102 BELL SYSTEM TECHNICAL JOURNAL 



In making permalloy we use the purest commercial nickel and 

 Armco iron. Our samples for laboratory study are prepared by melt- 

 ing these metals in a silica crucible, using a Northrup high-frequency 

 induction furnace. The particular furnace which we use will con- 

 veniently melt a charge of about six pounds. An analysis typical of 

 the resulting billets is as follows: 



Ni 78.23 



Fe 21.35 



C .04 



Si .03 



P trace 



S .035 



Mu .22 



Co .37 



Cu .10 



The presence of other elements than nickel and iron is of course to 

 be expected after any practical method of preparation. To determine 

 their effects, samples were prepared in which the usual impurities 

 were present in various proportions. It was found that their presence 

 does affect the permeability of the alloys and that carbon is espe- 

 cially harmful. Since, however, the variations produced by slight 

 changes in heat-treatment are very large compared with those due to 

 small quantities of impurities we have found it unnecessary for most 

 purposes to require higher purity than that indicated in the analysis 

 above given. 



In our laboratory studies we have made it a practice to reduce the 

 billets through the forms of rod and wire to tape 3.2 mm. wide and 

 0.15 mm. thick. Accordingly test samples are available in a variety 

 of forms and conditions. Thin narrow tape is particularly adapted 

 to use in experiments involving heat-treatment, since it possesses a 

 high ratio of area to volume and is easy to manipulate. Fortunately 

 the entire nickel-iron series can be mechanically worked if sufficient 

 care is exercised and we have thus been able to use samples of the 

 same size, shape, and mechanical condition in all measurements upon 

 which we have based comparisons between alloys. This practice has 

 also made possible strictly comparable micrographic studies through- 

 out the series. 



