The Iron Age 1912-01-04: Vol 89 Iss 1

eS New York, January 4, 1912 Established 1855 Vol. 89: No. 1 A Dispatching System for Works Management The Bullard Machine Tool Company’s Charts Showing at a Glance Costs and Progress, Past and Future, and Insuring Orderliness, and Exact Balance of Equipment BY JOHN NELSON The dispatching system has come fo be recognized as an indispensable adjunct of the well managed manufactur- ing plant. In practice, however, it varies greatly in effec- tiveness, and in the very important item, the cost of main- achieved by each department by weeks, all as indicated. The departmental chart, of which there is one for each department, is shown on page 3. It gives the apportion- ment of productive hours between orders and the detailed tenance. it is a source of actual misinfor- mation. To produce the kind of re- sults that count, a simple, practical, cheaply operated method must be devised. It must automatically af- ford quitk information as to the status in the works of each order and each part which enters into the product. The department must give, at all times, the location of work in the shop. The system should also show the time and money which have been expended upon a lot of machinery at any mo- ment during its progress through the plant, and a close estimate of the time and money which must be spent before the order is completed, The Bullard Machine Tool Com- pany, Bridgeport, Conn., has had in operation for some time a dispatch- ing system which was developed and has been carried on by J. L. Sydnor, the company’s efficiency manager. The methods employed are essen- tially efficient as well as quite unique in several respects, when compared to accepted practice, and have proved to be invaluable in the operation of the large shops, which build vertical turret lathes and bor- ing mills and their accessories. Moreover, the system is operated at relatively small cost because of its simplicity. It is a most essential factor in its relation to the bonus system of wages under which the works are operated. In some cases it is of doubtful value, in others WHAT THE CHARTS SHOW OF WORK IN PROCESS OF MANUFACTURE The relation of the work on an order in all departments in respect to one another as to time. The date of beginning and the es- timated date of ending of work of each department, on a given per- centage of shop capacity. The estimated number of pro- ductive hours which each depart- ment will require for producing a lot of machines. The progress of the work through each department, week by week, in productive hours. The condition of each depart- ment and the exact apportionment of productive hours among orders at all times. The time and money expended to date on each order as a whole or by departments. An average hourly wage affords a sufficiently accurate basis for calculation. The estimated time and cost re- quired to complete any “order in process, as a whole or by depart- ments. The limits of productive hours between which the works can be operated economically. progress of the department as a whole. The general chart, which shows the progress of all orders in the works by total productive hours, is presented on page 4. The dispatching system goes far in its operation in that it gives imme- diate information of the progress of each part of a machine through the shops, its location by departments being always indicated by a system of cards shortly to be described. All figures used in the illustrations are, of course, entirely fictitious but serve to illustrate the principles in- volved. Charts and cards do not bring results in themselves. They are the reminders which insist con- stantly that the management «hall carry through a system which will procure economy and expedition. The personal equation is the great factor in all efficiency work, but without the system the human ele- ment is seriously handicapped. Each item of each chart has two lines, the unbroken line indicating the estimated time, the dotted oo the progress of each order, ter being added to weekly. master chart, the line of represents the period eae ing and ending the work of department. It does not repr productive hours, excepting department, in order: to» work indicated by this line, quire a certain number of tive hours in order to bring its cll ee ans F t i i The basis of the dispatching system consists of three graphic charts, which in combination constitute a constant source of exceedingly valuable information. What they show is given in the accompanying enumeration. In every-day practice the dispatching department has ascertained, inside of half an hour, the exact cost to date of every machine in process in the works and the estimated per of money which would be required to complete alt orders. The Graphic Charts The master chart reproduced on page 2 shows for each order the relation of each department to all of the _ others, together with the date of beginning the work and the estimated date of completion, and the progress to completion on the given date. In the other charts, the line of estimate represents moe prod urs, and its beginning and a Be dates as shown in the master Gach: correatjons as may come from changing as gg poe apreegs ge fine umns representing weeks and the horizontal columns, de- partments. In ‘the departmental charts the iy AE I AEG RCE TE oe oe RR ee a reemeaae 5: eigen THE IRON JUNE JULY AUGUST AG™ January 4, 1912 SEPTEMBER OCTOBER JOB B Drawing Pattern Foundry Cutting off Cleaning Chucking Turret Turning Milling Boring Mill Drilling Forging Grinding Stock Commercial Scrape Drawing Pattern Foundry Cutting off Cleaning Chucking Turret Turning Milling Gear Cutting Planing Horiz. Boring Boring Mill Drilling Forging Grinding Stock Commercial Scrape Sub. Ereet Erecting THE MASTER CHART Showing the Co-Relation of all Departments in Elapsed Time Between the Dates of the Beginning and Ending of Work on Each Order. —— Standard Time Allowance. -=-=- Actual Time Taken quired for a department to complete its work on an order is, for example, 800 hr., the line of estimate is placed on the eighth of the ten division lines. The line of progress, plot- ted each week, begins at zero, and ascends with the ad- vancement of the work. When 