166 THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1957 



TUBES FOR THE NEWFOUNDLAND-SCOTLAND CABLE 



Early Development Considerations 



In Bell Telephone Laboratories, work on tubes for use in a proposed 

 transatlantic cable was started in 1933. This was preceded by a study of 

 what type of tube would best fit the needs of the various proposed ampli- 

 fier systems and by consideration of what might be expected to give the 

 best life performance. 



At the time this project was started, reasonably good tube life had 

 been established for the filamentary types used in Bell System repeaters. 

 Some groups of tubes had average lives of 50,000 or 60,000 hours (6 or 

 7 years) . Equipotential cathode tubes were not then used extensively in 

 the plant, and there was no long life experience with them. However, 

 there appeared to be no basic reason why inherently shorter thermionic 

 life should be expected using the equipotential cathode and there were 

 several advantages in its use. One was the greater freedom in circuit 

 design afforded by the separation of the cathode from the heater. Also 

 there was the possibility of operating the heaters in series and using the 

 voltage drop across the heaters for the other circuit voltages. It was felt, 

 in addition, that the overall mechanical reliability would be greater if 

 the cathode were stiff and rigidly supported. 



The first equipotential tubes made were triodes. They were designed 

 for use in push-pull amplifiers wherein continuity of service might be re- 

 tained in case of a tube failure. This circuit was abandoned in favor of a 

 three tube, feedback amplifier that was the forerunner of the present 

 repeater. The pentode was favored over the triode for this amplifier for 

 obvious reasons, and in 1936 the triode development was discontinued. 



Early in the development of the tube three basic assumptions were 

 made. These were, (a) that operation at the lowest practical cathode 

 temperature would result in the longest thermionic life, (b) that operat- 

 ing plate and screen voltages should be kept low, and (c) that the 

 cathode current density should be kept as low as practicable. 



The first assumption, concerning the cathode temperature, was based 

 on the observation of life tests on other types of tubes. While the data 

 at the time of the decision were not conclusive, there was definite indica- 

 tion that too high a cathode temperature shortened thermionic life. 

 Little was known about life performance in the temperature range below 

 the values conventionally used. 



The second assumption, concerning low screen and plate voltages, had 

 not been supported by any experimental work available at the time of 

 decision. Sixty volts was originally considered for the output stage; this 



