PRIVATE LINE DATA TRANSMISSION 



1477 



Table V — Use of Telephone Band by Various Data Systems 



System 



1. Proposed. . . 



2. Exploratory 



3. Exploratory 



4. Polytonic <i 



5. 40ClTeg. 



6. 43A1 Teg. 



toll . . 

 ocal 



Modulation 



VSB 



VSB 

 DSB 

 DSB 

 DSB 

 DSB 

 FS 



Max. No. of 

 Channels 



1 

 1 

 1 



5 

 5 



18t 

 17t 



Bits/Sec. 

 per Channel 



1600 



1650 



750 



100 



300 



74tt 



74tt 



Total Bits/Sec. 



1600 



1650 



750 



500* 

 1500* 

 1332tt 



1258tt 



* Not realizable with 2 out of 5 codes used. 



t Not realizable over some facilities. 



ft Based on 100 word per minute channel capability. 



All these products are necessarily even multiples of k and therefore al- 

 ways fall half way between carriers. This allocation, however does not 

 permit mitigation of third order modulation effects. 



3.2.5 Experience 



Table V reviews systems which have been discussed earlier in this 

 paper, to indicate the extent to which they use a general telephone 

 channel facility, in terms of the bit rate output. It is clear, of course, 

 that the various systems are not engineered to the same conservatism. 

 These differences have already been commented upon. 



The general conclusion which one can reach from the table is that the 

 use of the medium in a single channel gives possibilities of a higher bit 

 rate than subdivision. However, it is to be kept in mind that the tele- 

 graph facilities are conservatively engineered. Further as noted, they 

 can be used to the full extent indicated only over the broader band tele- 

 phone channels. For example, the full 18 telegraph channels can not be 

 used over a C carrier telephone channel. 



3.3 Noise 



A general theory regarding the influence of noise on digital systems 

 was presented in 1948 by Oliver, Pierce and Shannon.''' This was dis- 

 cussed further at a symposium.'' Several additional points are considered 

 here, relating to the effect of noise on a data transmission service of the 

 type contemplated. 



3.3.1 Effect of Channel Subdivision on Vulnerahility to Noise 



This has been suggested earlier at several points. A more detailed 

 discussion is given of the effects in Appendix II. 



The conclusions reached there are, broadly: 



1. Channel subdivision has comparati\'el3^ small effect on ^'ulnera- 



