Search Documents

By H. J. Hand

Statistical methods are becoming increasingly important for interpreting routine reports, or for analyzing special test data in industrial plants, such as steel plants. They have already become practically a necessity to the plant metallurgist. A paper by Chancellor, 1 written in 1933, gives many examples of the benefits the plant metallurgist can derive from the use of these methods. The present paper discusses applications of importance to many others who must analyze reports in the steel plant, as well as other examples of interest to metallurgists. Statistical Methods Logically Suited to Plant Problems.—Statistical methods are not new as a branch of mathematics. Until comparatively recently, however, practical applications were confined mainly to actuarial science, biology, and economics; fields in which an observed result is obviously influenced (and often determined) by a complicated set of variables more or less constantly tending to affect it. As pointed out by Daeves,2 a decade or more ago, problems in industry are basically similar to those in these other fields in that a given result is at least partly due to that chance combination of causes happening to be present at the time. When trying to do the same thing repeatedly, constant results are never attained; they vary about the desired result with more or less well defined mathematical regularity. Such a condition arises out of the great number of factors tending to affect the final result. If statistical methods are so valuable to the biologist or economist, why then should they not also be of great value to plant managers and engineers, and to all who must interpret reports from various sources in the industrial plant? Within the last few years, industry has become conscious of this situation, and the use of these methods has been rapidly expanding in such p1ants. Many Useful Statistical Devices Available.—Many statistical methods of attack are available for the analysis of plant data. It is the object of

Jan 1, 1938

By Leonard C. Grimshaw

Diffusion of carbon from gases into iron has been the object of much research, because of its long recognized importance in carburizing processes, but the direct diffusion of carbon from steel into iron has received relatively little attention. Because of the increasing use of clad metals in industry, direct diffusion from steel into iron is becoming of practical importance. Therefore, the object of this investigation was to study the diffusion of carbon from several different kinds of steel directly into relatively pure iron, realizing that elements in the steels other than carbon might influence the rate of carbon diffusion. The previous work on diffusion of carbon into iron has been considerable, and includes many reports in foreign languages. This work has recently been well reviewed by Mehll and Krynitsky12 and the reader is referred to their excellent bibliographies. Methods of Study There are two main methods by which the direct diffusion of carbon from steel into iron may be studied. In one method, steel is first decar-burized in part, and diffusion of carbon then allowed to take place from the steel into the decarburized steel. This method is open to two serious objections: (1) the decarburized steel is not iron, but contains elements such as manganese, silicon, chromium or tungsten, which may be an essential part of the steel under study, and which may seriously influence the diffusion of carbon into the decarburized area; (2) there is from the start of diffusion an unknown carbon-concentration gradient between the steel and the decarburized area. Another method, which possesses neither of these disadvantages, is to study diffusion from steel into a thick plating of electrolytic iron. Herc the chief difficulty is to secure an electrolytic deposit of iron upon steel without introducing an oxide film at the interface that will interfere with diffusion. The author is able to deposit iron upon any kind of steel without introducing an oxide film at the interface, so that diffusion of elements from the steel into the iron is free to proceed unhindered.

Jan 1, 1938

By C. O. Burgess, W. D. Forgeng

The results of an investigation of the ternary system iron-chromium-silicon were reported in 1936 by the present authors.l Partly for the sake of theoretical interest, and partly because of the possible commercial importance of these alloys, the present investigation was undertaken to determine the effect of some of the commoner elements likely to occur, deliberately or otherwise, in such alloys if made on a commercial scale. The elements thus added were carbon, nickel or manganese. The alloys studied were, therefore, parts of quaternary systems. This investigation was concerned primarily with the effect of these additional elements on the alpha-iron phase and the boundary between this phase region and the alpha plus sigma two-phase region. The sigma phase is a ternary solid solution based upon the compound FeCr. The work was performed mainly by X-ray methods, but a number of microscopic examinations were also made and are reported. While isolated alloys in the composition ranges studied may have been discussed in metallurgical publications, no systematic investigation of any assistance to us was found. Of the earlier work, the investigations on the iron-chromium-carbon system and on the chromium carbides by Westgren and his coworkers2'3 enabled us to identify the carbide phase. Several articles appeared after the present experimental work had been completed. Lucas and Wintrup4 determined the solubility of carbon in molten iron-chromium-silicon alloys. Tofaute and his coworkers5'6 have made detailed studies of the iron-chromium-carbon system, which will be referred to later. Materials and Methods Materials.—The iron, chromium and silicon used were essentially the same as those used in the investigation of the ternary system.l The

