Custom Metal Fabrication FAQ

Yes, a sample, photo or sketch can support early review or reverse engineering, but photos cannot reliably define wall thickness, internal geometry, tolerance or material condition. Provide the physical sample, six scaled views, at least three reference dimensions, mating location, service load, material direction and estimated quantity. For a worn or distorted sample, confirm whether to reproduce its current state or recover the intended design. Before production, create a revision-controlled 2D drawing; add a STEP model for complex geometry. The customer should also confirm its right to reproduce, modify or manufacture any third-party design.

A budgetary quote is possible, but it should be marked as an estimate rather than a final commitment. The minimum useful input is multi-angle photos or a sketch, major dimensions, material or service environment, function, quantity breaks, finish, delivery location and target date. Unknowns should be listed as assumptions—for example, “estimated as 2.0 mm mild steel, general tolerances and black powder coat.” Update to a firm quote after critical assumptions are closed. New information affecting material mass, weld length, tooling, inspection or packing can change price and lead time.

Yes. The 2D drawing should define material grade/condition, dimensions, tolerances, datums, weld symbols, finish, cosmetic surfaces, revision and special inspection requirements. The 3D model supports geometry and programming but does not automatically replace missing drawing tolerances or notes. A DFM review is completed before production. A DFM proposal does not authorize an undocumented design change. Any change to holes, material, welds, appearance or function requires an approved deviation or revised drawing.

Provide both a PDF 2D drawing and STEP/STP 3D model where possible. The PDF controls dimensions, tolerances, datums, material, welding and finish; STEP supports complex geometry, interference review and programming. DXF can define flat profiles, but state units, scale and whether it is an approved flat pattern or reference only. Part number, revision and units must match across filenames, title block and RFQ. If 2D and 3D conflict, agree the precedence in writing before order rather than asking production to guess.

Specified standards, grades, tempers and sizes can be used subject to sourcing—for example, 304/2B stainless sheet, 5052-H32 aluminum sheet or a defined structural-tube specification rather than a material family alone. State whether heat/lot traceability, certificate, origin or a named brand is required. Non-stock grades, unusual thicknesses and small usage may require a full sheet, coil, bar or mill minimum and may extend lead time. The quote should distinguish net consumption from minimum purchase and define ownership of surplus material.

Prioritize load, weight, corrosion environment, forming, welding, appearance and budget. Density provides a first screen: carbon steel is about 7.85 g/cm³, common austenitic stainless about 8.0 g/cm³ and aluminum alloys about 2.70 g/cm³. Aluminum is roughly one-third the weight of steel at equal volume, but stiffness, joints and local buckling prevent direct weight-for-weight substitution. Coated carbon steel often suits indoor structures; stainless is considered for wet, washdown or cosmetic service; aluminum suits weight and thermal needs. Marine, deicing-salt or chemical service also requires the correct grade, drainage, crevice control and maintenance—not merely “outdoor suitable.”

Manufacturing and sourcing recommendations can be provided, but the project should define load magnitude/direction, support span, allowable deflection, safety factor, impact/fatigue, joint method and environment. Without these inputs, a recommendation is only a starting point and does not prove load safety. For a simplified equal-width plate, bending stiffness scales approximately with t³. Tube design also requires second moment, slenderness, local buckling and welded-joint checks. Confirm the final size by calculation, FEA, prototype load testing or the applicable product standard.

A substitute can be evaluated, but similar chemistry is not complete equivalence. Compare product specification, minimum yield/tensile strength, elongation, hardness, thickness tolerance, formability, weldability, corrosion resistance, surface condition and certification. A stronger but less ductile grade may crack during bending. Before purchasing, document the original grade, candidate, differences and effects and obtain written approval. Never substitute for stock convenience without approval; safety-, regulatory- or customer-qualified parts may require revalidation.

