370 THE BELL SYSTEM TECHNICAL JOURNAL, MARCH 1954 



All previous tests of card drop time were made on Laboratories' built 

 models of the translators and since tests made after any changes in 

 model design have produced considerable differences in card drop times, 

 it was felt that this study should be made on a sample of the Western 

 Electric product. In addition, previous tests did not encompass all of the 

 known variables over their extreme ranges. 



For this study, a new translator was obtained from Western Electric. 

 This translator was one of the regular production and was selected to be 

 a representative unit. Foiu'teen hundred new 200A blanks were also ob- 

 tained for use in this translator. Of these, 10 were coded for use as test 

 cards to be observed. One hundred were coded to fill some bins with all 

 coded cards. The balance of the 1,400 cards was left uncoded. This trans- 

 lator was connected to a simplified test set which would cycle the ma- 

 chine through the operation of the 10 coded test cards. Although this 

 test set was simplified in its operation all of the pertinent time relation- 

 ships involving card dropping were the same as are used in the standard 

 translator circuits. 



This study considered all of the variables, with the machine working 

 in a normal cycle of operation as is currently used in the 4A system. 

 These variables are: 



1. Bin — Some differences in card dropping time had been ol)served 

 depending on the bin in which the operating card was located. 



2. Position of Card Within Bin — Earlier tests indicated that the 

 location of the card within its bin made a considerable effect on dropping 

 time. 



3. Code of Card — ■ Since both three and six digit cards will he used 

 in the translator the code becomes a variable which may affect dropping 

 time. 



4. Coded or Uncoded Cards in Working Bin — Previous tests were 

 made using only a few coded cards in the bin under obser\'ation anci it 

 was felt necessary to learn whether or not having all coded cards in a 

 bin made any difference in the dropping time. 



5. Load in Working Bin — This relates to the number of cards in 

 the bin under observation. A range of from 15 to 105 cards per bin was 

 used. 



6. Load in Machine — This has to do with the possibility of any 

 effect into the bin under observation from the cards in the other bins 

 of the translator. 



7. Consistencn of Data — This has to do with repeated measure- 

 ments to observe if the card dropping time is consistent over short 

 periods of time. 



