The 3-2-1 locating principle constrains all six degrees of freedom (DOF) of a workpiece using exactly three locators on the primary datum (3 points define a plane, removing 3 DOF), two on the secondary datum (removing 2 more DOF), and one on the tertiary datum (removing the final DOF) — no locator does double duty and none can be removed without introducing motion. For a 150 × 100 × 50 mm steel block, correctly sequencing these six locators before applying any clamp force typically reduces positional variation to under 0.02 mm across repeat setups; a reversed clamping sequence on the same fixture can lift the primary datum and introduce 0.05–0.15 mm of tilt error before the spindle turns.
Quick 3-2-1 Fixture Reference
| Problem / Goal | Primary Action | Expected Impact |
|---|---|---|
| Positional repeat error > 0.05 mm | Verify all 6 DOF are uniquely constrained; check for redundant locators | Repeat accuracy typically ≤ 0.02 mm in rigid setups |
| Workpiece lifts when first clamp is applied | Activate primary-datum clamps first, then secondary, then tertiary | Eliminates datum-lift-induced tilt of 0.05–0.15 mm |
| Chatter in a long-reach milling operation | Add rest support under the cutting zone (not a locator — support only) | Can reduce overhang deflection by 50–90% depending on support placement (deflection ∝ L³) |
| Thin-wall part springs away from locators | Reduce clamping force by 30–50%, use soft contact pads | Wall deformation after release < 0.03 mm for 6061-T6 at ≥ 3 mm wall |
| Over-constrained fixture jams on re-load | Remove one locator from the constrained DOF, switch to friction contact | Eliminates jamming and reduces reload time by 30–60 seconds |
| Secondary datum face not fully seated | Apply secondary clamp before tertiary; use feeler gauge to verify 0-gap | Ensures perpendicularity error < 0.01 mm per 100 mm |
The Six Degrees of Freedom and Why 3-2-1 Addresses Each One
A rigid body in free space has exactly six degrees of freedom: three translational (X, Y, Z) and three rotational (rotation about X, Y, and Z axes). Every fixture must remove all six — no more and no fewer — before a clamp is applied.
The 3-2-1 principle distributes those six constraints across three datum surfaces:
- Primary datum (3 points): The largest, most stable face. Three non-collinear locating pins or buttons remove one translational DOF (motion perpendicular to the face) and two rotational DOF (tip and tilt about the two axes in the plane). This leaves the workpiece constrained in three DOF, still free to slide along or rotate within the datum plane.
- Secondary datum (2 points): A perpendicular face, typically the longest available edge. Two locators along this face remove one more translational DOF and one rotational DOF — the workpiece can no longer rotate in the primary plane.
- Tertiary datum (1 point): A third perpendicular face, stopping the final translational DOF. The part is now fully located.
The primary datum carries the most locating responsibility because it contacts three points; for this reason, it should be the flattest, most accurately machined surface on the workpiece. ASME Y14.5 and ISO 1101 use the same three-datum hierarchy in GD&T datum reference frames: |A|B|C| maps directly to primary/secondary/tertiary under the 3-2-1 arrangement.
In practice, locators are hardened pins, spherical-tip buttons, or flat pads ground to within ±0.005 mm of a common height. For steel workpieces on 4140 Steel fixtures, hardened and ground locator pads are preferred because repeated loading of soft locators cold-works the contact area and shifts the effective datum measurably after several hundred cycles in typical shop practice.
Assigning Datums: Rules for Choosing Primary, Secondary, and Tertiary Surfaces
Choosing the wrong surface as primary is the most common fixturing error — it produces maximum positional scatter even when the locators are dimensionally perfect.
The primary datum should be the surface with the largest contact area and the tightest form tolerance on the workpiece. For a prismatic part, this is almost always the largest flat face. A datum selection that violates this rule — for example, using a narrow edge as the primary datum — amplifies angular error: a 0.01 mm bow across a 20 mm narrow face produces 0.5 mrad of tilt, which at a 100 mm feature distance translates to 0.05 mm positional error.
Three rules govern effective datum selection:
- Area governs stability. Primary datum contact area directly determines sensitivity to locator height variation. Increasing the primary contact triangle side length from 40 mm to 80 mm halves the angular sensitivity to a given locator height error.
