Chapter III: Machining and Heat-Treatment of Shrapnel Shells

Shrapnel shells are manufactured either from bar stock or forgings. The bar-stock method, however, is not considered as satisfactory as forging because of piping, so that the greater number of shrapnel shells made at the present time are turned out from forgings. The first step, therefore, in the making of a shrapnel shell is to cut off a billet of the required length from a bar of steel of the necessary constituents. In the making of an 18-pound shrapnel shell, the billet is cut off from a bar of 46-point carbon, 60-point manganese steel in machines of different types. One way of doing this, as shown in Fig. 2, is to use a Newton cutting-off machine having an air clamp for holding the bar in place while it is being cut off. A Hunter duplex saw, as shown in the illustration, provided with high-speed steel inserted teeth, performs the cutting operation. The billet for an 18-pound shrapnel shell is 3¹⁄₂ inches in diameter by 4¹⁄₂ inches long. It is then forged to shape, as has been previously explained.

Assuming that the forging has been completed, the following is a complete summary of the machining operations on the shell up to the point of assembling. In one plant where this work is being done, the shrapnel shells are put through in lots of 120, each lot being kept in three boxes, forty shells to a box. Out of every 120, one shell after heat-treatment is tested for tensile strength. The tensile strength before heat-treatment must be from 30,000 to 40,000 pounds per square inch, and from 80,000 to 90,000 pounds per square inch after heat-treatment. For facilitating transportation, trucks of various designs are used. One type of truck used for this purpose is shown in Fig. 3. This is built by the Chapman Double Ball Bearing Co. of Canada, Ltd., Toronto, Ontario, and has some interesting features, the chief of which are the ball-bearing swiveling head, ball-bearing wheels, and the means of releasing or raising the load with the handle in any position. This feature is valuable in using the truck in a crowded space.

=Trimming and Facing the Shell Forging.=--The first machining operation on the forged shell is to cut off the ragged end, which is generally from ¹⁄₂ to 1¹⁄₂ inch longer than that required for the finished shell. This operation is performed in many different ways, but one of the most common is to place it in a Hurlbut-Rogers cutting-off machine as shown in Fig. 4. For performing the cutting-off operation, two plain forged cutting-off tools made from “Sabine” extra high-speed steel are used. The forging is located in the proper position in the chuck by a plunger or stop _A_, sliding in a fixture _B_ clamped to the base of the machine. This plunger locates the shell from the bottom of the hole or powder pocket and forces the shell into the chuck against the resistance of an open-wound spring. The stop is then located by a gage _C_ that forms a member of the fixture and fitting ring _D_ on the stop. The chuck jaws are now clamped on the work and the cutting off commences. As soon as the excess stock is cut off, the stop is drawn back and the pressure of the jaws on the work released; the spring in the chuck then ejects the forging. The production of an 18-pound shell from one machine is about 140 in eight hours.

The next roughing operation is to face off the bottom or closed end of the forging, bringing the shell to approximately the correct length. There are also many ways of performing this operation. One method is to grip the forging in a chuck, as shown in Fig. 5, in an ordinary lathe and face off the end with a high-speed steel tool held in an Armstrong tool-holder. From to ¹⁄₄ to ³⁄₈ inch is faced off from the end.

=Rough-turning Operations on Shrapnel Forging.=--Practically every type of engine lathe and turret lathe as well as special machines are used for turning and boring shrapnel forgings, and in the following chapter each method will be dealt with separately. Before doing this, however, a complete summary of the methods of machining employed in a large plant turning out shrapnel will be described. In this plant, the first rough-turning operation is handled on a flat turret lathe, as shown in Fig. 6. For this purpose, the shell forging is held on an expanding arbor and is driven by a dog fastened to it and driven by the faceplate of the lathe. A multiple tool turner is first brought into position and takes a cut of about ¹⁄₈ inch from the diameter for practically the entire length of the shell. The next tool then faces off the end of the shell to length.

The shell forging is now ready for cutting the rifling band groove and producing the waves. This is handled in an ordinary engine lathe equipped with a special fixture, carrying grooving, waving and under-cutting tools. The shell forging, as shown in Fig. 7, is held in a chuck at one end and supported by a revolving center at the other. One part of the fixture is clamped to the bed of the lathe and the other to the carriage. The grooving and ribbing is accomplished with a tool held in holder _A_ at the front of the lathe, whereas the two under-cutting tools are held in holders _D_ and _E_ at the rear of the lathe. In operation the carriage of the lathe is moved toward the chuck, carrying the fixture to which are fastened cams _C_, _F_, and _G_. Cam _C_ forces in the holder carrying the combination grooving and ribbing tool, whereas cams _F_ and _G_ force in the holders carrying the two under-cutting tools, these being presented at an angle to the work. The required oscillations to the slide carrying the grooving and ribbing tool are secured through a face-cam _B_ clamped to a “Whiton” chuck. The face-cam operates against the tension of spring _H_ and gives the required oscillations to the tool-slide carrying the ribbing and grooving tool, shown at _A_.

