CARD TKANSLATOU EXi'EltlMEJMTS 



379 



Table III — Loading Patterns in Partially Loaded 



Card Translator — Balanced versus Unbalanced 



Load Test 



Note: All cards in Bin I are coded, see Table IV. Cards in Bins II-XII not 

 coded. 



and the three loads of cards. A suita})le design is the factorial design with 

 one factor, the loading pattern, at four levels (W, X, Y, Z) and the other 

 factor, the loads of cards at thr(>e lev(>ls (200, 400, 000) with ten test 

 card measurements taken at each of the twelve points. 



III. data 



The data on card dropping time were obtained l)y means of siiadow- 

 grams of the light output from two of the light channels of the translator. 

 Each shadowgram comprised the operation of the 10 test cards in se- 

 ([uence. Samples of these shadowgrams are siiown on Figs. I and 2. 

 Vov the purpose of these expei'iments the cai'd dro])i)ing time is defined 

 as the time from the release of the puU-uj) magnets until the full closure 

 of the light channels of the translator exclusive of any cai'd rebound. 

 The data from the various experiments were tabulated and ar(^ gixcn 

 in Tables V to XI, inclusive. 



