These notes make no pretense as an essay on steel. Most of their contents is already well known to experts. The writer's object is to lay before the service the facts bearing most directly on Naval Inspection. Technicalities have been avoided as far as compatible with understanding. A perusal of the clearly conceived and well worded specifications for the Maine and Texas will prove of assistance. It is essential to the purport to dwell at some length on the chemical constituents of steel. The importance of a chemical analysis cannot be overestimated.
STEEL and its CONSTITUENTS.
"Steel is an alloy, fusible and weldable."
"Steel is a mechanical mixture of carbon and other substances with pure iron."
"Steel is a solidified solution of carbon with pure iron."
The specifications call for mild steel. Generally, soft steels carry the carbon below .08 of one per cent; mild steels from .08 to .20; hard steels from .20 to .40; and extra hard, or spring and tool steels from .40 to .80 carbon, and even above.
High class steel should contain only iron and carbon. Every desirable quality— strength, ductility, hardness, toughness, and elasticity—can be secured by a proper proportion of carbon. All other elements entering into the composition of steel, whether incident to the ore, or added to secure certain desiderata, are impurities. The chief of the metalloids which thus enter are manganese, phosphorus, sulphur, silicon, and in minute proportions arsenic, copper, chromium, cobalt, and antimony. It is safe to say of them all that every useful quality gained by their presence is always counterbalanced by the loss of some valuable characteristic of the pure steel. It must not be understood that we are yet at a point where the steel-maker can totally neglect the empirical formulas by which he now obtains the desired results. But the purer the mixture of iron and carbon, and the greater reliance placed on intelligent mechanical and heat work, to obtain quality, the better the metal and the less possible, annoying and unforeseen failures. Every heating of a piece of steel leaves its record in the metal, and when various metalloids enter, the problem of so using heat and work as to benefit all and injure none, becomes extremely complex.
THE CHEMICAL ANALYSIS.
Carbon, the one essential alloy, varies closely with the strength desired. Carbon of .04 per cent gives a tensile strength of about 50,000 pounds per square inch. Carbon of .90 per cent gives say 150,000 pounds. In ship plates, angles, beams, etc., carbon runs from .10 per cent to .20 per cent. This ratio of carbon, with proper reductions in the rolls, will give T.S. 60,000 to 70,000 pounds, elongation 25 to 30 per cent in eight inches. No limit is placed upon carbon, and no uneasiness need be felt with carbon below 0.33 per cent. The required carbon point is obtained by re-carburization. The original carbon in the pig and ore being oxidized in the furnace, ferro-manganese containing 7 per cent carbon, 80 per cent manganese, and 13 per cent ferro, is generally used for re-carburizing, and is added to the heat during the run into the ladle.
Manganese, sometimes called treacherous, is largely used to exclude or neutralize more injurious elements. By thoroughly saturating the heat, manganese counteracts red-shortness, or brittleness in rolling. By uniting with the oxide of iron, manganese reduces it, becoming oxide of manganese in the floating slag. From 40 per cent to 75 per cent of the manganese disappears in this way. The presence of protoxide of iron is one of the worst causes of red-shortness, and the manganese acts as a "medicine" or purge of this impurity. High manganese produces high tensile. In ship-plate, manganese varies from 0.22 per cent to 0.60 per cent as a maximum. Very high manganese, say 8 to 12 per cent, produces an extremely hard and stiff steel.
Phosphorus is undoubtedly the most dangerous, insidious and undesirable of all the metalloids. The low cost of phosphoretic pig presents a constant temptation to the steel-maker. Apart from this, the economic effects of phosphorus are: Better and quicker melt in the furnace; more liquid pour of the heat at a lower temperature; higher elastic limit and tensile strength in the steel. The quicker melt means more time for repairs and less fuel; the liquidity at lower temperature secures less sponginess and fewer blow-holes in the ingot; higher tensile implies less mechanical work and fewer reductions. Each and all of these bear directly upon the cost.