200 hr. have been consumed, the curve has reached the second line; at 300 hr. the third, and so on until, when the work is completed, the line should practically coincide with that of the estimate. Laying Out the Charts When an order is about to enter the works, the necessary data are placed on the master chart. If the lot of machines is a duplicate of one or more that have gone before it, the time required by each department is easily procured from the cost records. If it is a mew type of machine or changes have been made, a certain amount of estimate will have to be included, but experience has shown that figures arrived at by this means are, for practical purposes, as cor- rect as those taken from cost cards. ° The dispatching department, studying the order, begins the master chart at a certain date with castings. In the ease of the Bullard Company, which operates no foundry, this line represents the time which the foundry company that supplies this raw material is given to make deliveries. The planing, of course, cannot commence until the castings are received, and therefore the line of the department does not start until later. On the other hand, the cutting off, the screw machine and certain other departments may eco- nomically begin their portions of the work almost imme- diately if their conditions warrant it. Every master chart is reproduced on a grand master chart which is suspended on the wall of the dispatching de- partment, so that every order is there shown. At the Bul- lard plant the roll of cloth paper makes provision for the record of four years without renewing, The same apportionment of time is made on the depart- mental chart. This is a very important record. A mo- ment’s inspection tells the distribution of work in hours in a department. If, for example, the department has 1000 productive hours, one order may have 300, another 100, an- other 200, and so on up to the capacity. This allowance may be changed from week to week. If a rush order for-a certain type of machine is received it is a very easy matter to concentrate on one particular lot in process and sacrifice the others for the time being, or an examination of the situation may demonstrate the value of putting on more men and increasing the capacity of one or all departments, as the case may be. The third chart, as has been stated, contains the progress of all orders, the lines representing the total number of productive hours which it is believed will be required and which have been expended. Of course, in works like those of the Bullard Company, which manufactures on a large scale, the orders are of such size that the totals of produc- tive hours are very large. Yet it has been demonstrated time and time again that the actual time expenditure checks up within a very small percentage of the estimate, which means that the actual cost is almost exactly that of the estimate. Even where a new model is built and the work is not based on experience the variation is not a wide one. Practically the only difference is the saving which results January 4, 1912 OEPT. 27 BALANCE IN DEPARTMENT bach Space = 3000 Hrs. | JOB A Each Space =300 Hrs. JOB B [Each Space = 500 Hrs. [~ JOB D Each Space = 600 Hrs. JOB E |Each Space 500 Hrs. JOB F |Each Space = 1000 Hrs, JOB G | Each Space = 1000 Hrs. | | AUGUST THE IRON AGE 3 SEPTEMBER OCTOBER Standard Time Allowance Actual Time put in wee ky STOCK SPECIAL ORDERS ACTUAL CAPACITY CAPACITY PER CENT THE DEPARTMENTAL CHART, Giving the Apportionment of Productive Hours Among Orders, and Their Dates of Beginning and Ending from the efficiency system in use in the works, which in- cludes the bonus system. Dispatching and the Bonus System The company is constantly seeking more economical methods of manufacturing and means of decreasing wasted energy, which effort naturally has its effect on costs. Asa matter of fact, one of the very important functions of dis- patching is its co-relation with the payment of rewards in the bonus system. These are dependent one upon the other to a large extent in producing the best results. It is an accepted principle that if a workman can have all of a lot of parts for assembling his task is more anne done than if he gets them a few at atime. The ing system serves not only to bring together the full num- ber of parts, but the right parts at the right time. This advantage is felt not only in the assembling department but | Fi rs * Pas aia 7. 5 ae ~mrcamegteenne inpte MEE c LARS DP IE USB Orage ‘ 5 ‘ AUGUST 13 20 THE IRON AGE January 4, 1912 NOVEMBER ~ 20 5 12 19 iN SHOP JOB C Each Space =10000 Hrs; an0at- =50004-4 NT 5000; 5000| VACATION Each Space =10000 H al JOB E Each Space=10000 Hrs. Note: STOCK i ’ ——— Standard Time Aliowance SPGTIM. ORDERS pot ew Actual Time put in weekly ters SHOP ea CAPACITY == a NON PRODUCTIVE : SHOP ne ACTUAL - a | } T T ; + APA CAPACITY . PER CENT | | | | | | amntanlneeaisees idee mal l l petite C ai THE GENERAL CHART, Telling the Progress of All Orders in the Works in Total Productive Hours, as a Whole and as Individual Orders also in the manufacturing departments during the progress of orders through the works. Naturally, the line of estimate sometimes has to be ex- tended on its chart, not because of additional productive hours required, but because the capacity of the works has been reduced, owing to, say, depressed business conditions throughout the country. On the other hand, the progress line may be contracted, as compared to the line of estimate. If 5000 hr. is the estimate, the original line, giving the dates of beginning and ending, has to be made on the basis of a certain shop capacity. That is to say, when the chart was first made out, the works were running at 75 per cent. ca- pacity. If a reduction is made to 50 per cent. capacity, of course the period of weeks would have to be extended, in order to give the 5000 productive hr., and this change would develop in the line of progress. Or if the capacity was increased to go per cent., the order might be filled more quickly. However, this makes no difference whatever in the result of the operation of the system. Every Saturday a letter, signed by the superintendent, is sent to the foreman of each department, instructing