The third machining operation is accomplished in a flat turret lathe, as illustrated in Fig. 8. This consists in facing the open end of the shell, boring the powder pocket and facing and boring the diaphragm seat, and also turning the angular surface on the external nose of the shell. First, a roughing drill is brought in to rough out the powder pocket. The turret is then indexed and a tool for turning the angle of the nose is brought into position. The machining on the nose is then accomplished by operating the cross-sliding head. Then a roughing cutter is brought in to rough-bore the powder pocket. The turret is again indexed and a finishing tool is brought in to finish the powder pocket and face the diaphragm seat. This finishes the machining operations on the shell previous to heat-treatment.

=Heat-treating Shrapnel Shells.=--As was previously stated, the tensile strength of a forged shrapnel shell after heat-treatment must be from 80,000 to 90,000 pounds per square inch, and in order to obtain the desired physical qualities, it is necessary that the heat-treating operations be properly conducted. Several methods of heat-treating employing different cooling solutions are used in the manufacturing plants making shrapnel shells. One method, as shown in Fig. 9, is to heat the shell in a Hoskins electric furnace that contains a barium-chloride bath, heated to a temperature of about 1480 degrees F. The shells are left in this furnace for half an hour and are taken out and dipped in a bath of cottonseed oil heated to a temperature of 113 degrees F. The temperature to which the shell is heated varies with the different constituents of the steel and practically every different batch of 120 shells requires a slightly different temperature. The proper temperature is determined by cutting out a section of a heat-treated shell and testing it for tensile strength. The next step is to draw the temper on the open end of the shell. In this operation a muffle gas furnace heated to a temperature of about 1000 degrees F., is used. The temper is drawn for about two-thirds of the length of the shrapnel shells.

=Testing for Hardness and Tensile Strength.=--One shell from a batch of 120 is now cut open in the proximity of the powder pocket and the cut-out section sent to the government inspectors to test it for tensile strength. Each one of the shells in the batch, in addition, is tested for hardness by a Shore scleroscope as shown in Fig. 10. Before testing for hardness, the shell near the band groove is polished so as to get a true reading, then placed in a fixture, and the hammer of the scleroscope allowed to drop on it. The reading should be between 40 and 50, indicating an elastic limit of from 80,000 to 90,000 pounds per square inch. The shell must not be ruptured at the point tested when the charge in it is exploded or when the charge in the case is set off. Should the shell upset near the rifling band groove when it is propelled out of the gun, it would tear out the rifling in the bore of the gun.

Experience with the scleroscope has disclosed the existence of a fairly definite relation between the hardness and strength of metal. In determining the strength of metal, two stages are recognized: First, the elastic limit, determined by the load required to produce a permanent set; second, the ultimate strength, determined by the load required to cause rupture. The hardness indicated by the scleroscope is intimately related to the elastic limit. The elastic limit increases more rapidly than the hardness from 43 to 45, this being the minimum index of the strength value required. As an elongation of 8 per cent in 2 inches is also required, there must necessarily be an upper limit to the hardness. On the steel used for shrapnel, which is generally about 50-point carbon and 60-point manganese, the maximum hardness should not be over 60 on the scleroscope.

=Tests relating to Heat-treatment of Shells.=--In the September, 1915, number of MACHINERY, Mr. J. M. Wilson, who has been actively engaged in heat-treating shells since the beginning of the war, and who has had to rely entirely upon his own resources in meeting and overcoming the troubles which seemed to arise on all sides, relates the results of his experiments.

The British government shell specifications call for a yield point or elastic limit, after heat-treating, of not less than 36 tons per square inch, a breaking point or ultimate strength not less than 56 tons per square inch, and an elongation not less than 8 per cent in ⁵⁄₈ inch. Officially there is no maximum specified for either of those three physical characteristics; but as a matter of fact any unusual condition which is not in conformity with recognized metallurgical practice may cause the chief government inspector for the district in which the manufacturer is located to reject a shipment. Reference has been made to certain points in the shell which must resist the strains due to firing. The nature of these strains and condition of the steel best suited to meet them will be understood from Fig. 11, which shows a cross-section of the British 18-pound shrapnel shell. When a shell is fired from a gun, the base _A_ is subjected to a blow, _i.e._, a sudden increase of pressure which almost instantly attains a maximum of from 12 to 14 tons per square inch, and imparts the initial velocity to the shell. The shell, being a body at rest, opposes this velocity with its own inertia, the result being that both compressive and tensile strains are set up in the shell body. The shell body assumes the conditions of a column which has a compressive load varying from nothing at the nose to a maximum at the base. The tensile load is due to the inertia of the bullets inside the shell. These bullets are subject to an increasing compressive load from the top down, the resultant strain being a bursting effort which attains a maximum in the region of the point _B_, known as the “set-up point.”