The evils of high phosphorus are: Unequal distribution through the ingot, with consequent segregation and lack of homogeneity in the plate or shape; cold-shortness, or brittleness when bending the steel cold; and insidious weakening of parts subject to reheatings, or great vibrations. All of these defects are so well known that phosphorus is limited, or conditioned, in all material for the Maine and the Texas. In ship-plates, shapes and hull rivets, no steels carrying above 0.06 per cent phosphorus are received. For boiler plates and rivets, stay-bolts, etc., the condition permits nothing above 0.03 per cent. As both these are maxima, the maker must leave a safe margin, or run the risk of rejections. To illustrate this, the hull material for the Maine averages 0.048 per cent phosphorus, though the condition exacts only 0.060 per cent. The great advance in quality of plates and shapes in the United States may be appreciated when it is recalled that less than one year ago ship-plate averaged 0.08 per cent with a maximum of 0.10 per cent phosphorus, and shapes in excess of the plate maximum.
High phosphorus produces high elastic limit. As loads are graduated to this, rather than to ultimate strength, the steel is accepted above its true value. That is, a high phosphorus steel, showing great elastic limit, will, when exposed to shocks, vibrations, or changes of temperature, deteriorate rapidly in use. Bridge builders now condition phosphorus to 0.05 per cent in tension and compression members. In gun metals a still lower condition should obtain. The fallacy of carrying high phosphorus with low silicon will find few adherents. In rails the percentage of phosphorus is o.io. This seems requisite to give a hard wearing surface and prevent scalping. Comparatively few rails break, but the ideal rail is one with a hard bulb, and tough web and flange.
I have dwelt particularly upon the dangerous characteristics of phosphorus, as there are not wanting those who assert that phosphorus up to 0.07 per cent is beneficial. One year ago these same "self-luminants" placed the limit at 0.10 per cent. I have been able to trace many "mysterious" breakages home to phosphorus, and unless the history of the heat is known, always suspect it. The only good word I can find to say for phosphorus is, that its metals are more adversely affected by cold rolling, and give unmistakable evidences of this serious defect—somewhat on the principle of administering poison to produce a rash.
Silicon, as an alloy, is useful in securing more solid castings, preventing blow-holes and sponginess. Its useful qualities beyond these are doubtful. A high percentage is now frequently found in tool steels without injury to the edge. In ship-plate, the silicon of the pig is usually consumed in the furnace, and the remaining percent is very low, say not exceeding 0.04 per cent.
Sulphur causes red-shortness or brittleness in hot working. This is so obvious and fatal a defect that it cannot escape notice. Sulphur rarely exceeds 0.03 per cent.
Arsenic, in large proportion, causes cold-shortness and great brittleness. Copper causes red-shortness. Antimony combines the evils of both arsenic and copper. Silver and cobalt are rarely present beyond a trace.
CHARGING AND MELTING.
Apart from the chemical knowledge of the heat, an inspector should have a good general knowledge of the composition of the furnace charges. Variations in the pig and mill irons, blooms, crop ends, scrap and ores that all go to make up the heat, are surely recorded in the resulting steels. Any sudden change in the character of the plate may be referred to the charge. To illustrate: a series of surface defects, snakes, pittings, and laminations was traced step by step from the roll train, through the rolls, heating furnaces, casting and melting, till the cause was finally located in the too great proportion used of acid-washed metal. The proportion of this was reduced from 50 per cent to 5 per cent and the defects ceased. The chemical analysis showed the same before and after. The inspector has no authority over what shall be charged, but the books are open to him, and should he suspect evil in any of the charges, the specifications provide for extra tests to assure himself that the steel is uniform. These additional tests will either justify his suspicions and reject the heat, or dissipate his distrust.
Ordinary open-hearth furnaces run two heats each 24 hours. The melt is usually tapped into a ladle previously heated to prevent chilling, and thence teemed into ingot molds. The ingots vary in weight from 500 pounds to 30,000 pounds, and the cross-sections from 4 X 6 inches to 36 X 48 inches. Ingots are either top cast, by pouring directly into the mold, or bottom cast by stand and runners. Each method has its advocates. For small ingots the bottom cast is best. Whilst the heat is teeming, and near the middle time, a test ladle is taken for chemical analysis. Should the inspector desire, test ladles will be taken at the beginning and end also. Wide variations in these tests show poor mixtures and segregation. Specimens pulled from such a heat represent only the ingots from which they are taken. In such a case, test each ingot independently.