Yes. A typical route may include tube/sheet cutting, drilling or punching, tube/sheet bending, fixturing, tacking, welding, straightening, finishing and assembly. Review tube-to-sheet interfaces, weld access, thermal distortion, drainage, coating coverage and final inspection datums. Avoid long dimension chains from raw edges; tube, sheet and final interfaces should share functional datums. Critical holes affected by welding may be drilled/reamed afterward or verified with a functional gauge.

Not always. Prototypes may use laser cutting, standard brake tools and manual fixtures to limit initial investment; stable production may use dies, dedicated gauges, robotic welding or batch finishing. A process change may reduce cost but must not silently change CTQs, material properties, welds or appearance. Before production, issue a prototype-to-production delta list and reconfirm CTQs, tooling, parameters, finish master and inspection plan. A major route change should trigger partial or full first-article inspection.

Yes, based on material, thickness, one/two-side access, load and service cycles. Self-clinching hardware requires the supplier’s minimum sheet thickness, hardness, hole and edge distance. Rivet nuts suit blind installation but need grip-range and spin-resistance validation. Weld nuts offer high load potential but add spatter, heat distortion and masking. For the first article, check relevant thread go/no-go, location, perpendicularity, push-out or torque-out characteristics and define installation torque for the mating bolt. Do not tack-weld an incorrectly selected clinch fastener as a repair.

Powder coating, electroplating, hot-dip galvanizing, anodizing, cleaning and subsequent hardware assembly can be integrated by project, after confirming substrate, appearance, environment, thickness, masking and inspection. Typical powder dry film is often about 60–120 μm; film on both hole walls reduces diameter by about twice the thickness, so threads, grounding points and telescoping fits need advance planning. Define which parts are finished before assembly, which hardware cannot enter an oven or acid bath, permitted touch-up and approved color/gloss limits. ISO 1461:2022 is relevant to hot-dip galvanized fabrications, while distortion, vent/drain holes and fits still require separate review.

Tolerance depends on material, thickness, size range, process, datums and function. Mark dimensions affecting location, sealing, safety and interchangeability as CTQs for individual agreement; use an agreed general tolerance for the rest. Applying ±0.05 mm everywhere may force secondary machining and 100% inspection without improving assembly. For patterns and interfaces, use functional datums and position tolerance rather than chained ± dimensions. ISO 1101:2017 covers geometric tolerancing and ISO 5459:2024 datum systems. State whether dimensions are inspected in final condition after welding, straightening and coating.

A control plan commonly covers incoming verification, first-off, in-process patrol, last-off and pre-shipment inspection. First-off prevents a bad start, patrol detects tool wear/process drift, and last-off defines the lot endpoint and next-setup reference. CTQs may receive increased frequency or 100% mistake-proofing; stable ordinary characteristics may be sampled. Where AQL applies, use the contract edition of ISO 2859-1:2026: lot size, inspection level, AQL and normal/tightened/reduced state determine sample size and Ac/Re. AQL does not mean “AQL% defects are allowed in this lot” and does not replace SPC or 100% checks for critical safety items.

Project documentation may include ballooned dimensional reports, first-article reports, material certificates, heat/lot traceability, welding or finishing records, coating/color data, functional results, certificate of conformity and shipment photos. Documentation affects inspection effort and price, so define it during RFQ rather than just before shipment. Reports should show actual values, gauge IDs and drawing revision—not only “PASS.” For third-party testing, full-dimensional FAI, PPAP or an industry-specific format, define the standard, submission level, sample quantity and approver.

First stop the affected operation and quarantine suspect product made since the last accepted check, retaining part, lot, quantity and defect evidence. Determine scope and choose rework, repair, scrap, 100% screening or a deviation request. “It can be assembled” is not authorization to ship without written customer approval. Repeated or major issues require root-cause and corrective action, such as 5 Why, fishbone analysis, validation trials and control-plan updates. Reworked parts must be reinspected to the original or approved repair criteria with lot and rework traceability retained.

