Geometric dimensioning and tolerancing describes a part by the geometry it must achieve rather than by dimensions alone. The system uses fourteen controls, grouped into five categories. This page lists all of them with what each actually controls.
It matters for balancing more than it might appear. A balancing machine holds a part on its datums, so the axis of rotation the machine works to is defined by those surfaces. Runout, perpendicularity and position errors on mounting features become unbalance directly, regardless of how well the part itself is distributed.
Form — four controls
Form controls apply to a single feature and need no datum. They constrain the shape of the feature against itself.
| Symbol | What it controls |
|---|---|
| Straightness | How far a line element or an axis may depart from a perfectly straight line. |
| Flatness | How far a surface may depart from a perfect plane. Applies to the surface as a whole, not to individual cross-sections. |
| Circularity | Roundness of a cross-section. Each individual cross-section must lie between two concentric circles a stated distance apart. |
| Cylindricity | Circularity, straightness and taper controlled together along the full length of a cylindrical feature. The most demanding of the form controls. |
Profile — two controls
Profile controls define a tolerance zone that follows the true shape of a feature. They can control form, orientation and location at once, which makes them the most versatile controls in the system.
| Symbol | What it controls |
|---|---|
| Profile of a line | A tolerance zone applied to individual cross-sections through a feature. |
| Profile of a surface | A three-dimensional tolerance zone that follows the entire surface. Widely used for complex or curved geometry where no simpler control will serve. |
Orientation — three controls
Orientation controls relate a feature to a datum by angle. They constrain tilt but not location.
| Symbol | What it controls |
|---|---|
| Angularity | A feature’s orientation at a specified angle to a datum. |
| Perpendicularity | A feature at 90° to a datum. Directly relevant to balancing: a mounting face out of perpendicular to the rotation axis tilts anything mounted on it, producing couple unbalance. |
| Parallelism | A feature parallel to a datum, at any distance. |
Location — three controls
Location controls place a feature relative to datums.
| Symbol | What it controls |
|---|---|
| Position | The most used control in GD&T. Locates a feature’s axis or centre plane within a tolerance zone relative to datums. Can be applied at maximum material condition to gain bonus tolerance. |
| Concentricity | Requires the median points of a feature to be within a zone about a datum axis. Demanding to inspect, and often better replaced by runout or position. |
| Symmetry | The median points of a feature must lie within a zone about a datum centre plane. As with concentricity, difficult to verify in practice. |
Runout — two controls
Runout controls are measured while the part is rotated about a datum axis, which makes them the closest of the fourteen to what a balancing machine experiences.
| Symbol | What it controls |
|---|---|
| Circular runout | Variation measured at a single cross-section as the part turns one full revolution about its datum axis. Controls circularity and coaxiality together at that section. |
| Total runout | Variation across the entire surface as the part rotates. Controls circularity, straightness, taper and coaxiality simultaneously. The tighter and more complete of the two. |
Why This Matters When Balancing
The axis a balancing machine works to is the axis established by the part’s datums — its bearing journals or mounting surfaces. Three consequences follow.
Runout on a mounting diameter displaces the part relative to the true rotation axis, producing static unbalance that no amount of correction on the part itself will fix. A face out of perpendicular tilts whatever mounts against it, producing couple unbalance. And positional error on a pilot diameter shows up as an eccentricity, which is the dominant unbalance source in assemblies such as axle drive flanges.
A part can be perfectly balanced on a machine and unbalanced in service if the datums it was held on differ from the surfaces it actually runs on. Where balance tolerances are tight, the GD&T on the mounting features deserves as much attention as the balance specification itself.
How Each Control Is Actually Measured
Specifying a characteristic and measuring it in production are different problems. The control tells you what the part must achieve; the gage has to establish the datums, present the feature and resolve the deviation inside a cycle time. How that is done differs sharply by category.
Form controls — flatness, straightness, circularity, cylindricity — need no datum, which makes them the simplest to fixture and the most demanding to sample. A single probe touching one point tells you nothing about form; the measurement is the variation across many points. Circularity and cylindricity are far faster taken dynamically, with the part rotating under a probe, than built up point by point at rest.
Orientation controls — perpendicularity, parallelism, angularity — and location controls are only as good as the datums. In a production gage the fixture is the datum simulator: the locating surfaces stand in for the datum features called out on the drawing. If the fixture locates on something other than the specified datum, or locates on it imperfectly, the measurement is accurate and wrong at the same time. This is the single most common cause of a gage disagreeing with a coordinate measuring machine.
Runout is the category where a purpose-built gage separates itself furthest from a CMM. Circular and total runout are conditions of a rotating part, and measuring them by rotating the part against a probe takes seconds. Reconstructing the same result from discrete touch points is slow and, on a surface with lobing, unreliable.
Profile controls need dense data across a surface, which is where non-contact laser measurement earns its cost over a contact probe.
One characteristic of production gaging cuts across all of these: measurement is normally comparative, against a master representing the nominal condition, rather than absolute. That cancels most thermal and fixture error and makes the cycle fast — and it makes the master part of the measurement, with its own control and verification interval.
BTI builds dimensional gaging systems that measure exactly these characteristics — flatness, straightness, circularity, cylindricity, perpendicularity, angularity, parallelism, profile of a surface and circular runout — using contact LVDT probes, non-contact laser heads, or both on one station, with static, dynamic and under-load measurement. Where the same part also needs flaw inspection or mass correction, gaging combines on a single machine with crack detection and balancing.
Frequently Asked Questions
How many GD&T symbols are there?
Fourteen geometric controls, in five categories: form (4), profile (2), orientation (3), location (3) and runout (2).
Which GD&T controls need a datum?
Orientation, location and runout controls all reference one or more datums. Form controls do not. Profile controls may or may not, depending on whether they are locating the feature or only controlling its shape.
What is the difference between circular and total runout?
Circular runout is measured at one cross-section at a time. Total runout applies across the whole surface simultaneously, and additionally controls taper and straightness.
How does GD&T affect balancing?
The balancing machine holds the part on its datums, so datum errors become unbalance. Runout produces static unbalance, perpendicularity error produces couple unbalance, and pilot eccentricity produces the quasi-static unbalance common in flanged assemblies.
Source
Adapted from the BTI GD&T reference chart, Balance Technology Inc. Download the original (PDF).
Browse all BTI technical resources or talk to an application engineer.
