Cutting, bending, forming, welding and stamping — and how parts are quoted flat

Sheet metal starts as flat stock and becomes a part through cutting (laser, punch, waterjet) and bending (press brake), with secondary forming, welding and hardware. Every quote is derived from the flat pattern — the unfolded geometry — so 3D models must be bend-aware. Standard tolerance is ±0.3 mm on cut and ±0.5 mm on formed dimensions; costs are driven by thickness, bend count, hardware and finish.

1. Cutting processes

Laser cutting is the default for most sheet parts: kerf 0.1–0.3 mm, edge quality good, thickness 0.5–12 mm in steel/aluminum/stainless. Punching (CNC turret) is faster for high-volume repetitive holes; waterjet cuts thick or heat-sensitive materials but slower and costlier.

Minimum hole ≈ sheet thickness (punch) or 0.8 mm (laser); hole-to-edge ≥ 1.5× thickness. Cut edges are slightly tapered (laser) or sheared (punch) — specify deburring or edge-breaking where fingers will touch.

2. Bending and forming

Press-brake bending forms the flat into 3D: minimum inside radius ≈ thickness for air bending, and bend length up to ~3 m on large brakes. Springback varies with material and temper — formed tolerances run ±0.3–0.5 mm. Bend reliefs (3–6 mm) prevent tearing at corners.

Forming operations add depth: hemming (rolled edges), countersinks, louvering, embossing and coining. Each adds a setup cost; minimize distinct bend radii on a part to keep the brake setup short.

3. Hardware, welding and finishing

Captive hardware (PEM® studs/nuts/standoffs) is pressed in after forming; specify the standard panel thickness range. Welding (MIG/TIG/spot) joins parts and weldments; weld distortion is real — keep weldment tolerances ≥ ±0.5 mm and put welds on non-visible faces.

Finishes: powder coat (25–100 µm), zinc plating, anodizing (aluminum), passivation (stainless), plus silkscreen labels. Finish covers cut edges, so deburr before coating for a clean line.

4. How sheet parts are quoted

The engine unfolds the flat, then prices: material area × thickness × density × price index, plus laser time, bend count × bend rate, hardware items, welding minutes, finish area and tooling (flat/blank dies for volume). Quantity tiers discount setup-heavy small runs.

Design for the flat: model bends with a consistent K-factor, avoid bends too close to holes (≥ 2× thickness), keep bend lines away from hardware, and let the drawing state bend radius and direction.

Quick reference

Parameter	Typical range
Thickness	0.6–6.35 mm (0.024–0.250 in)
Cut tolerance	±0.13–0.3 mm (laser)
Formed tolerance	±0.3–0.5 mm
Min bend radius	= thickness (air bend)
Max part size	~3000 mm (laser), 1250 mm bend
Min hole	0.8 mm laser / 1×t punch
Corner relief	3–6 mm
Finish thickness	Powder 25–100 µm

Sources & references

• ISO 2768-1:1989
• ASTM A1008 / A240 / B209 (stock)
• Xometry Learning Center — Sheet Metal Basics (public engineering reference)

How molds are built, why tooling costs, and what drives molded part price

Injection molding melts plastic pellets and injects them into a closed steel/aluminum cavity under pressure; the part cools, shrinks and is ejected. The mold is the investment — aluminum/prototype tools from ~$2k–8k, production steel tools $10k–100k+ — and unit cost collapses with volume. Design rules (draft, wall uniformity, gating) are locked at tool build, so DFM review before tooling is the highest-leverage engineering step.

1. The molding cycle

Each cycle: clamp the mold, inject melt (~0.1–2 s per mm of wall), pack under pressure, cool (the longest step), eject. Cycle time scales with wall thickness — a 2 mm wall part cycles in ~20–40 s; thinner walls cycle faster but need higher pressure and better flow.

Production volumes reach 100k–1M+ parts per tool with hardened steel; aluminum tools are good for 1k–100k and cheaper. T1 samples (first shots) are as fast as 5 working days on prototype tools — the platform quotes tooling separately, amortized in per-part comparisons.