When the time required for the fuse to act has elapsed, the powder charge is exploded, and the contents of the shell are blown forward in the usual manner. The contents are released either by the stripping of the thread of the brass socket, or else the walls of the shell yield at the point _C_, opening the threads sufficiently to free the socket. At _A_, (the base) the shell must be perfectly sound and free from flaws such as minute cracks, etc., which may allow the flame from the firing charge to strike through with disastrous results to the shell and gun. The metal in the base must not be too hard or it may fracture under the pressure of the explosion, and it must not be too soft or it may flatten out and spoil the rifling in the bore. At the point _B_ there is no maximum requirement so far as tensile strength is concerned, but any abnormal strength is viewed with suspicion unless it is accompanied by a generous elongation. At _B_ the metal is particularly liable to distension while the shell is acquiring velocity, and unless the shell is strong enough to resist the sudden bursting strain, and the amount of elongation is sufficient to cushion or absorb this strain at the instant of firing, the shell is liable to take a permanent set in the region of point _B_, with results mentioned above. The shell must not be too hard at the point _C_ as it may burst, thus neutralizing the real object of a shrapnel shell which is to project the bullets forward with increased velocity at the predetermined instant, being in fact an aerial gun arranged to discharge its contents at any desired point of its flight.

=Uniformity of Steel for Shrapnel.=--Having these requirements firmly established in his mind, the heat-treating expert is now confronted with a double problem: How is it possible to give steel the suitable strength; and having done so, how is it possible to know that the desired result has been obtained, without actually making test pieces from each shell. The principal condition upon which successful heat-treating depends is uniformity of material. Carbon and manganese are the principal substances which influence the results. The exact composition of steel specified by the government is not given to any manufacturers other than steelmakers. It is, however, generally understood to be a 0.50 per cent carbon, 0.60 per cent manganese steel. Allowing five points variation in carbon and ten points variation in manganese, the requirements would be approximately 0.45 to 0.55 per cent carbon and 0.50 to 0.70 per cent manganese. In one carload of forgings, one firm received shells from 23 different heats or melts, with carbon varying from 0.60 to 0.47 per cent, and manganese varying from 0.63 to 0.49 per cent, with all possible combinations and proportions between these limits. The number of forgings supplied from each heat varied from one up to 1200 so that the question of determining the best temperature for each carbon content was indeed quite impracticable. Many manufacturers at the present moment may be in a similar position, and the gravity of the situation, both from a financial and a military point of view, may justify a somewhat detailed description of the method which was followed in treating shells of such varying composition.

=Results of Tests.=--It is generally known to manufacturers that the highest tensile strength of steel is obtained by cooling it rapidly from a temperature slightly higher than the decalescent point or critical temperature. The degree of hardness resulting from this operation can be ascertained quickly, accurately, and repeatedly by means of the scleroscope. The degree of hardness thus shown is a reliable indication of the probable strength of the material; that is to say, after making due allowance for different makes of steel and varying proportions of the principal constituents, the scleroscope readings are a reliable indication of the results which may be expected when a tensile test is made of any given shell. In the opening months of the shell business, considerable reliance was placed on the accurate determination of the decalescence point. Forgings of varying analysis were received; the carbon being from 0.48 to 0.53 per cent, and the manganese from 0.54 to 0.69 per cent. All steels whose composition was within those limits showed a decalescence point of between 1390 and 1425 degrees F., and when quenched in water at 50 degrees F. above the decalescence point, such steels would have a scleroscope hardness number as high as 85; but when quenched in ordinary fish oil the hardness was only slightly over 50, the sample being 1 inch square and ¹⁄₈ inch thick. A complete shell quenched in fish oil would show a scleroscope hardness number at the set-up point of from 38 to 40. Test pieces from such a shell failed to reach the minimum breaking strength of 56 tons by the narrow margin of 0.6 ton, and this failure brought up the question of which was the best quenching medium. A series of experiments gave the results presented in Table I; all conditions were equal in each test, and the test pieces were all made from the same forging.

TABLE I. RESULTS OF TESTS TO DETERMINE THE BEST QUENCHING MEDIUM FOR SHRAPNEL SHELLS

+-------------+----------------+--------------------+---------------+
| Quenching | Quenching | Temperature of | Scleroscope |
|temperature, | medium | quenching | hardness No. |
| degrees F. | | medium, degrees F. | |
+-------------+----------------+--------------------+---------------+
| 1475 | Fish oil | 90 | 50 to 55 |
| 1475 | Coal oil | 90 | 65 to 70 |
| 1475 | Cottonseed oil | 90 | 70 to 75 |
| 1475 | Engine oil | 90 | 75 to 80 |
| 1475 | Oil of degras | 90 | 77 to 85 |
| 1475 | Water | 90 | 82 to 87 |
+-------------+----------------+--------------------+---------------+
_Machinery_