him as to the division of time among these several orders. A sample letter reads as follows: December 23, 1911. Department Foreman: Please arrange to spend time as follows for the ensuing week: Job A 200 hr. Job B 300 hr. Job C 250 hr. should finish Stock order and accessories 100 hr. For the week ending December 16, your department stood as follows: Capacity 85% Men’s efficiency 90% Time on bonus 93% The foreman is consulted in relation to these details. If he cannot conform to the instructions, he makes a delay report, a blank for which is shown among the illustrations. The foreman may not have the material, or there may be some question of expediency. These conditions are sup- posed to be attended to in advance, but, of course, a cer- tain percentage of error must creep into any system. The delay report constitutes a check which automatically oper- ates in the dispatching system to correct errors of omission and judgment. The Foreman’s Bonus The foreman’s bonus is an essential factor in the com- pany’s profit-sharing plan. It is based on three elements: 1.—Capacity, the ratio of the total capacity of a department to actual production, which factor is governed by general business conditions; 2—the efficiency of the workmen, the ratio of actual time consumed in production to the time al- lowed for the work under the bonus system (the more efficient a foreman makes his departmert through the abil- ity and application of his men, the greater his own share of the profits) ; 3.—the percentage of time on bonus, which is the percentage of productive hours under bonus as com- pared with regular day work. Capacity is reckoned as 40 points, efficiency of workmen, 20 points, and the percentage of time under bonus, 40 points. The foreman, therefore, participates in earnings when there are profits for the company. When the works are running light, his share is diminished automatically. The amount of his profits depends on the degree of efficiency of the men under him who are working under the bonus system and for the same reason his effort is to keep the percentage of bonus hours as high as possible. The com- pany has under advisement an additional schedule in which the importance of the foremen is recognized. In the deparmental chart and in that of the works as a whole, an important item is the line of balance of produc- tive hours. The balance is the name given the number of hours required to complete all orders in process. Exper- ience has shown that there are well defined limits in this respect between which the works as a whole, on the one hand, and each department, on the other, can be operated economically. This line is shown in the upper horizontal division of the departmental and general charts. In the departmental chart these limits are given as 4000 and 18,000. The footings of the chart show the weekly reduction from the balance in “hours consumed.” In the sample chart, starting from the 18,000 maximum, the first week reduces the balance by 1285, and the line takes a pro- portionate drop. The balance falls each week until Septem- ber 17, when a new job is started, adding 7000 hr., and the line of balance goes up correspondingly. It is like a bank account, increasing with each deposit and dropping off with each withdrawal. January 4, 1912 This chart shows also productive time consumed on “stock special orders” and time which’ is non-productive. It also gives the capacity by departments, which in comparison to actual time consumed, produces the percentage capacity at which the department is operated—the factor which enters into the foremen’s profit sharing. The general chart is a continuous productive labor in- ventory. It tells the management the amount of work it has undertaken to perform; in other words, the liability that has been assumed. The exact status of the shop is recorded each week by individual orders and for the entire product. The chart tells what the company has spent each week and where the money has gone. If costs are exceed- ing the estimate or if a saving is being made, the fact is revealed graphically. The actual labor cost on which prices are based is told in advance and so is the date of comple- tion. If delay has occurred, the fact is immediately shown, for in such a case the curve of progress ceases to ascend, and continues horizontally through the period of idleness. Keeping Track of Parts in Process Not only does the dispatching system give accurate in- formation as to the progress of orders as a whole, but it also keeps track of each lot of parts of machinery. At any moment in the day the department can tell where in the shops. ‘any parts may be found. As a great many parts have to go to a considerable number of departments the lack of such knowledge is certain to produce chaotic ignor- ance. The system begets orderliness, and this in a manu- facturing plant means economy because of increased effi- ciency. The elemental unit of the system which looks after the parts is known as the master card, shown on page 5. Each part has its master card on which are columns for four orders. That used as an illustration keeps accurate track of the progress of the saddle of the crosshead. It contains on one side the drawing number, material, number of pieces, and so on. It shows that the castings were or- dered from the foundry company, under a certain order THE IRON AGE pv. OO. No. /-60/ a.c Std. Time 5S Name daddly Mansng. complete Job No. 125 THE DISPATCHING CARD, [he Telegraphic Message of the Systein number and include 50 pieces. The date of the order and the date when delivery is due are‘ given, and that of actual delivery is included. Incidentally, this is a constant check on the foundries and serves to regulaté deliveries, so that castings are received in the order in which they are re- quired. Consequently, the stock of castings on hand, in other words, the money tied up in raw material in the yard and storehouse, is materially reduced. This card goes into a drawer which is subdivided by departments. It is the tell-tale of the dispatching system. It passes in the drawer from one department to another, just as the work passes along through the shop. Therefore, to find out where the work is one has only to locate the