From the ingots thus cast the inspector should select the four poorest for test plates. These ingots are charged, when cold, into the heating furnace. When raised to the proper working heat, as determined by the heater's eye, the color varying from light yellow to white, according to the carbon, the ingot is hauled from the furnace and placed upon the rolls. The ingot is usually coated with a loosely adhering scale, that cracks and falls away at the first pass through the rolls. Each roll-pass accomplishes a reduction in the vertical section, large at first, and decreasing in draught as the temperature decreases. Four passes are made longitudinally, then the ingot is turned and rolled transversely, then "cornered" to bring out square-shouldered plates; after which it is again placed longitudinally and finished to the required thickness, the transverse passes giving the proper width. These remarks apply only to plates, as all shapes are rolled entirely longitudinally. Mill scale is swept off by brooms, or, if very obstinate, is blown off by salt liberally sprinkled on the glowing surface of the plate. As the salt passes through the rolls the explosions dislodge the objectionable scale. Plates are very free from scale defects on the lower side; scale defects cause ridges and produce rough, corrugated faces.
Very much depends upon the heater's skill. A poor heater will burn or overheat the ingots. A burnt ingot has not only lost a per cent of steel oxidized, it has also changed the size and form of its crystals. These enlarged crystals cannot be reduced in a rolling mill, and will cause weak sections. Overheated and bleeding ingots produce pits and laminations that no subsequent rolling can eradicate.
Cold rolling comes either from insufficient heating or too rapid cooling below a working heat. Look for this defect in thin wide plates. The injury to the plate or shape is due to the draught of the rolls, causing a cold flow of metal, entailing injurious strains and stresses, and reducing elongation.
SURFACE INSPECTION.
An inspector provided with a long-handled light hammer, pointed at one end, should be always at the rolls. Most surfaces are covered in a short time with red or black oxides that effectually conceal serious defects. True, these defects come to light in pickling, but it must be humiliating to any inspector to depend too much on this. Therefore, inspect plates and shapes as hot as possible. The principal surface defects are laminations, hair cracks, scale marks, scabs or blisters, pits, snakes and cobbles. Probably 75 per cent of surface defects are due to pitting. Pits are conical cavities extending towards the centre. If their depths are at all considerable, they are fatal to acceptance. Their cause has been noted. Should cinder or pieces of fire-brick be rolled in, a spotted appearance will indicate their presence. A sharp tap of the hammer will dislodge them and show the extent of the injury. Bits of slag are discovered by fine hair lines marking their areas. The hammer discloses their size and shape. Scale cools more rapidly than the plate itself and produces hummocks. Cracks are found in the direction of the rolling, and give indications of blow-holes long drawn out. To ascertain their injury, test transverse specimens. Snakes, on the contrary, as their name implies, are twisted in every direction. It is accepted that they are caused by some foreign substance, such as peroxides or protoxides of iron formed in the heat, separating two masses of pure steel in the ingot. No amount of work can cause a true union or weld across this flimsy layer. Once in the ingot they reappear in the plate. So true is this that an ingot known to be snaked is cast aside for scrapping. A snake is fatal, and no material for any purposes should be accepted where snakes are evident. The application of a fine file will soon settle doubtful cases. In general, high carbons produce more snakes, but poor melting is the prime cause. It is a noteworthy fact that a record of two hundred and fifty basic heats, of thirty tons each, did not produce a single snaked plate.
Laminations occur both on the surface and at the edges. Surface laminations may usually be traced to chipping the ingot. The cavity thus formed is covered by overlapping edges. A few blows of the hammer will cause these overlaps to separate into sheets. If proper allowance is made in width for shearing, a properly rolled ingot will never laminate. All plates should be inspected after shearing for laminations. Cobbles are produced by unequal heating of the sides of an ingot. Under the same draught the hotter side creeps faster, producing ridges. This defect cannot be remedied. It will be recognized by the diagonal trend of its ridge or crest. Simple slight waves in a plate are not serious.
SHAPES.