Jan 1, 1938

By A. G. H. Anderson, Eric R. Jette

A few years ago, while the writers worked on the constitution of ternary and quaternary alloys consisting mainly of iron, chromium and siliconl,2 some information on microstructures and hardness, not incorporated in the published work, alas obtained. It has not been possible to do any further work on the properties of the alloys. The information possessed is hereby presented for the interest it may have in connection with the constitution of the alloys. Microstructures Referring to the iron-chromium-silicon diagram,' the regions that received most attention were those in which the sigma phase appears. It was found that the mode of occurrence of this phase varied much with the condition under which it mas formed, and the mode influenced considerably the hardness of the alloys. In some slowly cooled alloys, well within the alpha + sigma-phase boundary, sigma occurs in pabterns suggesting a eutectic. The structure of such an alloy is shorn in Fig. 1, showing it to be partly coarsely laminated and partly globular mottled. In other annealed and slowly cooled alloys, the lanlinated structure is entirely absent, and sigma appears in an irregular distribution of sharply cornered elongated areas (Figs. 2 and 3). These areas all possess a more or less mottled structure. A carbide phase is also visible in these two photomicrographs; it is represented by bright white arras of various shapes, sometimes apparently enveloping the sigma. Quenched and aged specimens, whose compositions arc: located near the phase boundary between alpha and sigma + alpha regions, have structures not unlike martensite (Figs. 5 and 6). Fig. 4 shows the structurc of alloy No. 35 as quenched from 1100" C.; it is a typical solid solu-

Jan 1, 1938

By A. G. H. Anderson, Eric R. Jette

During the period from 1910 to 1920, there was a lively interest in the subject of grain growth and many papers were published, followed by interesting discussions. Questions dealing with the fundamentals of grain growth—why and how it occurs—were treated in papers by Howe,' .Jeffries,² Sauveur,² Chappell,4 Beilby,5 Mathewson and Phillips,6 Mc Adam,7 Carpenter and Elam,8 Stead and Carpenter,9 Ruder,l0 as well as others. Since that decade, much indeed has been done regarding the practical control of grain size, and it is now possible to order material having a size sperified in the American Society for Testing Materials classification. Also, one hears the term "spliced boundaries" used in connection with certain tungsten lamp filamentsl1 wherein it is important to prevent "sagging" between consecutive turns of the helix.12 This practical use of a certain knowledge as to the things that influence or even control grain growth should not be taken as an indication that the subject has been thoroughly explored and that nothing more can be learned from fundamental studies. It is with this thought in mind that a report of some observations and experiments on abnormal grain growth is believed justified. Abnormal grain growth is met with in heat-treating high-carbon steels in the austenitic range,'? in the hardening treatments of high-speed steels,14'15 and in the low-temperature annealing of strained low-carbon steels as in sheet and wire material. Entirely aside from these instances of exaggerated grain growth, which some might consider as freakish, it is thought that abnormal growth is simply a case where the growth forces and the forces tending to prevent growth are almost, but not quite, evenly balanced. Information relative to grain-growth characteristics in general may be obtained from a study of abnormal growth conditions, which may be looked upon as vantage points. Reflection will show at once that if growth forces are great in comparison to inertia or resistance to growth, many crystals start to grow and, for that reason, none becomes very large, whereas if inertia is