card. On the reverse face of the card against each depart- ment is a column, as also here shown, giving the bonus standard time in hours and columns shewing the date of receipt of the lot of parts and the date of their completion, the two columns being known as “in” and “out.” Duplicate columns for the time are for use where the lot is split in two sections. The Blue Dispatch Card On this page is shown the blue dispatch card. It is the telegram message which is sent out from the office to notify the fore- man of a department to start work on a lot of parts. Its operation is analogous to the dispatching system of a well-managed rail- road. The railroad has a chief dispatcher who is responsible to the superintendent only. His division is divided into blocks, each of which has its station operator. The chief dispatcher directs the operation of all trains in all details. The responsibility of a station operator is to report the entrance and departure of trains in relation to his block. In the Bullard shops the superin- tendent’s office is the chief dispatcher. The shop clerks are the station operators. The dispatch cards- are the messages between the superintendent's office and the manu fac- turing departments. ume Gaoldly DIVISION bros Gal mince ne. /~60/ Cb The card tells the foreman the standard pwe.ne: //586 uarenia “Ooo. size ahaa time by hours for his bonus system and gives the operations required for the work ee 6 (yo * eV in the department: The moment the depart- 125 / ASF bo. ; |_ ORDERED FROM |_ ORDER No. |_No. ORDERED |_DATE ORDERED DATE DUE DATE and Ne. RECEIVED THE MASTER CARD, Which eg Track of Each Part Through ae usecase in Upper Half of the Cut Shows the Reverse Side o the Works. The the Master Card, Showing the Standard Ronte Through the Works, Giving Standard Time for Bonus System and the Actual Time Consumed jn Each Department ment has finished the order the card returns to the office as a notification of progress. It is an order to move the master card to the next department in the drawer, and to forward another dispatch card to the corre- sponding department in the works. The operation cf the dispatch cards continues until finally the master card has left the drawer, the work being completed. Wher all the cards representing an order have been assembled as finished the lot of ma- chines is ready for delivery to customers. The dispatch cards go to another draw- er, where they tell the story of the progress of the machine as a whole. The drawer is divided longitudinally into two divisions, one for work which is to be done, the other for completed work. The entire group representing an order is made ready before rt vONT FN eRe Hh CTE a Sok mace Medco Seth an sce ELT LT LE ALE TT TT ERTL RD Ea cs PINT _chines in order to do its share of the TRAIT ti eT a eT MES NAMES LLLP RNR RRS RAR aa 6 | | THE IRON AGE January 4, 1912 manufacturing commences, and the thick bunch of cards is placed in the left-hand side of the drawer. Upon their return, one by one, from their departments, they go - THE BULLARD MACHINE TOOL CO. DEPARTMENTAL DELAY REPORT. Dept. into the “finished” side of the drawer. A Wath in Gib Diparenent wi tel tetted Rew Seteate exlem Weiowins ented b ferebbed Hemet: onpen glance shows the progress of the machine. In the case in question, when all the blue cards having to do with the saddle have changed sides, the order for the saddle has been completed. Both master and dispatch cards are used over and over again, with each order containing the part. A tag, of which a reproduction to one- half size is here given, accompanies the work for the convenience of workmen and inspectors. As will be noted, it gives a good deal of information which is frequent- ly of value. In addition, there is a special replacement card and its master card, for use when it is necessary to replace any work where defects have developed. In this connection a card is provided as a “Spoiled or Defec- tive Material Report.” This card, as indicated, is filled out by the inspector or foreman who rejects work or material. Obtaining a Balance of Equipment In the beginning the system demonstrated immediately that the shop lacked balance in its equip- ment, a very common condition in works using machine tools. For example, the original chart lay-out for orders contem- plated and in progress indicated that one department would require additional ma- Job 125 THe B. M1. 0. ORDER DEPT. work. When the foreman of the depart- Name ment was told this he laughed and said that, on the contrary, he had too many ma- chines and that he was contemplating lay- ing eff a man or two. The system did not lie. The new machines were installed and the balance was found to have been ad- justed. The same condition existed to a greater or less extent in cther depart- ments. We have already spoken of the order- liness of the system and its results. An excellent proof is afforded in the erecting department, which had been suffering be- cause of the failure of the works to de- liver parts of machines coincidentally. Men were laid off some cf the time when they should have been working. The as- sembling department was rushed or idle by Style No. ivision Grou el Division Gross Gaul Piece No. /-60/ ac __ Hole Drawing No. //586 MATERIAL The Departmental Delay Report this means that instead of parts of machinery being stored at haphazard through the shop départments in various stages of completion, they are gathered together ready to be assembled by the erecting department. The erection of a machine is now never begun unless all of the units which enter into it are tested and in the stock room. Advantages in Determining Delivery Dates MATERIAL The effects of the dispatching system where works are rushed to 100 per cent. 50 capacity are worthy of special considera- iene" tion. One great evil of rush times is the inability of manufacturers to quote deliv- eries intelligently. They accept an order for shipment months ahead, and on the agreed date the completion of the machine is more or less remote. The resulting controversy with a customer is always un- pleasant. The trouble is that machinery builders often can only guess at produc- tion, which was demonstrated in numer- ous cases five