5.2 Inspection Setup and Datum Establishment
Key Takeaways
- There are three distinct entities: the datum (theoretical, perfect), the datum feature simulator (physical equipment or CMM algorithm), and the actual datum feature (imperfect real surface); only the simulator is physically set up.
- Surface plates, three-pin fixtures, expanding mandrels, collets, and V-blocks each establish a different datum type; the choice follows the datum feature's geometry and whether it is referenced at RMB or MMB.
- Setup repeatability depends on datum feature quality (form error, surface finish) and the modifier: RMB is sensitive to actual feature form, while MMB allows datum shift that can mask form error.
- Datum targets (points, lines, areas) are used to define a repeatable datum on unstable surfaces such as castings, forgings, or sheet metal; the simulator contacts only the target locations, not the whole feature.
- CMM datum alignment uses the same simulator logic but in software: the CMM samples the datum feature, fits the simulator per the modifier, and constructs the DRF in the measurement model.
Three Entities a Senior Must Distinguish
A recurring Senior-exam and shop-floor confusion is conflating three different things:
- The datum — a theoretically perfect point, axis, line, or plane derived from a datum feature. It is the origin of the DRF and exists only in the math.
- The datum feature simulator — the physical equipment (surface plate, mandrel, collet, gage pin) or the CMM software algorithm that represents the datum. It is of sufficiently true form.
- The actual datum feature — the real, imperfect surface on the part.
The datum is what the feature control frame references (the letter). The simulator is what you set up. The actual feature is what contacts the simulator. Senior questions frequently swap these terms in distractors—watch for it.
Equipment Choices by Datum Type
| Datum feature geometry | Typical simulator | Modifier behavior | Notes |
|---|---|---|---|
| Planar surface (primary) | Surface plate | RMB: high-point contact | 3-point contact defines the plane; flatness of the surface drives repeatability |
| Planar surface (secondary/tertiary) | Adjustable stop, side rail, or angled fixture | RMB: high-point contact; MMB: shift allowed | Stop constrains 1 translation; MMB on a width gives a virtual condition boundary |
| Cylindrical hole (RMB) | Expanding mandrel or adjustable pin | Expands to actual mating size | Translation modifier allows the mandrel to translate along its axis |
| Cylindrical hole (MMB) | Fixed-size pin at virtual condition | Pin is smaller than MMC; datum shift = MMC − actual | Part can shift on the pin by the clearance |
| Cylindrical shaft (RMB) | V-block or collet | Contracts to actual mating size | Collet repeats well; V-block introduces cosine-error on out-of-round shafts |
| Cylindrical shaft (MMB) | Ring gage or fixed bushing | Ring is larger than MMC; datum shift | Shaft can shift inside the ring |
| Unstable surface (casting, forging, sheet) | Datum targets (point/line/area targets) | Targets define the contact set | Whole feature is NOT contacted—only the target locations |
Surface Plate and Three-Pin Fixture
A surface plate is the simulator for a primary planar datum at RMB. The plate contacts the three highest points of the surface; those three points define the datum plane. Repeatability is governed by the flatness of the actual datum feature: a part with 0.5 mm flatness error on datum A will rest differently each time it is placed on the plate unless the same three high points are reused.
A three-pin fixture (three fixed pins protruding from a plate) is used when datum targets are specified. Each pin is a datum target point; the three pins define the primary plane deterministically and repeatably regardless of the overall surface form. This is standard for castings and sheet metal.
Expanding Mandrel and Collet
For a cylindrical hole referenced at RMB, an expanding mandrel is the simulator. It expands inside the hole until it contacts the actual mating envelope. The centerline of the mandrel is the datum axis. For a cylindrical shaft at RMB, a collet (or a chuck that closes concentrically) is the simulator; it contracts to the actual mating size and the collet axis is the datum.
A V-block is not a true RMB simulator for a shaft because it contacts only two lines and its indicated centerline shifts with the shaft's out-of-roundness (cosine error). V-blocks are acceptable for quick checks but not for certifying position to a datum axis at RMB.
Setup Repeatability and Modifier
| Modifier | Setup sensitivity | Why |
|---|---|---|
| RMB | High — sensitive to actual feature form, finish, and chip contact | Simulator conforms to the actual feature, so form error changes the DRF between parts |
| MMB | Lower — datum shift absorbs some form variation | Simulator is fixed at the virtual condition; clearance allows the part to shift, but the shift is bounded and must be reported |
| Datum targets | Highest repeatability on unstable surfaces | Contact is confined to defined target locations, isolating the DRF from the overall surface form |
A Senior must anticipate the setup error introduced by the modifier: at MMB, the part can move on the simulator, and that movement is datum shift, not bonus tolerance. Datum shift is a property of the datum reference frame, not of the controlled feature.
CMM Datum Alignment
A Coordinate Measuring Machine (CMM) establishes the DRF in software: the probe samples the datum feature, the CMM software fits a simulator (a plane for planar features, a cylinder for holes/shafts) according to the modifier, and constructs the DRF from the fitted geometry. The operator aligns the part to the DRF, then measures controlled features in that DRF.
CMM alignment mirrors physical setup but is more flexible: an RMB plane is a least-squares or minimum-zone plane fitted to the probed points, an MMB cylinder is a fixed-diameter virtual condition, and datum targets are simply probed at the specified locations. Chapter 12 covers CMM strategy in depth and Chapter 14 covers reporting.
Common Setup Errors a Senior Must Anticipate
- Wrong simulator for the modifier: using a fixed pin on an RMB hole (under-constrains) or an expanding mandrel on an MMB hole (over-constrains, no shift).
- Contacting the whole feature when datum targets are specified: this averages out form error and changes the DRF.
- Squaring the part to the drawing view instead of the DRF: covered in 5.1; produces false rejections on inclined datums.
- Ignoring datum shift at MMB: the shift is real movement of the part and must be accounted for in feature acceptance (see 5.3).
- Using a V-block to simulate a shaft datum axis at RMB: introduces cosine error on out-of-round features.
Which of the following is the correct relationship among the datum, the datum feature simulator, and the actual datum feature?
A casting has an unstable planar surface specified as datum A with three datum target points. What is the correct simulator, and why?
Why is a V-block not a true RMB simulator for a shaft datum axis?
Which statement correctly characterizes datum shift at MMB?