent in basic 
 stock. Practically it is not free from either, but the condi- 
 tion is none the less subject to control. Therefore in se- 
 lecting pig iron for basic melting the maximum percentage 
 
52 MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 of silicon is given as one per cent. The amounts of that 
 element in the various kinds of steel scrap are so low that 
 the figures are negligible. Assuming the total charge to 
 be equal parts of pig iron and steel scrap the initial silicon 
 carried in by stock will average about 0.5 per cent. It is 
 good practice to keep in the neighborhood of that figure, 
 but a slight variation or increase will not be a serious ob- 
 jection. If there should be on hand some "off basic pig 
 iron" that is high in silicon (over one per cent), or high in 
 sulphur (over 0.05 per cent), a small quantity can be 
 charged at the rate of about one per cent of the charge and 
 will have but a very slight effect upoij the desired initial 
 composition. 
 
 SULPHUR 
 
 In regard to sulphur and its position in the composition 
 of the charge, it does not seem necessary to be very rigid 
 as to how much can be considered dangerous. Actually 
 standard basic pig iron, while rated at 0.05 per cent maxi- 
 mum carries more than that because of the generally ac- 
 cepted method of analysis giving uniformly low results. 
 Were a more tedious method and a more accurate one em- 
 ployed, the figures on sulphur would generally be nearer 
 0.07 per cent than the standard stock requirements; hence 
 the amount of sulphur actually going into a charge is really 
 higher than shown by the individual analyses of the pig 
 irons forming part of it. The same differences between 
 the apparent analyses and actual content of sulphur holds 
 true on various grades of pig iron until very low amounts 
 are reached. In the light of recent practice, the influence 
 of sulphur in the finished product is not regarded as harm- 
 ful as it once was. There is therefore no good reason to 
 place a strict limit on how much shall go into a charge. On 
 an average the total amount will rarely go above 0.04 per 
 cent. Rail scrap frequently carries 0.07 per cent and over. 
 The pig iron will range from o.oi to 0.05 per cent, and 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 53 
 
 with many kinds of steel scrap to select from, ranging from 
 0.02 per cent and higher, it will not be difficult to keep the 
 initial sulphur at or about the average stated if so desired. 
 In the writer's experience there was a time when it was 
 considered bad practice to have the initial sulphur exceed 
 0.03 per cent, and the aim was to get lower. (At this point 
 the question as to the influence of sulphur will not be con- 
 sidered, but will be deferred to a subsequent chapter). 
 With a normal basic slag and proper conditions, much 
 higher ranges of initial sulphur can be safely handled. 
 There will be almost complete elimination in conversion, so 
 that specification for finished product can be kept within 
 easy reach even with an initial sulphur of 0.07 per cent. 
 
 PHOSPHORUS 
 
 Concerning phosphorus and keeping in mind the ulterior 
 limit in the analysis of the finished product, it is as well not 
 to make the entire pig iron charge all high phosphorus 
 material. While the figure on standard basic pig iron is 
 one per cent and was drawn for ingot practice for castings 
 it is better to have at least two kinds of basic pig of which 
 one may be low and one other standard. A brand carry- 
 ing about 0.25 per cent of phosphorus or less is not diffi- 
 cult to procure. With a charge of mixed brands of pig 
 iron, and the usual varieties of scrap, it will be possible 
 to keep the initial phosphorus of the charge low. During 
 the progress of conversion and with a strongly basic slag, 
 there will be practically a total absorption of phosphorus 
 by it from the bath. At the end of the heat in recarburizing 
 with ferro-silicon, etc., there will be a tendency on the part 
 of the phosphorus to re-enter the metal, depending in a 
 great measure upon the amount of that element charged, 
 and the basicity of the slag. There is always some re-ab- 
 sorption, and it is greater with all high phosphorus pig 
 than when it may be diluted with low phosphorus pig iron. 
 Hence the need of mixing the pig iron in regard to phos- 
 
54 MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 phorus content. Average practice will give a negligible 
 reabsorption if the initial phosphorus is about o.i to 0.3 
 per cent. 
 
 MANGANESE 
 
 In addition to the elements enumerated, there is always 
 more or less manganese carried in with the stock. In 
 basic melting it is useful in several ways. One function 
 is to assist in maintaining a necessary fluidity of the slag, 
 and it will exist in that body principally as an oxide, re- 
 sulting from the metal in the charge. It also aids in the re- 
 moval of sulphur in the bath, and at some period during the 
 melting unites with it, forming a readily fusible sulphide 
 of manganese which floats upwards, and either dissolves 
 in the slag or upon reaching the surface and becoming ex- 
 posed to the flame action, may be oxidized or volatilized. 
 The action of manganese, however, is not quite clear, but 
 from actual results there is no doubt that its influence in 
 de-sulphurizing is quite potent. In that respect it operates 
 in conjunction with the Ca O (lime) furnished by the lime- 
 stone, which is also an active de-sulphurizer. It has been 
 observed that a moderately high initial manganese assists 
 in washing the bath of dissolved oxides always present in 
 metal which has been subject to the impinging action of 
 the flame before the period of liquation has been reached 
 and the mass disappears below the slag. The deoxidizing 
 effect will not begin until the mass has partially 
 or entirely fused and is . covered with a layer of 
 slag. Deoxidizing will take place by the union 
 of silicon and carbon with the oxygen, and what- 
 ever manganese may be present in excess of that 
 which may unite with the sulphur will considerably aug- 
 ment the effect. The joint operation of the three elements 
 will give cleaned metal. With ordinary melting stock the 
 average or initial manganese will be about 0.5 per cent, but 
 if it is possible by the use of a high manganese pig iron to 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 55 
 
 raise it nearer one per cent, the results or deoxidizing ef- 
 fect will be more satisfactory. Or in the absence of that 
 kind of melting stock, the deficiency can be made up with 
 spiegeleisen, an alloy carrying 20 per cent of manganese 
 or a certain amount of manganese ore will be useful for the 
 same purpose, if charged at the proper time, before the 
 first portion of metal. It is recommended, however, that 
 the needed manganese be carried in by high manganese pig 
 iron, because it is possible to get a better distribution of 
 that element than by adding a small quantity of the alloy 
 mentioned to make up any deficiency. For that reason it 
 is good practice when purchasing melting stock, to have on 
 hand pig iron carrying between 1.5 to 2.5 per cent man- 
 ganese. 
 
 CARBON 
 
 Since pig iron generally contains 2.5 to 3.5 per cent total 
 carbon and with 50 per cent pig and 50 per cent steel scrap 
 the initial carbon of a charge will be from 1.5 to 2 per cent. 
 Whether the carbon may be in the combined or graphitic 
 form is not considered, because when the pig iron is liquid, 
 it is all in the former condition irrespective of the constitu- 
 tion before melting. It will be observed that there is a 
 difference between acid and basic melting in regard to the 
 percentage of pig iron in the charge. In the former there 
 will be 15 per cent, but in the latter 50 per cent. The 
 reason for the difference is due to the larger volume of slag 
 formed in basic melting, which requires a considerable por- 
 tion of the heat units to penetrate it, to promote the neces- 
 sary fluidity of the bath of metal below. As has been 
 pointed out, the amount of carbon present largely controls 
 the melting point of iron and steel and with a high initial 
 percentage of it in a charge of cold stock, it is possible to 
 acquire liquation at a relatively low temperature. Were 
 there a low initial carbon as in acid melting, there would 
 be much trouble with viscosity of the bath, because of heat 
 
56 MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 absorption in the heavy body of basic slag-. With the bath 
 melted at a high range of carbon, the necessary thermal 
 conditions are comparatively easy to attain. A high carbon 
 at melting, between 0.5 to 0.8 per cent, is considered good 
 practice ; further, it means a prolongation of that period in 
 which deoxidization or clarifying of the bath occurs. A 
 lower carbon at melting would shorten that period, and at 
 the same time there would be undesirable conditions likely 
 to arise, such as low temperature of the bath, and overoxi- 
 dized metal, both tending to give bad results in the product. 
 Should a heat melt soft (low carbon) it can be "doctored" 
 by adding cold pig iron, and the temperature raised by 
 flame adjustments, but recourse to such treatment is always 
 dubious as to the outcome. 
 
 SLAG COMPOSITION 
 
 The foregoing elements of the charge being adjusted in 
 accordance with general practice, the next step will be to 
 consider the slag composition, which is to play an impor- 
 tant role in the conversion and purification of the bath. In 
 contrast to acid melting, the slag will be one highly charged 
 with lime, a substance already shown to be objectionable 
 in an acid hearth, but desirable in treating basic stock. 
 The normal composition of a basic slag will not be reached 
 until the stock has melted and the slag attained its maxi- 
 mum fluidity, which period occurs during the second stage 
 of a heat that for convenience may be designated "Metals 
 Melted." Taking as an average the analyses of twenty 
 samples of slag selected from various normal heats in the 
 second stage, the following can be considered the usual 
 composition of a basic open-hearth slag: 
 
 Silica 16.00 per cent 
 
 Ferrous oxide 22.00 per cent 
 
 Lime 40.00 per cent 
 
 Magnesia 8.50 per cent 
 
 Manganous oxide 8.50 per cent 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 57 
 
 The foregoing analysis can only be regarded as applying 
 to the conditions of practice existing when they were made, 
 and not as a rule to follow in all cases, but under general 
 conditions they may serve as a guide. In order to form 
 a basic slag there are several important conditions to take 
 into account. First Enough lime must be added to com- 
 bine with or neutralize the silica, which will result from the 
 oxidation of the silicon carried in by the metals, and that 
 carried in mechanically as sand ; otherwise, there would be 
 a thin slag and an undue cutting of the hearth by the silica. 
 Second Too much lime must be avoided, as it tends to 
 form a thick, pasty slag that causes trouble in tapping the 
 furnace or brings about a heavy consumption of fluorspar 
 to flux it. 
 
 There is a medium to be sought between basicity and 
 fluidity and at the same time the function of the slag in ab- 
 sorbing and retaining phosphorus must not be overlooked. 
 The measure of the latter action mainly depends upon the 
 amount of lime in excess of that united with the silica, lime 
 in a basic slag being the active de-phosphorizing agent. 
 As to fluidity, that condition in addition to the influence 
 of silica will depend upon the varying amounts of man- 
 ganese in the stock which enters the slag as an oxide. Ox- 
 ide of iron operates in the same manner and is always pres- 
 ent independently of that added as iron ore. The iron 
 oxide of the slag will result from the rusty coating on the 
 stock and the impinging action of the flame during the first 
 stages of a heat. In general terms the higher the content 
 of silica the greater the fluidity, but at the expense of de- 
 phosphorization and hearth. Fortunately the medium can 
 be found because of the presence of iron and manganous 
 oxide, since a law seems to exist in slag formations that 
 fluidity or fusibility is controlled by the number of bases 
 present. So, assisted by manganous oxide and iron oxide 
 in the presence of lime and silica, desirable slag conditions 
 in basic melting can be attained without much difficulty. 
 
 In addition to the elements given in the slag analysis 
 
58 MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 there may be small negligible quantities of alumina and 
 alkalies. The amount of magnesia results mainly from 
 hearth cutting. To forecast the exact amount of lime to 
 be added solely on theoretical grounds is not easy, because 
 the total available silica cannot be accurately determined 
 owing to the uncertain quantity of sand adhering to the 
 stock. Whatever silica may be available in the stock will 
 require at least enough lime in the ratio of I to 3 to satisfy 
 it. Also the purity of the limestone, generally used as a 
 carrier for the necessary lime of the slag, must be taken 
 into account and any silica this may have will depreciate 
 the available lime in the same degree. Therefore, the near- 
 er a limestone approaches a pure carbonate of lime the 
 greater its efficiency. In actual practice the amount of 
 limestone required to form a normal slag will vary between 
 7 and 15 per cent of the weight of the charge. With 
 standard melting stock and a pure limestone it will be com- 
 paratively easy to get satisfactory adjustments. In the 
 hands of a skillful operator assisted by the chemist the be- 
 havior of a heat can be followed at any desired stage, so 
 that untoward conditions of purification will be within 
 control. Agencies are constantly at work in slag -forma- 
 tions following the laws of attractions and affinities which 
 present problems of scientific interest. To solve their com- 
 plexities would be a difficult task. 
 
 If an open-hearth furnace be regarded as a chemist's 
 laboratory experiment on a large scale, and viewed from 
 a chemist's standpoint, the delicacies of definite chemical 
 reactions are not obliterated because of the very hugeness 
 of a furnace. Whether the operation be conducted on a 
 minute scale or on a larger one, definite laws are omni- 
 present. Could the operator bear in mind that he is hand- 
 ling tons of material with forces equally potent as in the 
 chemist's ounces, he would stand in awe of them. Careful, 
 intelligent work combined with practical experience makes 
 it possible to put basic practice in the same plane of effi- 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 59 
 
 ciency as acid melting notwithstanding prejudice to the 
 contrary existing in some minds. 
 
 ORDER OF CHARGING 
 
 Having determined upon the amount of limestone neces- 
 sary to form the basic slag, the next step will be to consider 
 the order of charging. With a basic hearth prepared as 
 outlined in previous chapters the method of charging may 
 be conducted as follows : First the stone is spread over the 
 bottom in as regular layers as possible, followed by the 
 lighter sections of scrap, which should be more or less 
 rusted. If not rusty, some iron ore may be added with 
 the scrap to make up the lack of oxide to assist in subse- 
 quent slag fusibility. Then two-thirds of the pig iron 
 should be distributed as evenly as possible, which may be 
 followed by the remainder of the scrap. By this time the 
 furnace will be well filled with bulky stock, and it may be 
 necessary to allow the flame to partly melt it, thus forming 
 room for the final addition of pig iron. 
 