years ago. The Bullard dispatching system must obviate a great deal of this trouble. The department can lay out work for months ahead and come very close to the exact date of completion of each order. With a known producing capacity this is com- paratively easy of accomplishment. When works are running at less than 100 per cent. a variation of capacity must be taken into account in determining the date of A. intervals. Under the dispatchiug system The Tag Which Accompanies the completion. This constitutes an unknown this department, working continuously, so increased its output that it became neces- sary to add to the erecting force, which also works with no let-up caused by delay.in manufacturing departments. The erecting department stock room is a striking com- mentary on the benefits of orderliness. It had been found practically impossible to create a stock of assembled units. To-day a large room is required for their storage. Analyzed, THE BULLARD MACHINE TOOL CO. SPOILED OR DEFECTIVE MATERIAL REPORT. ‘Time Woy sa ek rae Ne te Ed a, _ Fore THIS CARD MUST BE SENT TO SUPT.’S OFFICE AT ONCE WHEN FILLED OUT. SPACES BELOW MUST NOT 58 FILLED IN BY FOREMAN. Order Through the Works quantity, of course. But where this fac- tor is assured, the date can be determined far in advance. The only change would be a possible earlier delivery, which would occur if capacity were increased beyond 100 per cent.; that is to say, if a night shift were put on and no effort made to increase the volume of prod- uct. If capacity were increased and if the volume of work going through were to be made larger, the dispatching system would tell accurately how much greater the volume could be made without disturbing deliveries. The company is mas- ter of the situation at all times because it knows the facts within very narrow limits. L. Vogelstein & Co., 42 Broadway, New York, furnish the following figures of Ger- man consumption of foreign copper for the 10 months ended with October, 1911: Tons. SUMTUS 2s oo oss: oa bree pee 157,934 BOGIES ae ccs s 60%ban0s' baa eee 7,507 EET en ee 150,427 This compares with a consumption dur- The Spoiled or Defective Material Report ing the same period in 1910 of 141,472 tons. Of this quantity 138,991 tons was im- ported from the United States. The Brinell Test for Automobile Steel An Investigation Showing That Hardness as So Determined Gives No Dependable Measure of Strength en BY DENISON The automobile industry has now become firmly estab- lished in the commercial world of to-day. Many of the smaller companies which sprang into existence with the first mushroom-like growth of the industry have fallen out of the race, due either to a lack of financial backing or to a poor car. The stronger and more reliable companies, having attained their stride and solved many of the me- chanical difficulties, are now devoting a part of their ener- gies to the standardization of parts and a study of the quality of the material used in the construction of the car. From the more recent efforts of the Society of Automobile Engineers we have a standard table of hardnesses for the testing of steel used in the construction of the car. In this article the author wishes to consider the application of the Brinell hardness test to the steel frames. j The two general methods of hardness testing in this country to-day are the Brinell test and the Shore sclero- scope method. The former measures the resistance to in- dentation when the metal is stressed beyond the elastic limit by a concentrated load. The latter is based upon the re- bound of a hard body from the surface to be tested. It is immaterial to this discussion whether we agree with Auchy pr with Shore as to the exact qualities which the re- spective machines measure, as it is generally agreed that the Brinell hardness results are indicative, in general, of ‘the static properties of the metal. But is the Brinell hard- ness test applicable to sheet steel as used in pressed steel automobile frames? The Steel Tested and the Method It is now generally recognized that the frame steel must be given a special heat treatment in order to relieve the strains set up in the rolling and later in the forming operation. This treatment, when exactly conducted, should 130,000} ; m | e Peg 120,000 >. ’ } ~ | . . ‘ 1 } om | . ‘~ Ls * s . ‘| Ps 83 = 110,000} 3 : Pc ac] , thie, i : re Seen meer & | fei 2 ab ae a . = | , oe = 100,000} ee 1 a Ta } : . ° 2 . t , ais : ‘ > : z —} - gs i } i st 2 itesesi | ~ ° . ? . ee : : } 90,000) ee rice 2" | 4 ? er a | : ek . ae 80,000} “ Fin te i*e | > 130 140 15 100 170 1% 19 200 210 20 Hardness Number Fig. 1—Hardness Number and Tensile Strength produce a finely crystalline, homogeneous structure, and bring. out the maximum static and dynamic properties of the steel. It was upon such heat-treated stock that the fol- The results are taken from some lowing tests were made. *Metallurgical Engineer, Reading, Pa. or Other Properties K. BULLENS* 3000 tests made during the past year in the laboratory of a leading automobile frame manufacturing company. They represent both experimental work and tests made on regu- lar side bars. The material varied between % in. and %4 Pee ! } i aremcteliagee ; el 7 - 110,000r- potree el | | x“ a oe . : . ’ * 7 } j ? *| e 100,000 | bo ~ Saal 2 * § | a4 z y .4 2 Tits a ad = 90,000/- T 1 “ | = + * 8] ee hoe | = Be ars z . | +4 ey | I p : | b ; } . ! i = 30,000) ote b | +4 5 2 } . os j - | > } : jo i i i i a } . } rs Stn ie —| e r 3 - le @ | ’ i oo? . Ts Sinner eg | | . ae L . , LJ : ee a ed eT hd ob eee tee] | Rao oe Bey 60,000}. ve r _| “an od . e Pe *| | 6 . e 4 | | + | ‘ = ———+ 1» 10 1 io 10 iO 19 BO 20 BO Hardness Number Fig. 2—Hardness Number and Elasti¢ Limit in. in thickness—a variance which had no effect upon the hardness results. The steel, as used in these tests, may be broadly classed as follows: » 1. Chrome-nickel steel, hot rolled. 2. Chrome-nickel steel, cold rolled. 3. Special carbon steel, hot rolled. 