Jan 1, 1938

By M. L. Samuels

In a recent paper the authors1 discussed the reactions to tempering of hypereutectoid steels quenched from 1000' C., as revealed by studies on changes in hardness, electrical resistivity, coercive force and metallo-graphic charaeteristics. The data, which related specifically to tempering temperatures of 550°, 650" and 710" C., for varying periods up to 125 hr., permitted the following important conclusions: 1. The rate of softening of the steels at the temperatures examined differed markedly although the steels were of "similar" general chemical composition. 2. The temperature of maximum softening spread over a range from about 650" C, to the eutectoid inversion. 3. The divergence in behavior was ascribed, at least in part, to a profound difference in tendency to graphitization on annealing at sub-eutectoid temperatures. 4. The mode of spheroidization of the cementite, as revealed by metallographic studies, also varied considerably in the different steels. The purpose of the present investigation is to pursue further the study of the difference in behavior of the steels on tempering over a wider range of temperature. To this end the following features have received particular attention: 1. Comparison of temperatures of maximum softening of the various steels when the quenched samples were tempered for 5 hi. and for 75 hr. followed by furnace cooling and by quenching in water. 2. Deterrnirlation of the effect of quenching temperature, prior to tempering, on the rate of softening. Quenching temperatures of 1000" and 850" C. were used in the investigations. 3. Examination of the microstructures obtained with the various heat-treatments employed.

Jan 1, 1938

By B. S. Norris, Charles R. Austin

The stresses which initiate deformation processes in metals are fundamentally important in the study of the mechanical properties of metals. A point of inflection in the load-elongation curves obtained for single crystals' has been rather too readily interpreted as a true yield point, and the critical resolved shearing stress on the plane of slip taken as a material constant. Even early work showed that the stress at this point varies with rate of loading, but it has not been generally appreciated that creep tests under static loads alone are suited for determining the existence or nonexistence of a yield point.* It has now been shown by Miller that single crystals of zinc3 when loaded statically begin to yield at loads too small to measure; as a result there is no case now known, apart from that cited in the present paper, in which a true critical stress for yielding has been demonstrated for single crystals of metals. The case of iron has especial interest: the existence of an upper and lower yield point and of an elongation at the yield point in ordinary iron and steel marks this substance as unusual, though the effects have been found to a very minor degree in other metals and alloys. $ Most alloys do not show such marked phenomena when appreciable yielding begins, and accordingly modern practice is to speak of the yield strength of such materials, a stress read from the load-elongation curve after elongation has begun, rather than yield point. The special behavior of iron suggests that it is more likely that single crystals of iron should have a true yield point than other metals. The study of the mechanical properties of metals and alloys requires a determination of the crystallographic mechanism of deformation, wlletller deformation proceeds by slip, twinning, or cleavage, or a combination of these, and a determination of the stresses that induce these modes of deformation. With this information, a study of the rates of work-

Jan 1, 1938

By R. F. Mehl, M. Gensamer

Creep investigators have made extensive studies to determine the interrelation of stress, temperature and the tensile creep rates of metals. It has been suggested that at small stresses the secondary or constant creep rates obey a simple viscous law. Building upon this suggestion, departure from pure viscosity as stress increases may be expressed mathematically as a consequence of reversibly altered viscosity or "flowability" in the same sense that temperature change reversibly alters viscosity. From the index to alteration of viscosity, a value for the energy of secondary creep may be derived having the magnitude of the free energy of vaporized metal. The suggestion is made that application of stress alters the energy of secondary creep. Strain energy is argued to be but a minor factor in the energetics of secondary creep. A deformation mechanism involving the large energy associated with secondary creep implies a sort of solid self-diffusion process. Evidence indicates that creep strain above the recrystallization temperature takes place by crystallographically directed self-diffusion at grain boundaries and points of crystal defects. A self-diffusion theory of secondary creep is proposed, embracing the principles of the Dushman-Langmuir theory of solid diffusion. An equation is suggested whereby the rate of tensile creep may be related directly to the physical and structural character of the solid metal. Basis of Investigation Much of the investigation that has been given to the creep characteristics of metals and alloys has been directed toward studying the dependence of creep rate upon stress. Great interest has centered upon the approximately uniform rate of flow that has been demonstrated for numerous metals and alloys. At temperatures above the strain-hardening ranges of materials, this uniform rate frequently is attained after a relatively short period of initial rapid flow, and the period during which it persists determines for most practical purposes the useful life of the