 The serious part of melting now begins and at this stage 
 the finished product in regard to quality will be greatly in- 
 fluenced by the manner in which the exposed stock may be 
 subjected to flame action. There will be a desire on the part 
 of the operator to get the heat out of the furnace in record 
 time and with speed in mind the temptation to use a sharp, 
 hot flame will be strong.- On the other hand, a soft, mellow 
 flame will not melt as quickly as a sharp one, yet the char- 
 acter of the product will be better than under the first 
 named conditions. In one instance the bath will be highly 
 charged, when melted, with dissolved oxides and in the 
 other instance less so. It is not possible under any condi- 
 tions to melt a mass of pig iron and steel scrap without 
 some formation of oxides, and their subsequent solution 
 in the bath, so long as it remains uncovered by a protective 
 layer of slag. Good practice requires that there be a divid- 
 ing line between speed and slowness, a sharp flame and a 
 
6O MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 soft one, and that object is a difficult one to reach. The 
 quantitative expression of dissolved oxides cannot be 
 stated, but' their influence can be detected in varying de- 
 grees. That such conditions do exist is a fact, and the con- 
 clusion is based upon careful observations and research, 
 the results pointing undoubtedly to the influence of the 
 flame action, whether gas or liquid fuel may be used and 
 also to the amount of initial carbon present in the stock. 
 Oreing during the second stage or "Metals Melted" also 
 may aggravate the evil. That it is an evil, will be pointed 
 out, and one not fully appreciated or understood. To it 
 can be traced the non-success and consequent prejudice 
 against basic melting, because, in some instances, it has been 
 found in certain physical requirements that basic steel was 
 not as good as acid steel for castings. In other words, ex- 
 pressing the difference between basic and acid steel in 
 regard to certain physical properties, basic steel is more 
 likely to carry dissolved oxides than acid steel, provided 
 care has been taken to guard against their introduction or 
 formation. Differences in chemical composition can only 
 exist in the amount of sulphur and phosphorus for a given 
 grade of product and the said differences being too small 
 to exert any great influence, no cause can be assigned other 
 than the one mentioned in studying the contrasting physi- 
 cal properties of both classes of steel. The condition of 
 over-oxidation is not so liable to present itself with a gas 
 flame, but even such a flame can be so adjusted as to abuse 
 the stock. With an oil flame the danger is greater, because 
 of the intense blow pipe effect of the oil burner. In ex- 
 perienced hands, however, quality duly regarded, normally, 
 the oxidizing effect with any kind of fuel need not be ex- 
 cessive. 
 
 MELTING 
 
 Under flame action in a fully charged furnace the stock 
 will soon begin to melt and it will be observed that the melt- 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 6l 
 
 ing of the pig iron goes on quite rapidly, dripping to the 
 hearth in small streams, while the steel scrap melts at a 
 slower rate, and in melting, the liquid steel frequently scin- 
 tillates or burns forming either vapors or solid oxide. 
 The dripping pig iron lessens that action, more or less, by 
 protecting the lower carbon steel stock. The melting pro- 
 ceeds gradually, and the mass decreases in bulk and passed 
 below a layer of rapidly forming slag. At this time, how- 
 ever, the slag formation is not entirely complete, but will 
 be found to consist mostly of silica and iron oxide with 
 comparatively little lime, the temperature of the furnace 
 at this point not being sufficiently high to thoroughly dis- 
 solve the lime still resting on the bottom of the furnace. 
 The composition of the slag at this period is practically 
 an acid one and results from the sand of the stock and the 
 silicon of the metals, since it is one of the first elements to 
 submit to oxidation. Samples taken from the first portions 
 of melted metal have shown a high sulphur content, much 
 higher than that carried in by the stock. The cause can be 
 attributed to the absorption of sulphur from the flame. 
 During the first stage but little information can be ascer- 
 tained as to the changes which may take place in the com- 
 position, because the mass is a conglomeration of pasty 
 semi-liquid pig iron, steel and slag. 
 
 CHARGING COLD STOCK 
 
 In charging cold stock about one half of the time re- 
 quired to work a fifteen to a twenty-ton heat is taken up 
 in melting. As that interval advances the slag is increasing 
 in volume and some changes in composition will take place 
 in the first portions of the metals melted. The slag is more 
 or less viscous with lumps of lime coming to the surface 
 as it becomes loosened from the bottom. There is a foam- 
 ing or bubbling in the slag caused by COa being liberated 
 from the limestone and also by the action between the ox- 
 ides of the slag and the carbon of the bath. As the lime 
 
62 MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 continues coming to the surface it increases the basicity of 
 the slag and with that condition there will be a rapid ab- 
 sorption of phosphorus from the bath. 
 
 The operator feels over the bottom of the furnace with 
 a long bar to determine whether the bath is entirely melted 
 or not. If no pasty masses are touched the bath is com- 
 pletely molten. A sample of metal is then poured into a 
 small oblong mold, quenched and broken to study the frac- 
 ture, the appearance of which indicates the carbon present 
 and also whether there may be any phosphorus remaining. 
 An experienced eye can readily read the carbon of the test 
 quite closely. If the grain is small and the fracture of 
 rather a dull color, the heat is said to have melted "hard." 
 If the grains are large and bright the heat has melted 
 "soft." Other indications confirm this in the way the first 
 test piece may break under the blows of a sledge, the 
 "hard" or high carbon steel snapping easily, while the 
 "soft" or low carbon steel bends or flattens before fracture. 
 
 To confirm the melter's judgment, as a rule, the first test 
 piece is sent to the laboratory for a quick test on carbon 
 and phosphorus. When the test piece indicates a high 
 carbon, the melter does not wait for the analysis, which 
 only occupies about ten minutes, but may throw into the 
 bath a few shovels of iron ore to assist in the lowering 
 of the carbon, and, at the same time, thin the slag. In a 
 few moments the chemist reports carbon usually 0.50 to 
 0.70 with only a trace of phosphorus. The melter is now 
 satisfied that his conditions are normal. Should the phos- 
 phorus have shown incomplete removal in the first test, a 
 second one would be taken after an interval of about fif- 
 teen minutes and submitted to another chemical analysis. 
 The addition of the iron ore immediately after the first test 
 would have increased the slag's basicity and fluidity and 
 no doubt would have eliminated what phosphorus might 
 have remained in the bath at the time of the first sample. 
 Usually the second test shows phosphorus out of harm's 
 way. 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 63 
 
 If the phosphorus resists removal, it is necessary to add 
 lime to the slag. This would have a tendency to cool the 
 bath and also raise the lime content of the slag and of 
 course intensify the de-phosphorizing action. If the tem- 
 perature of the bath at melting be very high, there may be 
 a tendency to retard the removal of phosphorus, since there 
 seem to be certain ranges of temperature wherein de-phos- 
 phorizing does not depend entirely upon slag conditions. 
 It has been observed that in cases where a normal slag re- 
 fused to completely de-phosphorize, an addition of a few 
 hundred pounds of pig iron caused a quick and perfect ab- 
 sorption of the phosphorus by the slag. The action could 
 only be explained on the basis that the pig iron so added 
 could exert no other influence regarding the phosphorus 
 than that of cooling the bath. This fact gives plausibility 
 to the theory of the thermal ranges where they may be 
 a change or reversal of affinities, or, in other words the 
 oxidizing action is more inclined to give its attention to 
 the carbon in preference to the phosphorus. 
 
 The activity in de-carbonizing is one that varies with the 
 temperature and this fact also seems to apply with equal 
 force inversely to phosphorus. The term "very hot" may 
 seem indefinite in reference to the bath when melted, but 
 in the absence of figures stating temperatures at the several 
 stages of an open-hearth heat, the use of vague expressions 
 concerning them cannot be avoided. It is doubtful if the 
 slight ranges of temperature covering the periods affecting 
 the removal of phosphorus can be observed by the eye even 
 with experienced operators. It is certain that the ranges 
 of temperature in question come within apparently normal 
 conditions and are therefore narrow. But, if the ranges 
 were wider, they would be readily detected without the 
 seeming need of pyrometry to fix them. 
 
 Assuming that the second test sample shows a low phos- 
 phorus, usually 0.005 - OI 5> further manipulation proceeds 
 smoothly. The slag will have nearly reached its full share 
 of lime; in other words, the lime will have come to the 
 
64 MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 surface from the hearth bottom where it was charged. If 
 the dose of iron ore has not given the desired fluidity to 
 the slag, an addition of two or three shovels of fluor-spar 
 will bring it about quite rapidly. Fluor-spar should be used 
 but sparingly because excessive doses tend to aggravate 
 the cutting action on the slag line around the hearth. 
 
 If the first or second preliminary test samples show low 
 phosphorus the examination for that element is not carried 
 out in subsequent tests. Only the carbon is carefully 
 watched and after it has reached about 0.25 no further ore 
 should be given to a bath. The carbon reduction can go 
 on without any assistance from that source because the slag 
 will, at that period, carry its full quota of iron oxide to 
 actively promote decarburizing. Excessive additions of 
 iron ore will surcharge the bath with oxide \vhich causes 
 blow holes and red shortness in the finished product. A 
 liberal use of iron ore will shorten the time of making a 
 heat and in a measure, increase the yield of metal because 
 the reaction between the carbon of the bath and the oxide 
 of iron furnished by the ore sets free an equivalent of metal- 
 lic iron which enters the bath. But good practice demands 
 discretion in the treatment of the bath with ore. 
 
 It generally follows that when a heat melts "hard," tem- 
 perature conditions at the end of the heat will be normal 
 because the initial amount of carbon will be high enough to 
 produce perfect liquation and the interval of time occupied 
 in its removal will be great enough to preserve proper ther- 
 mal adjustments by regular reversals of flame action, in- 
 flowing air and outgoing gases. Should a heat melt "soft," 
 the temperature of the bath is apt to be low and without a 
 gain in the carbon much difficulty would be encountered 
 with pasty steel. An attempt to raise the flame temperature 
 to get the fluidity of the bath would harm the roof and 
 brick work of furnace by scorching. But recourse is had 
 to extra doses of cold pig iron to furnish the necessary car- 
 bon and thus promote fusion and other desirable conditions. 
 Such practice is called "doctoring" but the outcome is 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 65 
 
 never so satisfactory as when the heat is worked naturally. 
 It is a safe rule that when a heat melts "soft" and conse- 
 quently dull or cold, that the first preliminary test will 
 show thorough slag absorption of phosphorus. This fact 
 gives plausibility to the theory of the temperature ranges 
 affecting the removal of that element. 
 
 When the carbon reaches a point where the test samples 
 are very tough, difficult to break and showing, upon frac- 
 ture, a fibrous structure, and chemical tests show them to be 
 between o.io to 0.14 carbon, preparations are made to finish 
 the heat. If a test spoon of liquid steel be taken from the 
 bath and poured over the lip, running freely and leaving 
 the spoon clean and free from skull or chilled steel, the tem- 
 perature of the bath is considered good. The next step will 
 be to quiet the bath by adding one-third of a weighed por- 
 tion of ferro-manganese. This material as shown in pre- 
 vious chapters takes precedence over the remaining carbon 
 in de-oxidizing and so lessens the ebullition of gaseous car- 
 bon. After an interval of a few minutes the heat of metal 
 is tapped into a pre-heated ladle, and before the slag comes, 
 the final doses of ferro-manganese and ferro-silicon are 
 thrown in with the stream of steel and the heat is said to be 
 finished. The furnace bottom is now drained of slag and 
 some remaining pools of steel. The slag line is patched 
 with raw dolomite and whatever holes there may be in the 
 bottom are filled with ground magnesite. The furnace is 
 then ready for a succeeding charge. 
 
 DETAILS OF RE-CARBURIZING 
 
 It will be noted that in basic melting no ferro-silicon is 
 added to the bath to deoxidize as in acid practice. To do so 
 would be to invite irregularities, such as a release of phos- 
 phorus from the slag, a re-absorption of it in the bath, a 
 cutting of the hearth, a great loss in available silicon and 
 uncertainty as to quality of product. The measure of the 
 
66 
 
 MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 power of the slag to take up phosphorus from the bath is 
 its basicity and any addition of silicon decreases that prop- 
 erty. The better practice is to make the silicon additions 
 
 FIG. 8. COMPOSITION OF A NORMAL BASIC OPEN-HEARTH HEAT 
 
 entirely in the ladle. If the ordinary grade of ferro-sili- 
 con (9 per cent to 13 per cent Si) is used, it must be melted 
 before it can be added to the ladle. 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 The melting may be done in a small cupola, or a reverba- 
 tory furnace fired with coal, or in a suitable oil furnace. 
 In any case, there are certain annoyances, such as hitches 
 in getting the ferro-silicon melted at the proper interval 
 and in condition to transfer the desired quantity to the steel 
 ladle when the heat has been tapped from the open-hearth 
 furnace. There are also losses in available silicon due to 
 oxidation when melting the charge of ferro-silicon by any 
 method. These losses are variable and cannot be avoided. 
 In making calculations allowances must be made for such 
 losses. It would not be practical to add the dose of ferro- 
 silicon, using the common grade, cold. The drawbacks just 
 mentioned can be avoided by the higher grade electrolytic 
 ferro-silicon which usually carries about 50 per cent of 
 silicon. The quantity of this latter grade for a dose is only 
 about one-fifth of the commoner kind, so that there is no 
 need to melt it before using, thus eliminating the difficulties 
 arising in handling the ordinary grade. The heat units 
 evolved in the reaction between the oxide of iron contained 
 in the molten steel and the silicon available in the high 
 grade ferro-silicon, more than offset any chilling effect set 
 up by the addition of a cold charge of that material. There 
 is always some consumption of silicon in de-oxidizing and 
 it is greater in basic melting than in acid. The loss varies 
 between 20 per cent to 30 per cent of the total available, 
 particularly in using the common grade of ferro-silicon. 
 
