414 BELL SYSTEM TECHNICAL JOURNAL 



possibilities of the device could not be realized. For these reasons its 

 use was limited almost entirely to comparatively short cables where 

 the strength and shape of the received signals were sufficiently good to 

 control the relay directly with only a small improvement in shape and 

 with no amplification. The original arrangement was later modified ^ 

 to adapt it to the operation of longer unloaded cables. 



One of the principal features of the cable multiplex herein described 

 is the synchronous vibrating relay ^ which was developed particularly 

 for high speed operation on long cables, and is a great improvement 

 over the Gulstad device. The vibrating impulses, instead of being 

 derived from an adjustable vibrating circuit, are generated by a 

 segmented commutator located on the receiving head of the distributor. 

 As the brushes on the receiving head of the distributor rotate in 

 nearly exact synchronism with the transmitting brushes, it is evident 

 that the rate of vibration of the relay will coincide exactly with the 

 frequency of the transmitted signals and by properly adjusting the 

 angular position of the vibrating segments, the time of closure of the 

 relay contacts with respect to the incoming signalling impulses can be 

 accurately fixed. The accuracy with which the missing impulses of 

 unit length in the received signals are reinserted by this means makes 

 it possible to realize the full speed possibilities of the vibrating relay 

 principle and obtain faithful reproduction of signals on a given cable 

 at almost double the speed obtainable through the use of ordinary non- 

 vibrating relays. 



Another important advantage gained through the use of the 

 synchronous vibrating relay is greater freedom from the efTects of 

 extraneous interference. The amplitude of the received signals is 

 sufficiently great to permit of its being reduced by the efifects of 

 interference to approximately half of the normal value before the 

 distortion becomes sufficiently great to cause errors in printing. 

 Likewise interference occurring during the zero intervals in the 

 received signals must attain a value of approximately half of the 

 normal received signal amplitude before causing errors. Interference 

 occurring during the intervals between vibrating impulses will, of 

 course, produce no effect upon the relay unless the amplitude of the 

 interference attains a sufficiently large value to operate the relay 

 directly. This ratio of interference to received signal amplitude 

 represents the absolute limit of operation and some margin must 

 obviously be allowed. It has been found that continuous satisfactory 

 operation can be maintained so long as the interference does not 



8 W. Judd, British Patent No. 9,768, April 25, 1913; G. R. Benjamin and Herbert 

 Angell, U. S. Patent No. 1,579,999, April 6, 1926. 



9 A. A. Clokey, U. S. Patents Nos. 1,521,870 and 1,522,865. 