2. Tooling classes and cost drivers

Mold cost scales with: number of cavities (1-cavity prototype vs 8+ production), complexity (side actions, lifters, inserts, textures), steel grade and size. A simple 1-cavity aluminum tool ≈ $2k–8k; a 4-cavity steel tool with slides ≈ $20k–80k.

The quoting engine prices tooling from part volume and feature complexity: toolingBase + toolingPerVol × part volume, plus per-feature charges for undercuts, threads and texture. Mold ownership transfers to the customer.

3. Part cost structure

Molded unit cost = resin cost (weight × price/kg) + machine cycle cost + finishing/assembly + margin, with quantity tiers discounting setup. Resin is usually a small fraction; wall thickness and cycle time dominate — every 1 mm of extra wall adds seconds per cycle.

At 10k+ units, per-part cost can drop below $1 for small parts; at 100k+, molding is typically 10–50× cheaper per part than machining the same geometry.

4. DFM checklist before tooling

Uniform walls (1–4 mm), draft 1–2° on all vertical faces, ribs 0.5–0.6× wall, no zero-draft undercuts unless side actions are budgeted, gates/ejector marks on non-cosmetic faces, generous radii at transitions, and shrinkage allowance per resin (ABS 0.4–0.7%, PP 1–2.5%).

Get a DFM review before cutting steel: the platform flags wall ratio, draft, and feature issues on upload, and recommends tooling class by quantity.

Quick reference

Tool class	Material	Shots	Lead time	Typical cost
Prototype (Class 105)	Aluminum / P20	1k–10k	5–15 days	$2k–8k
Bridge (Class 103)	P20 / H13	10k–100k	15–25 days	$8k–30k
Production (Class 101)	Hardened steel	100k–1M+	25–40 days	$20k–100k+
High-cavity (8+)	Hardened steel	500k–1M+	35–60 days	$50k–200k+

Sources & references

• DIN 16742 — Plastic part tolerances
• ISO 294-4 — Shrinkage
• SPI mold classification
• Xometry Learning Center — Injection Molding Basics (public engineering reference)

When a prototype process becomes a production process, and the crossover decisions

Prototyping optimizes speed and iteration; production optimizes unit cost and repeatability. The same part often starts printed or machined and migrates to molding/stamping/die-casting at volume. The crossover is arithmetic — tooling cost divided by the unit-cost delta — and the platform quotes both paths so the decision is explicit.

1. What each phase optimizes

Prototype: prove function, fit and manufacturability. Tolerances can be slightly looser; surface finish cosmetic; quantity 1–100. Speed and iteration count matter more than unit cost — printed nylon or machined 6061 dominate this phase.

Production: minimize unit cost at target quality. Processes that amortize tooling (molding, stamping, die-casting, investment casting) win at volume; tolerances and process control are locked in with the tool.

2. The crossover quantity

Crossover = tooling cost ÷ (unit-cost delta). A $5k mold with $2 saved per unit crosses at 2,500 units; a $20k mold with $5 saved crosses at 4,000. Printed→molded crossover is typically 300–1,000; machined→stamped/die-cast crossover is 1,000–10,000 depending on complexity.

Include secondary savings: molded parts may integrate fasteners (inserts), reduce assembly steps, and improve consistency — the true crossover is often lower than the pure per-part math.

3. Production-grade prototyping

Rapid tooling (aluminum molds, 3D-printed tool inserts) gives near-production parts at 1–10k quantities with molded properties and cycle times — the ideal validation before committing to hardened steel.

DMLS metal printing produces end-use parts for low-volume aerospace/medical where tooling is unjustified; MJF/SLS nylon does the same for consumer brackets. Prototype and production are not always different processes — they are different quantities of the same process.

4. Decision framework

Ask three questions: quantity (now and planned), tolerance/surface needs, and how fast the design will change. High iteration → print/machine; stable design + volume → tool; hybrid → rapid tooling first.

The quoting engine returns price/lead-time curves across processes for the same model, making the crossover visible instead of guessed.