Jan 1, 1938

By J. J. Kanter

There has been much research and commercial development in recent years in the use of chromium and nickel in steels of various types, including those intended for high-temperature service. By "high-temperature steels" are meant those specifically designed for service at elevated temperatures in steam plants, oil refineries and other industries where high-temperature operations are conducted. This paper is devoted to some preliminary data on the effect of silicon on certain of the most popular chromium-molybdenum steels now being used in quantity by the refining industry and includes a short statement on the influence of copper in steels intended for high-temperature service. Uses of Silicon Next to oxygen, silicon is the most widely distributed and abundant of the earth's substances. Although classed as a nonmetal, it is highly important in the metallurgy of all steels, having been used for years as an efficient and economical degasifier and deoxidizer. Silicon is added to various steels in the form of ferrosilicon, generally at the end of the heat so as to avoid oxidation losses. With acid steels it is added in the furnace (also in the ladle) and generally with basic steels is added in the ladle, so as to avoid reaction with the basic lining through high-temperature acidic reaction with silica, which may be formed by oxidation of the silicon alloy. Several grades of ferrosilicon are available, containing, respectively, 15, 50, 75 and 90 per cent Si. By far the largest tonnage produced is the 50 per cent grade although the 75 and 90 per cent ferro-silicons are used as ladle and furnace additions in the manufacture of steels of high silicon content. The introduction of silicon into steel as an alloying element is well known in connection with high-yield-strength constructional steel for bridges, ships, car construction, etc., and it is also well known as an alloy-

Jan 1, 1938

By H. D. Newell

Jan 1, 1938

By G. B. Gould

Every mathematical statement of a measure of anything (as distinguished from a count') is followed by a qualification—always implied if not explicity stated—-that the statement is only an estimate, which may vary from absolute accuracy within certain limits, according to the accuracy of the measuring instrument. This basic fact, so obvious to engineers, is not so well understood by many that make use of laboratory coal tests. The tolerances that must be allowed in measurements of coal quality are very important in both their commercial and engineering applications—and the significance of these tolerances depends upon the use made of the laboratory coal test. Only in recent years has there been any intensive study of this question of tolerances as applied to measurements of coal quality. There is very little in our published literature on the subject, and what has been published is not readily comprehended by those who sell coal, or buy coal and use it, unless they happen to have made a study of the subject themselves. There is great need for more widespread understanding of what lias been discovered and how it applies to the use of laboratory coal tests. What follows is an attempt to condense into a few pages the results of the major investigations so far reported, and to explain their meaning in simple nontechnical language, with the addition of some recent illustrative experiments. Determining Deviations in Quality It is necessary at the outset to distinguish between: (I) the accuracy with which one test indicates the quality on one lot of coal; and (2) the accuracy with which a test of one lot of coal represents the average quality of a commercial product, of which the buyer will presumably receive a series of shipments—perhaps many thousands of tons. It is necessary to make this distinction because coal is a variable material—a heterogeneous mixture of a variety of components—the quality of which will vary from one lot to another. Therefore an absolutely accurate determination of the quality (if that were possible) of one lot of a coal is an

Jan 1, 1938

By T. W. Guy

During the past three years the Surface Preparation Committee of the American Mining Congress Coal Operators' Committees has been collecting data on dewatering and drying washed coal, and on screening coal in representative plants in the Eastern and Central bituminous fields. Because of wide variations in bed moisture, such data can be compared only after deducting the bed moisture. It was found that practically every company used a different routine for moisture sampling and determinations. Some reported the moistures as total moistures, others as surface moistures, air-dried moistures, or oven-dried moistures. For most of the reports it was necessary to estimate, from a consideration of the bed moisture and other data available in a given case, corrections to be made in order to obtain surface-moisture figures as a common basis for comparison. At any plant, and regardless of bed moisture, the percentage of surface moisture remaining in a product when the dewatering, or a given stage of the dewatering, has been completed, is the real measure of what has been accomplished. The descriptions and data given with Fig. 1 show that surface-moisture figures show relationships that would not be shown by total moistures. Therefore bed moisture should be deducted, or, if more practicable, the surface moisture of the products should be measured direct. A preparation engineer finds one per cent, or even a part of one per cent, of surface moisture a very important consideration in the results of his efforts to find the most effective and economical methods of reducing and controlling the surface moisture in his products within the required limits. Therefore it is important for him to have accurate measurements of surface moisture with minimum uncertainty as to whether they sometimes include less than the surface moisture and at others all of the surface moisture plus an unknown part of the bed moisture. The Problem We know that pieces of coal have a capacity for holding water: (1) inside themselves, which is a natural characteristic of the bed from