- Datum selection must match the functional datum in the drawing. ISO 1101 and ASME Y14.5 specify that tolerances are measured relative to datum reference frames; if the fixture datum does not match the drawing datum, the part passes in the fixture and fails on the CMM.
- Cast or rough surfaces should generally not serve as locators without a machined datum pad when positional accuracy below 0.05 mm is required. A raw casting surface has 0.3–0.8 mm of form error; locating on it constrains the part to a randomly oriented reference that shifts with every load.
DIN 1870-1 guidance for fixture design recommends locating on functional datum faces wherever possible, so that the fixture coordinate system aligns with the part coordinate system used for tolerancing. For 6061-T6 aluminum parts, anodized datum faces are acceptable provided the anodize layer thickness (typically 15–25 µm) is accounted for in the locator height.
Clamping Sequence: Load Order to Prevent Datum Lift
Applying clamps in the wrong order is the second most common cause of fixture positional error — it is less visible than a wrong locator placement but equally damaging.
The correct clamping sequence activates clamps in the same order as the datum hierarchy: primary-datum clamps first, secondary-datum clamps second, tertiary-datum clamps last. This sequence presses the workpiece progressively into each datum before the next clamp can deflect it away.
The failure mode of reversed sequence: if a tertiary-face clamp is tightened first, its moment arm relative to the primary datum lifts the near corner of the primary face off its locators. A 500 N clamp force applied at a 150 mm moment arm from the primary datum generates a 75 N-m tilting moment. For three primary locators spanning a 100 mm triangle, the reaction force on the far locator reaches 750 N — enough to elastically deform the locator contact zone by typically 0.02–0.05 mm, leaving a seating gap that persists after all clamps are set.
Practical clamping sequence for a three-clamp setup on a prismatic part:
- Apply primary-datum clamp(s) — push the part against the three primary pins, tighten to 60–70% of final torque.
- Seat the secondary-datum face manually (push the part against secondary locators) before activating secondary clamps.
- Seat the tertiary-datum manually, then tighten the tertiary clamp.
- Return to primary-datum clamps and bring to full torque.
The final step (re-torquing primary clamps) compensates for any micro-movement caused by steps 2 and 3. Skipping the re-torque step on the primary datum can leave a residual gap of typically 0.01–0.03 mm on high-friction surfaces, contributing directly to part-to-part variation at the final inspection stage.
Best Practice
Verify primary-datum seating with a feeler gauge (0.02 mm blade) after all clamps are set and before running the first cut. A blade that slips under any primary locator indicates an unseated datum — re-sequence the clamping before proceeding.
Deflection Control Under Cutting Loads
Locating and clamping constrain the workpiece before cutting; deflection control addresses what happens when the cutting force is applied. These are separate problems with separate solutions.
Workpiece deflection under milling loads follows the cantilever beam formula: δ = FL³ / (3EI), where L is the unsupported overhang length. Deflection dominates: halving the unsupported span reduces deflection by a factor of eight (L³ relationship), while doubling the cross-section height only reduces deflection by a factor of eight through I ∝ h³.
For a 6061-T6 aluminum plate (E = 69 GPa) clamped at one end with a 10 N cutting force at 100 mm overhang: δ = 10 × 0.1³ / (3 × 69 × 10⁹ × I). For a 10 mm thick, 50 mm wide plate, I = 4,167 mm⁴ = 4.167 × 10⁻⁹ m⁴, giving δ = 10 × 0.001 / (3 × 69 × 10⁹ × 4.167 × 10⁻⁹) = 0.012 mm. Moving the clamp 25 mm closer to the cutting zone (L = 75 mm) drops δ to 0.005 mm — a 58% reduction from a 25 mm repositioning.
Three deflection control strategies, in order of implementation priority:
1. Add rest supports under the cutting zone. A rest support (jack screw or adjustable pad) underneath the workpiece at the cutting location is NOT a locator — it contacts the part after locating is complete and carries only vertical reaction load. Rest supports can reduce mid-span deflection from 0.05 mm to under 0.005 mm for thin plates. They must be set against the workpiece with zero preload; preloaded rest supports lift the primary datum and invalidate the 3-2-1 constraint.