4. Special carbon steel, cold rolled. The physical tests were made with a 100,000-lb. Olsen automatic-autographic testing machine. The hardness re- sults were obtained upon, an Alpha-Brinell hardness ma- chine, set for a pressure of 2830 kg. The diameter of the hardness impression was read with a specially constructed microscope and scale, by means of which a difference of 0.01 mm. could be readily estimated. As much has been said about surface decarbonization a number of tests were made to determine if this actually took place. Readings were made upon the surface and upon the steel taken at depths (obtained by grinding away the surface) varying to 0.030 in. No difference whatever could be observed in these readings. The only precaution necessary was to see that there was no hardened oil upon the steel. In order to give a clean surface, however, the spot at which the hardness test was to be made was slightly polished. Microscopic investigation confirmed the above results, no skin or area of decarbonization being detected. The hardness tests were made upon the ends of the pieces used in the pulling tests, care being taken to get entirely away from the strained structure. Hardness as Related to Other Physical Properties Figs. 1, 2, 3, and 4 show graphically the relation of the hardness numbers to the tensile strength, elastic limit, per cent. elongation, and per cent. reduction of area of some 250 tests upon heat-treated, hot rolled, chrome-nickel steel. 7 sears ceeneterreneeeenaatas OR eS ee ~~ ee a 1 Sal sayhareeny ; | ; : : eid ik Wi enbivevmoahnicurme IR ce ae Ps : 8 THE IRON AGE The last three plots show, as would be expected, no locus of points. There exists, of course, a varying relation be- tween the elastic limit and the tensile strength, so that the plot shown in Fig. 2 will bear a certain relation to that of Fig. 1. This would also be true to some degree of the per cent. of elongation as shown in Fig. 3. But as the Brinell test strains the steel beyond the elastic limit it is not to be expected that a definite curve would result. It is the relation of the hardness to the tensile strength, however, with which we are mainly concerned. Accord- ing to general experimental work upon other steel we would expect some well-defined curve or locus. From Fig. 1 we can readily see that this is not the case with the sheet steel under discussion; no such locus exists. There is only a broad area or band which shows a difference of 30,000 Ib. per sq. in. between the maximum and minimum readings. For example: a hardness number of 174 repre- sents a range of 86,000 lb. up to 116,000 lb.; number 187, 83,000 Ib. to 115,000 Ib.; number 205, 94,000 Ib. to 129,000 Ib. Similarly, 86,000 Ib. is represented by hardness numbers varying between 131 and 195; 94,000 lb. by 144 to 205; 103,000 Ib. by 156 to 224, and so on. All the results given in Fig. 1 are obtained from hor 26 ° acme on 24 . sh es 22 ° ° +20 =. . vo S $ - >—~—6-+-o-—--@ $ o ° . . a s = 18 + + - o- i2—- ° 56 , i ; e . . a . > . a = * ° - £ =a o- 7 a e = . . . * *“% *%e . . * 16 + ° wo jo—eo—| 2_<9 ls + . *-* 8 ° ee q ° ° ° ° +» ° joe . 14 . — 12 : : ra J 140 «6150 160 =—s:170 180 1890 200 210 220 Hardness NurLer Fig. 3—Hardness Number and Elongation rolled, chrome-nickel sheet steels of approximately the same chemical composition. The carbon will not show a variance of 10 points between the high and low values. The chromium and nickel are constant within limits of 20 points. The manganese values remain within 7 or 8 points. The phosphorus, sulphur and silicon are all low, so that thev have no bearing upon the subject under discussion. It was thought, however, that even these slight differ ences in chemical composition might possibly affect the steel as regards hardness. A series of tests was therefore made, using samples taken from the same part of the same heat, and rolled under exactly the same conditions. The chemical composition of each series of specimens was con- stant. It might also be stated that segregation was en- tirely absent in these high grade “quality” steels. A Transference Number In order to show the results obtained in a clearer light computations were made by which a “transference num- ber” resulted. This number was obtained by dividing the tensile strength by the hardness number. An average transference number was found for each heat. By using this new result, and the hardness number of each test, a set of new, or computed, tensile strengths was obtained. January 4, 1912 The difference between the actual and computed tensile strengths should show the degree of accuracy of the hard- ness test. A comparison of the average transference num- bers of the various heats should show the reliability of the hardness test as applied to steels of approximately the same chemical composition. Tables 1, 2, 3, and 4 show the results obtained. The average transference. numbers of the four lots of steel taken are 493, 549, 571 and 597, a difference of over 100 points between the first and last. With a hardness num- ber of 200 this would result in a possible error of 20,000 Ib. between the actual and computed tensile strengths. From experiments with numerous other heats of the same ap- prox'mate chemical composition the author has found aver- age transference numbers varying between 460 and 630. One can readily see where this will lead if that of any one heat, or even an average of several heats, is used. Taste 1.—Stocx SO. 933. ‘Roi Hardness Computed Diff. in Lb. Sq. In. No. Trans. No. a & ae l 82,900 163 508 80,400 — 2,500 2. 82,900 163 508 80,400 — 2,500 3. 84,200 1706 495 » 83,800 — 400 4. 84,500 178 474 87,700 + 3,200 5. 85,000 163 521 80,400 — 4,600 6. , 86,100 174 495 85,800 — 300 a 86,300 195 443 96,100 -+-10,200 8. 87,000 170 512 83,800 — 3,200 9. 88,400 178 496 87,700 — 700 10. 88,800 187 475 92,200 + 6,600 11. 89,000 187 476 92,200 + 6,800 12 89,300 178 502 87,700 — 1,600 13. 90,060 178 505 87,700 — 2,300 14. 90,800 182 498 89,800 — 200 15. 91,300 187 488 92,200 + 1,100 16 91,800 195 471 96,100 + 4,300. 17 92,600 187 495 92,200 — 400 18. 94,800 195 486 96,100 + 1,300 19. 96,300 191 504 94,200 — 2,100 20. 97,200 178 546 87,700 — 500 ai. 86,300 178 485 87,700 + 1,400 Average Trans. No., 493. TABLE 2.—Stocx 22,059. 7.3 Hardness Computed Diff. in Lb. Sq. In. No. Trans. No. a. Bs Tid. ] 73,000 156 468 85,700 +12,700 2 78,750 139 558 76,400 — 2,350 3 78,750 44 540 79,100 + 350 4 78,800 39 560 76,400 — 2,400 5 79,300 156 507 85,700 + 6,400 6 82,450 44 572 79,100 — 3,350 7 86,600 41 635 77,500 —12,100 Average Trans. No., 549 TasBLte 3.