# Info

1 i

Fig. 47A "Imperial" Dimensions (a) in fractions; (c) in decimals. The "Fractional Conversion" shown at (b) is NOT recommended - (a) or (c) is "preferred".

(d) The implications of the 'significant figures' in a decimal dimension must always be kept in mind. It is not sufficient simply to translate the fraction into a decimal. The dimension 5/8" is a rule dimension. Stated as 0.625" it immediately becomes a very close dimension indeed. The third significant figure implies that there might be a fourth, lying between 0.0005 below and 0.0004 above i.e. that it is between 0.6245 and 0.6254 rounded off to 0.625. If you start using 32s and 64s in decimals you can run into worse trouble - 7/64" is fair enough, but 0.109375" means that you must work to a millionth.

The RULE is that you should use no more significant figures than are absolutely necessary. Thus 5/8" might be 0.6" for a cored hole, 0.63" for a drilled one, and 0.625" only when reamed. Similar considerations apply to all decimal dimensions. The situation does not apply to a great degree in industry, as drawings are increasingly dimensioned in millimetres. For those who wish to, or who must, retain imperial dimensioning there are three alternatives. First, stick to fractions, applying tolerances as necessary. Second, convert to decimals but take very great care over significant figures. Or, third, abandon fractional equivalents altogether and change to true decimals of inches, just as we do in millimetres.

This last is the logical thing to do, but it is important to note that the whole design must be in decimals. Forget 3/8 - this now becomes 0.40 for a rule dimension and 0.400 for an exact one. (Decimal imperial rules are usually scaled in fiftieths or 0.02";

it is difficult to read one scaled in hundredths.) I have tried to illustrate this in Fig. 47A. At (a) is a lever dimensioned in fractions and (b) is the sort of result which appears all too often when a designer attempts to be 'with it' by using decimals. It is quite in order to quote '7/16 inch ream', as readers may not have truly decimal reamers. But simply to round off fractions to the nearest 0.01 inch results in very odd and sometimes difficult measurements. At (c) I have shown a similar lever as it might be designed in decimals from the start. You can see the difference immediately, although you may not agree with some of the degrees of accuracy implied by some of the significant figures which I have used (my sketch is intended only to show the general idea). My own practice is to use old-fashioned imperial fractions, sometimes with decimal tolerances, for most of my work, but to use purely metric design and dimensions where the class of work demands it. I deal with the conversion of imperial drawings in metric and vice versa on page 103.

In this connection, it is worth noting that there is absolutely no objection to quoting 'l/2-inch ream' on a metric drawing; this is just as reasonable as to quote '12mm drill' or even 'Drill No. 23' on an imperial drawing. It will be a long time before such crossover specifications disappear; not until all our reamers, mandrels, plug gauges and the like wear out. (4) Angles

The significant figure rule can apply to angles also. Thus 25-1/2° would be a simple protractor measure. The same dimension, stated at 25° 30' implies accuracy to one minute and would need a vernier protractor. Finer than this would need very special

1-15° INC