Quick reference

Phase	Quantity	Process	Key metric
Concept	1–10	FDM/SLA print	Speed, cost
Functional proto	10–100	SLS/MJF, CNC	Strength, tolerance
Pilot / rapid tooling	100–10k	Aluminum mold, 3DP tooling	Near-production parts
Production	10k–1M+	Steel mold, stamping, die-cast	Unit cost, Cpk
Bridge	varies	CNC + EDM	Volume without tooling

Sources & references

• ISO/ASTM 52900:2021
• SPI mold classification
• Xometry Learning Center — Prototype vs Production (public engineering reference)

Post-processing: deburring, blasting, coating, plating, heat treat and marking

Secondary operations convert a raw machined/printed part into a finished deliverable: deburring and edge-breaking, surface prep (blasting), protective or cosmetic coatings (anodize, powder, paint, plating), heat treatment, and identification (laser marking, silkscreen). They add cost per part but are what make parts functional, durable and traceable — specify them explicitly on the drawing.

1. Mechanical finishing

Deburring (vibratory, tumbling, manual) removes edge burrs from machining and stamping — free-ish on CNC, always required for safety and coating adhesion. Bead/sand blasting textures and homogenizes surfaces (Ra 0.8–1.6 µm typical) and hides tool marks; polishing takes surfaces to Ra 0.2–0.4 µm for optics and seals.

Shot peening compresses the surface for fatigue life — specified on springs, axles and aerospace brackets. Grinding and lapping correct geometry and achieve the tightest tolerances (≤ ±0.01 mm).

2. Coatings and plating

Aluminum: anodize Type II (5–25 µm, decorative, dyed), Type III hard coat (25–75 µm, wear); chromate conversion under paint. Steel: zinc plating (5–12 µm), nickel (hardness/anti-corrosion), black oxide (cosmetic), phosphate+oil. Stainless: passivation (citric/nitric) to restore oxide and remove embedded iron.

Polymers: powder coat (25–100 µm) and paint for appearance/UV; dyeing for printed nylon; clear coat for acrylic. Electroless nickel covers complex geometry uniformly; electroplating needs line-of-sight. Salt-spray ratings (ASTM B117 / ISO 9227 hours) define corrosion targets.

3. Heat treatment

Steels: quench-and-temper (4140/4340) to target hardness; case hardening (carburizing, nitriding) for wear surfaces; annealing before machining for softness. Aluminum: T6 aging for strength; stress-relief before final machining for stability. Titanium: annealing and stress-relief.

Specify hardness range (e.g. 28–32 HRC) rather than vague “hardened”, and note any distortion allowance — heat treatment moves thin features.

4. Marking and documentation

Laser marking (engraved, 2D DataMatrix, serial numbers) is permanent and preferred for traceability; silkscreen and pad printing for labels; embossing for plastics in mold. FAI and material certificates (EN 10204 3.1) ship with orders that require them.

Each secondary operation is a cost line on the quote; combining operations (anodize + laser mark) in one vendor pass saves money. Mark the drawing with every operation, its spec (e.g. MIL-SPEC anodize class, salt-spray hours) and the cosmetic face.

Quick reference

Operation	Typical spec	Effect
Deburr / vibratory	All edges	Safety, coating adhesion
Bead blast	Ra 0.8–1.6 µm	Matte, uniform surface
Anodize Type II	5–25 µm, dyed	Corrosion + color
Hard anodize III	25–75 µm	Wear, insulation
Zinc plate	5–12 µm + clear/chromate	Steel corrosion
Passivation	Citric / nitric	Stainless oxide restore
Q&T 4140	28–32 HRC	Strength
Laser mark	DataMatrix	Traceability

Sources & references

• ASTM B117 / ISO 9227 — Salt spray
• EN 10204 — Material certificates
• MIL-A-8625 — Anodize classes
• Xometry Learning Center — Secondary Operations (public engineering reference)

Sinker and wire EDM: hardened materials, tight corners and burr-free fine features

Electrical discharge machining erodes material with controlled sparks — a wire (wire EDM) or a shaped electrode (sinker EDM) — in deionized water or oil. It machines any conductive material regardless of hardness, with no cutting force and no burrs, making it the standard for hardened tooling, carbide, fine slots, tight internal corners and features no end mill can reach.

1. Wire EDM

A brass/zinc wire 0.1–0.3 mm diameter erodes along a programmed path, cutting through the workpiece. Accuracy ±0.005–0.01 mm, surface Ra 0.2–0.8 µm, and a minimum internal corner radius equal to the wire radius — far tighter than any milling cutter.

