744 THE BELL SYSTEM TECHNICAL JOURNAL, MAY 195G 



action by the write-sync pulses. The output transformer supplies the 

 writing current pulses, under control of the selector switch, to the chosen 

 magnetic head. Arrangements are provided for synchronizing an oscil- 

 loscope to display the writing current pulses or the voltage outputs from 

 the head at the selected address, as required. 



When a new translation item is to be entered, or an existing one al- 

 tered, the address corresponding to the desired slot-pair is determined 

 from a card-index, or ledger, listing all items on the drum. The address 

 keyboard is then set to the assigned number, thereby singling-out the 

 desired slot pair so that the writing operation can proceed as described 

 above. During this procedure, the monitoring oscilloscope may be used 

 for verifying the new entry, two cells at a time. Over-all verification is 

 accomplished by exercising the translator through facilities already avail- 

 able in the toll switching office. There is nothing about this procedure 

 which precludes the use of automatic facilities for performing the admin- 

 istration. There is also no fundamental need to take the translator out 

 of routine service during the administration operation, since each writing 

 operation disables the equipment for only a few microseconds and would 

 rarely delay a translation by as much as one drum revolution. 



CONCLUSION 



After short preliminary tests, the equipment described and pictured 

 was installed in the switching systems laboratory at Bell Laboratories. 

 A rapid-transfer arrangement permitted direct interchangeability with a 

 card translator in a skeletonized model of a toll switching office. 



A testing program was then begun entailing continuous 24-hours-per- 

 day operation of the magnetic drum translator for approximately one 

 year. After an initial shakedown period during which wiring faults and 

 other minor troubles were recognized and cleared, many millions of trans- 

 lations were handled with only a small proportion of failures. The accu- 

 mulated data on failure rate and cause was significant, being one of the 

 primary objectives of the experiment. An analysis of the data indicated 

 the desirability of certain simple design changes in the existing circuitiy 

 and established a basis for the selection of future designs. 



If, ill the future, consideration is given to the design of ciniipineiil of 

 this type for some specific application, new electronic developments must 

 also be taken into account. Many more types of transistors are now 

 available than when the present design was undertaken, and some of the 

 newer types have capabilities which make them obvious candidates for 

 many of the jobs now done in the translator with electron tubes. Such a 

 substitution would not only increase reliability and decrease power con- 