MOQ is not one fixed number for every part. It is affected by full-sheet/bar purchasing, minimum finishing line charge, programming/setup, tooling, destructive tests and packing. Even when one part is possible, its price may carry every one-time cost. Request prototype, first-lot and annual price breaks, such as 5/50/200/1,000 pieces. Unit cost can be viewed as material and cycle cost plus (programming, setup, first-off, tooling and minimum lot charges)/quantity. Do not overbuy solely for a lower unit price; inventory, revision and obsolescence also cost money.

Lead time depends on drawing approval, material arrival, programming, tooling, fabrication, outsourced finish, inspection and rework allowance. Samples requiring unusual material, dies, welding fixtures, custom color or third-party testing normally take longer than ordinary laser-cut and bent parts. The quote should define when the clock starts—for example, after drawing freeze, order/payment confirmation and receipt of all technical input. Break the schedule into DFM/approval, material, manufacturing, finish, inspection and transport; identify the critical path for a fixed project date rather than asking only “how many days?”

First define what the sample approval covers: dimensions, assembly, load, finish, color, packing or all of them. Then freeze drawing revision, material, BOM, process route, tooling, CTQs, inspection frequency, limit samples and packing as the production baseline. If production differs from the prototype process, submit a delta list and reconfirm the first article. Retain a signed or golden sample where useful, but it does not replace digital drawings and objective tolerances. One sample may sit near a tolerance limit; production must meet specification rather than match one unit exactly.

A repeat order should recall the same part number, drawing revision, material/BOM, approved master, tooling program, control plan and packing revision, while reviewing prior last-off or historical measurement data. Material lots, sources and finishing lines still create natural variation, so “identical” must become measurable dimensional, color, gloss and functional ranges. Changes to material, supplier, equipment, tooling, process, production location or drawing require review; resubmit a first article or color panel when CTQ or appearance is affected. The customer should also state the latest revision on the repeat PO so an old order does not override a new drawing.

Yes. Packing can be designed around size, weight, cosmetic grade, quantity, handling and the customer’s line. Options include individual bags, foam/bubble material, cardboard dividers, corner protection, formed inserts, accessory kits, cartons, crates or pallets, with labels for part, quantity, lot and barcode. At RFQ, state pack quantity/maximum gross weight, stacking orientation, forklift entry, whether plastic may contact the finish, unpacking sequence and waste constraints. Packing materials must be coating-compatible so plasticizer, moisture or rough separators do not mark the surface.

Identify the failure path first: part-to-part rubbing causes scratches, free space causes impacts, stacking load bends thin parts, straps concentrate force at edges and moisture starts corrosion. Apply surface separation, directional nesting, movement restraint, load spreading, moisture control and stacking limits rather than simply adding film. Separate heavy parts from fine cosmetic surfaces; support long welded frames at designed points without long suspended spans; allow powder-coated parts to cool and cure before packing. Build one pack-out sample, simulate handling and inspect appearance, dimensions and movement before freezing production packing.

Cartons, pallets, crates, moisture protection, labels and packing lists can be configured for sea, air or road freight and destination requirements. Raw-wood pallets, crates or dunnage commonly need approved treatment and the IPPC mark under the destination’s implementation of ISPM 15. Packaging made wholly from processed wood such as plywood or particleboard is generally outside its raw-wood pest scope, but destination and carrier rules still require confirmation. Before order, define destination, port, mode, maximum size/weight, stackability, treatment evidence, shipping marks and hazardous protrusions. Never apply an IPPC mark to unauthorized wood packaging.

Validation intensity should match part value, mass, route and handling risk. Basic review covers gross weight, voids, restraints, center of gravity, carton/crate deformation, stacking and unpacking. Higher-risk projects can use an agreed transport-simulation plan with drop, random vibration, compression, incline impact or temperature/humidity cycling. Define pass criteria before testing—for example, no functional damage, no new visible Class-A scratch, no CTQ change, no loosened hardware and a package still safe to handle. For the first shipment, tilt/impact/humidity indicators and arrival photos can provide data to improve the next lot.