Shapes, such as deck-beams, angle, T and Z bars are usually free from pits, scales and snakes. The blooms or billets from which they are rolled have small sections, and permit more perfect heating and better cleaning in the rolls. Owing to the small section, sand is frequently used to prevent slipping. Look for sand scabs, laminations caused by overlaps, grooves and wire edges. Many beams are rolled with web eccentric to the angle. Reject these. Measure carefully the width of angle, as beams run scant near the ends. The same rule holds good as to angle bars. Watch for shady backs. Reject all shapes which are not full to true section. In bulb beams especially the angles are frequently cold drawn by the more rapid travel of the metal in the web, due to its receiving a great deal of work after the flange is fairly formed. The billet is symmetrical, but the finished beam, unlike I beams, contains 25 per cent more metal below the centre of the web than above. The injurious effect of the cold flow causes wire edges, and sets strains, that in many cases fairly equal the strength of the metal. A beam thus rolled might easily shiver to pieces on being thrown from a car. Proper annealing would restore the crystals to an amorphous condition and relieve the strains. There would be an apparent loss of tensile strength, but the actual working strength would be increased. In angles of unequal sections and Z bars the above strains obtain, but to a less extent. The Navy demands peculiar sections. The manufacturer is loath to turn new rolls at great expense; he therefore resorts to old rolls approximating the shape. The steel thus shaped is badly rolled. More passes and smaller reductions cause cold rolling. It sometimes happens that unusually large sections stall the engines. For all these reasons, the inspector should see the more important shapes rolled. Test specimens for shapes are taken from cropped ends, which are poorer than the rest of the shape, but an inspector will do well to cut his tensile tests, and in all cases his bending tests from the angle, and not the web.
TEST SPECIMENS.
The plate or shape having passed surface inspection, the specimens are stamped with heat number, ingot number, and inspector's initials. These specimens should come from near the edges of plate or shape. Four tests come from each of the four test plates. This allows duplicates in case of faults. Ingots are tapered to facilitate stripping. Hence the bottom of the ingot has a greater cross-section than any other portion. Moreover, the steel is more compressed and solid at the bottom than at the top. Both these causes contribute to give a better steel at the bottom end of a plate than at the top. To obtain a specimen which shall show, not the average condition, but the poorest part, cut from the top part. Let me emphasize this; the inspector should always bear in mind that his search is for the worst features of the finished steel. The strength of a plate is the strength of the poorest cubic inch it contains. There is no average strength. The value of a plate is the value of its weakest section. "The weakest part must bear the strain."
As showing how tensility and elongation vary in the same plate, I cite a remarkably well rolled plate at the Homestead Steel Works, which showed T. S. rising in eighty inches from 49,000 pounds at the top end to 51,800 pounds at the bottom. The elongation fell in the same distance from 32.71 per cent, top end, to 29.8 per cent at the bottom end. The reduction of area fell from 66 per cent, top end, to 60 per cent bottom end. These results, in each case, were the mean of six tests taken side by side. In all thirty tests were pulled from this one plate. The plate gauged I of an inch, and was rolled from an ingot 10 X 18 inches. The taper gave 27 reductions for bottom end, and 26 for top end. The true value of this plate was T. S. 49,000, elong. 29 per cent, red. 59.6 per cent, or the value of the poorest test specimen.
In inspecting plates rolled from slabs, from large ingots, i.e. ingots weighing from 8000 pounds to 30,000 pounds, cut the specimens from the upper end of the slab, not only for the reasons above, but also for the graver one, that the segregation of the metalloids increases with great increase in size. The slower an ingot cools, the greater the segregation. Large ingots presumably cool slowest. Remember, however, that segregation is largely a question of temperature in pouring. The lower the temperature of the pour, the less the segregation. Large ingots can be cast very rapidly owing to being top poured from a large orifice in the ladle, consequently the temperature in the ladle is much reduced without danger of chilling. The ingot which developed the most segregation, of any I have seen analyzed, was a pigmy, 10 X 18 X 30 inches. Surprising variations occur, notably in phosphorus and carbon. The tensile strength, elongation and elastic limit in the same ingot vary widely. As an illustration I give here two tests from opposite ends of the same plate rolled from a slab:
Top end T. S. 64,500 lbs. Elong. 19 per cent. Red. 31 per cent.
