Free State vs Constrained Inspection: 7 Checks

Free state vs constrained inspection is not a choice between a loose method and a strict method. The two states answer different questions, while validated virtual fixturing can provide a third evaluation path for some flexible parts. The drawing, approved inspection plan, measurement conditions, uncertainty, and decision rule must agree before any result controls acceptance.

This guide is intended to assist design and manufacturing personnel, quality teams, and buyers in interpreting dimensional inspection requirements for thin-wall, sheet metal, plastic, seal, gasket, and other deformable parts. It does not provide a universal clamping force, torque, support pattern, temperature limit, or pass/fail formula.

Free State vs Constrained Inspection: The Short Answer

Required dimensions and features should be inspected in free state when the governing requirement calls for no external restraint. Use constrained inspection only when the drawing, specification, or approved functional method defines how the part is held. If the required design feature is not represented by either physical route, an approved and validated virtual-fixturing method may be used to evaluate the part.

The current ASME Y14.5 public listing identifies Y14.5-2018 (R2024) as the reaffirmed edition in effect. That page establishes the standard’s role in dimensioning and tolerancing, but this article does not reproduce protected clauses or drawings.

What Free-State Inspection Actually Measures

When free state is applied, the feature under consideration is interpreted without considering any forces that are intended to cause conformality of the part. Support, thermal expansion, residual stress, and measurement force remain part of the measurand; the method statement should define them rather than treat the part as floating without gravity or contact.

A 2019 peer-reviewed flexible-part inspection study describes free-state differences caused by gravity and residual stress. Therefore, “unclamped” is an inadequate method description, since the orientation of the part and its support state still affect the measurement condition rather than define part function.

Temperature is included in the definition. The official ISO 1:2022 catalogue page defines reference-temperature concepts for geometrical and dimensional properties and the measurand; the older ISO 1:2016 page is withdrawn.

When Constrained or Restrained Inspection Is Legitimate

In this guide, a constrained result is treated as eligible for acceptance only when the controlling drawing, specification, contract, or approved inspection plan defines a restrained condition relevant to design intent. The method must state how the part is located and restrained, not merely label the result “constrained.”

The current ASME Y14.5 listing identifies the applicable dimensioning-and-tolerancing standard; it does not define a part-specific restraint. For this guide’s decision framework, governing documentation must identify the external state and enough setup variables to reproduce it. Resemblance to final assembly is not adequate if the fixture adds unsupported contact, preload, or torque.

It’s the responsibility of qualified personnel in engineering and quality to resolve any gaps in the controlling documentation prior to commercial production. Blind acceptance due to the absence of controlling documents isn’t acceptable.

How Restraint and Support Create a False Pass, or a False Fail

A part can be erroneously approved if a restraint forces it toward nominal. Conversely, a part can be erroneously rejected if a restraint bends an acceptable part away from its required state. Force magnitude is only one variable; direction, sequence, contact, orientation, residual stress, and workholding and support geometry can produce systematic bias.

One open-access 2023 experiment comparing optical scanning and CMM measurement on thin-wall Ti-6Al-4V parts reported a maximum discrepancy of about 8% under its tested conditions. The study is evidence that method choice and part flexibility can matter; it isn’t an 8% correction factor for other parts.

If the outcome of a measurement is affected by support, load, sequence or orientation, the difference must not be considered random. It is evidence that the inspection state needs to be defined and validated.

The State-of-Inspection Decision Ladder: Seven Checks

The State-of-Inspection Decision Ladder moves from authority to decision. Uncertainty and conformity should be addressed only after the inspection state is defined and validated. Reversing this sequence leads to the setup being chosen only after a determination has been made.

