Prototype to Serial Production for Sheet Metal Parts

The prototype fitted on the bench and the first serial shipment did not. Nothing was redesigned. What changed was the process: a laser-cut, brake-formed sample became a tooled part, and the tool brought its own tolerances, its own bend sequence and a coating film that had never been on the sample.

Prototype to serial production for a sheet metal part is a process transition, not a drawing transition. The drawing can stay at revision C while the route moves from a laser and a press brake to a progressive die, a welding fixture and a powder line. Everything that the prototype absorbed by hand has to be written down before the tool is cut.

Why does the serial part stop fitting when the prototype fitted?

A prototype is a process of one. The blank is laser cut to the drawing, the bends are set on a brake, and any feature that comes out slightly wrong is adjusted before the part leaves the bench. Nothing about that route repeats at volume.

A serial part inherits four new sources of variation. The die sets its own hole positions relative to one datum. The strip carrier moves the blank outline. A fixture holds the part for welding in a position the prototype was never welded in. And the coating adds a film to every surface, including the ones that were checked before paint.

None of that is a defect. It is the difference between a part that was made and a part that is produced. The handover works when each of those four is specified before the tool is cut rather than discovered in the pilot batch.

What changes at each stage?

The quantity bands are guidance, not thresholds. A 300-piece-a-year programme may never leave the pilot column, and a 5,000-piece-a-year programme may sit on a single-stage tool for its whole life. What matters is that each gate has evidence, and that the evidence is produced against the same drawing revision that production will use. The 200–500 band is the pilot proper: it supports the 30-piece capability study and a full PPAP submission. A 50–200 piece soft-tool run is a pre-pilot, which proves the tool can make the part but is too small for the capability evidence the pilot carries.

What has to be frozen before the tool is cut?

Six items decide whether the tool survives the first year.

The datum scheme. Three datums, and a clear statement of which feature is primary. If the prototype was measured from a different face than the assembly locates on, the tool will pass its own inspection and fail at the customer.

The hole pattern. Hole position, diameter and tolerance. A hole added after the die is cut means a new station or a second operation, and a hole moved by 2 mm can move a whole station.

Material grade and thickness band. Not just "1.5 mm CRCA" but the grade and the permitted thickness tolerance. A tool built for nominal thickness will produce a different burr at the top of the band.

The bend sequence and bend reliefs. A progressive die forms in a fixed order, and a part that needs a bend inside another bend needs a relief that the prototype may have done without.

The coating specification. Film thickness, the surfaces that must be masked, and any tolerance that the film will close. This is the single most common cause of a serial part that will not assemble.

Packaging and orientation. A coated part that arrives scratched is a failed transition, whatever the dimensional report says.

What does the same bracket cost in time at each stage?

Take a bracket of 220 mm × 150 mm in 1.5 mm CRCA with four bends, 6,000 pieces a year.

As a prototype the part is laser cut and bent. The cut perimeter is about 1.2 m, and at 15 m/min on a fibre laser that is 4.8 seconds of cutting. Four bends at 20 seconds each, plus 20 seconds of handling, gives about 105 seconds per part. Twenty prototypes take about 35 minutes of machine time.

As a serial part the same bracket runs through a four-station progressive die at 25 strokes a minute, which is 2.4 seconds per part. The 6,000-piece annual volume is 14,400 seconds of press time, or four hours for the year.

The ratio is about 44:1. That ratio is the reason a tool exists, and it is also the reason the tool must be built for the part that will actually be ordered, not for the prototype that was actually approved.

What should a pilot batch prove?

A pilot batch is the cheapest place to find a problem. It runs on the production tool, at production speed, with production material, and it produces enough parts for statistical evidence.

Three things need to come out of it:

• A capability study. Thirty consecutive parts on the two or three critical characteristics, with a Cpk of 1.33 or better before release.
• A full dimensional report on a sample drawn from the batch, not from the first ten pieces.
• Parts through the customer's assembly, or through a representative fixture, including the coating. A pilot batch that is only measured and never assembled has tested the drawing, not the part.

For attribute checks on the batch, ISO 2859-1 at AQL 1.0 and general inspection level II asks for 80 pieces from a lot of 501–1,200, accepting 2 and rejecting at 3. That is a lot check; it is not a substitute for the capability study on the critical dimensions.

What does a PPAP submission contain for a stamped part?

A PPAP submission is the evidence pack that says the process can make the part repeatedly. For a stamped and coated assembly it typically carries:

• The design record and the authorised engineering change, at the revision being submitted.
• A process flow diagram, station by station, including outsourced operations.
• A failure mode and effects analysis for the process, with the high-risk items linked to the control plan.
• A control plan that names the characteristic, the method, the sample size and the reaction plan.
• A measurement system analysis for each gauge and fixture used on a special characteristic.
• Dimensional results on the parts tested.
• Material and performance test results, including coating thickness and adhesion where specified.
• The part submission warrant, signed.