From the results of the tests presented in Table I, oil of degras, commercially known as “No 2 soluble quenching oil,” was selected as the quenching medium and operations were commenced on forgings supplied from two separate heats. The results were all that could be desired until forgings were received from a certain heat, which would not respond to treatment based upon the results of preliminary experiments. Investigation yielded the results presented in Table II. While water-treatment of the forgings from “Heat No. 3” gave satisfactory strengths under test, the liability of shells to crack, owing to their thin walls contracting more rapidly than the base, was a fatal objection to this method. Attention should be called to the fact that while the temperature at which quenching should be done is specified by the government at 1560 degrees F., manufacturers are not tied down to this particular temperature. What is required is that the manufacturers shall so treat the material that it will fulfill the requirements already stated. If, when fulfilling these requirements, the treatment should prove detrimental to the shell in other respects, then it must be changed accordingly.

TABLE II. RESULTS OF TESTS CONDUCTED TO SECURE GENERAL DATA ON HEAT-TREATMENT

+------------------------------------+---------+---------+-----------+
| Heat No. | 1 | 2 | 3 |
+------------------------------------+---------+---------+-----------+
|Carbon, per cent | 0.45 | 0.52 | 0.50 |
|Manganese, per cent | 0.68 | 0.62 | 0.47 |
|Decalescent point, degrees F | 1400 | 1425 | 1390 |
|Quenching temperature, degrees F | 1450 | 1475 | 1450 |
|Temperature of oil, degrees F | 160 | 160 | 120 |
|Resultant hardness, scleroscope No |65 to 75 |65 to 75 | *39 |
|Temperature of water, degrees F | | | 75 |
|Resultant hardness, scleroscope No | | | 55 to 60 |
|Tempered until showing a scleroscope| | | |
| hardness of | 48 | 48 | 52 |
|Yield point, tons | 47.8 | 48.6 | 46.5 |
|Breaking point, tons | 67.9 | 65.4 | 66.2 |
|Elongation, per cent | 14.5 | 16.9 | 17.4 |
+------------------------------------+---------+---------+-----------+
_Machinery_

* Note: This shell was then reheated and quenched in water with results shown.

Referring to results presented in Table II, “Heat No. 3,” it will be observed that the manganese is only 0.47 per cent with carbon 0.50 per cent. Comparing “Heat No. 3” with “Heat No. 1”, it is evident that an increase of 5 points carbon is more than offset by a reduction of 21 points in the manganese. Increase of temperature seemed to offer the greatest possibilities and sample shells were drawn every 12¹⁄₂ degrees up to 1675 degrees F. The greatest hardness was obtained at 1637¹⁄₂, scleroscope readings of from 50 to 55 being the average. This was not considered satisfactory, and the oil-circulating pump was speeded up. Scleroscope readings as high as 65 were frequently obtained at a quenching temperature of approximately 1635 degrees, and when the shell was tempered to read 48 to 52 on the scleroscope, three test pieces from one shell gave the results presented in Table III. A careful study of this data revealed the fact that, while a low-carbon, low-manganese steel hardens satisfactorily within a limited range of temperature, a medium steel has a wider range, and a high-carbon steel, a still wider range of hardening temperature.

TABLE III. RESULTS OF TESTS ON SAMPLES TAKEN FROM A SHELL WITH A SCLEROSCOPE HARDNESS NUMBER OF FROM 48 TO 52

+--------+---------------------+-------+---------+-----------+
|Heat No.|Scleroscope reading |Yield |Breaking |Elongation,|
| |on test piece after |Point, | point, | per cent |
| | machining | tons | tons | |
+--------+---------------------+-------+---------+-----------+
| |Outside 52--53--50 | | | |
| 1 |Inside 55--55--55 | 55.8 | 73.3 | 14.3 |
| | | | | |
| |Outside 52--54--50 | | | |
| 2 |Inside 55--57--53 | 53.8 | 72.4 | 17.4 |
| | | | | |
| |Outside 57--57--49 | | | |
| 3 |Inside 60--62--51 | 52.8 | 77.3 | 12.7 |
+--------+---------------------+-------+---------+-----------+
_Machinery_

When the shipment of mixed heats previously referred to was treated, the method pursued was to take 0.50 per cent carbon and 0.50 per cent manganese as a base composition which hardened at 1600 degrees F. to show 55 to 65 hardness on the scleroscope. Then: (a) If, for every point of carbon below 50, there be present 1 or more points of manganese above 50, the steel should harden satisfactorily at 1600 degrees F. (b) If, for every point of manganese below 50, there be present 2 or more points of carbon above 50, the steel should harden satisfactorily at 1600 degrees F. (c) If both carbon and manganese be below 0.50 per cent, increase the hardening temperature 12¹⁄₂ degrees F. for each point of manganese short of 50, and 6¹⁄₄ degrees F. for each point of carbon short of 50. (d) If both carbon and manganese are above 0.50 per cent, a hardness number above 55 will probably be obtained at a quenching temperature of 1600 degrees F., but the maximum hardness, _i.e._, from 75 to 80, will be obtained at a somewhat lower temperature, the exact temperature being most easily found by starting at 1500 degrees F. and trying a couple of sample shells every 25 degrees F. until a maximum hardness is obtained. Forgings containing from 0.50 to 0.55 per cent carbon and from 0.54 to 0.62 per cent manganese in any varying proportions may be hardened at 1600 degrees F. to show a hardness number of from 55 to 75; and when tempered to give a hardness number of from 48 to 52 they will yield the following results: yield point, 45 to 50 tons; breaking point, 65 to 70 tons; and elongation, 14 to 20 per cent.