Jan 1, 1938

By H. F. Yancey, M. R. Geer

The importance of the size composition of coal is reflected in the difference in price of the various sizes of the same coal and in the large number of primary sizes and mixtures of sizes produced by the modern preparation plant to meet the demands of a market that in late years has grown increasingly conscious of coal size. The importance of size in the industrial and domestic uses of coal has its counterpart in the role played by size composition in the laboratory examination of coal. Investigation of many of the physical properties of coal in their relation to its uses requires an evaluation of size composition. In fact, the Standards on Coal and Coke issued by the American Socaiety for Testing Materials in 1936 contains 12 standard methods of testing that involve some use of size composition. The screen analysis is, of course, the standard method of determining size composition, whether the material is coarse run-of-mine coal or fine pulverized fuel. Custorriarily, the analysis is reported as the weight percentages retained on or passing each of a series of testing screens, the number of which may range from only two or three to as many as 20. When more than two or three sieves are used, which is usual, the number of weight percentages obtained is so great as to make the analysis difficult to comprehend. Direct comparison of analyses made 011 the same sieves is hard, and comparison of analyses covering the same range of sizes but made on different sieves is next to impossible. Consequently, a frequent purpose of sizing tests-—-correlation of size with properties and charactcristics related to size-—is rendered difficult. Less cumbersome means of expressing size composition have been attempted with various degrees of success. An average particle size, calculated by any of several formulas, is useful for some purposes, also the weight percentage retained on a single sieve finds some application as a

Jan 1, 1938

By L. E. Young

The purpose of this paper is to start a discussion on: (I) methods of supporting the immediate roof, particularly of local areas of poor roof, during the mechanical loading of coal; (2) methods of increasing the rapidity of the movement of the mining faces so that, without the use of cribs, jacks, park walls, and filling, there will be no danger to men, equipment, and working places from the subsidence of the overlying strata. In the Pittsburgh district the application of mechanical loading devices in mines working in the Pittsburgh seam has been retarded by the difficulties of holding and handling the draw slate. Normally the thickness of this slate averages about 8 in. but the hardness varies, and a cut of coal cannot always be loaded without placing posts or crossbars to support the slate. Locally, the slate may be thicker and special methods of support may be required. If crossbars are placed to hold the slate when the coal averages 5 ft. in thickness, the standard methods of transporting the coal cannot be followed, and the use of conveyors may be desirable. Very little information is available on the effect of increasing the rate of movement of advancing or retreating break lines or longwall faces. A few British publications have presented some data but all such data relate to longwall faces that have moved not more than 6 ft. a day. The most rapid rate of movemerit of longwall faces in Germany is 6 ft. per day. In the ordinary hand-loading mine in western Pennsylvania break lines on full retreat will move about 2 ft. per day. If loading machines are used under similar conditions the rate of retreat may reach 4 ft. per day. In a mine where coal is loaded by hand on conveyors three shifts per day, the rate of retreat has ranged from 11 to 16 ft. per day, with an average of 13 ft. If mobile loading machines could be used to load on conveyors on break lines, the rate of retreat should be increased substantially During 1935 and 1936, the United States Bureau of Mines, under a cooperative agreement with the Pittsburgh Coal Co., carried on studies of roof movement at Montour To. 10 mine. The work war done urider the direction of Mr. H. P. Greenwald, Supervising En neer, Experi-

Jan 1, 1938

By Samuel M. Cassidy, John A. Garcia

In some districts of the bituminous coal field the problem of constructing bulkheads to seal off water under pressure is becoming increasingly important. Recently this matter has been brought very much to the front by the loss of a large Illinois mine due to failure of a bulkhead that separated it from several old, flooded mines. As this great misfortune has provoked much inquiry about proper bulkhead design, it is felt that a description of the successful and extensive water-sealing practice at three mines that have different natural conditions will be of considerable interest to those connected with properties having a present or future water problem and to engineers that may have to design safe bulkheads of economical construction. These three mines are the Saxton and Dresser, near Terre Haute, Ind., and the Hegeler, near Danville, Ill. All three have water conditions that make it necessary to regularly seal off all worked-out sections of the mine, and there are considerably over a hundred bulkheads in each. AS far as the writers know, these three mines are the only ones in the bituminous coal fields of the United States with anything like so large a number of water seals and experience, although a number of other mines are known to have a few bulkheads. More detail will be included herein on the Saxton mine than on the others, as the writers are more familiar with it. General Mining System All three of the mines use the same general panel room-and-pillar system, which consists of rooms turned both ways off room entries and a solid block of coal left on all sides of the completed panel except on the cross-entry side, where a heavier pillar is pierced by two entries for haulage and air. On completion of the panel the two entries arc sealed and the panel is then isolated from the rest of the mine. Owing to water conditions, no pillars are pulled; recovery inside the panel barrier pillars is approximately 65 per cent.