 The greater part of the ferro-manganese addition can be 
 made in the ladle with but a slight loss in de-oxidizing and 
 with no chilling effect upon the liquid steel. 
 
 If the finished product is to show a final analysis as fol- 
 lows : 
 
 C 0.20 to 0.25 per cent 
 
 Si 0.30 to 0.35 per cent 
 
 Mn 0.65 to 0.85 per cent 
 
 the calculations for the recarburizers, as an illustration, will 
 
68 MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 be made on the basis of molten FeSi with 9 to 13 per cent 
 Si and also on FeSi carrying 50 per cent Si, the weight of 
 the charge being 36,000 pounds. Taking the lower limit of 
 silicon, in the final analysis there will be required 
 
 36,000 X 0.003 1 08 pounds Si. 
 
 With the ferro-silicon averaging n per cent silicon the 
 charge of that material to furnish 108 pounds silicon is 
 108 X ioo 
 
 = 982 pounds, 
 
 ii 
 
 but there may be a combined loss of the synthetical silicon 
 in melting and in de-oxidizing of about 20 per cent ; it will 
 be necessary to increase the charge of ferro-silicon accord- 
 ingly to 982 X 1.2 = 1,1783/2 pounds. The carbon carried 
 in by that quantity will be 
 
 1,1783/2 X 0.015 17.67 or 18 pounds 
 divided by the weight of the charge 
 18 
 
 x ioo = 0.04, 
 
 36,000 
 the percentage of available carbon. 
 
 To get the manganese in the final analysis to, say, 0.75 
 per cent it will be necessary to add : 
 36,000 X 0.0075 ~ 2 7 pounds Mn. 
 
 Standard ferro-manganese carries 80 per cent, then to have 
 the synthetical manganese it will require 
 270 X ioo 
 
 = 337/ / 2 pounds FeMn. 
 
 80 
 
 The residual manganese in the bath at the time of the ad- 
 ditions will be an increment and it will nearly equal the 
 loss mentioned in available manganese. From the ferro- 
 manganese there will be carbon furnished to the extent of 
 
 337^ X 0.06 = 20^4 pounds. 
 21 
 
 X ioo = 0.05 per cent C. 
 
 36,000 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 69 
 
 The addition of ferro-manganese is divided into two por- 
 tions, 1-3 in furnace and 2-3 in the ladle. 
 
 Summarizing the recarburizers there will be available in 
 total elements. 
 
 C 0.12 + 0.04 0.05 0.21 
 
 Mn 0.15+ 0.75 0.90 
 
 Si 0.02 + 0.36 0.38 
 
 The foregoing available composition in regard to the car- 
 bon will be very close to the required analysis and the com- 
 plete analysis compared with the total available elements 
 will depend in a great measure upon the condition of the 
 bath in regard to the quantity of oxide formed during the 
 conversion and refining of the charge. 
 
 If the electrolytic ferro-silicon be preferred to the com- 
 mon grade, it will not change the details in regard to the 
 dose of ferro-manganese, which will remain the same as 
 shown in the preceding paragraphs. The figures for sili- 
 con can be taken as follows : 
 
 36,000 X 0.003 1 08 pounds Si. 
 108 X 100 
 
 = 216 pounds Fe-Si. 
 
 50 
 
 " Since the electrolytic silicon being in a condensed form 
 the de-oxidizing effect is more pronounced and efficient. 
 The losses are much less than in the ordinary grade. There 
 arises no particular need to take into account the losses or 
 differences between the synthetical analysis and the final. 
 It is not necessary in adding silicon as a deoxidizer to work 
 extremely close to figures for finished analysis, if enough 
 is added in the first place to accomplish the purpose. The 
 amount of carbon furnished by the electrolytic silicon is 
 
 0.54 
 
 quite small 216 X 0.0025 0.54 pounds and X IQO 
 
 36,000 
 = o.ooi per cent or practically nothing at all. From the 
 
70 MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 ferro-manganese there will be 0.05 per cent which added to 
 the residual, assumed to be 0.12 per cent, would make the 
 available carbon 0.17 per cent, a figure rather low for 
 ordinary castings where the tensile strength required may 
 be 60,000 to 70,000 pounds per square inch. 
 
 Two methods may be followed to raise the carbon to 
 0.20 0.25 per cent. One is to finish the heat at a corre- 
 spondingly higher residual carbon. The other is to make 
 up the difference between 0.17 and 0.20 0.25 by an addi- 
 tion of solid carbon in the form of coke or anthracite at 
 the time of dosing in the ladle. 
 
 Concerning sulphur, there is no great effect upon it at 
 the time of recarburizing. The figures for the residual and 
 the final analyses are practically the same excepting occa- 
 sionally a small decrease coming within the common errors 
 of chemical tests. If any differences occur outside of them, 
 the lessening of the sulphur can be assigned to the influ- 
 ence of manganese in combining with some sulphur and 
 passing off into the slag as a manganese sulphide. No 
 gain in sulphur at this stage has been observed in the writ- 
 er's experience. 
 
 If the initial phosphorus is kept low in the charge, about 
 o.io to 0.15 per cent and with only 0.005 0.015 per cent 
 remaining at the time of finishing the heat, the synthetic 
 silicon in the action of de-oxidizing in the ladle changing to 
 silica passes into the super-natent slag, combines with the 
 lime, liberates or releases some phosphorus held there, 
 which immediately reverts to the molten metal and raises 
 the analysis of that element. Under such conditions, tak- 
 ing a sample when about one-half of the metal from the 
 ladle has been poured, there will be found a marked in- 
 crease of the phosphorus over the residual so that the 
 finished steel will show 0.02 0.03 phosphorus. The re- 
 phosphorizing in the ladle goes on continually so long as the 
 metal remains liquid and if a sample were taken from the 
 last portion of metal in contact with the slag, there would 
 be found a still further gain in phosphorus. If a much 
 
MANIPULATION OF HEATS IN BASIC PRACTICE 
 
 higher initial phosphorus were charged there would be 
 a still greater revision, and castings poured at the end of a 
 heat might be found with a content of phosphorus near the 
 point of being objectionable. This feature of re-absorption 
 is peculiar to the basic process and is more noticeable in the 
 steel castings practice than in the production of ingots for 
 rolling purposes, because in the latter product the silicon 
 added at the end of a heat is much less than in the former. 
 Hence the importance of not charging large quantities of 
 highly phosphoritic stock in the manufacture of basic steel 
 castings. 
 
 The melting losses depend upon the character of the stock 
 and the conditions of practice and may vary from 6 per cent 
 to 12 per cent of metal charged. In representative plants 
 with established and extensive experience a yield of 93 per 
 cent in the ladle can be considered a fair average. 
 
CHAPTER VI 
 
 CHEMICAL ANALYSES AND PHYSICAL TESTS. DETERMINA- 
 TIONS OF SILICON, PHOSPHORUS, MANGANESE, 
 
 SULPHUR AND CARBON IN STEEL AND 
 i PIG IRON. PRELIMINARY 
 
 TESTS. PHYSICAL 
 TESTS. SOLU- 
 TIONS 
 
 The appended methods of chemical analyses are the result 
 of several years' experience and have proven to be accurate 
 and reliable. In several instances, the methods may be 
 considered as standard, and are in daily use in many lead- 
 ing establishments. 
 
 In the manufacture of basic steel castings, where it is 
 highly important to know the extent of dephosphorization 
 as quickly as possible in preliminary tests taken during the 
 progress of a heat, a suitable centrifugal machine is neces- 
 sary, which can be purchased through any house dealing 
 with chemical supplies. 
 
 DETERMINATION OF SILICON IN STEEL 
 
 Five times the factor weight (2.3510 grams) of the drill- 
 ings are weighed off, placed into a 5 % -inch porcelain dish, 
 and a watch glass placed concave side down. Thirty cc. 
 
CHEMICAL ANALYSES AND PHYSICAL TESTS 73 
 
 of silicon mixture are added. The dish is heated until the 
 solution is evaporated, and fumes of sulphuric acid are 
 given off. After cooling, 45 cc. hydrochloric acid, (i part 
 hydrochloric, 2 parts water) are added. Heat until all is 
 disolved, except the separated silica. Filter, wash alternately 
 with hot hydrochloric acid (i-i) and hot water, until free 
 from iron and acid. Burn and weigh. One fifth the weight 
 represents the silicon percentage. Time required, fifty min- 
 utes. 
 
 DETERMINATION OF PHOSPHORUS IN STEEL 
 
 Two grams of the drillings are weighed off, placed into 
 a 250 cc. Erlenmyer flask, and 35 cc. nitric acid (specific 
 gravity 1.2) added. The flask is heated until the drillings 
 are in solution, and no brown fumes are given off. Ten 
 cc. of a five per cent solution of potassium permanganate 
 are added, and the heating continued until the solution is 
 colored brown, when a few drops of saturated cane sugar 
 solution are added. The solution is cleared, and after heat- 
 ing a few minutes the flask is removed from the heat, and 
 cooled in water. Neutralize with concentrated ammonia, 
 and acidify with concentrated nitric acid until the solution 
 assumes an amber color. To this solution at a temperature 
 of 70-80 degrees Cent., add 50 cc. of ammonium molybdate 
 solution, and a few drops of ammonia to hasten the pre- 
 cipitation. Shake well, and allow to settle in a warm place 
 for ten minutes. Filter, wash with cold water until free 
 from acid. Remove paper to original flask, add from 3 to 
 5 cc. or more if necessary, of potassium hydroxide solution. 
 Add a few drops of a one per cent alcoholic solution of 
 phenolphthalein. Shake thoroughly, allow to stand for five 
 minutes, and titrate excess of alkali with nitric acid solu- 
 tion. The number of cubic centimeters of potassium hy- 
 droxide solution, neutralized by the phosphomolybdates rep- 
 resents hundredths per cent of phosphorus in the drillings. 
 Time required, forty minutes. 
 
74 CHEMICAL ANALYSES AND PHYSICAL TESTS 
 
 DETERMINATION OF MANGANESE IN STEEL 
 
 One-tenth dram of the drillings is weighed off and 
 placed into a test tube (i x 8 inches), and 15 cc. nitric acid 
 (specific gravity 1.2) are added. The best tube is heated 
 over a sand bath until the drillings are in solution, and the 
 nitrous oxide fumes are expelled. About 0.7 grams lead 
 peroxide (PbOa) is added to the test tube taken from the 
 sand bath, after which it is replaced and heated for one 
 minute, then filled two-thirds with hot water. Boil a few 
 minutes and settle by means of a centrifugal machine. 
 Pour the clear solution into a flask, and titrate with sodium 
 arsenite solution. The number of cubic centimeters of 
 arsenite solution required, represents tenths per cent of 
 manganese in drillings. 
 
 Time required, twenty-five minutes. 
 
 DETERMINATION OF SULPHUR IN STEEL 
 
 Fifty times the factor weight (6.88 grams) are weighed 
 off, and placed into a 500 cc. Erlenmyer flask: The de- 
 livery tube is attached leading into a 100 cc. Erlenmyer 
 flask, which contains 10 cc. hydrogen peroxide, 10 cc. 
 concentrated ammonia, and 20 cc. water. Eighty cc. of hot 
 hydrochloric acid (i-i) are added, and heat applied. 
 When solution is complete, disconnect the apparatus, make 
 the solution slightly acid with hydrochloric acid, bring to 
 a boil, and add 10 cc. of a ten per cent solution of hot 
 barium chloride. Boil five minutes, filter, using ashless 
 pulp, burn and weigh barium sulphate, and make correction 
 for sulphur in hydrogen peroxide. The weight of barium 
 sulphate, divided by 50, represents thousandths per cent 
 sulphur. Time required, fifty minutes. 
 
 DETERMINATION OF CARBON IN STEEL 
 
 Weigh 0.3 grams of drillings, and 0.3 grams of stand- 
 
CHEMICAL ANALYSES AND PHYSICAL TESTS 75 
 
 ard into separate test tube. Add 5 cc. nitric acid (speci- 
 fic gravity 1.2). Heat in a water bath until solution is 
 complete. Cool in water and compare. Time required, 
 thirty minutes. 
 
 DETERMINATION OF SILICON IN PIG IRON 
 
 One gram of the sample is washed off into a 324-inch 
 casserole, and a watch glass placed concave side down. 
 A few drops of concentrated hydrochloric acid are added, 
 then 15 cc. of silicon mixture. The casserole is heated 
 until sulphuric acid fumes are given off. After cooling, 
 add 30 cc. hydrochloric acid (1-2), and boil until all is dis- 
 solved except the carbon and silica. Filter, wash alter- 
 nately with hot hydrochloric acid (i-i), and add hot water 
 until free from iron and acid; burn and weigh. If high 
 in silicon, treat with hydrofluoric acid. The weight mul- 
 tiplied by 0.47 represents the percentage of silicon. Time 
 required, one hour. 
 
 DETERMINATION OF PHOSPHORUS IN PIG IRON 
 
 From high phosphorus pig weigh off 0.2 gram and from 
 low phosphorus pig as much as is best suited for the de- 
 termination. To the 0.2 gram add 25 cc. of nitric acid 
 (specific gravity 1.13). Boil until the drillings are in so- 
 lution, adding more acid if necessary. Filter off graphite, 
 add potassium permanganate, and proceed as in the deter- 
 mination of phosphorus in steel. Time required, fifty min- 
 utes. 
 
 DETERMINATION OF MANGANESE IN PIG IRON 
 
 Proceed as in the determination of manganese in steel. 
 