Looking back, (c) offers a basis for charting the hardening points in a fairly approximate manner, to form a guide as to where the best hardness may be obtained. Such a chart is shown in Fig. 12. By following the horizontal and vertical lines from the carbon and manganese content until they intersect, a diagonal line will be found which will indicate the temperature at or about which the maximum hardness will be obtained. This does not prevent the use of 1600 degrees F. as the average temperature for the majority of shells, provided they are strong enough when hardened at that temperature; but where shells do not harden satisfactorily at 1600 degrees F., the chart offers an alternative method subject to such variation as may arise due to the use of steel from different makers, etc. Probably the best practice is to make careful scleroscope readings of each piece before pulling. Care must be taken to have a uniform surface on both sides, all tool marks being removed with fine emery cloth. The points tested are shown at _A_, _B_, and _C_ in Fig. 11. After the test piece is made, the value of the hardness number increases as a result of the piece being solidly supported in the scleroscope, whereas, when the reading is made on the shell, the arched form of the wall acts as a spring, and absorbs the shock to some extent. Readings thus increase from 2 to 10 points after the test piece is finished.

TABLE IV. DATA ON THE HEAT-TREATMENT AND STRENGTH TESTS OF SHRAPNEL SHELLS

+--------+----------+------------+-----------+------------+-------+----------+-------------+
| | | Quenching | Tempered, | | | | |
|Carbon, |Manganese,|temperature,|scleroscope|Readings of | Yield | Breaking | Elongation, |
| per | per | degrees | hardness |scleroscope | point,| point, | per |
| cent | cent | F. | No. | | tons | tons | cent |
+--------+----------+------------+-----------+------------+-------+----------+-------------+
| 0.50 | 0.47 | 1635 | 51 | 60-57-57 | 48.3 | 69.9 | 16.9 |
| | | | | 47-48-48 | | | |
| | | | | | | | |
| Three pieces from one shell | | 60-56-53 | 45.2 | 70.6 | 19.1 |
| | | 48-52-58 | | | |
| | | | | | |
| | | 63-56-57 | 51.6 | 74.6 | 16.9 |
| | | 51-55-54 | | | |
| | | | | | |
| 0.48 | 0.65 | 1565 | 49 | 51-54-52 | 47.3 | 67.4 | 15.9 |
| | | 48-53-50 | | | |
| | | | | | |
| Three pieces from one shell | | 51-52-49 | 48.2 | 67.9 | 15.3 |
| | | 53-51-51 | | | |
| | | | | | |
| | | 52-55-50 | 49.2 | 70.7 | 15.4 |
| | | 50-55-47 | | | |
| | | | | | | |
| 0.50 | 0.57 | 1600 | 50 | 50-52-50 | 46.0 | 64.8 | 19.0 |
| | | | | 49-50-49 | | | |
| | | | | | | | |
| 0.50 | 0.57 | 1600 | 50 | 56-60-57 | 55.8 | 77.8 | 14.3 |
| | | | | 54-56-54 | | | |
| | | | | | | | |
| 0.50 | 0.57 | 1600 | 50 | 59-60-56 | 60.7 | 82.2 | 12.7 |
| | | | | 55-59-56 | | | |
| | | | | | | | |
| 0.60 | 0.57 | 1600 | 50 | 60-61-55 | 57.8 | 80.0 | 12.6 |
| | | | | 60-62-57 | | | |
| | | | | | | | |
| 0.60 | 0.57 | 1600 | 52 | 57-57-56 | 48.2 | 69.7 | 17.5 |
| | | | | 54-56-53 | | | |
| | | | | | | | |
| 0.50 | 0.57 | 1600 | 50 | 48-52-50 | 44.2 | 64.3 | 17.4 |
| | | | | 49-52-49 | | | |
| | | | | | | | |
| 0.50 | 0.57 | 1600 | 50 | 52-55-55 | 44.7 | 65.2 | 14.7 |
| | | | | 60-51-52 | | | |
+--------+----------+------------+-----------+------------+-------+----------+-------------+
_Machinery_

