## Tolerance

Tolerance represents the total amount a dimension may vary. It is basically defined as the difference between the upper and lower limits. Working to absolute or exact basic dimensions is impractical and unnecessary in most instances; therefore, the designer calculates in addition to the basic dimensions an allowable variation. Work must therefore be implemented within the limits of accuracy specified on the drawing. A clear understanding of tolerance and allowance can go a long way toward preventing small but potentially critical errors.

Tolerance is shown on a drawing as ± (plus or minus) a certain amount, either as a fraction or decimal. Limits are the maximum and/or minimum values prescribed for a specific dimension, while tolerance represents the total amount by which a specific dimension may vary. Tolerances may be shown on drawings in a number of different ways. Figure 7.1 shows three examples: A. The unilateral method, which is used when variation from the design size is permissible in one direction only; B. The bilateral method, where the dimension figure shows the variation in either direction that is acceptable; and C. The limit-dimensioning method, where the maximum and minimum measurements are both stated. Figure Figure 7.1 Three methods of indicating tolerances: A. Unilateral method, B. Bilateral method, and C. Limit dimensioning method (source: Blueprint Reading and Sketching, Navedtra 14040).

7.2 illustrates a typical method used to show tolerances for holes and shafts. Surfaces being toleranced have geometrical characteristics such as roundness or perpendicularity to another surface. Figure 7.3 demonstrates typical symbols used in lieu of or in conjunction with notes to state the geometric characteristics being toleranced.

If tolerances are not actually specified on a drawing, certain assumptions can be made regarding the anticipated accuracy by applying the following principles: For dimensions ending in a fraction of an inch, such as 1/8, 1/16, 1/32, or 1/64, the required accuracy will be to the nearest 1/64 inch. When the dimension is given in decimal form, the following principles should be followed: If a dimension is given as 2.000 inches, the accuracy expected is ±0.005 inch; if the dimension is given as 2.00 inches, the accuracy expected is ±0.010 inch. The ±0.005 is called in shop terms "plus or minus five thousandths of an inch." The ±0.010 is called "plus or minus ten thousandths of an inch." 