188 BELL SYSTEM TECHNICAL JOURNAL 



Further statistical analysis of this type of situation is needed together 

 with experimental determination of the distribution of dimensions actually 

 encountered in specific cases. 



The distribution of dimensions in a product arises from a variety of causes. 

 One type of cause is the variations such as those in the dimensions and 

 physical properties of raw material which may produce different product 

 dimensions even from a particular tool. A different and more systematic 

 type of cause is the change in the dimensions of tools as a result of wear. 

 The practice followed in establishing tool wear allowances will therefore 

 affect the limits and statistical distribution of part dimensions during the 

 life of the tool. Some designers and some tool makers consider that the 

 specification of a nominal value with plus and minus variations requires a 

 different handling of initial tool dimensions and wear allowances than does 

 the specification only of minimum and maximum limits for a part dimen- 

 sion. Equally good authority maintains that a manufacturer recognizes 

 no difference. Establishment of standard practices in such matters is a 

 needed step in determining the distribution of dimensions to be expected in 

 machined parts. In the present absence of standards or of any consistent 

 attitude on the subject it is necessary for designing and manufacturing 

 engineers to reach an agreement in specific cases where this factor is im- 

 portant. 



Such are the factors which determine the tolerances which can be obtained 

 economically or which perhaps will be unavoidably encountered. It is 

 necessary for a designer to keep informed of the interaction of these factors 

 as his design crystallizes and he must also determine the effect of such toler- 

 ances upon functioning in order to complete a design which will function 

 properly when assembled in quantity production. 



Functional Dimensioning 



Efecl of Tolerances 



If apparatus parts are minute or have complicated relative motions it is 

 recognized that manufacturing drawings to the usual scale have serious 

 limitations to their usefulness in the analysis of the effects of combinations of 

 tolerances. In such cases designers frequently make layouts to larger 

 scales or large scale adjustable models to investigate the effect of variations 

 on functioning. Illustrations of this practice are numerous in the expe- 

 rience of most designers of small apparatus. 



Even in large parts which are stationary in use the application of toler- 

 ances, in effect, establishes several possible positions for each element and 

 may present problems similar to those involving motion. These are not 

 easily recognized because of a curious limitation inherent in small scale 