 DETERMINATION OF SULPHUR IN PIG IRON 
 
 Five grams of the drillings are weighed off into a 500 
 
76 CHEMICAL ANALYSES AND PHYSICAL TESTS 
 
 cc. Erlenmyer flask, and a delivery tube connected which 
 leads into a glass containing 10 cc. cadmium chloride solu- 
 tion, and 50 cc. water; 70 cc. hot hydrochloric acid (i-i) 
 are added through the thistle tube, and heat is applied until 
 the drillings are in solution, and steam has driven off all 
 other gases. Disconnect, add a few cubic centimeters of 
 starch solution, 40 cc. hydrochloric acid (i-i), and titrate 
 immediately with iodine solution. The number of cubic 
 centimeters of iodine solution represents hundredths per 
 cent of sulphur. Time required, thirty-five minutes. 
 
 PRELIMINARY TESTS 
 
 CARBON. Weigh off 0.2 gram of the drillings, and 0.2 
 gram of the standard into separate test tubes. Heat until 
 solution is complete and compare colors. Time required: 
 Four minutes. 
 
 MANGANESE. Proceed as in the final test. Do not boil 
 after adding water. Time required, six minutes. 
 
 PHOSPHORUS. One and one-half grams of the steel are 
 weighed off into a 300 cc. Erlenmyer flask, and 35 cc. nitric 
 acid (specific gravity 1.2) are added. When violent action 
 ceases, place flask on heat, boil until drillings are in solu- 
 tion and no brown fumes are given off. For every eight 
 points carbon, add one cubic centimeter of concentrated 
 solution of chromic acid. Boil one minute. Transfer to 
 a graduated bulb containing 50 cc. molybdic acid solution, 
 shake violently, and allow to settle in centrifugal machine. 
 Each small division on bulb represents hundredths per cent 
 phosphorus. Time required, five minutes. 
 
 SOLUTIONS 
 SILICON MIXTURE: 
 
 Water 1,300 cubic centimeters 
 
 Nitric acid 435 cubic centimeters 
 
 Sulphuric acid 275 cubic centimeters 
 
 Add sulphuric acid slowly. 
 
CHEMICAL ANALYSES AND PHYSICAL TESTS 77 
 
 STANDARD ACID AND ALKALI FOR PHOSPHORUS DETERMI- 
 NATION : 
 
 i per cent solution potassium hydroxide, 
 i per cent solution nitric acid. 
 
 100 milligrams oxalic acid crystals equals 10 cc. potas- 
 sium hydroxide. 
 Standardize with steel of known phosphorus contents. 
 
 MOLYBDATE SOLUTION: 
 
 Stock Solution. 
 
 Molybdic acid 250 grams 
 
 Ammonia 1,000 cubic centimeters 
 
 To use, 90 cc. stock solution in 338 cc. nitric acid 1.20 
 specific gravity. 
 
 Keep acid cool, and stir while adding molybdic acid. 
 
 SODIUM ARSENITE SOLUTION : 
 STOCK SOLUTION. 
 
 Sodium Carbonate 15 grams 
 
 Arsenous Oxide 5 grams 
 
 Water 500 cubic centimeters 
 
 Boil and filter. 
 
 To use, take 68 cc. stock solution to 2,000 cc. water, 
 standardize against steel of known manganese contents. 
 
 CADMIUM CHLORIDE SOLUTION: 
 
 Cadmium chloride 35 grams 
 
 Water 1,200 cubic centimeters 
 
 Ammonia 800 cubic centimeters 
 
 IODINE SOLUTION: 
 
 Iodine 9 grams 
 
 Potassium Iodide 18 grams 
 
 Dissolve in a few cc. of water and dilute to two liters. 
 Standardize against American Foundrymen's Association 
 pig iron standards. 
 
CHEMICAL ANALYSES AND PHYSICAL TESTS 
 
 STARCH SOLUTION: 
 
 Mix a tablespoon full of starch with a little cold water, 
 and pour into 100 cc. of boiling water, and continue the 
 boiling for some time. Shake before using. 
 
 NITRIC ACID, SPECIFIC GRAVITY, 1.2: 
 
 Nitric acid 1,250 cc. 
 
 Water 185 cc. 
 
 NITRIC ACID, SPECIFIC GRAVITY, 1.13: 
 
 Nitric acid 645 cc. 
 
 Water 1,800 cc. 
 
 PHYSICAL TESTS 
 
 The specifications prepared by the American Society for 
 Testing Materials, and shown in the following table, are 
 
 FIG. 9. STANDARD TEST BAR 
 
 quite generally recognized as being fair to both producer 
 and customer. An objection to be raised is the costly 
 style of the test bars recommended (Fig. g). 
 
CHEMICAL ANALYSES AND PHYSICAL TESTS 79 
 
 When a large number of tests are to be made in the ma- 
 chine, cutting the threads greatly increases the cost. 
 
 When no specifications for physical tests are imposed 
 the type of the bar shown in Fig. 10, will answer for works 
 
 FIG. 10. TEST BAR FOR WORKS* TEST 
 
 tests, solely as a guide to the quality of product, and where 
 the customer does not insist upon the test-bar with the 
 threaded ends as illustrated. 
 
 The specifications, mainly, are drawn to cover the physical 
 properties and they alone should be considered, for the rea- 
 son that there is no distinct relation between the chemical 
 composition and the physical behavior. 
 
 The method of the treatment of castings is the prime 
 consideration, and with a given chemical composition, it is 
 possible to get a wide variation in physical tests with dif- 
 ferent heat treatments. From the standpoint of the manu- 
 facturer, specifications should only be applied to physical 
 tests, leaving to him the adjustments of the processes of 
 manufacture to harmonize with the expected physical re- 
 quirements. 
 
 Test bars can only at best, show the condition their 
 metal may be in, both in regard to the composition and the 
 treatment to which they may have been subjected. 
 
 It would not follow that, because a test bar represent- 
 ing a given heat of steel, gives poor results in tensile or 
 bending stress, the entire heat was poor also in that. A 
 test-bar can only show what might be expected of a finished 
 casting of equal composition if it were to receive the same 
 treatment as did the bar. 
 
 Tensile and bending tests serve useful purposes, but still 
 
80 CHEMICAL ANALYSES AND PHYSICAL TESTS 
 
 leave something lacking. If a test-bar under such stress 
 gives excellent results or if a finished casting be subjected 
 to the drop test, and possibly bends without fracture, prov- 
 ing great ductility, the evidence is not necessarily a posi- 
 tive guide as to life in service. Quite generally a steel cast- 
 ing is set up in service in such a manner that it may be 
 constantly under alternating stresses or rapid and repeated 
 shocks. Practical results from service data do not con- 
 firm the theory that ductile steel should resist such strains. 
 Failures under such stresses frequently fall short of expla- 
 nation in the light of recognized tests projected to prevent 
 such failures. There then arises another want to be filled, 
 a gap vacant because chemical analyses, physical tests and 
 microscopical examinations do not enable the metallurgist 
 to throw around the fruit of his labors the fullest possible 
 precautions intended within the scope of such tests. Possi- 
 bly fuller information is to be found by supplementing 
 already established tests with other tests to approach as 
 near as possible the conditions to which steel castings are 
 subjected in service, namely vibratory stresses. A study 
 of such stresses will no doubt give much valuable infor- 
 mation and change certain conditions now thought to be 
 essential. 
 
CHAPTER VII 
 
 RELATldN BETWEEN COMPOSITION AND PHYSICAL PROPER- 
 TIES CARBON CAST STEEL TENSILE STRENGTH 
 DUCTILITY ELASTICITY ELASTIC LIMIT 
 HARDNESS 
 
 CARBON In order to understand the matter under con- 
 sideration it would be well to study the definition of steel. 
 Principally it is a combination of iron and car- 
 bon, and certain other ingredients to be regarded 
 further, without any sharply defined limits as to 
 the amount of carbon alloy. On the basis of 
 composition alone the lower amounts may equal 
 those found in wrought iron, and going to the other 
 extreme it is difficult to say where steel ends and cast iron 
 or pig iron begins. The composition of wrought iron in 
 regard to carbon may range from traces to such amounts 
 as 0.07 per cent, yet the softer steels can be 
 produced with as low a carbon as 0.07 per cent, 
 but the physical condition may be quite different as 
 compared with wrought iron. As the carbon in- 
 creases until a point is reached at or about 2.5 per cent 
 the combination is then regarded as cast iron, the latter 
 being the crudest^form from which by various processes of 
 purification and decarbonization many kinds of steel are 
 produced. The process of manufacturing wrought iron 
 is also a purifying and decarbonizing one with an essential 
 difference whereby, in a sense, the former is a product of 
 
82 COMPOSITION AND PHYSICAL PROPERTIES 
 
 a "dry" method, while steel with any range of carbon is the 
 result of a "wet" one. In other words, the temperature 
 of the puddling furnace in which wrought iron is made 
 is not sufficiently high to maintain fusion or fluidity com- 
 pletely during the full interval of the progress in conver- 
 sion and purification. The initial charge is composed 
 mainly of pig iron with a total carbon of between 2% and 
 4 per cent, which readily melts and because of certain 
 active influences gradually loses its carbon but gains in 
 plasticity because of temperature limitations and the type 
 of furnace used, so that at the end of the operation there 
 will be a mass of almost carbonless iron which can be di- 
 vided into several bulky, spongy balls and worked into 
 shapes. On the other hand, steel making is the result of a 
 process conducted under temperature conditions wherein 
 fusion or liquation is maintained from the time of melting 
 cold stock until such time when decarbonization has pro- 
 ceeded to almost any desired extent, so that with a proper 
 type of furnace and necessary thermal ranges it is possible 
 to equal in composition the commoner analyses of wrought 
 iron. Thus, it will be understood that the content of car-, 
 bon in the lower ranges will not suffice to mark the 
 division between iron and steel. Another difference pres- 
 ents itself in, perhaps, a mechanical way. That is, wrought 
 iron may contain slag or cinder enveloping its fibres vary- 
 ing in quality between 0.2 per cent and 2.0 per 
 cent, and which will play an important part upon 
 its physical properties. The presence of cinder 
 is due to the sponge-like formation of plastic 
 wrought iron in its last stage of manufacture en- 
 closing slag or cinder. Steel being produced by liquation 
 permitting a separation of the slag and liquid metal by 
 gravity, the finished product will be practically free from 
 mechanically contained cinder, thus securing a greater 
 continuity or intimacy within itself and therefore more 
 strength than wrought iron containing an equal amount of 
 carbon. 
 
COMPOSITION AND PHYSICAL PROPERTIES 83 
 
 The metallurgy of steel is based principally upon the 
 influence of carbon which controls the familiar properties 
 of strength, ductility, malleability, weldability and most of 
 ^all, the property of being hardened and tempered. The 
 presence of carbon in iron is purely accidental, due to the 
 use of carbonaceous fuel as a source of heat in smelting 
 iron ores. Had the primitive iron makers stumbled upon 
 some other substance that had a similar power of separat- 
 ing iron from its earthy matters, we might today be 
 working with steel, the most important and useful of all 
 metals, founded upon another basal element. Nature, in 
 her bounteousness, has given us coal in abundance. It 
 seems the simplest fundamental fuel for smelting iron 
 ores; but owing to the affinity that exists between carbon 
 and iron for each other there results a product highly car- 
 bonized, yielding a metal limited in its capacity for 
 strength and ductility. 
 
 Carbon is recognized as existing in iron or steel in sev- 
 eral forms with a variety of subdivisions ; generally it is 
 regarded as two constitutional divisions, one as combined 
 carbon and another as free or graphitic carbon, the former 
 being a definite chemical compound of iron and carbon, 
 expressed by the symbol FesC, and the latter (graphite) 
 nearly pure carbon. It is stated by authorities that iron 
 can carry theoretically, about two per cent of combined 
 carbon and still be regarded as high carbon steel. Should 
 the carbon be carried beyond that point and if heated and 
 cooled slowly the pure or free form of carbon would appear 
 in a small degree and the alloy would then be on the bor- 
 derland of cast iron. As is well known, wrought iron and 
 the very softest steels are both malleable, ductile and weld- 
 able, but as the content of carbon increases, those proper- 
 ties gradually diminish and the skill to work them be- 
 comes of a higher order than that required in the softer or 
 milder grades. Cast iron is practically unforgable or 
 weldable merely because of the excess of carbon present 
 either in the combined or free state. Summarizing the 
 
84 COMPOSITION AND PHYSICAL PROPERTIES 
 
 general explanations, steel may be regarded as a matrix 
 of iron in which is dissolved or alloyed varying amounts 
 of combined carbon, with an absence of it in the free or 
 graphitic form; a freedom from slag or cinder, the prod- 
 uct of a process of liquation or complete fusion through- 
 out the operation of refining and conversion. 
 
 CAST STEEL is- a substance distinguished from wrought 
 iron, in that during certain stages of manufacture it is suf- 
 ficiently liquid to pour into such receptacles as metal or 
 sand molds. The composition of the product may cover 
 wide ranges. Wrought iron as already stated is worked 
 up from plastic masses of partially or completely decar- 
 bonized cast iron. Open-hearth castings are the product 
 of liquid steel with limited composition poured into molds 
 composed chiefly of silica sand. The physical properties 
 must conform to certain requirements of strength, tough- 
 ness, ductility and soundness. The attainment of such ob- 
 jects is brought about by adjustments of chemical compo- 
 sitions during the process of refining the raw material 
 and also by certain methods of heat treatment applied to 
 the finished product. Steel castings are divided into three 
 grades: "soft," "medium;" and "hard." Standard speci- 
 fications do not define the chemtcal composition of each 
 grade but give attention mainly to the physical properties 
 as cited in the preceding pages. Generally stated the fol- 
 lowing figures will cover the composition of each class of 
 castings : 
 
 Carbon. Silicon. Sulphur. Phosphorus. Manganese. 
 
