AN OSCILLOGRAPH FOR TEN THOUSAND CYCLES 



83 



hundred cycles above or below that value. The change from low to 

 high amplitude or vice versa, although apparently instantaneous, 

 actually lasts about a hundredth of a second. Although no satis- 

 factory explanation has been reached, this phenomenon may be as- 

 sociated with the method of supporting and stretching the string. Its 

 study and elimination would become important, however, only should 

 some means become available of permitting the string to carry several 

 times the present maximum current without overheating — as might 

 be possible if an alloy should be developed with the mechanical proper- 

 ties of duralumin and the conductivity of copper. 



The amount of phase distortion present with these various methods 

 of damping and equalizing is difficult to measure directly for frequen- 

 cies much above Fq. A measurement up to about 4000 cycles was 

 obtained with a two-string galvanometer arranged for somewhat shorter 

 strings than those usually employed with an Fq of 3200 cycles. One 

 of the strings was stretched to an Fq of 6000 cycles and left undamped 

 except by air friction. Its phase distortion was computed and is 

 plotted as the lower curve of Fig. 9. The other string was stretched 



1200 1600 2000 2400 2800 3200 

 FREQUENCY IN CYCLES PER SECOND 



3600 4000 4400 



Fig- 9 — Phase distortion with equalized and resistance damped strings up to about Fo. 



to an Fq of 3200 cycles, and equalized for the fundamental and third 

 harmonic as already described. Both strings were fed from the same 

 oscillator, and a series of oscillograms taken from 50 to 4000 cycles. 

 The phase shift between the strings was then measured and is plotted 