Jan 1, 1938

By George Watkin Evans

The need for mechanization in coal mining wherever it can be successfully introduced is evident when we take into account the fact that during recent years the coal-mine employee's wages have increased and his hours have been shortened. Coal is no longer King, and a brief study of statistics covering the production of coal, petroleum and natural gas and the development of water power will convince anyone that coal has had very powerful competitors during the past 10 years. Petroleum production in the United States between 1918 and 1936 increased from 100 per cent to 309 per cent. Water power increased from 100 per cent in 1918 to 270 per cent in 1936. Bituminous coal slipped from 100 per cent in 1918 to 75 per cent in 1936. The anthracite industry slipped from 100 per cent in 1918 to 50 per cent in 1932, but climbed to 55 per cent in 1936. In 1918, the State of Washington produced a little over four million tons of coal and the production of California oil, at that time, was a little less than one hundred million barrels per year. By 1929 the production of California oil had increased to a little over 292,000,000 bbl. or nearly 300 per cent as compared with 1918. By 1934 the production of coal in the state of Washington had dropped to about 1 1/3 millions of tons per year. Our neighbors in British Columbia had their coal production reduced from a little over 2 1/2 million tons in 1920 to less than 1 1/3 million in 1933. The North western Improvement Co., a subsidiary of the Northern Pacific Railroad Co., owns coal properties in various parts of Washington and Montana. The Northern Pacific is one of the very few railroads using any considerable quantity of coal in its operations. Practically all of the American railroads in the Pacific Northwest, except the Northern Pacific, are using cheap California fuel oil or electricity. This problem, of course, is not a new one in coal mining, and all of the progressive coal companies in the country have tried to meet competition from an overproduction of petroleum and an overproduction of natural gas as well as enlarged developments of hydroelectric plants.

Jan 1, 1938

By Phillip B. Bucky, R. S. Taborelli

The term "longwall mining" is best known to coal men, although modifications of the method are continually being used in other fields. Longwall mining is of interest today because it makes for greater inechanization and better supervision at the working face, and therefore lower production costs and increased safety. It is an approach to the factory system of production, the high efficiency of which as compared to that in mines is well known. There is available an accumulation of experience regarding this work but coordination of experience is lacking and this series of experiments was undertaken to help serve this purpose. Model experiments were used for reasons of safety, cost and time, and because it has been previously shown1,2,3 that a scalar inodel built of the same material as the prototype will behave in a manner similar to that of the prototype, with time effects the same, if the effective weight of the model is increased in the same proportion as its linear dimenbions are decreased. This is accomplished by substituting a centrifugal for the gravitational field; i.e., placing the model in a centrifuge and controlling the strength of the centrifugal field by varying the rate of rotation. Many factors affect the behavior of the geologic material overlying a bed, therefore results and conclusions contained in this paper apply only to the conditions presented herein. Longwall Mining What Is Longwall Mining?—Agreement should be reached as to what constitutes longwall mining and therefore what constitutes a longwall face. Let longwall mining be defined as the mining procedure in which the roof and overlying bed breaks tend to parallel the working face. In the ordinary stope or room-and-pillar mining the roof breaks tend to

