The combination of geometric, dimensional, and tolerancing values, also known as GD&T, determine how multiple parts fit together into an assembly to form a product. The “T” part of this acronym, or tolerancing, specifies how much parts or assembly can vary from their nominal geometric or dimensional values. Also referred to as tolerances, tolerancing is typically expressed with upper and lower limit values.
Dimensions refer to a size in a particular direction, like length, width, or diameter. Geometries refer to shape, parallelism, tilting, position, run-out, concentricity, and so on.
Tolerances that are specified by designers to be excessively tight can be too expensive to manufacture to be competitive. Tolerances that are too loose can easily cause problems in assembly or in the ability of a product to perform as intended, which can be very costly. Tolerances that are just right, help ensure that the product assembles and performs correctly at the lowest possible production cost.
Determining tolerances that are optimal is a very important aspect of a mechanical design process. Designers use tolerance analysis for making tolerancing decisions.
This article focuses on the tolerancing part of GD&T and overviews:
- Brief history of tolerancing
- How are tolerances measured?
- How are tolerances specified?
- What is GD&T?
- GD&T examples
Brief history of tolerancing
The formalization of tolerances for mechanical products has been cited by historians to have its roots with the concept of interchangeable parts, starting in the 1700s. For example, a Swedish clock maker was producing interchangeable parts for his clocks and other machinery in the 1720s. A “uniformity principle” for light artillery was promoted by a French general in the 1760s, which helped American colonists quickly disassemble, reassemble, and move the artillery supplied by France during the American revolution against Britain. In the 1850s the Springfield Armory in Massachusetts, U.S., gained worldwide recognition for demonstrating disassembly and random reassembly of their muskets.
In the late 1800s and very early 1900s the evolution of standards and measurement technologies, such as micrometer calipers, began to allow some production of interchangeable parts to be made in different locations from a set of written specifications. For instance in 1902 the Newall system appeared as a standard for limits and fits of shafts and holes. At least in the U.S., designers did not begin writing tolerances on drawings until just after 1900.
With the rise of the automobile industry starting in the early 1900s, with its huge scale up in mass production, tolerancing standards, tolerance analysis, and metrology technologies to support true interchangeability of parts within assemblies took off. In the 1980s, formalization of standards for worldwide use accelerated with publications issued from the American Society of Mechanical Engineers and the International Standards Organization. See below for more details on these standards.
How are tolerances measured?
Dimensional tolerances, also sometimes known as size tolerances, are typically measured in millimeters (or inches in the United States), usually to thousandths of a millimeter for parts within mechanisms, such as those for machinery, appliances, automobiles, or medical devices.
Geometric tolerances vary according to a parts’ feature of interest. For example they could be measured as a position, flatness of the surface, radius of curvature, angularity, cylindricity, etc.
How are tolerances specified?
There are multiple standards for specifying mechanical tolerances within engineering drawings. These include ISO (International Standards Organization), ANSI (American National Standards Institute), ASME (American Society of Mechanical Engineers), DIN (German Institute for Standardization), JIS (Japanese Industrial Standards), BS (British Standards), and many others.
These standards describe mechanical tolerancing systems that provide engineers with a symbol-based language for specifying dimensional and geometric tolerances. While these standards may differ in their details, the underlying tolerancing concepts are largely the same.
Enventive’s experience is that the two most widely used standards globally are ISO’s Geometric Product Specifications (GPS) and ASME’s Geometric Dimensional and Tolerancing (GD&T) specifications. It should be noted that the term GD&T is widely used generically to describe the specification of design parameters within technical drawings.
What is GD&T?
As noted above, GD&T is short for geometric dimensioning and tolerancing. It revolves around specifying tolerances in reference to geometric features.
A quick internet search will show that the roots of modern GD&T are attributed to a Stanley Parker who worked in a torpedo factory in England just before World War II. He realized that dimensional tolerancing, based on X and Y measurements, led to usable parts being rejected. His idea was to instead specify tolerances relative to the position of one surface to another, which in today’s modern standards is represented through the use of datums. Parker’s concept gained acceptance in the military and in the 1950s grew to commercial use. It soon incorporated other geometric concepts such as flatness, parallelism, runout, and more. Various specifications for GD&T from different organizations began being published during this time and were steadily expanded.
Broad industry-wide standardization of GD&T as a mark-up language can be thought of as accelerating with the first publication of the ASME Y14.5 standard in 1982. New releases of this standard have been issued every 10 to 15 years, with the most recent being 2018. Shortly afterwards, in 1985, ISO published its first in a series of its GPS standards, its ISO 8015: Geometrical product specifications (GPS). The latest version was published in 2011 and confirmed as being up to date in 2021.
The ASME standard, as well as other GD&T standards like the ISO’s GPS, incorporates symbology, feature control frames for visually representing a geometric tolerance, and datum reference frames which are planes, lines, points, axes, or other geometric features from which measurements should be made.
For symbols, examples include those for parallelism, perpendicularity, concentricity, and symmetry, as seen in the figure, as well as straightness, angularity, flatness, cylindricity, surface profile, and more.
Feature control frames visually combine in a single rectangular box the arrows, symbols, datum references, values, and other information that define the tolerance. See the adjacent figure for examples of ISO symbols and an illustration of a position symbol, simple frame, and datum as applied to a 2D cross section of a part.
Applying GD&T within a design starts with the identification of the datums to define the positions of parts. These are usually the surfaces that will be in contact with other components.
3D Flange Assembly Example
The adjacent 3D drawing of a flanged assembly is an illustration from a CETIM handbook for ISO GPS showing two datum references, A and D, which are located on the mating surfaces of the flanges. The two feature control frames specify nominal and tolerancing values related to the four bolt holes and datum surfaces.
2D Shaft Example
The drawing of a shaft below is a 2D GD&T example from Enventive’s Concept software for tolerance analysis. Concept supports GD&T markups using either ISO or ASME standards — this example applies ISO. You can see that the shaft has six different cylindrical surfaces, two of which have diameters of 17 and 20 mm. Those surfaces are for mounting ball bearing assemblies, which then sit in a housing of a rotating machine.
Both surfaces have a tolerance specified such that the diameter cannot be more than .01mm above its nominal values and cannot be any smaller — if the diameters were any smaller then the bearings would not fit because they would be too loose. The surfaces precisely control the orientation of the shaft and are defined as datum A1 and A2. A constructed line common to A1 and A2 is further defined as first datum A. All three datums are highlighted with circles.
By incorporating geometric tolerancing to physically position surfaces, such as with concentricity of the A1 and A2 datums relative to datum A, problems like misalignments of the mounted bearings are avoided for which a series of simple dimensioning tolerances would overlook.