 Soft 0.17-0.20 0.25-0.35 0.015-0.050 0.020-0.04 0.50-0.75 
 
 Mediumo.2O-o.3O 0.25-0.35 0.015-0.050 0.020-0.04 0.50-0.75 
 
 Hard 0.30-0.40 2 0.015-0.050 0.020-0.04 0.75-1.00 
 
 TENSILE STRENGTH is understood as the resistance of a 
 body to a stretching force steadily applied. The force 
 necessary to' produce a rupture is expressed in pounds per 
 square inch of section. 
 
UNIVERSITY ) 
 
 V n F / 
 
 ^^UFOR\^X 
 COMPOSITION AND PHYSICAL PROPERTIES 85 
 
 DUCTILITY is that property which yields to a tensile 
 stres,s and produces a permanent deformation or elonga- 
 tion with or without rupture. It is expressed in "percent- 
 age of elongation" and is measured between two pre-deter- 
 mined points. It may also be indicated by the contraction 
 of area when a specimen is measured, before and after the 
 feature, in cross section. Not only will a specimen in- 
 crease in length under a certain stress but at one point it 
 will decrease in area so that upon fracture there may be a 
 distinct "necking" and the rate of decrease in sectional 
 area is expressed in "percentage of contraction," which, 
 together with the measured elongation, can be regarded as 
 indicies of ductility. 
 
 ELASTICITY is that point where a permanent deformation 
 under a load begins or where a specimen fails to assume 
 its original length or sectional area when the load is re- 
 leased. This definition can also be applied to describe the 
 arrival of a "permanent set." The determination of elastic 
 limit is a delicate one and in engineering problems is very 
 important, representing the practical value of a casting. 
 In making calculations for service loads the elastic limit of 
 the metal entering into castings is taken as a basis, not the 
 tensile strength. For general purposes 50 per cent of the 
 ultimate or tensile strength is taken as the value of the 
 elastic limit. An increasing ratio between elastic limit and 
 tensile strength is a prime object to reach in high grade 
 products. 
 
 HARDNESS is understood as the property of a body to 
 resist the static penetration of another harder body, or it 
 may mean the ability to resist attrition caused by the 
 movement of another harder body. 
 
 Unfortunately there are not known any precise data as 
 to the influence of carbon in steel castings upon the tensile 
 strength and ductility of the metal in them. The influence 
 can only be stated in general terms in the light of practice. 
 To get exact values would entail difficult researches; it 
 would be necessary to have conditions fixed in order to 
 
86 COMPOSITION AND PHYSICLA PROPERTIES 
 
 study the influence of successive increments of the ele- 
 ment. The furnace practice, casting temperatures and 
 heat treatment would have to be constant in each case for 
 strict comparisons. The contents of silicon, sulphur, phos- 
 phorus and manganese would have to be subject to definite 
 control with only variations permissible in the carbon of 
 each specimen before reliable information could be de- 
 ducted. Experience has proven as a recognized fact that 
 carbon is both a hardening and strength giving agent. 
 Reference to the physical requirements of the three grades 
 of castings and their corresponding chemical analyses will 
 bear out the general assertion, but owing to other disturb- 
 ing influences, more or less potent, the ranges of analyses 
 are not concordant with the tensile strength and ductility 
 for each class. That is, it is indefinite that a 0.17 to 0.20 
 carbon may show about 60,000 pounds per square inch, 
 and a given elongation and contraction. As the carbon 
 increases there are gains in strength but decreases in duc- 
 tility in the remaining grades. Campbell in his admirable 
 work on the "Manufacture and Properties of Structural 
 Steel," gives formulas for the calculation of the physical 
 properties on the basis of chemical composition and in a 
 later treatise 1 revises them, giving values for several ele- 
 ments common to steel, making a distinction between acid 
 and basic practice. His figures apply to rolled material pro- 
 duced under fairly regular conditions, but it would seem 
 inconsistent to apply the same formulas to steel castings 
 manufactured under totally dissimilar methods. Kent 3 
 quotes Webster giving values of carbon in rolled stock in 
 conjunction with three other elements, sulphur, phosphorus 
 and manganese, as a constant of 800 pounds per square 
 inch with each o.oi per cent of that element with a base 
 value of 34,75 pounds per square inch for carbonless 
 
 journal British Iron and Steel Institute, No. 2, 1904, pp. 21 to 62. 
 "Mechanical Engineers' Pocketbook. 1904, p. 389. 
 3 Iron and Steel and Other Alloys, 1903, pp. 192 to 162. 
 Manufacture and Properties of Structural Steel, 1896, p. 329. 
 
COMPOSITION AND PHYSICAL PROPERTIES 87 
 
 wrought irpn. Prof. H. M. Howe 3 shows that the harden- 
 ing power increases with the carbon content and also 
 roughly plots a curve giving a direct increase in strength 
 varying with the carbon until a point is reached near 1.20 
 per cent when the strength or tenacity rapidly decreases, 
 with a further gain in carbon up to three per cent, where 
 the strength then remains fairly constant up to 4.50 per 
 cent of that element. The maximum strength at 1.20 per 
 cent being about 140,000 pounds per square inch falling 
 to about 25,000 pounds at three per cent and at 4.50 per 
 cent to about 18,000 pounds. Campbell 4 assumes a base 
 value of wrought iron as 38,000 to 39,000 pounds per 
 square inch and constructs formulas for both acid and basic 
 steel compositions with o.oi per cent carbon giving 1,210 
 pounds per square inch for each point gained in the first 
 named grade and 950 pounds per square inch for each 
 point of carbon in the latter grade, but which, as already 
 stated, apply only to rolled material. In steel castings the 
 distributing factors are so numerous that it seems impos- 
 sible to formulate precise mathematical computations cov- 
 ering the physical properties, while taking into account 
 the composition and values of each element. The furnace 
 practice, perhaps, the greatest factor entering into the 
 equation and since no two heats are handled exactly alike 
 in all particulars the problem becomes more and more ab- 
 struse. Character of stock, flame action, oreing, casting 
 temperature, recarbonizing and the methods of doing the 
 same, rate of cooling in the sand, either in dry or green 
 sand molds, all need consideration as do methods of heat 
 treatment, leading one into a maze of speculation. Gen- 
 eral terms, therefore, cannot be avoided. Certain authori- 
 ties have been cited, and the reader is at liberty to form his 
 own conclusions as to which can or may be followed in 
 attempting to give a value to carbon and its influence upon 
 open-hearth steel castings. 
 
 LICON. To the gray iron founder silicon conveys a 
 ifferent impression than to the steel founder. In one case 
 
88 COMPOSITION AND PHYSICAL PROPERTIES 
 
 it is regarded as a softening agent because of its property 
 in releasing graphitic carbon in the cooling of gray iron 
 when liquid and thus lowering the combined carbon which 
 tends to make iron castings hard. In the other case, the 
 carbon being low and entirely in the combined state with 
 no graphitic carbon liberated, the function of silicon be- 
 comes a different one. To a slight extent, about one-tenth 
 that of carbon, the effect is to harden steel castings. Pri- 
 marily the purpose of adding silicon to open-hearth steel 
 is to promote solidity, but an anomalous condition some- 
 times arises in unsoundness coupled with an amount of 
 silicon in the finished product usually considered enough 
 to produce the opposite and desired effect soundness. In 
 general practice 0.30 per cent is enough to give freedom 
 from blow holes, but if unsoundness still presents itself an- 
 other condition is operating which will be explained later. 
 To greatly exceed that figure is wasteful and would tend 
 to induce brittleness in the castings which, however, may 
 be more or less modified by heat treatment. 
 
 Silicon in steel casting practice is depended upon mainly 
 as a deoxidizer and the action may be understood by the 
 following equations: 
 
 (1) Solid Solid Solid Gaseous 
 FeO'+ C = Fe + CO 
 
 As already explained the foregoing reaction always oc- 
 curs in a bath of molten steel until some agent is introduced 
 that possesses a greater affinity for the oxygen combined 
 with iron as ferrous-oxMe.\ Silicon being available for the 
 purpose reacts as follows: 
 
 (2) Solid Solid Solid Solid 
 2FeO + Si = 2Fe + SiO 
 
 and under normal conditions stops the gaseous formation 
 assisting the production of sound castings. The SiO 
 (silica) being lighter than iron floats upwards to the sur- 
 face and becomes part of the slag. 
 
COMPOSITION AND PHYSICAL PROPERTIES 89 
 
 If a test-spoon of liquid steel be taken from the bath 
 when the action is most lively, the metal, as soon as it be- 
 gins to solidify, will emit volumes of minute sparks giving 
 evidence of some gas-forming action or release of some 
 gases in conformity with equation (i). If the specimen 
 when cold is separated by fracture, a sponge-like texture 
 will be noticed as the result of the afore-mentioned escap- 
 ing gases ; and which will suggest what might be expected 
 were such metal poured into castings without a deoxidiz- 
 ing or solidifying treatment. Just what the composition 
 of the gases are is not clearly known, but it is certain that 
 the reaction in equation (i) is largely responsible for the 
 greater quantity of them. Some authorities anticipate the 
 presence of such gases as hydrogen and nitrogen. Their 
 presence may be possible in pneumatic processes, but 
 scarcely in open-hearth steel wherein fluidity of the bath 
 is maintained by the heat of a flame action radiated through 
 a protective layer of slag and entirely away from the pos- 
 sible contamination of such gases subject to introduction 
 with the atmospheric air necessary for flame combustion 
 and which do not come in direct contact with the metal 
 below the slag. Proof is ample that perfectly sound cast- 
 ings can be made, depending solely upon deoxidizers 
 which possess no attraction for hydrogen or nitrogen; 
 therefore if they should be present they are not sensibly in- 
 dicated by porosity. Silicon is accredited with the addi- 
 tional property of increasing the power of steel to dis- 
 solve or occlude gases. The question is largely conject- 
 ural because from practical observations the evidence in 
 support of such a theory is wanting. It is difficult to con- 
 ceive of a solution of a gas in a solid without some inti- 
 mation of pores; if such pores do exist they can only at 
 best be minutely microscopic. The writer, in his experi- 
 ence, has not observed any condition attributable to the oc- 
 clusion of either free hydrogen or nitrogen in open-hearth 
 steel castings. 
 
 The question of the influence of silicon beyond solidity 
 
90 COMPOSITION AND PHYSICAL PROPERTIES 
 
 has no bearing on weldability or forgability, such proper- 
 ties not being considered in steel castings. 
 
 SULPHUR. Perhaps there is no element which is so 
 strongly stigmatized as an enemy in steel casting practice. 
 No one as yet has claimed that it is harmful in the finished 
 casting, yet specifications usually state that it shall not ex- 
 ceed 0.05 per cent, and for what reason is yet not under- 
 stood. The manufacturer may be more concerned in see- 
 ing the sulphur excessive rather than the customer. Wheth- 
 er a casting may carry 0.05 per cent or more cannot af- 
 fect its value in service. Whatever harm may follow an 
 excess of sulphur ought to manifest itself before the cast- 
 ing is stripped from the mold and thus prevent it reaching 
 the customer. Authorities state that sulphur tends to make 
 metal "red-short," a condition existing in iron and steel ex- 
 hibiting itself by the metal crumbling or cracking when 
 being worked, rolled, forged or welded at temperatures 
 suitable for such operations. The effect- is not always due 
 to the presence of sulphur, which may be masked or modi- 
 fied by other constituents. In steel castings it is said to 
 produce "red-shortness" also and which may be described 
 as property of the metal to separate or crack at points 
 where the contractive force is greatest during a period 
 when the metal is passing from a liquid to a viscous semi- 
 plastic state. The condition may or may not occur in con- 
 cordance with the amount of sulphur present. "Red- 
 shortness" may appear with a very low sulphur content or 
 it may be absent when the sulphur is considered high, so 
 that it is not possible to scan a chemical analysis and recon- 
 cile the varying degrees of "red-short" effects with them. 
 It is possible to have castings of a given design molded 
 repeatedly under as near as possible like conditions yet 
 have some "red-short" and others absolutely free from any 
 such flaw with identically the same steel in composition in 
 each case. Similar designs may be cast in acid and basic 
 steel separately. The acid casting may not be "red-short," 
 but the basic badly so, or vice versa, yet in the first instance 
 
COMPOSITION AND PHYSICAL PROPERTIES 
 
 the sulphur will be nearly 0.05 per cent and in the other 
 0.25 per cent or less. The repetition of such evidences all 
 tend to discredit the belief that sulphur is the main cause 
 of castings being "red-short" in the mold. The presence 
 of manganese exerts an effect to neutralize sulphur's "red- 
 shortness" without changing the ultimate analysis of sul- 
 phur. An excess of the last named element beyond 0.07 
 per cent in steel castings may give some trouble, but gen- 
 erally the ranges are between 0.015 to 0.05 per cent, and 
 within them it is practically inert. The common practice 
 is to carry the manganese at or about 0.75 per cent and 
 with that quantity the usual content of sulphur will exist 
 entirely as a harmless sulphide of manganese. Were the 
 manganese much lower the sulphur then might exist as an 
 iron sulphide, the latter combination being more active as 
 a "red-shortner" than the former. If "red-shortness" per- 
 sistently appears in steel castings with normal analyses an- 
 other condition is active, a remedy for which is beyond 
 any effort to change the composition to control it. The 
 conditions under which the metal is treated in the melting 
 and the design of the pattern or details of molding often 
 play a more important role in producing "red-short" cast- 
 ings than sulphur/ 
 
 PHOSPHORUS<^-This element, like carbon, is a hardening 
 agent, but unlike it, its hardening influence is not subject 
 to any great modification by heat treatment. An excess of 
 the element produces the effect of "cold-shortness," the 
 opposite of "red-shortness," a property of weakness under 
 shock or impact; a condition of brittleness. s Where cast- 
 ings are subjected to severe strains the phosphorus should 
 be kept below 0.05 per cent, particularly with the higher 
 ranges of carbon. Ordinary castings with no special re- 
 quirements can carry as high as 0.08 per cent. Phosphorus 
 can replace carbon as a hardener and when the carbon is 
 
 American Standard Specifications For Steel Castings. 
 