Wire EDM cuts hardened steel, carbide, titanium and exotic alloys with zero burr and zero force: delicate tabs, fine slots (0.1 mm+), and tool steel die details are routine. It is slower per hour than milling, so it is used where milling cannot go.

2. Sinker EDM

A CNC-machined electrode (graphite or copper) is sunk into the workpiece to burn the inverse shape: cavities, deep ribs, ribs in hardened molds, and blind features with sharp corners. It also forms threaded holes (with orbiting electrodes) that no tap can cut in hardened steel.

Sinker EDM achieves ±0.01–0.02 mm with fine finishes (Ra 0.4 µm polished). Electrode wear means multiple electrodes for deep cavities; cost scales with cavity volume and finish requirement.

3. Design rules for EDM

Specify EDM only for: hardened material, sharp internal corners (≤ 0.25 mm radius), burr-free edges, thin slots (< 0.5 mm) and deep ribs. Everywhere else, milling is faster and cheaper.

Add a minimum of 0.1–0.2 mm stock for EDM on pre-machined surfaces; note the spark gap (~0.02–0.05 mm) that the process compensates automatically. Avoid EDM on aluminum/copper (milling wins) — EDM shines on steel, carbide and hard alloys.

4. Cost and lead time

EDM is billed per hour with a material/accuracy multiplier: wire EDM ~2–3× CNC hourly, sinker more for electrode work. Small EDM features add $20–100 per feature; full EDM-cut profiles are quoted by cutting length.

Lead time adds 2–5 days over standard machining for most parts; hardened-steel EDM inserts for molds are typically 7–14 days.

Quick reference

Metric	Wire EDM	Sinker EDM
Best for	Through cuts, slots, profiles	Cavities, blind features, threads
Accuracy	±0.005–0.01 mm	±0.01–0.02 mm
Surface	Ra 0.2–0.8 µm	Ra 0.4 µm (polished)
Min corner	= wire radius (0.05–0.15 mm)	Electrode radius
Materials	Any conductive	Any conductive
Cost index	2–3× CNC	3–5× CNC

Sources & references

• ASTM E407 (metallographic)
• ISO 1101 — GD&T for EDM parts
• Xometry Learning Center — EDM Guide (public engineering reference)

The decision rule: geometry complexity, quantity, and where printing beats machining or molding

3D printing wins when geometry complexity is high, quantity is low, or lead time is critical — the classic “prototype or complex bracket” case. It loses to machining on tight tolerances and fine finishes, and to molding on unit cost above a few thousand parts. The crossover quantity is usually 1–100 parts for printed nylon vs machining, and roughly 500–5,000 for molded parts depending on tooling cost.

1. Where printing dominates

Internal channels, lattices, conformal cooling, impossible undercuts and integrated assemblies print in one piece where machining would need EDM, assembly or welding. SLS/MJF prints self-supporting geometry at no extra cost; DMLS adds metal capability for ducts, manifolds and lightweight structural brackets.

Lead time is the second win: functional prototypes in 2–5 working days, iterated nightly, compress development cycles that would take weeks with tooling or complex machining setups.

2. Where printing loses

Tolerances: printing holds ±0.2–0.5 mm versus ±0.05–0.127 mm machined — mating features, press fits and bearing bores need machining or post-processing. Surfaces are layered (Ra 6–20 µm polymer; finer in DMLS but still textured) and need blasting/polishing for cosmetics.

Unit cost is flat with quantity: printing cost scales almost linearly, so above a few hundred parts, molding (plastic) or stamping/die-casting (metal) drops unit cost 5–20×. Strength is anisotropic in the layer direction — design around it.

3. The crossover math

Compare: printed PA12 bracket ≈ $20–150 each with zero tooling; machined 6061 ≈ $15–80 each (material+setup); molded nylon ≈ $1–8 each after a $2k–10k tool. The crossover for molding is typically 500–5,000 units; below that, printing or machining wins on total cost.

Hybrid is often optimal: print the prototype, validate, then mold or machine the production version. The platform quotes all three so the decision is arithmetic, not opinion.