—Srocx 6959. ] 76,800 56 492 89,100 + 12,300 2 81,600 150 543 85,700 + 4,100 3. 83,800 170 492 97,100 +-13,300 4 89,300 44 620 82,250 — 7,050 5. 90,500 156 579 89,100 — 1,400 6. 93,400 170 549 97,100 + 3,700 7. 102,000 60 637 91,400 —10,600 8. 128,800 205 623 117,200 —11,600 9 129,200 214 602 122,300 — 6,900 Average Trans. No. 571. Taste 4.—Stocx 5838. aw Hardness Computed Diff. in Lb. Sq. In. No. Trans. No. te Bi “bate 1 93,100 163 571 97,200 + 4,100 2 94,200 153 571 91,200 — 3,000 3. 94,200 182 517 108,800 -+- 14,600 4. 94,500 160 590 95,400 + 900 5 95,100 163 583 97,200 + 2,100 6. 95,200 168 595 95,400 + 200 7 98,200 166 591 99,000 + 600 8. 100,000 163 613 97,200 — 2,800 9. 100,000 170 588 101,500 + 1,500 10. 101,600 160 635 95,400 — 6,200 ll. 102,300 156 656 93,200 — 9,100 12. 103,000 170 606 101,500 — 1,500 13. 104,760 170 616 101,500 — 3,200 14. 105,000 170 617 101,500 — 3,500 15. 105,000 178 590 106,300 + 1,300 16. 108,000 174 619 104,000 — 4,000 17. 108,100 176 613 105,100 — 3,000 18. 108,500 187 579 111,800 + 3,300 Average Trans. No., 597. Variation in a Single Heat Let us now consider the error found within the heat itself. Take the largest variance in each heat. Table 1 shows a range from — 2500 lb. to + 10,200 Ib.; Table 2 from — 12,100 lb. to + 12,700 lb.; Table 3 from — 11,600 lb. to + 13,300 lb.; Table 4 from — 9100 Ib. to + 14,600 Ib. These show a possible error either way of 15,000 lb. be- tween the computed and actual tensile strengths. The tabulated results have been arranged according to ascending tensile strengths. The columns headed “Dif- ference in Tensile Strengths” show no corresponding or regular arrangement of positive or negative errors accord- ing to the strength of the steel. The error is quite as likely to be a plus as a minus quantity. a i , January 4, 1912 What has been said of the hot rolled, chrome-nickel stock is also true of the cold rolled steel. The “contrari- ness” of the Brinell hardness test for this class—or thick- ness—of matérial is well shown by Table 5. The “A,” or first set of tests, and the “B,” or second series of tests, were made on the same lot of steel, the latter being made about a week after the former. The difference in the aver- age transference numbers is 144 points. This would Taste 5.—Stocx CR. SO. 930. ae Hard. Trans. Hard. Trans, Poe No. No. Ti i No. No. 1. 122,900 205 598 81,000 205 395 2. 129,900 224 579 87,900 195 450 3. 131,000 224 584 88.500 187 473 4. 134,700 224 601 89.400 195 458 5. 137,000 214 640 92,700 195 474 6. 143,500 232 618 98.000 195 ~ 503 70° 4a ws sere poe 103,400 224 461 Av. Trans. No., 603. Av. Trans. No., 459. give an error of 42,600 lb. if the transferences number of the first series had been used with (1) of the second series. This needs no further comment: it speaks for itself. Results obtained upon hot and cold rolled special car- bon steel are given in Tables 6 and 7. The hot rolled stock has shown a nearer approach to a dependable hard- ness and transference number than all the other steels tested. But even this steel has many exceptions and often gives incongruous results. The Hardness Test in Practice \s an example of the application of the Brinell hard- ness test to the purchase of frames, take the following case: An automobile company purchases a large number of frames from the manufacturer, stating that certain limits of hardness must be observed. The frame-maker will probably use several heats of steel for the order. The data previously given in this article have shown that there exists a possible error of at least 20,000 lb. by using one general transference number or even an average of several, and to be added to this is another possible error of 15,000 Taste 6.—Srocx 3HR. To Hardness Computed Diff. in Lb. Sq. In No. Trans. No. T;.3. T,.3. 73,700 153 482 79,700 + 6,000 74,800 139 538 72,400 —— 2,400 3 75,800 160 473 83,350 + 7,550 4 77,000 153 503 79,700 + 2,700 5. 77,200 163 474 84,900 +" 7,700 6 82,000 163 503 84,900 + 2,900 7 82,000 174 471 90,650 + 8,650 8 83,500 163 512 84,900 + 1,400 ‘ 87,600 178 492 92,750 + 5,150 10 88,500 163 543 84,900 — 3,600 ] 92,500 182 508 94,800 + 2,300 l 93,500 191 573 99,500 + 6,000 13 95,000 178 533 92,750 — 2,250 14 95,500 182 524 94,800 - 700 15 96,400 182 529 94,800 — 1,600 16 97,300 182 533 94,800 ~ 2,500 17. 99,300 178° 557 92,750 — 6,550 18. 104,300 199 524 103,700 — 600 19, 105,600 182 580 94,800 —10.800 20 115,600 205 563 106,800 — 8,800 Average Trans. No., 521. Taste 7.—Stocx 3CR. T..3. Hardness Computed Diff. in Lb. Sq. In. No. Trans. No. 7. a 2s 1. 63,750 144 442 74,200 4+-10,450 é. 65,700 133 494 68,500 + 2,200 69,200 131 528 67,500 — 1,700 72,800 150 485 77,300 + 4,500 5 74,000 131 565 67,500 — 6,500 ¢ 76,150 150 507 77,300 + 1,150 7 81,500 156 522 80,300 — 1,200 8. 84,800 174 487 89,600 + 4.800 . 93,300 187 498 96,300 + 3,000 10 93,700 195 498 100,500 + 6,800 11, 94,000 174 540 89,600 — 4,400 12 96,600 187 ; 516 96,300 — 300 3 109,700 174 630 89,600 10,10) Average Trans. No., 515. lb. from each heat in itself. This makes a possible cumu- lative error of 35,000 lb, The manufacturer heat-treats the steel according to a precise treatment worked out by the metallurgist. Under careful supervision, this treat- ment gives a side bar of uniform elastic limit, tensile strength and per cent. of elongation, with a minimum limit in each case. This is done regardless of hardness, as the manufacturer wishes to give his customer the best that is in the steel. The purchaser may refuse certain frames, stating that they do not fulfill hardness specifica- tions. What is the result? The automobile company either accepts the word of the manufacturer that the frames are good, or else demands a re-treatment. If the THE IRON AGE 9 frame is re-treated in order to make it answer to hard- ness specifications an inferior article usually results. It is a case of careful metallurgical work, with constant test- ing, as opposed to the uncertainty of the Brinell hardiness test. Practical examples of this condition of affairs have come to the author's attention. Frames under discussion have been cut apart and test pieces cut