Sheet Metal Bracket: Rework, Respecify, or Scrap
A fabricator runs a 3 mm thick 5052-H32 aluminum bracket on a CNC press brake. The drawing calls for a 90° bend with a ±1° tolerance and a 12 mm inside bend radius. After a 400-piece lot, inspection finds the bend angle averages 91.6° with a standard deviation of 0.4°, and the inside radius measures 9.5 mm on most parts. The press brake has a fixed punch and a V-die whose opening is 40 mm. The customer's mating assembly uses a locating pin that passes through a 10.2 mm hole 25 mm from the bend line; the angle error shifts that hole roughly 0.3 mm per degree of deviation. The customer will accept parts only if the pin still engages with a 0.5 mm positional clearance. Material lot certificates are in order and no cracking is visible.
1. Diagnose the two out-of-spec conditions (angle and radius) and state which is a process-capability problem versus a tooling/design mismatch.
Read the worked answer
Angle: mean 91.6° against a 90° ±1° spec means the process is centered 1.6° high with σ = 0.4°. Upper spec limit is 91.0°, so the mean is already ~1.5σ above USL - essentially the whole lot is out of tolerance. This is a setup/capability problem: the ram depth or bend allowance is not dialed in, and even a perfectly centered process would have Cpk ≈ (1.0)/(3×0.4) ≈ 0.83, below the usual 1.33 target. It needs a ram-depth correction and a springback/angle-correction pass, not a redesign.
Radius: 9.5 mm actual against 12 mm specified. For air bending, the natural inside radius is roughly a fraction of the die opening; a 40 mm V-die on 3 mm 5052 will not naturally produce a 12 mm inside radius - the tooling cannot form it. This is a tooling/design mismatch, not a capability issue: the die opening and punch tip radius must be changed (or the part coined/bottomed) to hit 12 mm. Confirm with the bend-radius-to-die-opening relationship and the material's minimum bend radius before choosing new tooling.
2. Given the hole-position shift, is the lot usable if the angle is corrected? Show the clearance math and recommend a disposition.
Read the worked answer
The 0.5 mm clearance budget is consumed at roughly 0.3 mm per degree. At the current mean deviation of 1.6°, shift ≈ 0.48 mm - right at the edge of the 0.5 mm budget, leaving almost nothing for hole-position tolerance, material thickness variation, and pin fit. Even if every part were corrected to exactly 91.0° (the worst in-spec angle), shift ≈ 0.3 mm, consuming 60% of the budget.
Disposition: Do not ship as-is. Correct the ram depth to bring the mean to 90° and re-measure; if the corrected mean holds within ±0.5°, the shift drops to ~0.15 mm and the lot becomes usable with margin. Parts already formed at 91.6° mean should be sorted: those within 90.5° can be reworked (re-strike) and re-inspected; the rest are scrap or concession candidates. Also fix the radius tooling before the next lot, since the 9.5 mm radius is a separate nonconformance that will recur.
3. What verification steps would you require before releasing reworked parts?
Read the worked answer
- First-article re-inspection on a sample spanning the rework batch: angle (with a digital protractor or CMM), inside radius, and hole-to-bend-line distance.
- Functional gauge check: build a go/no-go fixture replicating the 10.2 mm pin and 0.5 mm clearance to confirm engagement, since the angle/position interaction is what the customer actually cares about.
- Capability re-check: run 30+ consecutive parts after the ram-depth correction and compute Cpk against the ±1° spec; require ≥1.33 before full release.
- Radius conformance: verify the new tooling produces 12 mm ± drawing tolerance across the lot, and check for cracking/thinning at the bend (especially with the tighter radius).
- Documentation: update the setup sheet, record the tooling change, and note the disposition of sorted/scrapped parts for traceability.