752 BELL SYSTEM TECHNICAL JOURNAL 



Method of Handling Work 



The round-robin tests on the Rockwell hardness tester, standard 

 test blocks and calibration of the machines were made by the American 

 Brass Co. (Waterbury, Bufifalo and Kenosha Branches), Bell Telephone 

 Laboratories, and the Western Electric Co. The tension tests were 

 also made by these laboratories. All of the rolling series were manu- 

 factured by the American Brass Co. Experience data on shipments 

 of non-ferrous metals were obtained jointly by the American Brass 

 Co. and the Western Electric Co. 



Methods of Test 

 Tension Test: 



As stated previously, the tension test is considered the basic me- 

 chanical test for cold worked materials. Complete data on tensile 

 properties were therefore obtained on the rolling series. All tests 

 were made in accordance with the standard testing procedure developed 

 for use in the Bell System and given in Appendix II. This method 

 has been in use over 4 years and has been found satisfactory to all 

 concerned. Three tension test specimens were made from each 

 sample. The tensile strength, proportional limit, percentage of 

 elongation in 2 in. and modulus of elasticity were determined for 

 each sample of the high-brass rolling series. The specimen 14 in. in 

 length was used which allowed the use of an 8-in. gage length Ander- 

 son ^ extensometer. The shorter specimen with the 2-in. gage length 

 was used for the clock brass and all later work. The tests were made 

 on an Amsler tension testing machine. The machine was carefully 

 calibrated, the load indications being correct to less than | of 1 per cent. 



The tensile strength and percentage of elongation on the high-brass 

 sheet rolling series are plotted in Fig. 8 against per cent reduction by 

 rolling. Nearly a straight line relation exists between the tensile strength 

 and percentage reduction by rolling and the values are grouped closely 

 about this line. The curve is plotted to show the average result, but 

 separate curves could readily be drawn for each thickness. This test 

 also has little sensitivity in the harder tempers where the curve becomes 

 asymptotic. The proportional limit is difficult to measure with any 

 degree of accuracy on account of the personal factor involved in 

 interpreting curves. The results in this case showed no definite 

 relation to any of the other values. 



In order to evaluate the test results, it is necessary to establish 



control limits to determine whether any variations in the data 



3 H. A. Anderson, "Tension Tests of Thin Gage Metals and Ligh Alloys," Proc. 

 A. S. T. M., Vol. 24, Part II, p. 990 (1924). 