2. Reposition clamps closer to the cutting zone. As the L³ relationship shows, moving a clamp 30% closer to the load application point halves the effective stiffness arm and reduces deflection by 66%. This costs nothing and requires no additional hardware.
3. Increase the contact area of the primary locators. Replacing three spherical-tip locating pins with three flat pads (30 mm diameter) reduces Hertzian contact deflection at the datum surface by 40–60% under identical clamping forces, because Hertzian contact stiffness scales with the square root of contact area.
Avoid This
Avoid using a rest support as a datum locator for a second setup. Re-seating a workpiece on a rest support that was adjusted during the first operation typically introduces a datum shift of 0.05–0.2 mm, because rest supports are not lapped to a common height reference. Treat rest supports as structural stiffeners within a single setup only.
Over-Constraint and Under-Constraint: Recognizing and Correcting Both
An over-constrained fixture has more than six locating contacts — it forces the workpiece to deform to satisfy conflicting geometric constraints, and repeatability degrades with every reload. An under-constrained fixture has fewer than six — the workpiece retains at least one free DOF and drifts under cutting force.
Over-constraint is far more common in practice. The classic failure mode: a machinist adds a fourth pin to the primary datum for "extra stability." The four pins cannot all be at exactly the same height (grinding tolerance ±0.003 mm), so the workpiece rocks on the highest three, with the fourth either not contacting or deflecting the part. Positional scatter across 20 reloads typically increases from ±0.01 mm (correct 3-2-1) to ±0.04–0.08 mm (four-point primary datum).
Recognition test: can any locator be removed without giving the workpiece a new free motion? If not, the fixture is over-constrained.
Correction for an over-constrained primary datum: replace the fourth fixed pin with a spring-loaded equalizing pad — it compensates for height variation and maintains full contact without imposing a conflicting constraint.
Under-constraint is typically caused by omitting the tertiary datum locator — a common shortcut when the tertiary face is inaccessible. The workpiece then retains one translational DOF (sliding along the secondary datum). Under horizontal milling forces, this produces drift of 0.1–0.5 mm per clamping cycle. The fix is to add a tertiary stop, even if it is a simple hardened button clamped to the fixture plate after the part is loaded.
Practical Application: Sizing Locators and Clamps for Common Workpiece Materials
Locator and clamp sizing must account for both the required constraint forces and the allowable contact stress on the workpiece material.
For hardened steel workpieces (4140 Steel, HRC 38–42), spherical-tip hardened locating pins (HSS or carbide, R = 8–10 mm tip radius) are standard because they maintain near-point contact regardless of workpiece surface flatness variation. Hertzian sphere-on-flat peak contact stress p_max = (3F)/(2πa²), with a = (3F·R/(4·E*))^(1/3); at F = 500 N, R = 8 mm, E* ≈ 110 GPa (HSS-on-steel) the resulting peak stress is approximately 2,000 MPa — above general elastic limits for medium-hard steel and high enough to cold-work soft (< HRC 30) workpiece surfaces after 100–200 cycles, but acceptable on case-hardened 4140 at HRC 38–42. Reducing pin radius to 3 mm raises peak stress to ~5,000 MPa and is generally avoided for repeat clamping.
For soft aluminum workpieces (6061-T6, yield 276 MPa), flat locating pads (25–30 mm diameter, hardened steel) are preferred over spherical pins because they reduce peak contact pressure by roughly three orders of magnitude compared with a 3 mm spherical pin at the same load. A 25 mm diameter pad at 500 N clamp force gives a contact pressure of 1.0 MPa — well below the 6061-T6 compression yield of approximately 276 MPa, producing no measurable indentation over 10,000 cycles.
Clamp force sizing follows directly from the cutting force analysis described in workholding clamping force calculation procedures:
- Minimum clamping force ≥ (Cutting force × Safety factor) / Friction coefficient
- Safety factor: 2.0 for stable roughing, 3.0 for interrupted cuts, 4.0 for brittle materials or interrupted heavy cuts
- Friction coefficient: 0.10–0.15 for smooth pads on aluminum, 0.15–0.25 for smooth pads on steel, 0.40–0.60 for serrated jaws on steel
For 6061-T6 at typical finishing feeds (f_z = 0.08 mm/tooth, 4-flute 12 mm end mill, a_p = 3 mm), tangential cutting force runs 150–250 N; a smooth-pad fixture requires 1,500–2,500 N of clamping force at a safety factor of 3 and friction coefficient of 0.15. Standard modular vise clamping force of 25,000–40,000 N (per workholding-clamping-force-calculation.md and modular-vs-sine-vs-toolmaker-vise.md) provides a 10–27× margin, which is adequate for all but extreme interrupted-cut scenarios.