92 COMPOSITION AND PHYSICAL PROPERTIES 
 
 Hard Medium Soft 
 
 Tensile strength, pounds per square inch 85,000 70,000 60,000 
 
 Yield point, pounds per square inch 38,250 31,500 27,000 
 
 Elongation, per cent in two inches 15 18 22 
 
 Construction, per cent 20 25 30 
 
 high the phosphorus should be kept low because an excess 
 of two or more hardeners will produce disagreeable brittle- 
 ness, so to preserve toughness it is a good practice to keep 
 all hardeners as low as possible, depending upon one ele- 
 ment only for the necessary strengthening effect. The 
 value of phosphorus as a strengthening agent is about 900 
 pounds per square inch for each o.oi per cent. 
 
 No effect is known traceable to phosphorus upon the 
 condition of "red-shortness." 
 
 MANGANESE. This element is one of the most useful 
 in steel casting practice and in all steel making. It works 
 in conjunction with silicon as a deoxidizer and assists in 
 the removal of gases in a very similar action by forming 
 a fusible slag with the oxygen combined with the ferrous- 
 oxide^ When found in castings it is the result, usually, of 
 an addition of ferro-manganese used at the end of the re- 
 fining operation and with it will be carried in a certain 
 amount of carbon. Roughly stated five-eightieths of the 
 manganese in excess of the residual amount of that ele- 
 ment in the bath at the finishing period represents the 
 quantity of carbon furnished by the ferro-manganese addi- 
 tion. So that as the manganese increases there will also 
 be a gain in strength until a point is reached at or about 
 1.50 per cent when a disagreeable weakness and brittleness 
 result. ; If the manganese is kept about 0.75 per cent, 
 other elements being low, brittleness is absent, but if it be 
 carried above that to about i per cent it is necessary to 
 carefully anneal the castings because without that they will 
 be more or less brittle. Heat treatment, however, improves 
 them, conferring toughness, elasticity, ductility, and 
 strength, provided other conditions are normal. In the 
 higher ranges manganese has a peculiar and contradictory 
 
COMPOSITION AND PHYSICAL PROPERTIES 93 
 
 action. When a content is reached at or about 1.50 per 
 cent the metal is hard and brittle, that condition remaining 
 until about 7 per cent is reached; the metal then becomes 
 both tough and hard. Between 7 and 14 per cent a peculiar 
 alloy is obtained, known as Hadfield steel, which possesses 
 the striking property of unusual toughness and hardness 
 combined. If manganese is kept between 0.70 to 0.75 per 
 cent in finished castings good results ought to follow. An 
 excess complicates conditions, while falling below might 
 cause blow-holes, "red-shortness," and other weakness in 
 products. 
 
CHAPTER VIII 
 
 BLOW HOLES IN STEEL CASTINGS DISCUSSION OF CAUSES 
 
 A common source of annoyance in steel castings is 
 found in porosity or in unsoundness. The ingot manufac- 
 turer finds it comparatively easy to get around such draw- 
 backs by liberal discards or end cropping of ingots usually 
 more or less spongy or blow-holed at the upper end. Not 
 so with the production of castings. The problem of pour- 
 ing liquid steel into the green or dry sand molds has a 
 distinctive complexity in comparison with casting into in- 
 got molds made of iron or dry sand with simple lines or 
 shapes. 
 
 Blow holes in a steel casting may be due to a variety of 
 causes. They must not be confounded with "shrink 
 holes" or cavities due to a difference in the rate of cooling 
 of parts contrasting in dimensions. A light section in con- 
 junction with a heavier one will cool faster and in doing 
 so will draw upon the heavier and more liquid section, thus 
 depleting it of steel, with the result that in the heavier sec- 
 tion there will be some voids. Hence the important reason 
 for forming larger headers or reservoirs to hold a supply 
 of steel from which the casting below can make drafts to 
 offset the contraction in cooling. Frequently in separating 
 such headers rough irregular cavities will be found at the 
 joint of the header and the castings which are known as 
 "shrink holes," (see Fig. 11) and may be formed even in 
 cases of the casting being free from "blow holes." The 
 
CAUSES OF BLOW HOLES 
 
 95 
 
 position of cavities or pores, due to either gases or shrink- 
 age, will determine their identification. 
 
 Blow holes may be found in any part of a machined or 
 fractured casting. Shrink holes occur only at certain 
 points. They are irregular in shape, with rough surfaces, 
 but upon close examination will be found covered, with 
 delicate, minute crystals. The walls of a shrink hole are 
 usually the same color as the metal itself, provided no air 
 was present while the casting was hot. Blow holes are al- 
 ways the result of air, vapor or gas. Their shape as a rule 
 is oblong, lenticular or spherical. If oblong in shape they 
 are due to metal being imperfectly deoxidized or "killed" 
 
 FIG. 11. SHRINK HOLE. DOTTED OUTLINE SHOWS POSITION OF 
 
 RISER 
 
 and they will always be found with their axes or longest 
 dimensions at right angles to the cooling surface of the 
 body of metal (see Fig. 12). If they are globular or 
 spherical they are caused by vapor or air and not by gas 
 or gases, the result of chemical action within the metal, 
 but by damp sand, imperfect venting details of molding. 
 While it is perfectly possible (and is regularly done) to 
 get sound, solid steel machinery castings from green sand 
 
CAUSES OF BLOW HOLES 
 
 molds, accidental causes may make the sand too damp. In 
 that case there may be blow holes because of an excess of 
 steam or aqueous vapor formed by the hot steel. 
 
 ]P art in, 
 
 00000000 
 
 O o 
 
 O o 
 
 o o 
 
 00000 000 
 
 lAne 
 
 FIG. 12. BLOW HOLES CAUSED BY GASEOUS STEEL IMPERFECTLY 
 
 DEOXIDIZED. 
 
 If the green sand mold should possess the right "temper" 
 and globular, spherical holes still be found, there may be 
 two causes effective. One may be that the sand, green or 
 dry, is too close or too strong, preventing a free escape of 
 the gases as they are formed by the heat of the inflowing 
 steel upon such binders as may be used to give body to the 
 
 Parting 
 
 00 O C 
 
 o o o < 
 o o o 
 
 o o o o 
 
 Line 
 
 Drag 
 
 FIG. 13. BLOW HOLES CAUSED BY DAMP SAND, VAPORS FROM 
 CORE RENDERS IMPERFECT VENTING 
 
 sand and cores. Evidence of that may be seen in fractured 
 castings, which will have a solid drag side, but a more or 
 less porous cope side (see Fig. 13). A casting blown be- 
 
CAUSES OF BLOW HOLES 97 
 
 cause of gaseous or improperly "killed" metal will be more 
 or less spongy at all points of both drag and cope. Poro- 
 sity is often caused by failure to freely vent the highest 
 points of the cope which will cause air pockets to be 
 formed which the liquid steel may not be able to displace 
 and such conditions as described can occur in spite of dry 
 and good green sand molding or properly deoxidized steel. 
 Thus it will be understood that in spite of the highest skill 
 manifested in the melting department, it may be nullified 
 by conditions on the molding floor. A study of the daily 
 discard in the scrap pile as it is broken up will be a good 
 guide, as to what errors are being made in practice, and 
 with a slight knowledge of the causes of porosity one can 
 greatly modify faults. 
 
 If the chemical analyses are normal, that is, the figures 
 on manganese and silicon, are within usual limits, but blow 
 holes of the oblong character are persistent, the cause can 
 be traced to the furnace platform. There may be too sharp 
 a melting flame, too much air admitted for flame combus- 
 tion, or a too liberal use of ore in refining. Any one of 
 these conditions will surcharge the metal with oxide be- 
 yond the influence of the usual addition of the deoxidizers, 
 manganese or silicon. Such practice may be covered up 
 or lessened by an increase of deoxidizers, but always at 
 the sacrifice of the physical requisite in steel castings, 
 toughness. The writer knows from his experience in ad- 
 vanced steel casting practice in either green or dry sand, 
 acid or basic steel, that it is possible to produce sound cast- 
 ings free from pin holes or blow holes which will satisfy 
 the most exacting demands. Still there is no royal way to 
 the production of sound castings. It is fraught with pa- 
 tience, skill, study and industry. 
 
CHAPTER IX 
 
 DISCUSSION OF THE CAUSES OF CRACKS IN STEEL 
 CASTINGS 
 
 One of the most common sources of weakness in steel 
 castings is the liability to crack in the mold. The condi- 
 tion is the result of "red-shortness." It is an annoyance 
 and a continual point of contention between the melter and 
 the molder, each blaming the other for his share in the 
 cause, the molder claiming the steel as it leaves the furnace 
 is not just what it should be and the melter saying that the 
 metal is faultily cast in the desired forms without any con- 
 sideration as to the proper distribution of metal in the light 
 and heavy sections of a casting. From the standpoint of 
 the metallurgist the melter is at times to blame and at other 
 times the molder, or, going further, the engineer who sub- 
 mits the designs to be cast. 
 
 The condition of ''red-shortness" or cracks is far from 
 obliging and may present itself at the most unexpected 
 and inopportune times. If in a given heat of steel there 
 should be found a number of discards among various de- 
 signs because of cracks and the trouble continues for a 
 long period, it is safe to say that the metal is not receiv- 
 ing the proper treatment in the furnace. If cracks appear 
 only in one design in a given heat of steel among other and 
 different designs, the trouble is due to some fault either in 
 a molding detail or the lines of the casting. In other words 
 a few discards in keeping with an average loss of bad 
 castings cannot be blamed upon the metal. 
 
CAUSES OF CRACKS IN CASTINGS 99 
 
 As discussed in previous sections, if "red-shortness" is 
 troublesome attempts are frequently made to reach a low- 
 er sulphur analysis, but not with success. It has been 
 thought that in doing so the trouble might abate, because 
 the element sulphur is given the credit for "red-short" 
 effects and that a correction could only be found in the 
 composition. If cracks are numerous the cause is mainly 
 due to the flame character in melting and the way the re- 
 fining is conducted. Through such conditions the metal 
 becomes contaminated with an oxide of iron that acts in 
 a measure precisely the same as sulphur is said to do, and 
 what may seem strange, also, is that it is possible to deoxi- 
 dize the metal to the extent that blow-holes caused by 
 gaseous steel are practically absent, yet "red-shortness" 
 will still be causing trouble. Numerous instances have 
 been observed in practice which point to the fact that there 
 must be an indefinable form of oxide (iron) that does not 
 submit to the cleansing action of silicon and manganese 
 as final additions. In such a case as soon as the furnace 
 manipulation could be brought under control a very seri- 
 ous campaign of cracked castings was stopped without any 
 change in design of castings, methods of molding or analy- 
 sis of finished product. There could be noticed also a 
 change in the fracture of the metal in regard to the ap- 
 pearance of the crystals, another evidence which could be 
 traced back to furnace manipulation, which is in a great 
 measure responsible for "red-shortness" in castings, in- 
 dependently of the amount of sulphur initially or finally. 
 
 In regard to molding conditions and their influence upon 
 cracks the trouble often lies in several directions. Steel in 
 cooling contracts much more than gray iron cast from a 
 similar temperature. There is a point where the steel has 
 lost its fluidity and is more or less viscous, but is without 
 stability and will crumble under pressure. When the steel 
 passes to a lower temperature it seems to increase in den- 
 sity and becomes more or less malleable. At the viscous 
 point should there be any resistance offered to the metal 
 
100 CAUSES OF CRACKS IN CASTINGS 
 
 while cooling and contracting, because of improper dis- 
 tribution of metal or absence of fillets at angles, or of a 
 flask bar, hard core, core arbor or hard molding sand, there 
 is danger of a separation or a crack (a crumbling of the 
 metal) at that point where the contractive force could not 
 overcome the resistance of the obstacles mentioned. 
 Should the metal be "over-oxidized" at the range of tem- 
 perature where viscosity is manifest, the liability to crum- 
 ble is aggravated and the castings will crack with very 
 slight resistance from the causes mentioned. 
 
 It is not to be supposed that should the metal be abso- 
 lutely free from oxide it will not crack even under extreme 
 resistance of mold parts. The conditions of crumbling 
 under slight pressure at a high heat, near the melting point, 
 are peculiar to all carbon steels. Therefore, if the steel 
 were ever so pure it would crack if held rigidly while cool- 
 ing and contracting. 
 
 Let the metal be poorly handled in the furnace and many 
 cracks will appear in spite of care on the part of the mold- 
 er. If, however, due care is observed in the position of 
 the gate to allow a uniform cooling of the metal in the 
 mold, cores are made of such a mixture that under the 
 heat of the liquid metal they will crumble to dust or non- 
 resisting masses, ample sand space between flasks, bars 
 and projections on castings provided, and molding sand 
 used of such a texture that when subjected to a high tem- 
 perature it will become non-resistant, then, with good met- 
 al, cracked castings need not cause much worry by a low 
 yield of salable product. 
 
CHAPTER X 
 
 HEAT TREATMENT AND ANNEALING CONSIDERATION OF 
 THE RELATION BETWEEN STRUCTURE, HEAT TREAT- 
 MENT AND PHYSICAL CONDITION 
 
 Under heading- of heat treatment and annealing it is not 
 proposed to advance arguments as to whether or not steel 
 castings should be annealed since some opinions are held 
 that it is not necessary with certain compositions chemical- 
 ly. Rather the remarks herein will be an explanation of 
 what occurs when cast steel is given various heat treat- 
 ments. 
 
