Verify material, tooling and program before batch release
A trial bend shows whether the proposed material, tooling and program can produce the drawing’s angle, flange dimensions, radius and surface finish. Make a representative part, measure it after unloading, correct the setup as needed and confirm the result on new specimens before releasing the batch.
Agree sample material, drawing and measurement requirements
Before bending, agree the drawing revision, target unloaded angle, radius, flange datums, dimensional and surface limits, measurement positions, repeat-piece count and person authorised to release production. Missing requirements remain open items.
Use the production grade, condition, batch, actual thickness, surface and coating. Record rolling direction relative to the bend line, measured blank length, bend-line position and a unique specimen ID. Prepare instruments and an agreed measurement method.
Check tooling, support and approved setup before bending
Check the machine and program revision, punch/die IDs, V opening, flange support, clearance and removal path against the approved setup. Check machine, tools and clamping-chain ratings for the actual arrangement and contact length using the manufacturers’ instructions; total force is not a load-per-length rating.
Qualified personnel should complete setup, seating, clamping and safety checks under the machine and tooling procedures. Record the stroke and backgauge settings, then run the first numbered specimen.
Measure unloaded angles, radius and flange lengths
For the same unloaded geometry, α is bend rotation and β is the included angle between the legs: α + β = 180°. Retain the controller’s original angle convention. Keep program command, model target and actual measurement separate; compare only after angle convention and loading state have been aligned.
Record the unloaded angle at agreed positions, with instrument, datum and units. Both ends and the middle can be useful positions where the part geometry permits; follow the agreed plan. Measure inside radius and flange dimensions, identifying outside dimensions to the virtual sharp corner versus tangent straight lengths.
An unloaded reading alone cannot quantify springback. A loaded angle requires a separately defined, manufacturer-approved safe measurement method; do not reach into the loaded tooling to obtain a reading.
Check deviations, adjust one factor and confirm on new samples
Compare each result with the agreed requirements before changing anything. A symptom prioritises checks; it does not identify one cause.
| Observation | Next check |
|---|---|
| Consistent angle offset | Check angle convention and measurement, then review compensation under the approved process. |
| Angle varies along the bend | Compare all positions; review setup and support rather than hiding variation with one global offset. |
| Angle correct, legs wrong | Check datums, blank, locating and dimension convention before the unfold rule or K. |
| Piece-to-piece drift | Check material, locating, tooling, measurement and program revision; do not assume one cause. |
Change one approved factor at a time and record before/after revisions. Use differently numbered new specimens for before/after comparison and confirmation; repeatedly bending the same piece is not repeatability evidence. Use the repeat count and limits agreed before the trial.
Record sample identity, settings and measured results
Keep the setup and actual readings together for each numbered specimen so the result can be checked and repeated. Record unmeasured characteristics explicitly and keep calculated targets separate from measurements.
| Record group | Minimum information |
|---|---|
| Identity | Specimen ID; drawing/revision; grade/condition/batch; actual thickness; rolling direction; blank/bend-line geometry. |
| Requirements | Angle/radius/flange and surface targets; datums; limits; positions; agreed repeat count; release owner. |
| Approved setup | Machine/program revision; tool IDs/V opening; approved settings and relevant machine condition. |
| Raw measurements | Specimen/position; command and model target separately; actual value/unit/loading state; instrument/datum; evidence location; operator/date. |
| Decision and change | Pass/fail/not measured; correction and revision; new specimen result; reviewer; open actions and release. |
Once the trial result and remaining questions are recorded, compare Press brakes against the demonstrated part requirements.
Stop for unsafe or unstable results and reassess changes
Stop under the site safety procedure for instability, cracks, unacceptable surface damage, interference or unknown applicable load limits. Investigate material, locating, tools, measurement method and program revision rather than repeatedly changing an offset.
When material batch, condition, thickness, rolling direction, tools, die opening, blank, locating datum, program, sequence or machine condition changes, assess its effect and repeat the relevant checks. Confirm the agreed number of new specimens, resolve open actions and have the designated person release the batch.
Use the trial to check bend deduction
The measured blank length and finished dimensions can help refine the unfolding model for this material and setup. Keep the original readings with the drawing, tooling and program revision so the calculation can be repeated.
For a single-bend specimen with no other forming or trimming, use measured blank length L and outside virtual-sharp dimensions A/B from that same specimen. Only with consistent length units for L, A, B, R and t, measured inside radius R, t > 0, 0 < α_rad < π and an applicable circular-bend model may you use BD = A + B − L; BA = 2(R + t)tan(α_rad/2) − BD; K = (BA/α_rad − R)/t. Here α_rad is the unloaded bend rotation in radians.
Use an effective K only when the measurements are complete, the angle is stable and the circular-bend model fits the specimen. It applies to that model and setup; assess springback separately. The bending article gives the worked derivation.