4. Practical selection rule

Choose 3D printing if: complex internal geometry, quantity ≤ 100, lead time ≤ 5 days, or iterative design. Choose machining if: tight tolerance, fine finish, metal with stock sizes. Choose molding/stamping if: quantity ≥ 500–5,000 and the geometry suits the tool.

When in doubt, quote both paths: the engine returns price and lead time for each, and the difference is usually decisive.

Quick reference

Factor	3D printing	CNC machining	Injection molding
Best quantity	1–100	1–5,000	500–1M+
Typical tolerance	±0.2–0.5 mm	±0.05–0.127 mm	±0.3% (molded)
Surface finish	Ra 6–20 µm (layered)	Ra 0.8–3.2 µm	Texture per tool
Lead time	2–5 days	3–10 days	T1 5 days + production
Tooling cost	$0	$0 (setup only)	$2k–10k+
Geometry freedom	Highest	Medium	High (with draft)
Strength	Anisotropic	Near wrought	Near datasheet

Sources & references

• ISO/ASTM 52900:2021
• ASTM F2924 / F3184 (DMLS)
• Xometry Learning Center — When to 3D Print (public engineering reference)

Comparing powder sintering, resin curing and multi-jet fusion for plastic parts

Three powder- or resin-based plastic processes dominate functional prototyping and low-volume production: SLS (laser sintering of nylon powder), MJF (multi-jet fusion, also nylon powder with an ink-based thermal process) and SLA (laser-cured liquid resin). SLS and MJF produce durable, isotropic-ish nylon parts without supports; SLA gives the finest detail and surface at the cost of brittleness and support cleanup.

1. SLS (Selective Laser Sintering)

SLS fuses nylon (PA12) powder with a CO2 laser, layer by layer, inside a heated chamber. Un-sintered powder supports overhangs and internal channels, so no supports are needed and complex lattices print freely. Layer height 0.1–0.12 mm, tolerance ±0.3% (±0.3 mm), UTS ≈48 MPa with good fatigue.

SLS parts have a matte, slightly porous surface (Ra ~8–12 µm) and are strong enough for functional brackets, enclosures and snap fits. Dyeing and media blasting improve cosmetics; sealing reduces fluid permeability.

2. MJF (Multi Jet Fusion)

MJF deposits fusing and detailing agents on PA12 powder, then passes an infrared lamp to fuse selected regions — faster than laser sintering and with slightly finer detail (0.1 mm layers) and a more uniform surface. Mechanical properties are comparable to SLS; MJF edges are often crisper.

MJF is usually a few percent cheaper per part than SLS at identical geometry and quantity, with marginally better surface finish (Ra ~8 µm). Both are true production processes for end-use nylon parts, not just prototypes.

3. SLA (Stereolithography / resin)

SLA cures liquid photopolymer with a laser (or DLP projector): layers down to 0.025–0.05 mm, tolerance ±0.2% (±0.2 mm), and surface finish Ra 1–3 µm — the best of any polymer printing. Standard resin (UTS ≈44–65 MPa) is rigid but brittle-ish; engineering resins add toughness, heat resistance or flexibility.

The trade-offs: SLA needs support structures that leave marks, parts are photo-degrading and weaker in thin sections, and build size is limited (~300 × 335 × 200 mm typical). Use SLA for visual models, flow testing, molds (for low-volume casting) and fine features; not for high-impact or high-temperature duty.

4. Choosing between them

Durable end-use nylon with complex geometry → SLS or MJF (MJF for finer detail at similar cost). Fine detail, smooth surfaces, optical clarity options → SLA. High temperature or chemical duty → consider PA12-GF, PA11, or switch to metal DMLS. FDM remains the budget option when finish and accuracy are secondary.

All three quote on the platform with material, finish and quantity multipliers; upload the same model to all three to compare price and lead time directly.