The submission is only as good as the control plan behind it. A PPAP that lists a 100% inspection on a characteristic the shop actually checks once a shift will not survive the first audit.

How much does coating move a dimension?

More than most drawings allow for. Powder coat builds a film on every surface the gun can reach, and the film is measured dry, so a 70 µm coat takes 0.07 mm off each side of an opening.

A 6 mm slot becomes 5.86 mm after a 70 µm film on both faces. A 12 mm hole becomes 11.86 mm. A shaft that needed 0.2 mm of clearance now has 0.06 mm. None of these fails a dimensional check taken before coating, which is exactly why the drawing has to say where the measurement happens.

The drawing fixes it in one of three ways: state the dimension as pre-coat and add a note, open the feature by twice the film thickness, or mask the surface. Masking costs handling time per part and is worth it only where the fit is critical or the surface has to conduct, such as an earth stud or a weld pad.

Measure coating thickness with a magnetic gauge on the finished part, not on a test coupon from the same batch. A coupon on the jig and a part in the corner of the booth do not see the same film.

What goes wrong most often in the handover?

Five failures account for most prototype-to-serial problems, and each one can be designed out before the tool exists.

• A hole pattern that drifts because the prototype was drilled and the serial part is pierced from one datum. Specify position from the datums, not from the edge.
• A missing bend relief where two bends meet. The prototype was bent in a different order and got away with it; the die cannot.
• A weld that moves the frame because the fixture clamps on a formed face and pulls the part square. Locate on datums, clamp on the flat.
• A coating that closes a fit, as above. Decide pre-coat or post-coat measurement on the drawing.
• Packaging that damages the finish on the first shipment. Specify the dunnage with the part, not after the complaint.

The cheapest insurance is bridge stock: hold 50–100 pilot parts from the approved batch so that production can start while a tool correction is made, instead of stopping the customer's line to prove a point.

What Ganesh runs across the transition

Ganesh has manufactured since 1992 and runs the whole route in-house: a 12 kW fibre laser with a 40 ft × 16 ft bed for prototype blanks, mechanical presses from 20 T to 300 T with a 250 T punching press for the serial tool, a 200 T, 5-axis press brake with a 4,450 mm bed, three CLOOS robotic MIG cells and 30-plus manual welding stations, plus seven-tank pretreatment, shot blasting and an 18 ft × 14 ft powder booth.

Because the tool room, the fixtures and the checking gauges are all built in-house, the datum scheme agreed at the prototype stage is the one carried into the serial tool. Custom parts run from 0.8 mm to 40 mm at ±0.5 mm typical tolerance, with first article inspection and PPAP documentation available, and the plants hold ISO 9001:2015 and ISO 14001:2015. The route supplies programmes for Mahindra, Mahindra Powerol, Godrej & Boyce, TKE, Eros Elevators, Minda, Gunnebo and Indian Railways.

The stamping capability page covers the press range and the tool room, and the finished coated assemblies page covers what changes when the part is painted before dispatch. The stamping and press tool hub collects the related design rules.

What to put in the RFQ for each stage

• The drawing revision, and a note on which features are still open.
• The stage you are buying: sample, pilot or serial.
• Annual volume and release size for the serial stage.
• The datum scheme, or a request for us to propose one.
• Material grade, thickness band and the standard reference.
• Coating specification, masking and the surfaces that must conduct.
• Tool ownership, and whether the tool must be portable to another site.
• The evidence you need at each gate: first article, capability study or full PPAP.

The build-to-print guide follows a drawing through quotation, tooling, production and dispatch, and the RFQ checklist keeps the stage-specific fields in one place. Use the weight calculator to sanity-check material consumption per part before the volume conversation, then send the revision through the quotation route.

When the transition should stop at the prototype

A prototype to serial programme is not always the right plan. If the annual volume is under about 500 pieces, a tool may never pay back, and the part should stay on the laser and brake with a fixture for welding; the low-volume stamping guide sets out where that line sits.

The same applies when the design is still moving. A prototype tool built against a revision that changes twice is a tool built twice. Keep the sample on the laser route, run the laser route into the first small series, and cut the die when the drawing and the volume both hold still.

And if the prototype needs a process we do not run, the honest answer is to say so at quotation rather than to bridge it with a fixture that will not scale. The laser cutting for prototype and production quantities guide covers where the laser route stops being competitive.