A shaft can measure within the stated diameter limits and still refuse to enter its mating bore. Two mounting holes can each have the correct diameter yet fail to line up with the machine. In both cases, the size result may be valid; it simply does not describe the whole geometry needed for assembly.
Size tolerance limits a feature’s size. Form, orientation and location controls describe other aspects of that feature and its relationship to the part. Which controls apply depends on the actual drawing standard, symbols, datums and modifiers. A caliper reading across one section cannot silently stand in for all of them.
What a matching diameter establishes
Suppose a drawing gives limits for a cylindrical shaft diameter and an inspection finds an allowed value at one cross-section. That result concerns the distance contacted by the instrument at that location. It does not automatically show that the shaft is round at every section, straight along its length or aligned with the mounting face.
Imagine two deliberately exaggerated shafts with the same average measured diameter. One has a straight axis and a consistent circular form. The other bends slightly, so a caliper can still find the expected diameter at individual sections while the overall feature cannot pass through a close, aligned bore. The assembly failure is not evidence that diameter no longer matters. It shows that size and axis geometry answer different questions.
The same distinction appears in a two-hole mounting pattern. Each hole may accept a pin of the intended diameter when checked separately. If the centres are too close together, too far apart or displaced from the mounting datums, a rigid mating part with two pins will not enter both holes at once. Increasing confidence in each hole diameter does nothing to resolve their location.
Keyence’s introduction to GD&T separates dimensional controls from geometric controls such as form and position. It also notes differences between ISO and ASME rule systems. The practical consequence is that the drawing’s stated standard matters; rules learned from one convention should not be assumed on another drawing.
Geometry describes relationships that size leaves open
Form controls address the shape of an individual feature. Roundness, for example, concerns how closely a circular cross-section follows the allowed form. Cylindricity considers the cylindrical surface more broadly. A few diameter measurements may reveal obvious variation, but they are not automatically the inspection method required by the drawing.
Orientation controls connect a feature to a reference. A shaft axis intended to be perpendicular to a mounting face can have an acceptable diameter yet lean relative to that face. When the part is bolted down, the lean can shift the far end of the shaft or load a bore unevenly. The diameter value remains unchanged while the working relationship changes.
Location controls address where a feature belongs relative to defined references. Hole position is a common example. The theoretically exact location may be expressed with basic dimensions, while a feature-control requirement defines the allowed positional variation under the drawing’s rules. The basic dimension is not an independent plus/minus tolerance to invent during inspection.
Runout can combine surface variation observed as a part turns about a referenced axis. It may matter where a roller surface, shaft and mounting interface need a controlled rotating relationship. A static outside diameter at one angle does not describe what the surface does through a revolution.
These terms are useful only when the drawing actually invokes them. A product photograph or a buyer’s expectation does not create a datum system. Nor should a general explanation generate a new tolerance for a part whose engineering requirement is unavailable.
Datums tell the drawing where “correct” begins
A datum reference establishes the origin or orientation from which another requirement is interpreted. On a roller or mounting component, a face may establish seating, an axis may establish rotation, and another feature may orient the bolt pattern. The order and meaning come from the drawing, not from whichever surface is easiest to reach with a caliper.
This is why a matching loose-part dimension can fail after mounting. If the measured feature is correct relative to itself but located incorrectly from the seating face, installing the part brings the discrepancy into the assembly. A measurement taken from an unfinished edge or a worn surface may also use a different reference from the drawing.
Modifiers can change how size and geometry interact. Under some ASME conditions, the envelope principle associated with Rule #1 relates perfect form at maximum material condition to the size limits; exceptions and other standards change the interpretation. The GD&T Basics discussion of Rule #1 is useful precisely because it shows that “size never controls form” is too broad. The drawing standard and feature callout must be read together.
Likewise, a positional tolerance may use material-condition modifiers that change the available tolerance with actual feature size. A buyer should not add “bonus tolerance” from memory to a drawing whose notation or governing standard has not been confirmed. The task is to obtain the reported result under the drawing’s actual rule set.
Diagnose the failed fit from the interference
The way assembly stops can direct the next dimensional question. A shaft that begins entering but binds as insertion increases suggests checking straightness, axis relationship, taper, damage and bore geometry in addition to size. A mounting face that seats while the second bolt will not enter directs attention to the pattern’s location and orientation. A part that rocks before seating raises a face or form question.
These observations are clues, not substitutes for inspection. Forcing the assembly can damage the evidence and the parts. Where heavy undercarriage components require support or alignment, use the machine-specific procedure and appropriate equipment rather than using bolts to pull a questionable fit into place.
For an AFTparts track roller enquiry, describe the failed relationship in the same terms used by the drawing. “Shaft diameter measured 50.02 mm” is useful if it names the section and conditions. Add whether the shaft begins entering the bore, whether the mounting faces seat, and which holes do not align. That information identifies whether more diameter readings, an axis check or a hole-location result is actually needed.
Read the inspection result as a set, not a single winner
A controlled comparison brings together the drawing identity and revision, governing standard, feature size results, geometric requirements, datum setup and the method used for each result. The inspection method must be capable of evaluating the requirement. A handheld caliper may be suitable for a broad size comparison and unsuitable for a tight position or runout decision.
Keep published drawing dimensions distinct from measurements of a worn installed part. Wear, coating, deformation and inaccessible datums can change what the field reading represents. A close match at one surface is valuable identification evidence, but it does not approve interchangeability with an unverified candidate.
The GD&T Basics guide to reporting basic dimensions illustrates why a theoretically exact value and the measured departure used to evaluate position belong to different parts of the report. In a purchasing record, copying only the basic value can make the result look perfect while omitting the actual positional deviation.
Return to the two-hole example. The useful outcome is not “diameters pass, part fails.” It is “both diameters meet their stated limits, but the measured hole relationship does not meet the drawing’s positional requirement from the specified datums.” That statement preserves the valid size results and names the geometry responsible for the failed fit. It also gives the supplier or inspector one defined discrepancy to resolve.