Quick reference

Metric	SLS	MJF	SLA
Material	PA12 / PA11 / GF	PA12 / PA11 / TPU	Standard & engineering resins
Layer height	0.1–0.12 mm	0.1 mm	0.025–0.05 mm
Tolerance	±0.3% (±0.3 mm)	±0.3% (±0.3 mm)	±0.2% (±0.2 mm)
Surface	Matte, ~8–12 µm	Smooth matte, ~8 µm	Ra 1–3 µm (smoothest)
Supports	None needed	None needed	Required
Strength	≈48 MPa UTS, tough	≈48 MPa UTS, tough	44–65 MPa, brittle-ish
Best for	Functional nylon parts	Fine-feature nylon parts	Visual, detail, molds

Sources & references

• ISO/ASTM 52911 (powder bed fusion)
• ASTM F2921 (additive designations)
• Xometry Learning Center — SLS vs SLA vs MJF (public engineering reference)

How 3/4/5-axis milling, turning and wire EDM differ, and which to choose

CNC covers three dominant material-removal routes: milling (rotating tool, stationary part), turning (rotating part, stationary tool) and wire EDM (electric discharge through a wire). Milling wins for prismatic, multi-feature parts; turning for axi-symmetric parts at high speed and low cost; EDM for hardened steel, tight internal corners and thin slots that cutting tools cannot reach.

1. Milling: 3, 4 and 5 axes

3-axis milling positions the part on a fixed table and moves the spindle in X/Y/Z — the workhorse for plates, housings and brackets. 4-axis adds a rotary axis (usually around X) for features on the side of a part. 5-axis tilts the spindle or table, allowing undercut-free machining of sculpted faces, deep cavities and turbine-type geometry in one setup.

Choosing more axes does not always mean better parts: 3-axis is the most rigid and cheapest per hour; 5-axis reduces setups and improves surface quality on complex forms but costs more. For most production parts, 3-axis with occasional 4-axis features is the economic optimum.

2. Turning: CNC lathes

Turning spins the workpiece while a fixed tool cuts the profile: shafts, pins, bushings, pulleys and fittings are turned in seconds with excellent concentricity and finish. Live tooling adds milling/drilling of cross features in one operation (mill-turn centers), collapsing what used to be two machines into one part.

Turning tolerances are tight (±0.05 mm routine, ±0.025 mm with care) and surface finishes are fine (Ra 0.8–1.6 µm). Bar-fed lathes make high-volume parts economically; Swiss-type lathes hold ±0.01 mm on long slender parts for medical and electronics.

3. Wire EDM

Wire EDM cuts by spark erosion: a thin wire (0.1–0.3 mm) erodes the workpiece along a programmed path, with the part immersed in deionized water. It machines any conductive material regardless of hardness — hardened tool steel, carbide, exotic alloys — with no cutting force, so delicate features survive.

EDM achieves ±0.005–0.01 mm accuracy and Ra 0.2–0.8 µm surfaces, with minimum corner radius equal to wire radius. It is slow and costly per hour, so reserve it for hardened materials, burr-free fine slots, and features no end mill can reach.

4. Choosing between them

Prismatic part with holes/features → milling; axi-symmetric (shaft, bushing, fitting) → turning; hardened or burr-critical detail → wire EDM; hybrid parts → mill-turn. The quoting engine uses your part envelope and features to suggest the process and its tolerance/finish multipliers.

All three share the same design rules: avoid deep narrow cavities, thin walls and sharp internal corners unless EDM is justified, and keep tolerances on the features that actually mate.

Quick reference

Process	Motion	Best geometry	Tolerance	Surface Ra	Cost index
3-axis milling	Spindle moves X/Y/Z	Prismatic, plates, housings	±0.127 mm	1.6–3.2 µm	1.0
4/5-axis milling	+ rotary/tilting	Complex faces, 1-setup parts	±0.05–0.127 mm	0.8–1.6 µm	1.3–1.6
Turning	Part rotates	Shafts, fittings, pulleys	±0.05 mm	0.8–1.6 µm	0.8–1.0
Mill-turn	Both	Hybrid one-setup parts	±0.05 mm	0.8–1.6 µm	1.1–1.4
Wire EDM	Wire erodes path	Hardened, fine slots, corners	±0.01 mm	0.2–0.8 µm	2.0–3.0

Sources & references

• ISO 230 — Machine tool test codes
• ASME B5.57 — CNC axis performance
• Xometry Learning Center — CNC Overview (public engineering reference)