To check loose fasteners in finished products, verify the hardware against the approved parts list, inspect each selected joint for poor seating or movement, and check tightening with the product's approved torque-audit method. Then test assembly function and security under the specified conditions. Record the original findings before anyone adjusts the product.
A screw that will not turn is not proof of a secure joint. It may be cross-threaded or bottomed out before the parts are clamped. Do not use a universal gap limit, an extra quarter-turn, or hand feel as the acceptance rule.
Set the Inspection Requirements
Before a pre-shipment inspection, confirm the drawing revision, parts list, assembly instructions and buyer checklist. For each important joint, the inspection plan should state:
- Hardware: screw diameter, thread pitch, length, material or strength grade, and any required washer, spacer or locking feature.
- Tightening: installation torque, units, tolerance, lubrication condition and the separate audit method and limits used to check finished assemblies.
- Assembly security: permitted gaps or movement, function checks, and any specified load, pull force, duration or operating cycles.
If a safety-related joint has no approved tightening or test criteria, flag it for engineering review. Do not fill the gap with a generic torque table for wood, plastic or metal.
Check Hardware and Fit
Count the screws, nuts, washers, spacers and inserts against the parts list. Compare similar-looking screws separately; matching the head shape does not confirm the thread, length or strength grade. Open the supplied hardware bag when checking a flat-pack product.
Assemble the agreed number of units using the supplied instructions and tools. In a furniture assembly check, this includes hole alignment, component matching, door and drawer movement, and missing fittings. Do not drill new holes, force misaligned panels together or substitute hardware to make the sample pass.
| Check point | What to examine | Evidence to record |
|---|---|---|
| Screw head and contact surface | Uneven seating, damaged drive recesses, cracked plastic or crushed board | Joint location and close-up photo before adjustment |
| Nut, washer and spacer | Correct parts, order and orientation; specified thread engagement | Part identification and any difference from the drawing |
| Insert or cam connector | A rotating or pulled-out insert; connector engagement against its instructions | Observed movement and the assembly condition |
| Connected parts | Relative movement, open contact faces, binding or unexpected noise | Measured gap, test position and video where useful |
Verify Tightening Without Hiding Defects
Preserve the Original Condition
Photograph suspect joints before tightening them. On a supported, unloaded product, check for unexpected relative movement using the approved procedure. Do not sit on a suspect chair, load an unstable shelf or reach beneath an unsecured assembly to locate the fault.
A rattle can help locate a joint, but it cannot tell you the gap size or prove that a screw is loose. Check the parts at that location. A missing spacer, damaged insert or misaligned hole may remain unsafe even after the screw is tightened.
Use the Approved Torque Audit
Use a torque tool with current calibration and a suitable measurement range. Confirm whether the requirement is in N·m, lbf·in or lbf·ft. Follow the approved procedure for tool direction, adapters, application speed and how the result is recorded.
Installation torque and residual torque checks are not interchangeable. Thread friction, coatings, lubrication and locking features affect the relationship between applied torque and clamp load; NASA's Fastener Design Manual explains these variables.[1]
Do not routinely loosen a fastener to “feel” whether it was tight. NASA's Mechanical Fastener Torque Guidelines, 540-PG-8072.1.2B, explains why loosening torque cannot establish installation torque. Its cover lists an effective date of July 7, 2020 and an expiration date of July 7, 2025. It is cited here only as historical technical background, not a current 2026 operating procedure or a consumer-product acceptance standard. A breakaway or restart check must follow the product's current, approved procedure and acceptance limits.
If a click-type tool reaches its setting without visible rotation, record that observation—not an invented measured torque. It does not, by itself, establish the original installation torque or rule out over-tightening. Stop if the fastener turns without advancing, the insert rotates, or the surrounding material cracks.
Check Locking Features
Verify any specified locknut, locking washer, threadlocker, retaining clip or pin. Do not add glue or extra washers as an unapproved repair. A painted witness line can reveal relative rotation later, but an aligned line does not prove that clamp load has been maintained.
Test Assembly Security
Operate doors, drawers, hinges, folding locks and adjustment mechanisms through the agreed checks. Record binding, unintended release and movement at the joint. For load or cycle tests, state the load, direction, duration, cycle count and support arrangement. Do not replace these conditions with an improvised push or shake.
Product-specific safety tests remain separate requirements. For example, the U.S. clothing storage unit rule covers certain free-standing bedroom clothing furniture at least 27 inches tall, weighing at least 30 pounds, with at least 3.2 cubic feet of enclosed storage. It excludes bookcases and entertainment furniture. Covered units must meet the specified stability tests without relying on the anti-tip kit.[3]
Separate product inspection from laboratory testing when a check needs controlled conditions, specialist equipment or destructive testing. A successful on-site assembly check is not a product safety certificate.
Apply the Sampling Plan Correctly
Agree the lot size, model coverage, inspection level, sampling standard and acceptance criteria before selecting units. Specify the quantities for visual inspection, assembly, torque audits and load tests separately; they need not be identical. Random sampling supports a lot decision, not a guarantee that every unexamined unit is defect-free.
Use the selected plan's acceptance number (Ac) and rejection number (Re). Do not compare the observed sample percentage directly with an AQL number. NIST describes single sampling plans in terms of sample size and the permitted count of nonconforming items.[4] The agreed sampling tables provide the applicable counts.
Illustrative example: If a selected single sampling plan specifies Ac=5 and Re=6, five nonconforming units meet that category's count limit; six reject it. This is a counting example, not a plan for a particular lot. Other safety and test requirements must still pass. If critical defects have a separately agreed Ac=0 rule, one observed critical defect triggers rejection under that rule; finding none does not prove the lot contains none.
Keep a Traceable Finding Record
For each failed joint, record the model, carton and sample identifiers, joint location, requirement and document revision, method, tool ID and calibration status, result, and photo or video reference. Keep the initial finding and the post-rework result separate. Mark inaccessible or unperformed checks explicitly.
Illustrative inspection record—not an actual UTS client case: Sample C-07 has four bolts, four washers and two spacers specified for its leg bracket. Only one spacer is installed. One bolt head appears seated, but the bracket moves during the approved function check. The record includes a whole-product photo, a bracket close-up and a parts-list reference. Tightening the bolt does not resolve the missing-part finding. Any approved repair and subsequent verification receive separate entries.
Report Re-inspection Findings for the Buyer’s Decision
When findings exceed agreed limits or reveal a safety concern, UTS records and reports the actual findings to the buyer. The buyer or an expressly authorized representative decides shipment disposition; the supplier identifies the affected production scope and carries out authorized containment or correction. The inspection report does not prescribe corrective-action recommendations.
Verify approved rework against the original requirements, including any required replacement of damaged or single-use hardware. Apply the agreed re-inspection plan rather than presenting only the repaired sample. The release record should identify the lot, verification results, remaining limitations and the person authorized to approve shipment.