A careful study of the data presented in Table IV reveals the fact that results are not always consistent. With an increase of carbon, one occasionally finds an increase in elongation and _vice versa_; and the results due to variations in manganese content are similarly unreliable. In order to secure a degree of uniformity in hardness, which will be sufficient to insure test pieces standing up successfully, it is necessary to have the shell hard inside as well as outside, and a method of doing this is referred to later. Assuming now that the shell has been tempered, it is rough-polished on a canvas buffing wheel around the outside of _B_, Fig. 11, for a width of at least 1 inch. Readings by the scleroscope are made on a zone ³⁄₄ inch wide, and if they are between 46 and 52 the shell may be relied upon to show good results in the tensile test. In making test pieces, it is desirable to cut the piece from a spot which reads 48 to 50; and in machining the test piece, care should be taken to remove an equal quantity of metal from either side of the wall so that the test piece is a true specimen of the average wall structure. Where a shell is carelessly quenched, and the test piece so machined that the surface on one side is practically the same as the inner side of the wall, the results would not be a true indication of the real average strength, and a lot of shells might possibly be rejected on account of a slight oversight in this respect. Reference has been made to the base _A_, Fig. 11. Forging defects show up here occasionally and in such cases the shell is at once condemned. These flaws take the form of small cracks, from the width of a hair up to ¹⁄₁₆ inch. They seldom can be detected until after heat-treating, and are most easily observed by polishing the base on a disk grinder. Losses in this respect vary, but might average about 0.20 per cent. The hardness of the base itself may vary from 38 to 50, which insures an ample degree of toughness and avoids all possibility of the shell cracking under fire.

=Heat-treating Department.=--Many methods of heating, quenching, annealing, and cleaning are in use by the different firms engaged in shell making. For rapidity of output, cleanliness of the resulting product, ease and economy of operation, and uniformity and control of results, the lead bath seems best for hardening, and the semi-muffle furnace for annealing. In one case the use of a lead bath by a skilled operator yielded excellent results both as to economy and uniformity, but, when the output exceeds 500 shells per 12 hours, a semi-continuous furnace meets the requirements to better advantage. The lay-out of a hardening room for an output of 12,000 shells per week is given in Fig. 13. The lead baths consist of a rectangular pot of suitable capacity, resting on a 4¹⁄₂ inch hearth built of common firebrick and heated by either oil or gas burners below the hearth. They are built in pairs with a common wall between, which is thick enough to provide a flue to carry off products of combustion. The quenching tanks are rectangular, water-jacketed, and provided with two quenching cradles each. These cradles are arranged to swing lengthwise in the tank, and, when the carrier holding the shell is lowered into the oil, a pipe is automatically extended downward into the shell and introduces cold oil in the inside of the shell, while the operator swings the cradle back and forth in the tank, thus cooling the outside of the shell at the same time. This method of quenching made it possible to harden shells which, by reason of low carbon and manganese, defied all conventional methods of dipping and swinging back and forth with tongs. The output per man with this apparatus is largely in excess of any hand method, while the uniformity and degree of hardness is all that could be desired.

The oil pump draws the oil from a depth of 6 inches below the surface and pumps it through 100 feet of 1-inch copper pipe arranged in two 50-foot coils in parallel. The cooled oil is delivered into an overhead reservoir, the overflow being connected to both tanks equally. After quenching, the shells are set on draining racks, and then washed in boiling water and sal-soda, placed on another draining rack and then brushed with wire brushes previous to tempering. The tempering furnace is of rectangular form, and consists of a long flat hearth with rails laid lengthwise on it. At each end a space is partitioned off from the body of the furnace, by means of vertical sliding doors; and a rack holding a number of shells is deposited on the rails at the front end of the hearth, the door is elevated and the rack is slid into the main chamber. After a suitable lapse of time another rack is introduced, and so on until the first rack is ejected at the rear end of the furnace. The shells are now hot enough to loosen all foreign matter on the surface, and a few seconds brushing with a wire brush cleans out the driving band groove, and leaves the shell with a delicate brown oxidized finish. The shell is now spotted on three places with a canvas buff and tested for hardness. Fig. 14 shows the arrangement of the scleroscope. The shell is supported on a single narrow V-block with hardened edges, situated immediately under the set-up point. A narrow strip supports the open end of the shell, thus giving a three-point support, while a vertical stop at the back of the shell maintains it in a position tangential to the radius of the swinging arm. The usual rubber bulb was soon dispensed with as being quite unsuited for such hard service, and a small pump cylinder substituted. The piston in the cylinder is operated by a downward pressure of the heel on the pedal to give compression, and a spring inside the cylinder gives the necessary pull when the scleroscope hammer is to be raised by suction. After being tested the shells are ready for “nosing-in.”

