TO SHIP CONSTRUCTION. 231 



A few months ago, the Welding Research Sub-Committee sent out questionnaires by 

 means of which it obtained statements of the current employed at that time by various con- 

 cerns for work of various dimensions and shapes. In the following table are given die values 

 employed by five different undertakings for welds equivalent to the welds in this series of 

 tests, and in the last line are given the values indicated by this study of the results from the 

 Wirt-Jones tests. 



Current for Welding Half-Inch Thick Ship Plate. 



Current 

 recommended Kind of 



(in ani/'crcs) . electrodes. 



British practice 75 to 120 Flux-covered 



Atchison-Topeka and Santa Fe Railroad 115 to 125 Bare 



New York Central R. R 110 to 150 Bare 



Chicago, Rock Island and Pacific Railroad 80 to 135 Bare 



Indications from some advance results from the 



Wirt-Jones tests, at least probably 175 to 200 Bare and covered 



It was interesting to note, from the data obtained from the 20 welds of the Wirt-Jones 

 series, that the operators who employed the largest current did not always employ electrodes 

 of the largest cross-section. In other words, no consistent relation was observed between the 

 current employed and the size of the electrodes, different operators employing different cur- 

 rent densities. In fact the current densities employed for these tests ranged all the way from 

 6,000 amperes per square inch up to 14,000 amperes per square inch. In Mr. Cox's paper is 

 given a table of recommended sizes of electrodes for various currents foT covered electrodes. 

 I have analyzed these values and found that the recommended current densities range in most 

 cases from 4,000 up to but little over 6,000 amperes per square inch. For covered electrodes 

 as employed in England, the use of lower-current densities than are necessary for bare elec- 

 trodes is claimed as an advantage. I doubt whether they are right on this point, and I am 

 also inclined to believe that they use too large electrodes. Although they obtain good welds, I 

 doubt if they are obtaining as good welds with the use of these low currents and 

 low-current densities as could be obtained if they should redesign their electrodes so as to main- 

 tain an equal rate of consumption of the flux covering and of the core when working with 

 the high-current densities. With an American design of covered electrode represented in 

 the curves by four points, high-current densities are employed and with decided advantage, 

 so far as relates to the mechanical characteristics of these four welds. 



The four lower curves of the chart represent an attempt to ascertain whether any con- 

 clusions could be deduced with reference to the effect of the current density on the mechanical 

 characteristics of the weld. I call the attempt by no means successful ; nevertheless, a great deal 

 of interest can be learned by inspection of these curves. Thus we observe that the current 

 densities range, as already mentioned, from 6,000 amperes per square inch upwards, whereas 

 British practice is confined to value below 6,000 amperes per square inch. Obviously, if for 

 making a given weld, with a given current, an operator uses an electrode of half the cross- 

 section employed by another operator, the electrodes will bum away at twice as great a rate- 

 In view of the present practice of employing straight, short lengths of electrodes, this would 

 mean as frequent renewal of electrode and a great loss of time in putting new electrodes in the 

 holder. Such considerations have probably contributed to the use of lower-current densities 



