{"id":7870,"date":"2026-10-05T17:53:58","date_gmt":"2026-10-05T17:53:58","guid":{"rendered":"https:\/\/chinametalfoundry.com\/?p=7870"},"modified":"2026-10-05T17:53:58","modified_gmt":"2026-10-05T17:53:58","slug":"casting-vs-fabrication","status":"publish","type":"post","link":"https:\/\/chinametalfoundry.com\/fr\/blog\/casting-vs-fabrication\/","title":{"rendered":"Moulage ou fabrication : quand une pi\u00e8ce moul\u00e9e l'emporte sur une pi\u00e8ce soud\u00e9e"},"content":{"rendered":"
\n
\n

Every buyer who has sent out a package for a machine base, a bracket or a pump housing has hit the same fork in the road: build it as one piece of cast metal, or cut plate and weld it into an assembly. The casting vs fabrication<\/strong> decision is almost never about which process is “better” \u2014 it is about which route lands on your drawing, your annual volume and your delivery date at the lowest total cost.<\/p>\n<\/div>\n

Most casting vs fabrication comparisons go wrong for a boring reason: the two quotes are not for the same thing. This guide puts real numbers on both sides of that choice: weld-metal hours, tooling amortisation, machining stock, stress relief and inspection. The cost models come from production parts we quote both ways; the material and fatigue behaviour is cross-checked against published guidance from the Soci\u00e9t\u00e9 am\u00e9ricaine de la fonderie<\/a> and the fatigue categories in AWS D1.1<\/a>. If you want a refresher on the casting side first, start with our Guide sur la fonderie des m\u00e9taux<\/a>.<\/p>\n

\"casting
Same function, two routes: a single cast pedestal with cast ribs on the left, a welded plate assembly on the right.<\/figcaption><\/figure>\n

Casting vs Fabrication: What Is Actually Being Compared<\/h2>\n

The casting vs fabrication question gets easier once you separate geometry from economics. A casting is made by pouring molten metal into a mould and letting it solidify into a near-net shape, so most of the geometry you see in the finished part was formed by the mould, not by a cutting tool. A fabrication \u2014 usually called a weldment \u2014 starts as mill plate, bar or structural section; the pieces are cut, formed, fixtured and joined by welding into an assembly that performs the same job.<\/p>\n

That single difference cascades into everything else in the casting vs fabrication comparison: tooling cost, lead time, buy-to-fly ratio, distortion behaviour, inspection method and how much a design change costs you in week three of the project.<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
Factor<\/th>\nCast route<\/th>\nWelded fabrication<\/th>\n<\/tr>\n<\/thead>\n
Starting form<\/td>\nMolten metal into a mould<\/td>\nCut plate, bar, structural section<\/td>\n<\/tr>\n
Up-front tooling<\/td>\nPattern and core boxes, several thousand dollars<\/td>\nWeld fixtures and jigs, usually low cost<\/td>\n<\/tr>\n
First-part lead time<\/td>\n5\u20139 weeks (tooling + sample)<\/td>\n2\u20134 weeks<\/td>\n<\/tr>\n
Geometry freedom<\/td>\nCurved walls, internal cavities, variable sections<\/td>\nLimited to what plate and section can approximate<\/td>\n<\/tr>\n
Buy-to-fly ratio<\/td>\n1.1\u20131.3 (gates and risers are remelted)<\/td>\n1.4\u20131.8 (skeleton drop is scrap)<\/td>\n<\/tr>\n
Weld length per unit<\/td>\nZero<\/td>\n10\u201340 m on a typical machine component<\/td>\n<\/tr>\n
Distortion control<\/td>\nRigging, risers and mould design<\/td>\nFit-up, weld sequence, post-weld heat treatment<\/td>\n<\/tr>\n
Inspection basis<\/td>\nVisual, ASTM E446 surface comparators, RT or UT on critical sections<\/td>\nMT or UT on welds per AWS D1.1<\/td>\n<\/tr>\n
Cost of a design change<\/td>\nPattern rework, 2\u20134 weeks<\/td>\nEdit the nest file, hours<\/td>\n<\/tr>\n
Economic volume band<\/td>\nRoughly 50\u2013200+ per year<\/td>\n1\u2013100 per year<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Where a Casting Beats a Weldment<\/h2>\n

Four mechanisms drive the win on the casting side of the casting vs fabrication comparison, and they are all about geometry and material behaviour rather than about the metal price.<\/p>\n

Part consolidation removes weld hours<\/h3>\n
\n

A fabricated bearing pedestal might be 14 pieces: base plate, two side webs, a bore boss, four gussets, six stiffeners. Every one of those pieces has to be cut, tracked, fixtured, tacked and welded. The same part as a casting is one piece with the ribs already there. On the cost model below, 18 m of fillet weld disappears entirely, and that is where the money is \u2014 not in the steel price.<\/p>\n

Damping and vibration behaviour<\/h3>\n
\n

Gray iron dissipates vibrational energy roughly an order of magnitude faster than a carbon-steel weldment, because the graphite flakes interrupt and damp the elastic wave. On machine-tool bases, compressor housings and pump frames that difference shows up as lower noise, better surface finish on the machined part and longer bearing life. It is the reason our fonte grise vs fonte ductile<\/a> comparison gets read by anyone specifying a structural base.<\/p>\n

Metal only where the load path needs it<\/h3>\n
\n

Plate comes in standard thicknesses, so a weldment tends to carry a uniform 12 mm or 20 mm wall even where 6 mm would do. A casting can taper a wall from 20 mm at a loaded boss down to 7 mm in an unloaded web, and core out pockets that plate cannot reach. On the pedestal example used later in this article, purchased weight drops from 68 kg of plate to a 48 kg casting while the finished part stays at 42 kg.<\/p>\n

Leak paths and fatigue at the joint<\/h3>\n
\n

Leak-tightness is the shortest argument in any casting vs fabrication debate. A pressure boundary built from welds is a boundary with a continuous seam, and every metre of that seam is a candidate for MT indications, repair and re-testing. Castings have no seam, which is why valve and pump bodies<\/a> are still overwhelmingly cast. Fatigue behaves the same way: AWS D1.1 gives an as-welded fillet joint loaded transverse to the weld a constant-amplitude fatigue threshold of about 69 MPa (Category C), and a joint with a severe geometric discontinuity drops to roughly 31 MPa (Category E). A cast radius in the same corner carries the fatigue strength of the parent metal, because there is no weld toe to initiate a crack.<\/p>\n

Where Fabrication Wins \u2014 and It Often Does<\/h2>\n

Roughly half of the parts that get quoted both ways should stay welded, because the other half of the casting vs fabrication equation is tooling, not metal. Here is when.<\/p>\n