Modular Vise Selection
For prismatic workpieces requiring ≤ 0.02 mm repeat positioning, modular precision vises with hardened and ground locating surfaces are the fastest path to a calibrated 3-2-1 datum — the vise fixed jaw acts as the primary datum, the floor as secondary, and a stop pin as tertiary. See the modular vs sine vs toolmaker vise comparison for jaw width and clamping force data across vise types. Calibrate the fixed jaw parallelism to the spindle axis to within 0.01 mm per 100 mm before use.
Summary
Locate on the largest, flattest surface first — then clamp in datum order.
Apply the 3-2-1 principle by placing three non-collinear locators on the primary (largest) datum, two on the secondary (perpendicular), and one on the tertiary. Always clamp in the same sequence: primary first, secondary second, tertiary third, then re-torque primary. Add rest supports (not additional locators) under the cutting zone to limit mid-span deflection. Avoid over-constraint — more than six locating contacts degrades repeat accuracy from ±0.02 mm to ±0.04–0.08 mm.
Sources
- ASME Y14.5-2018 Dimensioning and Tolerancing — Datum Reference Frames, Section 4
- ISO 1101:2017 Geometrical product specifications (GPS) — Geometrical tolerancing
- Machinery's Handbook 31st Edition — Jigs and Fixtures chapter
- Boothroyd & Knight, Fundamentals of Machining and Machine Tools, 3rd ed., CRC Press
- Hoffman, E.G., Jig and Fixture Design, 5th ed., Delmar Cengage Learning
What is the 3-2-1 locating principle in fixture design?
The 3-2-1 locating principle constrains all six degrees of freedom of a workpiece by placing three locators on the primary datum (removing 3 DOF), two on the secondary datum (removing 2 DOF), and one on the tertiary datum (removing the final DOF). It is the minimum contact configuration that fully positions a rigid body without over-constraint, typically achieving repeat accuracy of ≤ 0.02 mm in well-maintained fixtures.
Which surface should be the primary datum in a 3-2-1 fixture?
The primary datum should be the largest, flattest face on the workpiece, because it accommodates the three-point contact triangle needed to define a stable plane. A larger contact triangle reduces angular sensitivity to locator height variation — doubling the triangle side length halves the tilt error produced by a given height discrepancy. The primary datum must also match the functional datum referenced in the engineering drawing per ISO 1101 or ASME Y14.5.
Why does clamping sequence matter in a 3-2-1 fixture?
Activating clamps out of datum order can lift the workpiece off its primary locators before secondary clamps are set. A 500 N clamp applied 150 mm from the primary datum generates a 75 N-m tilting moment, enough to create a seating gap of typically 0.02–0.05 mm under the primary locators. Always clamp primary first, secondary second, tertiary third, then re-torque the primary clamps to close any residual gap.
How do you control workpiece deflection under milling loads in a fixture?
Add rest supports (adjustable jack screws or pads) directly under the cutting zone after locating — they are not locators and must carry zero preload. Because deflection scales with the cube of unsupported length (δ ∝ L³), moving a clamp 25% closer to the cutting zone reduces deflection by nearly 58%. Rest supports can bring mid-span deflection from 0.05 mm down to under 0.005 mm for thin aluminum plates without affecting datum accuracy.
What is over-constraint in a fixture, and how do you fix it?
Over-constraint occurs when more than six locating contacts are applied to a workpiece, forcing it to simultaneously satisfy conflicting geometric constraints. The most common case is a fourth pin on the primary datum: since all four pins cannot be at exactly the same height, the part rocks on the highest three, degrading repeat accuracy from ±0.01 mm to ±0.04–0.08 mm. Fix it by replacing the redundant fixed pin with a spring-loaded equalizing pad that accommodates height variation without adding a conflicting constraint.
