Balancing Evidence Review — sample report
This is what the €1,900 fixed-scope review actually delivers: an evidence matrix, a decision classification, and an explicit release-boundary statement — shown here on a realistic but fictional rotor case.
1 · The decision under review
Principle: the cheapest moment to catch a balancing-evidence gap is before metal is cut and fixtures are made. Once the manufacturing route, correction process, and assembly fixtures are fixed, balancing may be forced to compensate for decisions that should have been handled earlier in design review.
The manufacturer intends to release a 65,000 rpm aerostatic grinding spindle for optical-component finishing. The rotor (0.85 kg, integral shaft with tool adapter) has been balanced to grade G0.4 per ISO 21940-11 and passed a factory vibration check. The release question submitted for review: is the balancing evidence sufficient to support release for the intended surface-finish-critical application?
2 · Rotor regime screening
The first gate: which balancing methodology is even applicable. The rotor is treated by the manufacturer as rigid across the operating range. Supplied documentation contains a first-bending-mode estimate from an early design calculation (reported margin ≈ 2.1× above operating speed, or roughly 136,000 rpm for a 65,000 rpm operating point), but the calculation predates two design revisions of the tool adapter, and no updated modal estimate or test data was supplied. The rigid-rotor assumption is therefore plausible but currently unverified for the as-built configuration — and every downstream balancing conclusion inherits this status.
Numerical sanity check: for a 0.85 kg rotor at 65,000 rpm, G0.4 corresponds to an order of only ≈0.05 g·mm total permissible residual unbalance before allocation to tolerance/correction planes. At this level, the measurement setup, centring, drive, filtering, and repeatability are not secondary details; they are part of the evidence.
3 · Evidence matrix
| Evidence item | Status | Basis found | Consequence for release |
|---|---|---|---|
| Rotor regime (rigid assumption) | ASSUMED | Pre-revision design calc, margin ≈ 2.1×; not updated for current adapter | If margin has narrowed, low-speed balance may not represent operating-speed state |
| Balance grade selection (G0.4) | ASSUMED | Company practice for air-bearing spindles; no rotor-specific derivation from vibration sensitivity or bearing-integrity limit | Grade is plausibly conservative, but its sufficiency for surface-finish criticality is asserted, not derived |
| Pre-balance manufacturing baseline | SUPPORTED | Incoming rotor concentricity and journal roundness records present; rotors rejected above threshold before balancing | Balancing is not being used to mask manufacturing defects — gate before the gate is in place |
| Correction path | SUPPORTED | Two-plane correction, material removal, plane locations and per-plane amounts recorded per rotor | Correction path is traceable and repeatable |
| Component vs assembly state | UNVERIFIED | Balance certificate covers rotor without grinding-wheel adapter fitted; assembled-state residual not measured | Assembled rotor may not reproduce component-level result; the release-relevant state is the assembled one |
| Measurement apparatus | UNVERIFIED | Soft-suspension machine identified, but filter settings, runout compensation, and repeatability spread not recorded | Reported residual cannot be separated from measurement-chain contribution; tolerance consumption unknown |
| Tolerance-plane allocation | UNVERIFIED | A total grade-derived value is shown; tolerance-plane coordinates, centre-of-mass basis, allocation, and correction-plane mapping are absent | The reported residual / limit comparison is not yet plane-resolved |
| Machine/process capability | UNVERIFIED | No task-specific minimum achievable residual, reduction ratio, plane-separation, angular, or correction-capability record is attached | The process has not demonstrated that G0.4 can be measured and achieved with adequate margin |
| Rotor-state repeatability | UNVERIFIED | No vector spread across repeated starts, relevant thermal states, or reassembly is available | The requested tolerance may sit below the physical evidence floor |
| Assembly transfer | UNVERIFIED | The final adapter was not present and no before/after vector separation was performed in the future planes | Component balance cannot be transferred silently to the final assembly |
| Modal sensitivity | ASSUMED | The early critical-speed estimate does not cover the current adapter, damping provenance, or transient speed path | A favourable separation value alone cannot close operating-speed sensitivity |
| Operating-speed vibration | SUPPORTED | Factory vibration record at 65,000 rpm, unloaded, within house limit | Confirms acceptable unloaded factory behaviour |
| Operating envelope (customer site) | ASSUMED | Factory check is unloaded; process loads, duty cycle, and mounting at customer site differ; no loaded data | Factory vibration evidence does not automatically transfer to process conditions |
4 · Decision classification
Classification: REVIEW. The balancing work itself shows good process discipline — a real pre-balance quality gate and a traceable correction path, which is more than many shops can evidence. The classification is REVIEW rather than PASS because the two strongest claims in the release case (component-level G0.4 and unloaded factory vibration) are attached to states that differ from the release-relevant state: the assembled rotor under process load. The classification is REVIEW rather than HOLD because no evidence found contradicts the release case — the gaps are absences, not conflicts. A targeted assembled-state check is therefore required to upgrade confidence to PASS, but the existing evidence does not justify stopping the programme.
5 · Release-boundary statement
What it is not yet allowed to prove: that the residual is inside correctly allocated tolerance-plane limits after uncertainty; that the machine-process-rotor combination can demonstrate G0.4; that the assembled rotor meets the balancing intent; that the rigid-behaviour assumption holds for the current adapter; that modal sensitivity is acceptable over the complete speed path; or that factory behaviour transfers to loaded customer-site operation.
6 · Next-step actions (in priority order)
- Tolerance-plane definition: define the tolerance planes, centre-of-mass basis, total permissible residual, allocation rule, and mapping to the actual correction planes.
- Machine capability and guard band: attach task-specific minimum achievable residual, reduction-ratio evidence, plane separation, angular capability, and uncorrected total deviation for each tolerance plane.
- Assembled-state check: repeat the balance verification with the final adapter fitted and preserve the complete assembly fingerprint.
- Vector repeatability record: capture Ux/Uy or magnitude/phase over repeated starts, relevant settling states, and any required reassembly; report mean vector and spread.
- Modal-sensitivity refresh: update critical speeds, mode participation, damping provenance, and the acceleration/deceleration path for the current adapter revision.
- Grade rationale note: link G0.4 to surface-finish sensitivity, bearing integrity, or another rotor-specific consequence.
- Transfer and reassembly check: if the adapter is installed incrementally or may be replaced, separate the added vector and demonstrate repeatable reassembly.