 From a theoretical standpoint all steel castings should be 
 annealed. Practically it is difficult to properly undertake 
 the operation, particularly when the tonnage may be large, 
 as then the process presents commercial considerations. If 
 specifications are rigid it is important to carefully anneal, 
 and it is at this point that the nub of the question arises. 
 To keep the output parallel with deliveries would involve 
 an extensive array of annealing furnaces, and since most 
 foundries are without capacity to anneal their entire output, 
 an attempt to treat all castings produced would result in 
 a slighting of the necessary care to get the best out of the 
 process. It is better not to anneal at all than carry it out 
 without each casting getting a proper treatment under 
 skillful conditions. Annealing is a waste of time and 
 money without this proper and skillful care in the work. 
 
 In general most metallurgists view heat treatment as 
 
IO2 HEAT TREATMENT AND ANNEALING. 
 
 primarily a method to equalize or lessen strains and 
 stresses set up in a casting during cooling in the mold, es- 
 pecially when the shape may be complicated by intricate 
 parts or light and heavy sections combined. True a re- 
 heating will tend to adjust these strains but really heat- 
 treatment is a method to procure in a casting the best pos- 
 sible conditions in the internal structure consistent with the 
 physical properties. 
 
 To understand what is involved in the more advanced 
 practice necessitates a consideration of the relation between 
 structure, heat-treatment and physical conditions. 
 
 The internal structure or crystalline formation depends 
 mainly upon the casting temperature and the rate of cool- 
 ing from that temperature. It is known that the physical 
 properties reflect in a large degree the size, shape and 
 character of the crystals formed in steel castings. If a 
 freshly fractured surface of a steel casting cooled normal- 
 ly from casting temperature be examined, to the eye the 
 grain will be coarse and large. If the same casting be re- 
 heated to a much lower temperature than that at which it 
 was cast, say a bright red, cooled and again fractured it 
 will be noticed that grains or crystals are much smaller 
 and closer than in the original piece. It will also be no- 
 ticed that after the re-heating it was more difficult to pro- 
 duce fracture than in the first instance, that the metal 
 seemed tougher. These facts give but a hint of the potent 
 changes set up. 
 
 The refinement of the crystals when subjected to varying 
 ranges of temperature and rates of cooling offer interesting 
 features which to fully appreciate, requires a delving into 
 their details by means of a microscope. The microscopic 
 examination of metals has developed a comparatively new 
 science known as "Metallography" and with it a number 
 of terms which apply to crystalline formations in metals 
 not visible to the unaided eye. 
 
 Before going into a recounting of the constituents visible 
 through the microscope and formed in cast steel, the 
 
HEAT TREATMENT AND ANNEALING 
 
 I0 3 
 
 changes produced by heat treatment upon the physical 
 properties will be considered. 
 
 If a piece of cast steel be allowed to cool freely from a 
 casting temperature there will be a grain or crystal growth 
 proceeding to a certain point when the growth ceases. The 
 metal in that condition will not possess its maximum duc- 
 tility expressed as elongation and contraction of area. It 
 will be more or less brittle, depending upon the carbon 
 content. If that same piece of steel be re-heated to or 
 about the point at which the grain growth stopped and al- 
 lowed to cool slowly, the coarse grain developed during 
 the first cooling will be greatly modified, broken up or ob- 
 literated and replaced by a finer grain than it had origi- 
 nally. In this new condition the ductility will be greatly 
 improved; the metal will be tougher and better fitted for 
 service conditions in any case than without a re-heating. 
 Such a re-heating is properly speaking "annealing." The 
 object then in annealing is to so affect the grains or crys- 
 tals as to develop the maximum degree of toughness that 
 a casting of a given composition is capable of developing. 
 
 By referring to Fig. 14, an idea will be suggested as to 
 
 FIG. 14. SHOWING STRUCTURAL CHANGES 
 
 the changes taking place in structure by heating. The ex- 
 amination was made on a piece of 0.25 carbon steel pos- 
 sessing a coarse structure common to steel of that compo- 
 sition cooled normally from casting temperature. The ar- 
 rows indicate the temperature to which the specimens were 
 
IO4 HEAT TREATMENT AND ANNEALING 
 
 heated and immediately allowed to cool. The heat was ob- 
 tained in an electrical muffle furnace and the temperature 
 measured by a Le Chatelier pyrometer. The structures 
 were noted microscopically. The range "W" refers to re- 
 calescence or refining temperature. It is that range ther- 
 mometrically at which all coarse crystallization acquired in 
 cooling from temperatures above "W" and near the melt- 
 ing point is changed and replaced with a finer structure or 
 as fine as it is possible to get in an ordinary piece of cast 
 steel. A re-heating below "W" does not accomplish any- 
 thing in cast steel. A re-heating greatly above "W" causes 
 the grain to again grow after it had been refined while 
 the piece was passing through the recalescence period. 
 
 To grain-refine a piece of cast steel it is necessary to 
 pass through and slightly above "W," and after that point 
 has been reached nothing is to be gained by a prolonged 
 or higher heating. That is to say, if the piece is heated 
 throughout at the needful refining temperature the fire 
 may be drawn or if the shape will permit it, it can be air- 
 cooled immediately. In doing so all stresses will be mini- 
 mized and no matter what the previous structure may have 
 been, the irregularities of original stress and grain will 
 be removed by heating the mass to a uniform color (slight- 
 ly above "W") and then allowing it to cool uniformly. 
 
 The equation of time for grain refinement depends large- 
 ly upon the size and shape of the piece; a wire may be 
 brought to the right heat in a few moments. Then it 
 should be withdrawn because a longer heating is needless. 
 A cube 8 inches or more in section might take 
 several hours to refine it and heat it through, 
 but as soon as uniformly heated and refined there is abso- 
 lutely nothing to be gained by heating further. Longer 
 heating would result in grain enlargement, a decrease in 
 ductility, toughness, a heavy scaling of the piece and a su- 
 perficial de-carbonization as in malleableizing if "W" is 
 greatly exceeded. To sum up the question of temperature 
 and its effect upon grain size Prof. Sauveur says: 
 
HEAT TREATMENT AND ANNEALING 
 
 105 
 
 (1) "When a piece of steel, hardened or unhardened, 
 is heated to the temperature 'W,' all previous crystalliza- 
 tion however coarse or however distorted by cold working, 
 is obliterated and replaced by the finest structure which 
 the metal is capable of assuming." 
 
 (2) "The higher the temperature above 'W from 
 which the steel is allowed to cool undisturbedly the larger 
 the grains." 
 
 (3) "The slower the cooling from a temperature above 
 'W the larger the grains." "The Metallographist," Vol. 
 2, pp. 265 and 266. 
 
 So much for temperature and corresponding grain 
 growths. The matter in the following table is selected 
 from averages obtained in researches conducted some four 
 years ago by the writer upon the effect of grain size as af- 
 fecting the physical properties of cast-steel : 
 
 Tensile 
 
 strength, Elonga- Contrac- 
 pounds, tion per tion per 
 sq. in. cent. cent. Treatment. 
 
 Series I. 
 
 80,385 13.26 16.2 Metal as cast. 
 
 78,767 27.20 40.4 Heated to "W." 
 
 79,422 14.80 15.3 Greatly above "W." 
 
 Series II. 
 
 77,779 26.5 28.5 Metal as cast. 
 
 74,504 25.0 48.8 Heated to 830 degrees Fahr., 
 
 quenched, reheated to 750 and 
 
 air cooled. 
 78,792 25.0 34.3 Heated to 815 degrees Fahr., 
 
 one hour and air cooled. 
 74,058 25.7 31. Heated 24 hours at 850 degrees 
 
 Fahr. Cooled in furnace. 
 
 Four hours in heating. Eleven 
 
 hours cooling. 
 73,376 24.2 26.7 Heated 36 hours between 850 
 
 to 900 degrees Fahr. Heating 
 
 up 3 hours 15 min. Cooling 
 
 down 9:45. 
 90,400 2.5 3. Heated to 1,200 degrees Fahr., 
 
 and quenched. 
 
io6 
 
 HEAT TREATMENT AND ANNEALING 
 
 In the foregoing Series I are the averages of a number 
 of bars from one heat of steel and Series II from another 
 heat. A study of the figures will readily show what can 
 be done in arriving at different physical properties by vary- 
 ing the heat-treatment. Reference to diagram Fig. 15, 
 which is a record of some extreme tests carried out by 
 the United States government at Watertown arsenal, will 
 
 020,000 
 
 1 2 
 
 FIG. 15. RECORD OF GOVERNMENT TESTS 
 
 still further enlighten one as to the contrasting behavior 
 physically affected by heat treatment. These experiments 
 in conjunction with others on record show, that, unless 
 the heat-treatment is a constant and other conditions nor- 
 mal in steel casting practice, it is not possible to readily 
 forecast, by means of formulas, the physical properties tak- 
 ing into account the chemical composition. 
 
 The ranges of ductility and tensile strength seemingly 
 vary with temperature and rate of cooling. The treatment 
 
HEAT TREATMENT AND ANNEALING IO7 
 
 that will give the maximum degree of elasticity combined 
 with the maximum degree of ductility is the one that should 
 be aimed at in high grade product. 
 
 The strength or elasticity depends upon the amount of 
 carbon present and the form that it assumes as a result of 
 the heat treatment it may receive. 
 
 Ductility depends upon the smallest possible degree of 
 refinement or non-crystalline formation structurally that 
 the carbon-iron alloy is capable of assuming. These re- 
 finements control the annealing or heat-treatment methods 
 and satisfactory results cannot be obtained in practice with- 
 out an observance of the laws governing them. The pro- 
 cess can best be studied with the aid of a microscope. Pho- 
 tomicrograph Fig. 1 6 is a view showing the structure of 
 0.25 carbon cast steel magnified 190 times. Fig. 17 shows 
 the same steel heated to 1,200 degrees Fahr. and cooled 
 slowly in the furnace. Fig. 18 is the same steel heated to 
 1,200 degrees Fahr. and air cooled. In Fig. 19 it was 
 heated to 825 degrees Fahr. and air cooled. All photomi- 
 crographs are magnified the same, and a study of the vari- 
 ous formations in conjunction with the physical properties 
 as tabulated will show plainly what is accomplished in cast 
 steel by heating and cooling differently. Figs. 16, 17 and 
 18 are all coarse and more or less brittle. Fig. 19 shows 
 the refinement obtained at "W" and with it will be found 
 physically a great improvement in ductility. In each case 
 the composition is precisely the same since each specimen 
 was cut from one bar of steel. 
 
 Some attention will now be given to the constituents rec- 
 ognized microscopically in steel. 
 
 First in order comes "Ferrite" which may be seen in the 
 submitted photomicrographs by the white areas. It is 
 nearly pure iron, that is, carbonless (plus Si, S, P, and 
 Mn,). It is soft, weak and ductile. 
 
 Next in importance is "Cementite," and which is not free 
 or visible separately in ordinary cast steel. It is the car- 
 bon-iron alloy and is expressed definitely as FeaC. It is 
 
IO8 HEAT TREATMENT AND ANNEALING 
 
 that constituent which confers hardness, elasticity and 
 strength upon steel. 
 
 Finally, there is "Pearlite," which is distinguished from 
 "Ferrite" by the dark areas shown in the photographs. It 
 is a mixture or combination of "Cementite" and "Ferrite" 
 in the proportion of i to 6. The constituents as named are 
 the only ones that enter in the problems of annealing cast 
 steel. (There are others, such as martensite, troosite, etc., 
 which are only found in steels that may be hardened and 
 tempered.) Cementite is not structurally free until the 
 carbon exceeds 0.9 per cent. Pearlite is recognized as ex- 
 isting in three forms. In photomicrograph Fig. 16 it is 
 called "lamellar" and is always found in steel slowly 
 cooled from a high temperature (in this case from a cast- 
 ing temperature of nearly 1,600 degrees Fahr.). Fig. 17 
 indicates also cooling but from a lower temperature. The 
 structure is still "lamellar" but not distinctly so. Fig. 18, 
 cooled at a quicker rate (air quenching) but from the same 
 temperature as Fig. 17 is called "sorbitic" pearlite. The 
 physical properties accompanying such a structure would 
 be slightly stiffer than in Fig. 17 with about the same duc- 
 tility. In Fig. 1 6 the ductility is low but with a higher 
 
 FIG. 16. SPECIMEN OF CAST FIG. 17. SAME STEEL HEATED TO 
 STEEL IN CONDITION AS 1,200 CENT. (2,192 FAHR.) 
 
 CAST (Am COOLED) AND COOLED SLOWLY IN 
 
 THE FOUNDRY 
 
HEAT TREATMENT AND ANNEALING 
 
 tensile strength than Figs. 17 and 18. Fig. 19 gives a view 
 of "granular" pearlite and is one sought when ductility 
 and resistance to shock are necessary. The tensile strength 
 with such a structure is slightly lower than those of the 
 preceding. In the last structure we have a view of the 
 marked change that has occurred by a heating to "W" when 
 the previous structure was as shown in Fig. 16 and also 
 in the other photographs what occurs in grain growth 
 when the temperature is carried far above "W." Were 
 the temperature increased to or about 1,500 degrees Fahr. 
 there would have been in Figs. 17, 18 and 19 about the 
 same formation as in Fig. 16. 
 
 FIG. 18. SAME STEEL HEATED TO 
 1,200 CENT. (1,192 FAHR.) AND 
 AIR QUENCHED 
 
 FIG. 19. SAME STEEL HEAT- 
 ED TO 800 CENT. (1,472 
 FAHR.) AND COOLED IN 
 AIR. AN IDEAL STRUCT- 
 URE FOR STEEL CASTINGS 
 
 A study of Fig. 16 reveals a reason why cast steel unan- 
 nealed is more or less brittle and snaps or fails suddenly 
 on shock or impact. The dark areas or the carbon com- 
 pound is comparatively hard, while the light or carbonless 
 areas are weak and there is an uneven distribution of the 
 strong and weak parts so that the ferrite areas offer planes 
 of cleavage under stress. 
 