1. 1. Freeze the authority. Record the drawing, specification, revision, feature, datum references, and any free-state modifier or restrained-condition note.
2. 2. Define the functional state. Describe what the part experiences in service or assembly without assuming that every assembly force belongs in inspection.
3. 3. Characterize part behavior. Check gravity, orientation, support sensitivity, residual stress, springback, material stiffness, and time-dependent change.
4. 4. Select the evaluation path. Choose free state, authorized physical restraint, or validated virtual fixturing and document why it matches the authority.
5. 5. Control the conditions. Fix supports, contacts, sequence, load or displacement, alignment, temperature, stabilization, equipment, and program.
6. 6. Evaluate the result and uncertainty. Distinguish repeatability from systematic effects and confirm that traceability and the uncertainty chain are relevant to this measurand.
7. 7. Apply the agreed decision rule. Keep the measured result distinct from the acceptance decision and record who owns false-accept and false-reject risk.

NIST explains that metrological traceability is a property of a specific result and doesn’t, by itself, guarantee fitness for purpose. The NIST guidance therefore supports an important boundary: calibrated equipment is necessary evidence, not proof that the chosen inspection state is valid.

3-Path Inspection-State Framework for Flexible Parts

For this guide’s comparison workflow, preserve a free-state record before adding an authorized physical or validated virtual evaluation when that record is relevant to the feature. This keeps a fixture or model from erasing useful comparison information without claiming that free state always controls acceptance. The acceptance path still comes from the governing requirement, not from the sequence itself.

A 2019 CIRP Annals paper proposes a fixtureless method that superposes gravity and fixturing-restraint effects before profile deviation is evaluated. This shows that physical clamping isn’t the only technical pathway, but the paper doesn’t turn every point-cloud transformation into an approved production method.

An earlier peer-reviewed ASME journal study evaluated a fixtureless method with simulated manufactured-part models. That is method-development evidence, not universal validation for every geometry, material, or production program.

The three-path map supports analysis as well as decisions. A large difference between paths can reveal springback, local forcing, alignment sensitivity, or a model that is absorbing genuine manufacturing deviation.

What the Drawing and Inspection Plan Must Say

The controlled documents must make the inspection state reproducible. A reviewer should be able to identify the authority, rebuild the support or restraint, repeat the sequence, reconstruct the result, and see which decision rule converted that result into a disposition.

For a drawing-specific review, send the revision, critical features, expected inspection state, and reporting requirements through Zhenling Metal’s metrology and inspection resource. The page is a project-review route, not independent proof of a part-specific uncertainty or acceptance result.

Reading Conflicting Free-State and Constrained Results

In the case of contradictory results, don’t take the mean. Don’t take the passing result unless justified. Verify the authority, both setups, systematic effects, uncertainty, and decision rule. Take the method permitted by the governing requirement.

The result and the conformity decision are different artifacts. A NIST publication on decision rules and risk analysis explains that uncertainty evaluation is technical, while choosing the decision rule also involves cost and the balance between accepting a bad product and rejecting a good one.

If a disagreement exposes an unclear drawing or plan, engineering clarification should precede final acceptance. A supplier-side fixture change cannot silently repair a requirement gap.

Situations That Do Not Justify Constrained Inspection

Constrained inspection is not justified merely because the part is flexible, the free-state result fails, or the fixture improves repeatability.

An affirmative answer to the Invalid Restraint Test does not automatically mean the part is nonconforming. A “yes” means the method is not yet ready to determine acceptance. The next step is to resolve the requirement or validate the method, not to adjust the clamp until the result changes; the NIST publication on conformity decision rules explains why the measurement result and the rule used to decide conformity must be treated separately.

Buyer Handoff Checklist Before Production

Prior to tooling and production, the buyer should transfer the required inspection state for the part or feature under consideration. The assignment should include the evidence to support the evaluation, the variables to control during the evaluation, and the type of evidence to be reported; any missing field should remain an open technical question, and no exemption should be assumed. Supplier discretion shouldn’t apply to an unevaluated or missing field.

Shanghai Zhenling Hardware Co., Ltd. manufactures non-standard machined components from customer drawings and samples. For a meaningful review, provide the part function and inspection-state requirements with the drawing rather than asking for a generic “CMM pass.” NIST metrological-traceability guidance treats traceability as a property of a specific measurement result, so the handoff should request the result, method, uncertainty, and traceability evidence together.

Frequently Asked Questions