=Closing-in the End of the Shell.=--On some makes of shells, particularly the British, the nose is closed in before performing the third series of machining operations. The closing-in is generally accomplished in a hydraulic or power press. Fig. 15 shows the closing-in operation being performed in a vertical hydraulic press capable of exerting a pressure of 800 pounds per square inch. Before closing the open end of the shell, it is heated in the lead bath, shown to the left of the illustration, which is kept at a temperature between 1450 and 1500 degrees F. The steel diaphragm, which is larger in diameter than the nose of the shell, is first thrown in. Then the shell is placed in the press, and a cone-shaped die descends, closing-in the nose to the proper shape and diameter. The third machining operation consists in finishing the radius on the nose, both inside and outside, and cutting the thread. This is done, as shown in Fig. 16, in an ordinary engine lathe with a turret on the saddle. The boring is done with cutters held in boring-bars and the thread cut with a Geometric collapsible tap. The thread on the 18-pounder is 2.94 inches in diameter, 14-pitch, Whitworth type.

=Grinding Shrapnel Shells.=--The exterior surface of a shrapnel shell is straight for a portion of the length and then curved on the nose. While the limits required are not extremely close, it is necessary, where large production is required, to accomplish the finishing operations on the exterior of the shell in some way by which fairly close dimensions can be secured as well as large production. Grinding has, therefore, been recommended for finishing the exterior of the shell. One method of grinding shrapnel shells, in which a wide-faced wheel is used that covers the entire ground surface, is shown in Fig. 17. This machine is built by the Ford-Smith Machine Co., Hamilton, Ont., and carries a wheel about 8¹⁄₄ inches wide by 20 inches in diameter. The grinding wheel is rotated at 1200 R. P. M., and the work at 50 R. P. M. The depth of the cut is about ¹⁄₃₂ inch, and the time to complete one shell varies between two and three minutes. For grinding, a plug is screwed into the open end of the shell. This is held on the tailstock center and a chuck holds and drives the shell from the other end.

It is necessary, of course, that the wheel be kept the correct shape, and for this purpose an interesting type of wheel-truing device, differing considerably from that shown in Fig. 17, is now used. Referring to Fig. 19, it will be seen that this comprises a combination wheel guard and bracket, the latter being used as a base for the wheel-truing device proper. The diamond _A_ is carried in a holder _B_ that operates in a slide in the face of the traversing wheel-truing slide _C_. The diamond holder carries a cam point _D_ which is kept in contact with the guide or former cam _E_ by means of a spring _F_. The wheel-truing slide _C_ is traversed by a triple pitch screw _G_ so as to give a rapid movement to the slide in order to produce what might be termed a “rough-truing” of the wheel. For change in diameter, and also for bringing the diamond in contact with the wheel, a vertical slide _H_ is provided that is operated by handle _I_. In order to observe the diamond when truing the wheel, a trap door _J_ is provided in the wheel guard, which can be dropped down into place when the actual grinding of the shell is being done.

=Pressing on the Rifling Band.=--In order to rotate the shrapnel when propelling it out of the howitzer, it is necessary to put on a rifling band to take the rifling grooves of the gun bore. As a rule, these rifling bands are made from copper tubing and are simply cut off in a hand screw machine or turret lathe. The next operation is to close in the rifling band on the shrapnel shell. The ring is dropped over the shell and a fixture is used to locate it in the correct relation to the groove in the circumference of the shell. Then a slight pressure is exerted on it to align it properly in the groove. It is now placed in the banding machine shown in Fig. 18. This particular machine is provided with six dies as shown in Fig. 20, and back of each one is a hydraulic cylinder operated by water pressure. Two squeezers are necessary to close the rifling band properly into the groove, the shell being given a half turn after each squeeze.

There are several different machines on the market for performing this closing-in operation on the rifling band. Another machine, built by the West Tire Setter Co., Rochester, N. Y., is shown in Fig. 21. The principle upon which this machine operates is almost identical with that previously described, but in this case oil is used as a pressure medium. It is forced into the machine by means of a belt-driven pump shown to the left of the illustration, which drives the oil from the oil tank and carries it to the center of the base of the press. An oil head is located at this point from which the pipes are run to each of the six rams or cylinders. The amount of pressure required for compressing the copper band depends largely upon the width and thickness and the amount that the band must be spread to fill the grooves, rather than upon the diameter of the shell. The machine shown in Fig. 21 is capable of exerting a pressure of 30 tons on each cylinder or a combined pressure of 180 tons on all six cylinders. It has a capacity for compressing at least two bands per minute.

=Machining the Rifling Band.=--One method of machining the rifling band to the correct shape is shown in Fig. 23. Here a Fox lathe is used which is provided with a chuck for holding the shell and which carries in the turret a revolving center for additionally supporting it. The machining is done by form tools which are of the correct shape. Before any other machining operations can be accomplished it is necessary to put in the tin powder cup, brass fuse tube, bullets, and resin. This cup is slipped in past the steel diaphragm, then both parts are allowed to drop to the bottom and the fuse tube is screwed into the diaphragm. The required number of lead bullets, which for the British 18-pound shrapnel is about 375 per shell, is then poured in. The bullets are held in a tank and are allowed to flow out upon the opening of a stopcock. In order to pack the bullets solidly, a compressed air ramming device forms the base upon which the shell rests while the bullets are being poured in. This is operated three or four times for the filling of each shell and arranges the bullets compactly.