Bottom end T. S. 55,400 lbs. Elong. 26 per cent. Red. 48 per cent
Evidently the value of this plate is
T. S., 55,400 lbs. Elong. 19 per cent. Red. 31 per cent,
a poor result for ship or boiler plate.
The chemical analyses of these specimens disclosed:
Top end, Car. 0.31 Phos. 0.075
Bottom end, Car. 0.17 Phos. 0.050
Heat test. Car. 0.15 Phos. 0.044
The bottom cold bending piece closed on itself without a crack, whilst the top specimen broke short off at 170'', with a fracture indicating high carbon, with large weak crystals.
For test specimens for ship and boiler plates, and shapes, the piece is cut 16 inches long, with parallel planed sides, and a cross-section from 0.5 inch to 0.8 inch. The witness marks are eight inches apart. For ordnance and protective deck plate a filleted round is prepared. The ordnance adheres to two inches between witness mark, which secures an elongation about 30 per cent above that of the eight inches. Rounds generally give better results than flats.
Specimens are measured in three places to give mean section, these measurements being made with micrometer gauges which register the thousandth of an inch. Twelve one-inch witness marks should be punched on the edge. The specimen is then placed in the grips and the initial stress applied. Additional loads are added at short intervals, the beam being kept in equilibrium. Elastic limit is marked by unsteadiness of the beam ending in a sudden drop. The cracking of the mill scale is also a good, but not infallible, indication. From this time the loads are added until the ultimate strength is reached, as shown by the beam refusing to rise. These last loads should be added very gradually. From this on, the specimen stretches and necks until fracture occurs. The fractured ends are carefully fitted and the specimen measured as before. From this second measurement is obtained elongation and reduction of area. Examine carefully the nature of the fracture. Cup fractures, with fine crystallization and uniform gray color, indicate homogeneous, well rolled, metal. Sliding fractures occur frequently, but are not as favorable. Irregular, jagged fractures are bad. Mottled colors, parti-colored streaks, bright spots and dark patches all indicate poor steel, and should arouse suspicion however high the ultimate strength or the elongation. Coarse crystallization betokens insufficient work, high phosphorus, or burned metal. Bright patches and streaks denote segregation and cold rolling. Dark, irregular blotches evidence overheating.
Heat the quenching specimen to a dark cherry red in a fire free from smoke, or better, in a small gas furnace. When of a uniform color, plunge into water warmed to 82°, then bend cold till the centre doubles around a curve three halves the thickness of the piece in diameter. Quenching properly performed will detect cold rolling, brittleness and segregation due to phosphorus. A well rolled plate will bend double after quenching without cracks.
Cold bending requires the piece to close on itself without cracking. Very few specimens fail under this test. Thin plates will fold twice without fracture. Rivet steels are required to be cold flattened and hot flattened under a hammer. One specimen is additionally bent like a hook. With low tensile and low phosphorus, any failure in these easy conditions should be fatal.
TESTS FOR SHAPES.
The tensile and elongation requirements are the same as for plates, also quenching and cold bending. In view of the additional tests, this latter might safely be omitted. The additional tests are cold opening, cold closing, and drop test. I have already called attention to the unwisdom of accepting tensile tests from the web. These are uniformly superior to those from the angle. The opening and closing test of beams is one operation. Closing one angle opens the other. Much trouble was experienced in this test, of the IX-inch battery bulb beams, for some of the cruisers. Most of the beams broke short off in the web, close to the fillet, whilst a few cracked straight across the flange. These beams were rolled from unconditioned Bessemer steel, and the fractures disclosed large, weak, fiery, bright crystals. Over two hundred beams were rolled to supply one lot of thirty-eight. In the Maine and Texas, the phosphorus must not exceed 0.06 per cent, and the material used must be ordinary open hearth. Angles must both open and close. Of these, few fail in opening, the closing test being the more severe. Failures occur at the fillet near the jaw or apex. The present drop, or shocking, test affords an excellent criterion as to the amount of brittleness. A 5-inch X 3-inch X 9-pound reverse bar O. H., for the Maine, stood thirty-three blows from a 640-pound weight dropped five feet. The bar was inverted after each blow. The deflections up and down varied from five to seventeen inches. The bar showed fatigue at the twenty-fifth blow, and fine hair cracks developed in the heel at the thirtieth blow. Rupture occurred at the thirty-fourth blow, by the steel tearing, not splitting or cracking, half across the narrow angle. Three more blows were necessary to carry the fracture across the heel of the angle. The bar was by this time so twisted and flattened as to have lost all semblance to its original section. Similar experiments, with punched bars, showed that the fractures rarely extended into the holes, though grazing their edges. Ships, framed with such material, will stand a deal of bumping, battering and ramming before breaking.
