386 



BELL SYSTEM TECHNICAL JOURNAL 



Test Specimens. Accordingly, a large number of straight bars of 

 magnet steel were used for test specimens. Each of the bars was 12 

 inches in length. In the lot were bars of 3^ by 13^ in., /4 by 3^ in., 

 3/16 by ^/i in., 34 by 13^^ in., M by ^ in., and ^ by 3^ in. cross sec- 

 tions. All of the types of steel available were used, including 0.9 to 

 3.5 per cent chromium steel, 0.85 per cent manganese steel, 5 per cent 

 tungsten steel, and 20, 25 and 36 per cent cobalt steels. Along with 

 this range in composition a range in coercive force of from 40 oersteds 

 to 260 oersteds was available, accompanied by a variation in residual 

 induction of from 6,000 gauss to 11,000 gauss. 



Fig. 2 — The Babbitt permeameter. Used for making magnetic measurements on any 



kind of magnet steel. 



Experimental Procedure. The test bars were hardened in the man- 

 ner appropriate to each type of steel, except that some bars were 

 purposely quenched from a high temperature in order to secure low 

 values of Br. After hardening, the demagnetization curve for each 

 bar was determined by means of permeameter measurements, using 

 the Babbitt ^ permeameter and a Grassot fluxmeter with lamp and 

 scale. The control circuit of this permeameter is the conventional 

 circuit for a ring test, as shown in Fig. 3. The magnetic data are all 

 comparable since accurate measurements may be made with this 

 permeameter on all of the above mentioned types of magnet steel. 