The resin is now poured in, as shown in the center of Fig. 22. This is carried in the tank which is heated by a gas furnace and is poured in almost level with the top of the bullets. The shell is then placed on the scale in the immediate foreground and weighed. One dram plus or minus is allowed as a variation, and in order to not exceed this, more or less resin is poured in until the correct weight is obtained. The brass fuse socket is now screwed in as shown to the left of the illustration, and upon the completion of this operation the shell is ready for the fourth and last machining operation. This last operation consists in machining the brass socket on the outside diameter to conform to the radius on the nose of the shell, and boring on the inside and threading to fit the fuse body. These operations are handled in a Fox brass working lathe. Upon the completion of the machining operations the plug is screwed in, the shell stamped, cleaned, weighed, and inspected by government inspectors. After this, the shell is given two coats of paint and a red band is painted around the nose. It is now packed in boxes holding six shells and is ready for shipment. This completes the manufacture of the shrapnel shell.

=Gaging Shrapnel Shells.=--The machining operations on shrapnel shells are required to be held within certain limits, and government inspectors watch these closely. Some of the principal gaging operations on the shrapnel shell body are shown in Fig. 24. Fig. 25 shows the 18-pound shrapnel shell in section, and gives the principal dimensions together with the limits; it will be seen from this illustration that the range allowable is in most cases large. The Wells Bros. Co., Greenfield, Mass., has made a large number of shrapnel gages, some of which are shown in the accompanying illustrations. In the three upper views of Fig. 24, the Wells Bros, standard thread gage is illustrated. This is used for all diameter measurements by substituting flat gaging pins for the V-points used when gaging thread diameters.

=Gages for British Shrapnel Parts.=--Fig. 26 illustrates typical gages for gaging such parts of the British shrapnel as body diameters, diaphragm seat, powder pocket, fuse socket, thread diameters, and fuse parts. Fig. 27 shows the application of several different types of shrapnel shell gages. At _A_ is the gage for the over-all length. At _B_ is the gage used for measuring the thickness of the closed end. The outer arm of this gage can be swung away to allow the placing of the gage on the standard. At the extreme lower left-hand corner of the gaging arm is a slight shoulder on the rod and the height of this acts as the limit. _C_ shows the application of outside diameter and thread gages. _D_ shows three form gages for checking the shape and dimensions of the wave ribs, the diameter and shape of the undercut in the band groove, and the shape of the nose of the shell. _E_ shows the gage used for checking the thickness of the wall of the shell at different distances from the mouth. _F_ shows the application of a powder pocket gage, and also a gage for checking the shape of the finished rifling band.

=Gages for American Shrapnel Shells.=--Fig. 28 shows a miscellaneous collection of gages used in checking the dimensions of the American shrapnel shell. Gages, _A_, _B_, _C_, and _D_ are for measuring the diameter of the diaphragm seat. _E_ is for checking the distance from the diaphragm seat to the mouth end of the shell, and gage _F_ is for the outside diameter of the shell. Gage _G_ is used for the rifling band groove. Gages _H_ and _I_ are for the thread in the mouth of the shell, _H_ being a “not-go” and _I_ a “go” gage.

The gage at _J_ performs several gaging functions on the American shell. It consists of a standard having two upright posts across which a bar is mounted. The purpose of the bar is to gage the over-all length of the shell, and its lower surface is provided with two steps giving the limits. This gage is also used for measuring the depth of the powder pocket, rod _K_ and block _L_ performing this function. Two rings are cut around the rod _K_ registering with the top surface of the bar, the purpose being to show the accuracy of the work.

Another interesting gage is shown at _M_. This is for gaging the concentricity of the shell and consists of an arbor mounted so that it can be swung on a pivot. The arbor carries two collars _N_ and _O_ that fit in the shell. Collar _P_ is merely a sizing plug and when the gage is in use this plug is removed. A gaging finger _Q_ rests against the shell when it is on this arbor, and a standard type of indicator _R_ shows the variation in concentricity when the gage, collars, and shell are rotated on the arbor.

=Marking Shrapnel Shells.=--All shrapnel shells are marked on their circumference with five or six lines of lettering, as shown in Fig. 29. This indicates the size of the shell, the series, muzzle velocity, name of the manufacturer, date completed, etc. Two types of machines for producing the stamping, built by Noble & Westbrook, Hartford, Conn., are shown in Figs. 29 and 30. The machine shown in Fig. 29 is of the hand-operated type. The figure block _A_ is held in a slide that is moved longitudinally by pulling down handle _B_, rolling the shell, and at the same time stamping it. The shell is located on the table in the two positions by gages _C_ and _D_.

The “Dwight-Slate” stamping machine shown in Fig. 30 is power-driven, and the work is held on an elevating table. The stamp is held in a slide operated by an eccentric and connecting-rod. In this machine the shell is not distorted.