NOTES ON CHAIN CABLES. 167 



badly, and subjects it to internal stresses incident to the cold working and damages 

 it in general to an extent that could not be permitted in high-grade chain. More- 

 over, the process cannot be applied to sizes above i}^ inches. 



Numerous patent machines have been devised to form chain links of large size, 

 all of which are open to some objection. Descriptions of these machines may be 

 obtained from the Patent Office. 



Side-welded links are more easily formed than end-welded links, as the long 

 ends to be thrown together furnish ample leverage for accomplishing this, whereas 

 with the end-welded links the ends are so short that the force applied must be so 

 great as to crush the material while bending it. 



A machine for forming side-welded links by winding on a rotating mandrel is 

 in successful operation at the Bradlee Chain Works, Philadelphia, Pa., and at the 

 Lebanon Chain Works, Lebanon, Pa. A machine of this type was purchased some 

 years ago by the Boston Navy Yard, and is now installed in the chain shop at that 

 yard. Experiments were made with a view to adapting it to the winding of end- 

 welded links, and a mandrel was designed for this purpose. This experiment was 

 partly successful, but not entirely so, for the reason that the short ends were not 

 thrown in as far as desired, and the link was somewhat bruised and drawn at the 

 quarters, the scarf curled up and made somewhat concave. 



The general objection to all present end-weld link-forming machines is that 

 they damage the bolt by bruising it at some point, or by drawing away the mate- 

 rial in such manner as to reduce the section (this is shown in Fig. 20, Plate 96). 

 It became apparent, therefore, that rather than try to bend the short ends of the 

 link, it would be better to form them by a forging process whereby the end of the 

 link, being forced into a die, might be given any form that is desired, even to the 

 extent of reinforcing rather than weakening the link at the quarters. 



This was finally accomplished in the Ajax upsetting machine in the manner 

 shown in Figs. 21, 21 A, Plate 98, the operation requiring one heat and one stroke 

 of the machine for each end of the link. 



Scarfing. — The process for forming the end of the link should further pro- 

 vide for the proper forming of the scarf. This is a matter of most serious con- 

 sideration, in that it affects the quality of the weld. As shown in Fig. 16, Plate 

 94, the scarf as formed by the hand process is hollow, so that when the two scarfs 

 are brought together a pocket is formed in the center, which tends to collect slag 

 and dirt, and to prevent the surfaces of the scarfs coming together at the center of 

 the link (see Fig. 22, Plate 99). It thus happens that many of our links fail by 

 parting in the weld, showing the improper contact at this point. A desirable form 

 of scarf is shown in Fig. 22A, Plate 99. It should be noted that the surfaces of this 

 scarf are convex rather than concave, so that the first contact in the beginning of 

 the weld is made in the center of the bolt in such a manner as to force the slag and 

 dirt outward as the weld is worked toward the surface. Furthermore, the edges 

 of the scarf are somewhat rounded in order to prevent their burning away or 

 growing cool too rapidly, as occurs with a sharp-edged scarf such as formed by 

 shearing or by sawing. 