BOILER PLATES.
The remarks, as to ship plate, apply equally to boiler plates. High boiler or shell plates are required, in the Maine and Texas, to have an elongation of 25 per cent in eight inches. This is a march, not a step, in advance. The previous specifications called for only twenty per cent. Owing to the great size of the ingot, and greater difficulty in proper heating, there are more pits and laminations. The remedy for this lies in narrower plates and more of them. The shell plate for Cruiser No. 5 boilers are 1 3/32 thick, and weigh, finished, 4900 pounds each. More failures occur in shell plate than all other material combined. Flange plates (low boiler) are required to have an elongation of 29 per cent, and rarely fail. Boiler plates are planed, not sheared, to true dimensions. Be especially careful as to laminations in boiler material.
The desire to secure increased tensile strength in boiler plates is to be deprecated. No high steels are as dependable as mild steels, and any gain in pressure in this way is dangerous.
CONCLUSIONS.
The specifications are as rigidly drawn as the Articles of War or Navy regulations. To the precisian, they mean reject on any failure, or accept on any passing. To others, many doubtful cases will arise. The fault or success may be a quality of the specimen, and not of the piece itself. A careful inspector will want all the data obtainable, before rejecting or accepting. Attention is called to the wonderful persistence of like chemical constituents producing like physical qualities. If space permitted I could attest this statement with records of five hundred heats. Look first to the analysis, and after to physical errors for defects.
All structural steels for the Maine and Texas are ordinary open hearth. Boiler plates also open hearth. Rivets, either Clapp-Griffith (a modification of Bessemer converter) or open hearth. The steel castings for the Maine will be Bessemer; those of the Texas open hearth.
The significance of the specification, "ordinary open hearth," will not escape notice. This clause shuts out "basic open hearth," now being extensively used. The Steel Board has preferred to hold fast to that which is proven; but the trade is not so conservative. Basic material is fast making its way into every channel where steel is used. The Baltimore & Ohio and Pennsylvania railroads use it extensively for fire-box steel for locomotives. The large Globe Boiler Works have adopted it. It has been my fortune to witness over two hundred basic heats. I have seen the charging, the melting, casting, reheating, slabbing, rolling, testing, and bending. The ingots seem more porous than those from ordinary open hearth, but the holes weld in rolling, and the plates are homogeneous, and remarkably free from snakes, pits, and laminations. Physical tests compare favorably with ordinary open hearth. All of these statements refer to basic steel below 0.20 per cent carbon. Owing to the greater purity of the heat, basic steel requires at least two points more of carbon to produce tensile equal to acid heats.
Time was when crucible steel alone was considered reliable; open hearth is now a standard. Bessemer is still struggling to establish its reputation as a dependable metal. Its bastard, Clapp-Griffith, has come into Navy circles ahead of its progenitor, on rivets. Basic steel is still shut out, but its process is so rational and its product so reliable, that it will not be long ere it comes to the front.
There can be no shadow of doubt that the Navy is now obtaining for its latest additions, a material superior, in every good quality, to any other ever used in any ship. I make no exception whatever. It is a subject of congratulation, that from the Advisory Board of 1883, to the present day, the Navy has taxed the resources of the steel-makers to produce a quality of metal superior to their best. The requirements have been severe, the inspection rigid; but it is gratifying to note how the steel has successfully advanced to meet both. The specifications for Cruisers 1 to 5 and the four gunboats were denounced as impracticable and absurd; but now the much more severe specifications, for the Maine and the Texas, are accepted and carried out with thoroughness and cheerfulness. All attempts to set aside or reduce the qualifications have been firmly resisted by the Secretary of the Navy, and the Steel Board is laying the foundations for ships which will float or fight, with the best material on any ocean or under any flag.