CARD TRANSLATOR KXPERIMENTS 



393 



Table XIV — Mean Card Dropping Times avitii Allowances 

 Summarized with Regard to Variables 



Variable 



Looks 



C.S. III. 



C.S. Out 

 Codes 



C.y. In . 



C.S. Out 

 Idlers 



C.S. In . 



Graeco -Latin 



Square 

 Runs 



C.S. In .. 



C.S. Out. 

 Bins 



C.S. In... 



C.S. Out. 

 Positions 



C.S. In... 



C.S. Out. 

 Loads 



C.S. In... 



C.S. Out. 



Graeco-Latin 

 Sq. & Codes 



C.S. In 



C.S. Out 



Idlers & Bins 



C.S. In 



Idlers & Runs 



C.S. In 



No. 



Readings 

 in each 

 Mean 



200 

 50 



400 

 100 



200 



160 

 40 



160 

 40 



160 

 40 



160 

 40 



16 



4 



40 

 40 



Mean Dropping Time Milliseconds 



A 

 35.2 

 26.8 



3 difiit 

 35.0 

 26.6 



cards 

 36.5 



D 

 34.9 

 26.8 



100 

 34.6 



5 

 34.8 

 25.1 



V 

 35.5 

 27.5 



fg 

 34.9 

 26.5 



105 

 35.3 

 28.3 



Min 

 33.6 



16.8 



34.0 

 33.4 



Max. 



37.6 



37.5 



38.2 



36.8 



Range of 



Means 



.'Mlowance 



Xogligible 

 Negligible 



Negligible 

 Negligible 



2.1 



1.8 

 3.5 



0.7 

 5.6 



0.6 

 2.9 



0.5 



7.7 



4.0 

 20.7 



4.2 



3.4 



±0.4 



d=0.5 

 ±3.0 



±0.5 

 ±3.0 



±0.5 

 ±3.0 



±0.5 

 ±3.0 



±2.5 



±8.4 



±1.6 

 ±1.6 



nificance when operating in the normal manner were the load in the 

 machine and the nimil)er of coded and uncoded cards in the working bin. 

 In a lightly loaded machine the flux from the puU-np magnet is concen- 

 trated in the few cards and therefore in each card there is a greater 

 amount of flux to decay before the card is free to drop. When a bin is 

 loaded with all coded cards the dropping time may be increased because 

 the cards ai-e slightly deformed by the coding process which in turn in- 

 creases the effective thickness of the cards thus reducing their freedom 

 in the l)in. 



Although these two variables were significantly large they still are 

 included in the range of dropping times mentioned above and are not 

 of sufficient magnitude to warrant si)ecial consideration in loading the 