from them. In every instance the careful heat treatment has won out against the hardness test. In some instances it was found impossible to harden the steel sufficiently to reach a certain hardness number, and yet the steel showed up perfectly satisfactory when tested in the pulling machine. It has even been found that steel answering to the same chemical specifications, but purchased from different steel com- panies, would give largely variant Brinell hardness tests Again, heats purchased from the same company, rolled from the same temperature after soaking, with the same number of passes and per cent. reduction but having a large variance in chemical composition, have been found to give approximately the same transference numbers, Accounting for Variations It is very difficult to explain the seeming ‘contradic- tions which experience has shown. However, the large number of contributing causes might be broadly classed as follows: (1) Inherent; (2) mechanical; (3) structural. 56 at 5 — Reduction of Area, per cent 4 pie iw lo it io 0 wo Ww wo 20 2 Hardness Number Fig. 4—Hardness Number and Production of Area Inherent causes include the grade of the material used in the manufacture of the steel, and the chemical composi- tion. Mechanical causes include the temperature and length of soaking, the rapidity of reduction, and care used in rolling. Structural causes may be divided into anneal- ing and heat treatment, each influencing the actual condi- tion and proportions of the various microscopic constitu- ents of the steel. The annealing may or may not remove all the effects of rolling and previous heating. The ordinary furnace anneal will simply give a soft steel. The effects of any overheating in the soaking operation or too rapid reduc- tion in the rolls are not generally removed. This is due to the fact that a long and continued heating at the higit temperature necessary would result in such a degree of oxidation of the finished steel shects that they would be practically useless. Although these strains are generally relieved by the further heat treatment, a few of the worst sheets (structurally speaking) may still retain hard or soft spots which will influence the hardness test. The general microscopic constituents of heat-treated, chrome-nickel steel frames are troostite, sorbite, mar- tensite and pearlite. The troostite is generally the largest in amount, and the others in the order given. The pre- dominant constituent will be determined, first, by the chem- ical composition of the steel, and, second, by the heat treatment necessary to give certain physical tests. In chrome-nickel steel the eluments having the greatest in- filxence upon the physical properties are carbon, chrome, Net meee te 2 mere SARE NPN TA AEE ES PU MES RU tle — . P| B F 10 THE IRON AGE nickel, and manganese, in the order given. ‘ These also in- fluence the recalescence point of the steel, which becomes a very important factor in the correct quenching tempera- ture. With all these variables introduced, and with stand- ard physical specifications in view, there will necessarily result varying proportions of the microscopic constituents. And as these constituents have each a different hardness the complexity of the problem is readily recognized. The Brinell hardness test as applied to the purchase or sale of heat-treated, sheet steel automobile frames is not dependable. It is an excellent supplementary aid to labora- tory work, but it necessitates a thorough knowledge of the steel in question and constant bearing in mind that the test is only generally approximate and far from infallible. The American Brass Company With the opening of the new year the American Brass Company, Waterbury, Conn., becomes an operating com- pany instead of a holding company as heretofore. Its sub- sidiaries, the Ansonia Brass & Copper Company, Benedict & Burnham Mfg. Company, Coe Brass Mfg. Company, and eer 8.G. TRACK SHIPPING-.BUILDING & January 4, 1912 the industry in Waterbury, Ansonia, and surrounding towns decided to amalgamate their interests with a view to sim- plicity and economy in production. Consequently, at the formation of the American Brass Company, it acquired not only the control but the entire capital stock of the constituent corporations. It was first capitalized at $15,000,000, but this has twice been increased by $2,500,000, so that the present capitalization is $20,000,000. Of this, however, only $15,000,000 has been issued. Most of the other companies in the merger were taken in soon after its formation, and the recent policy of the directors has been not to increase their percentage of control over the brass industry. The present plan was authorized by the directors, who are also the principal stockholders, under authority vested in them by the by-laws. The annual meetings of the con- stituent companies take place on the first Tuesday in Feb- ruary. The American Brass Company is the only stock- holder, and it is expected that measures will be taken at that time for the final dissolution of the several corpora- tions. The new arrangement will make no change in the chief 8.G. TRACK 8.G. TRACK Waterbury Brass Company, become branches known by the names of the old corporations. Each branch will have its own accounts, as the companies do now, but checks will be drawn to and by the American Brass Company. Besides the corporations that are thus extinguished, the American Brass Company controls the Chicago Brass Com- pany, Kenosha, Wis., the’: Waterbury Brass Goods Corpora- tion, and the Ansonia Land & Water Power Company. They are owned by the present subsidiaries, and under the new régime they will become the property of the American Brass Compariy through stock ownership. They may even- tually be consolidated with it, as their parent companies are about to be, but that step has not yet been reached. Holmes, Booth & Hayden, another Waterbury brass house, was merged with the Benedict & Burnham Company some time ago and now becomes merged with the main corpora- tion. The American Brass Company was incorporated in Con- necticut in 1903 to hold the stock of the four principal con- stituent companies. The brass industry had long centred in the Naugatuck Valley, and the stock of