IIO HEAT TREATMENT AND ANNEALING 
 
 When the structure is broken up as in Fig. 19, the 
 crystals are very small with an even intermingling of the 
 several constituents with cleavage planes practically re- 
 duced to nil. A fracture occuring in cast-steel with a 
 coarse structure always follows the ferrite areas and it is 
 not known that the line of separation passes through the 
 pearlite. A coarse microstructure usually accompanies a 
 coarse fracture while a fine microstructure will show a fine, 
 silky fracture. 
 
CHAPTER XI 
 
 REPAIR OF STEEL CASTINGS WITH THERMIT " 
 ON" OF METAL 
 
 Since the advent of chemistry in foundry practice there 
 is no metallurgical invention that has proved so useful as 
 the thermit process as perfected by Dr. Hans Goldschmidt. 
 Its flexibility and simplicity appeal strongly to the steel 
 foundrymen for supplying liquid steel in almost any quan- 
 tity quickly and at any time. The process has been too 
 well promulgated in the technical press to need any ex- 
 planation here as to its composition and action. Rather 
 the remarks herein will relate to the methods of applying 
 it in the practice of making repairs of the defects com- 
 mon to steel castings. The consideration from the foundry 
 standpoint is mainly appearance and in this regard ther- 
 mit is superior to all means, in the writer's knowledge and 
 experience, of remedying unsightly flaws which can cause 
 rejection without depreciating the strength. 
 
 Whether a casting should or should not be repaired 
 with thermit will depend upon the cost of the operation. If 
 the repair cost should exceed the molding cost, it will be 
 cheaper to re-melt it, because the casting will always have 
 a credit value on the basis of the market price of scrap 
 and the loss because of the flaw will only represent the 
 labor in molding and handling. The question of delivery 
 frequently offsets other considerations. 
 
 Many castings are of such a size that they can be readily 
 
112 REPAIR OF CASTINGS WITH THERMIT 
 
 repaired in blacksmith's fire at a nominal cost for fuel and 
 labor. Thermit is extremely useful in heavy, valuable 
 castings. The commonest flaws in them are shrink-holes, 
 under-headers, sand-holes, miss-runs and cracks. 
 
 If the defect is a shrink-hole or sand-hole and where 
 there is no machined surface, the method of filling is quite 
 simple. There may be some grease or oil in the hole 
 which must be removed before it can be welded. It is not 
 always possible to pre-heat or burn out, because the oper- 
 ation is slow and tedious, particularly in the case of large 
 castings, and there may not be a large torch or oil burner 
 convenient. The simplest way to remove the grease is to 
 pour into the hole some lose thermit, ignite it and allow 
 the heat of the reaction to burn out the oil, grease and dirt 
 which it will do most effectively. There will be a violent 
 sputtering in the hole and a mass of slag and spongy metal. 
 The slag can be chipped out readily. The hole is then 
 ready for the final treatment. A gate-pin can be secured 
 which must be larger in diameter by at least I inch than 
 the hole. It should be placed over it in an upright position, 
 and some green molding sand packed around it. With- 
 drawing the pin leaves a mold into which the thermit can 
 be placed. The depth of the mold should be at least 2^ 
 inches. When the thermit is ignited, and while burning, 
 some loose thermit should be poured upon it for the reason 
 that the mold will not hold in the first instance enough 
 thermit to completely fill the hole. As soon as the action 
 has ceased and the slag solidified the mold can be broken 
 away together with the mass of slag. The button of metal 
 will be white hot and as soon as it reaches a yellow or 
 bright red it can be forged by a hand hammer. This 
 treatment will effectively and cheaply fill any hole of the 
 character described on surfaces that will not require ma- 
 chining. The button of metal will protrude more or less, 
 but it can be ground down or chipped, preferably ground. 
 
 In case such defects as just mentioned are on surfaces 
 that may require machining, a different procedure is neces- 
 
REPAIR OF CASTINGS WITH THERMIT 
 
 sary because blow-holes must be absent from such sur- 
 faces. The same method of removing grease and dirt can 
 be followed as in the foregoing paragraph and a charcoal 
 or coke fire can be built over the flaw. Sand must be re- 
 moved by chipping to expose clean metal. The mold must 
 be dry sand. It can be made from a slab-core by cutting 
 a recess or hole in it with a file. The mold must be as large 
 in diameter inside in excess of diameter of the flaw, as in 
 the previously described method, and must be carefully 
 placed over the flaw, weighed down and the joints daubed 
 with moist clay or dough and backed by green sand. A 
 thermit crucible is then placed directly over the mold, 
 charged, ignited and tapped in the usual manner. It is not 
 recommended in this case to attempt to forge or hammer 
 the button while hot. Let the parts cool down normally. 
 Afterwards grind the button down flush. Grinding is bet- 
 ter than tooling because the button, if the proper quantity 
 of thermit has been used, will stand up some distance when 
 the casting is placed upon a planer, and if the leverage of 
 the tool is too great and there is danger of taking too heavy 
 a cut, thus tearing the soft thermit metal out. Because of 
 its softness it gathers under or crowds the tool and also 
 prevents or interferes with free cutting, increasing the re- 
 sistance and danger of tearing. A light cut can be taken 
 and when the button decreases in size, of course the lever- 
 age decreases, too. Should the metal be porous from any 
 cause, or should the weld be imperfect because of an in- 
 sufficiency of thermit in the first instance for welding, the 
 operation can be repeated since thermit metal welds beau- 
 tifully to the same metal and a second treatment usually 
 satisfactorily finishes the job. 
 
 In case of shrinkage cracks, the metal must be cut away 
 along the line of the opening to the extent of at least J4 
 inch or more. A mold of core material must be rammed 
 on the casting to get the contour of the parts, allowing a 
 mold space to form a band overlapping the edges of the 
 opening at least ^ inch on each side of it and from 24 to 
 
114 REPAIR OF CASTINGS WITH THERMIT 
 
 I inch in thickness. In welding such flaws the incoming 
 thermit metal must enter the mold at the lowest point and 
 overflow at the highest, using enough to get a good circu- 
 lation. When cool the band of metal can be removed to 
 the finished surface. 
 
 No fixed rule can be laid down as to the quantity of 
 thermit to use. Usually there are no scales in the cleaning 
 room of a steel foundry, and the operator must use his 
 own judgment how much to use for a given operation. 
 Practice alone will govern and there is always the satisfac- 
 tion of knowing that if a weld be imperfect from any cause 
 it can be repeated. 
 
 All foundrymen are familiar with the well known "burn- 
 ing-on" of metal. Thermit practice is essentially in many 
 details the same operation with the advantage that a "burn" 
 can be made at any convenient time if the compound is 
 available. The same rules will apply as in "burning-on" ex- 
 cepting that allowance must be made for a voluminous 
 slag which always accompanies an action of thermit. 
 
 The operation of "burning-on" by ordinary liquid steel 
 can be augmented by adding thermit to the ladle of steel 
 taken for the purpose. An addition of 5 per cent by weight 
 will greatly increase the temperature of the steel and of 
 course make the operation sure because of the gain in heat. 
 
CHAPTER XII 
 
 COST OF EQUIPPING FOUNDRIES FOR THE MANUFACTURE 
 OF OPEN-HEARTH STEEL CASTINGS STEEL DEPART- 
 MENT FOR GRAY IRON FOUNDRIES ESTIMATED 
 PROFITS 
 
 Owing to the widespread interest manifested in the 
 erection of steel casting plants there has been considerable 
 speculation as to the cost of their equipment. The follow- 
 ing estimate is made on ,the basis of daily output of sound 
 and salable castings, per ton of capacity: 
 
 Furnaces only $1,200 
 
 Gas producers and gas mains, or oil storage tanks and 
 
 accessories 600 
 
 Buildings, 800 square feet of floor space per ton, at 
 
 $1.25 per foot 1,000 
 
 Power, machinery, cranes, hoists, molding machines', 
 
 tools, drying ovens, etc 2,300 
 
 Total cost per ton $5,100 
 
 These figures are based on a plant having an estimated 
 capacity of 150 to 160 tons per day, or 4,000 tons monthly, 
 with an equipment of 5 open-hearth furnaces costing ap- 
 proximately $750,000. A single open-hearth furnace 
 lined for acid melting costs very nearly $1,000 per ton ca- 
 pacity. One for basic melting costs $1,200 per ton capac- 
 ity. These figures cover excavation, brick work, castings 
 
Il6 COST OF EQUIPMENT 
 
 and structural material, including stack, but do not cover 
 the platform or facilities for charging. 
 
 STEEL DEPARTMENT FOR GRAY IRON FOUNDRIES 
 
 The average capacity of a furnace for open-hearth steel 
 casting work is 20 tons. The number of furnaces in oper- 
 ation in a given shop with two or more furnaces will vary 
 according to the demand. When castings are the only 
 product, at least one furnace should be kept in reserve, 
 pending repairs or a shut-down in the active furnaces. 
 This will equalize deliveries and production. Gray iron 
 foundries using steel castings will undoubtedly find it 
 profitable to install an open-hearth furnace which should 
 be of a size that can conveniently fit in with the existing 
 equipment for handling ladles of hot metal. Would recom- 
 mend a basic lining, since it permits the purchase of cheap 
 iron and almost any kind of steel scrap. The basic process 
 has reached such a stage of development that any foundry- 
 man with intelligence can successfully acquire the skill for 
 its profitable operation. Modern foundry chemistry is fatal 
 to mysterious information. There are no secrets or esote- 
 ric systems known only to the few. The principles of basic 
 practice are an open book. 
 
 With a basic furnace in an active gray iron foundry of 
 anything like a modern character it would be perfectly 
 feasible to manufacture steel castings in moderate, profit- 
 able quantities in conjunction with the regular product. 
 About the only change necessary in the molding end of the 
 practice would be a supply of silica sand and a sand mill 
 to prepare the molding mixtures. A furnace of convenient 
 size would have a capacity of 5 tons per heat. Such a fur- 
 nace could easily produce four heats in 24 hours and could 
 be depended upon to regularly make at least three, and if 
 desired, only one heat per day. In the meantime, however, 
 there would be a steady consumption of fuel to keep the 
 furnace hot. The fuel should be either natural gas or fuel 
 
COST OF EQUIPMENT 1 17 
 
 oil. Producer gas is often uncertain, irregular in com- 
 position, and requires additional labor for attendance and 
 maintenance. 
 
 COST OF A FIVE-TON FURNACE 
 
 A 5-ton basic open-hearth furnace would cost approxi- 
 mately $6,000 if erected in a modern active iron foundry. 
 As an additional outlay to cover bottom pour ladles, fur- 
 nace platform, oil storage tanks, pumps, oil piping, burn- 
 ers, etc., a liberal figure would be $4,000, making a total 
 expenditure of $10,000, not including, however, the build- 
 ings. 
 
 Charging 5 tons per heat, consisting of 50 per cent pig 
 iron and 50 per cent scrap, there should be produced under 
 normal practice, 12^ tons daily at the rate of three heats 
 in 24 hours. The loss or shrinkage is estimated at 15 per 
 cent which includes the melting loss, gates, risers, sculls 
 and defective castings. The melting loss in the furnace is, 
 on the average, 7 per cent. This low loss is an important 
 economical factor. 
 
 PROFITABLE INVESTMENT 
 
 With a normal demand for castings an average profit of 
 YT. per cent per pound could be expected, or $10 per net 
 ton of product, which would be equivalent to $127.50 per 
 day. There would be times when the furnace would be out 
 of commission for repairs, periods which should not exceed 
 one month at the most, but under ordinary conditions two 
 weeks should cover general overhauling. With proper 
 care a basic furnace is capable of producing 400 heats be- 
 fore undergoing general repairs, or a campaign of 133 
 working days, in round numbers four months. Assuming 
 the active period for production would only be nine months 
 in the year at the rate of 25 working days per 
 
Il8 COST OF EQUIPMENT 
 
 month, there would be a productive period of 225 
 working days. At the rate of 12^4 tons per 
 day or a yearly total of 2,868^4 tons at $10 per 
 ton, would yield a profit of $28,687.50, figures which 
 look attractive from a promoter's view point, and might be 
 vivid to embody in a prospectus. But, there would be 
 times when the yield would shrink considerably, owing to 
 errors in practice, break-downs, delays and other detrac- 
 tive conditions, which would seriously decrease the dif- 
 ference between manufacturing costs and selling prices. 
 As an extreme case, it will be assumed that the yield over 
 metal charged was only 50 per cent good castings for the 
 entire productive period as estimated; that the average 
 profit was decreased to 3^ cent per pound and that the ton- 
 nage was only 7^ tons daily or 1,687^ tons yearly. The 
 profit then would be $4,218.75, and charging off 25 per 
 cent for interest, replacements, depreciation, etc., there 
 would be a net return of $3,164.07 on an investment of 
 $10,000 for a five- ton furnace in an active iron foundry. 
 This would be equivalent to a profit of 31 6-10 per cent on 
 the outlay, which, in view of the extremely unfavorable 
 conditions considered in the estimating, makes a steel foun- 
 dry as profitable as any foundry enterprise. 
 
 An open-hearth steel foundry, with intelligent practice 
 and normal times or demand, offers an attractive venture 
 to the investor and no doubt will receive the attention of 
 gray iron founders who are interested in steel casting 
 manufacture. 
 
 UNIVERSITY 
 
 OF 
 
. 
 
 Ta re BOOR 
 
 ^ 
 
 MAR 2 1933 
 
 2 9 J948 
 
 2Jun$3KH 
 
 ,6Ja'59ES| 
 
 ct> 
 
 JAU 21959