ORIGAMI-LINE_
A converter needed a folded insert produced at line rate with a repeatable crease. Hand-folding held tolerance about two thirds of the time.
- Cycle
- 4.20 s/part
- median of 2000 parts
- Yield
- 96.4%
- n = 2000 parts, inspected against a go/no-go gauge
- Stations
- 6
- Build
- 9 d
- Brief → Six stations, continuous
- △3 · 02-04
- Published sheet, sections 00–06.
- △2 · 01-20
- Index beam retimed; cycle 5.1 → 4.2 s.
- △1 · 01-18
- Station 4 blade radius 0.8 → 1.2 mm after tearing at 200 gsm.
- Crease position
- 0.40 mm · calipers, n = 60
- Footprint
- 2.40 m
- Parts
- 268
What we were asked for.
The client folds a paperboard insert that has to seat inside a moulded tray. Hand-folded, it seated about two thirds of the time, and the failures were all in one place: the second crease drifted by a couple of millimetres depending on who folded it.
They did not want a production line. They wanted to know whether the part could be folded repeatably at all, and roughly what a cell would cost, before committing to tooling.
The targets, and what we actually hit.
| Constraint | Target | Achieved | How measured |
|---|---|---|---|
| Cycle time | < 5 s/part | 4.2 s | median over a 2000-part run |
| Crease position repeatability | ± 0.5 mm | ± 0.4 mm | calipers against a datum edge, n = 60 |
| Yield | > 95% | 96.4% | go/no-go gauge, n = 2000 |
| Footprint | Fits a 2.5 m bay | 2.4 m | tape |
| Changeover between part sizes | < 10 min | 14 minMissed | timed, two operators, three changeovers |
Targets were set at kickoff, before the first part was cut. A target we missed stays on the sheet with the number we actually got.
How it moves.
Six stations on a shared indexing beam, each doing exactly one fold. Splitting the folds rather than building one clever multi-axis head is what made nine days possible — every station is the same two parts with a different blade profile, so five of the six were the same fixture cut five times.
The crease is set by a hard stop, not by a servo position. A servo would have been more flexible and less repeatable; the stop is a ground pin against a machined face and it does not drift.
Measured, and where it fails.
Where it fails
- Stock outside 180–220 gsm creases inconsistently. The cell does not detect this; the operator has to.
- Humidity above roughly 60% RH softens the fibre enough that station four starts tearing rather than folding.
- A jam at station two is not detected until station five, which wastes three parts.
What it did.
A 2000-part continuous run at 4.2 s per part with 96.4% yield against a go/no-go gauge. The client took the cell, the CAD and the controller source, and decided against tooling for a second part size.
Changeover missed its target by four minutes. We told them it would, and why: the hard stops that make the crease repeatable are the same thing that makes a changeover slow. Fixing it means giving up the stop.
Elapsed days between recorded milestones, drawn to scale from the dates. The overall dimension runs from the brief to the last entry.
What changed since.
| Rev | Date | Change | By |
|---|---|---|---|
| △3 | 2026-02-04 | Published sheet, sections 00–06. | DZ |
| △2 | 2026-01-20 | Index beam retimed; cycle 5.1 → 4.2 s. | RM |
| △1 | 2026-01-18 | Station 4 blade radius 0.8 → 1.2 mm after tearing at 200 gsm. | DZ |
What we would do next
Put a photo-interrupter at each station so a jam is caught at the station that caused it rather than three stations later.
The changeover is a quick-release problem, not a control problem. Kinematic mounts on the blade holders would take it under ten minutes without touching the stops.
Who built it, and when.
- Entry
- CLIENT
- Client
- A packaging converter, mid-size, one plant
- Status
- ● SHIPPED
- Brief
- 2026-01-12
- Demo
- 2026-01-21
- Elapsed
- 9 d
- Built by
- DZ, RM
- MEC
- Mechanical design
- FAB
- Fabrication and finishing
- FRM
- Firmware
- CTL
- Controls and motion