Jan 1, 1938

By Charles T. Holland

Miners have observed for many years that as pillars are removed in mining operations the pillar adjacent to those mined out frequently shows evidence of being compressed. Although this has been known for a long time, little or no information is available concerning how far from the mining area the pillars are affected in this manner, or how great is the magnitude of the compression, or concerning the manner or characteristic way in which the compression acts, as the mined-out area approaches the pillar under consideration. The writer made a study of pillar action in the Mona mine of the Arkwright Coal Co., partly to get some quantitative data on the deformation of coal-mine pillars. The study was not completed but it seems best to present the information that was obtained. Method of Measurement Fig. 1 shows the details of the apparatus used for measuring the deformation. This is permanently attached to the pillar. Fig. 2 shows the details of the gauge that records the deformation. This gauge is removable and was used to measure the deformation at all the stations. The recording dial is graduated in thousandths of an inch and records measurements up to one inch. Fig. 3 shows how the removable gauge is placed in the instrument. When installing the instrument in the side of the pillar, all loose coal near by was removed and a trench 4 to 6 in. deep, and approximately 1 ft. wide, was dug in the pillar. In this trench near the top and the floor, the holes holding the instrument were bored. These holes were then partly filled with thin cement and the rods were inserted. The remaining space in the holes was tamped full of cement. At this time, by carefully manipulating the permanent rods and clamps and adjustable pin, it was possible to adjust the instrument so that the removable gauge would stand true and measure any vertical deformation between the top and bottom rods. Fig. 4 is a photograph of one of the instruments installed, showing the removable gauge in place.

Jan 1, 1938

By H. Landsberg

In a previous report1 it was pointed out that a successful attack on roof troubles has to be preceded by extensive scouting. As Lord Kelvin once said, scientific progress can be made only if accurate data are available: "unless you measure, you do not know." The Convergence Recorder For the observation of roof movements in mines the convergence recorder has shown that it will enable us to get some of the necessary information. Results with convergence recorders were first published by Winstanley and other English investigators for rapidly moving longwall faces. In their investigations the speed of work and the heavy overburden produced a considerable amount of absolute values of convergence in a short time. Under the conditions frequently encountered in our investigations in Central Pennsylvania, under shallow, overburden and in mines with slowly proceeding extraction, the movements observed were much less in absolute value, for an equal interval of time. This mas partly because the mines in which our studies were carried on frequently had idle days, sometimes four or five in succession. Therefore continuous records were not employed and a simple convergence gauge was built. In its first models, it consisted only of two telescoping tubes pressed against roof and floor by a coil spring. At the side of the lower tube a rule divided into 1/2-mm. divisions mas fixed and a knife-edge, which was attached to the upper tube, mas slid over the scale in case of convergence between roof and floor. While the larger motions could be read on this device, it seemed to be desirable to increase the accuracy by inserting between the two tubes a dial meter reading in 1/1000 in. Such a gauge was first built in April 1937 and installed for experimental purposes in a bituminous coal mine in Pennsylvania (Fig. I). It was read at least once daily, 24 hr. apart, and if feasible at shorter intervals. After several months of experience it can be said that this type of gauge answers the scientific as well as the practical needs. It is easier to handle than the recorder; no clock is used, there are no pen troubles. The price of

Jan 1, 1938

By P. Nicholls

Patent records show that the attempt to improve the burning of fuel by mixture or pretreatment with chemicals dates back to the early years of the last century. By 1845 English and French patents had included the specific beneficiations of elimination or reduction of smoke, soot and clinker; reduction of the ashes; elimination of sulphur fumes; increase of the rate of burning; increase of the time a given charge of fuel would last; and, in most patents, a saving of fuel. The earlier patents specified additions up to 80 lb. or more per ton of fuel, but after 1860 the quantity was seldom more than 4 lb. There is no record of the extent to which these patents were commercialized, but after 1874 many American patents were assigned to companies specializing in treatment of coal, so it is probable that the products were marketed on a considerable scale at least as early as that date. From 1900 on, many proprietary products were marketed under the general name of "fuel savers," the principal sales being in small packages for the use of householders; savings of from 10 to 40 per cent of the coal were promised. Starting with 1927, the use of proprietary or other mixtures of chemicals was extended to power plants and to treatment of coals at the mines, the claims of saving of fuel being subordinated to those of general improvement and, in particular, to reduction of clinker-ing. Such expectations are attractive and a number of companies experimented with or adopted the practice. There were reports of improved operation, but, apparently, few were confident enough that these improvements were due to the salts to publish the results of tests or their records of operation. There are no data on the yearly tonnage of coal treated but indirect reports suggest that it has been relatively large. Since its establishment in 1910, the U. S. Bureau of Mines has continually received inquiries about these products. At the request of other governmental departments it made tests in 1910, 1916 and 1930, using certain proprietary products applied according to the instructions of the vendors; these tests were made with eastern coals in domestic and

Jan 1, 1938