What Drives the Cost of Heavy Frames and Skids
Buyers new to heavy fabrication usually start by comparing steel prices. It is the wrong lever. On a typical multi-tonne machine base or equipment skid, material is rarely more than a third of the invoice, and every credible shop in the region buys plate from a similar pool of mills. The spread between quotes comes from hours, not tonnes.
Here is where the money actually sits, roughly in order of size, and what you can do about each line.
The cost stack
| Line | Share of a typical job | How much you control it |
|---|---|---|
| Fit-up and welding hours | 30–45% | A lot, joint design decides this |
| Machining | 15–30% | A lot, number of faces and setups |
| Plate and sections | 20–30% | Very little |
| Surface finish | 5–15% | Some, the system you specify |
| Stress relief | 3–8% | Some, method and whether it subcontracts |
| QC and documentation | 2–6% | Some, what you genuinely need |
| Packing and inland freight | 5–12% | Some, piece size and Incoterm |
Percentages move with geometry. A thin, fussy conveyor frame is nearly all welding. A thick, simple sole plate with four machined pads is nearly all plate and machining.
Welding hours are the main event
Weld metal is expensive twice: once in consumables and wire, and again in the arc time to deposit it. Anything that reduces deposited volume reduces cost roughly in proportion.
The usual offenders:
- Full-penetration welds specified everywhere. They are appropriate where the joint carries load in tension or fatigue. On a bracket that locates a cover panel, a fillet is fine and costs a fraction as much.
- Oversized fillets. A 12 mm fillet holds roughly twice the metal of an 8 mm one. If the calculation says 8, specify 8.
- Poor joint access. A weld a torch can reach in one clean pass is cheap. The same weld inside a closed box, reached through a hand hole, can take four times as long, and it will be harder to inspect.
- Closed sections with no access at all. If the internals cannot be welded or inspected, the shop has to change the build sequence, which usually means more fixturing.
On heavy plate, bevelling or double-bevelling a joint saves substantial weld metal against a square groove, which also reduces distortion, see welding distortion control on heavy plate.
Machining is the second-largest line, and the easiest to over-specify
Machining cost scales with the number of setups, not the number of features. Four pads on one face is close to the cost of one pad on that face. Four pads on four different faces is four times the work, plus re-datuming between each.
Two questions worth asking before the drawing is frozen:
- Does this face actually need machining, or does it need to be flat? A blasted and ground surface is far cheaper than a milled one, and is often adequate for a non-critical mounting.
- Can the machined features be brought onto one plane? Shimming a bracket by 3 mm in the design is free. Machining a second face is not.
Tolerance follows the same logic. A flatness call-out copied from an older drawing can force a whole extra finishing pass. The detail is in flatness tolerance on large machined weldments.
Plate: where you have least room
Material is a commodity. What you can influence is nesting, how efficiently your parts drop out of a standard plate. Two things help:
- Round dimensions to the material. A part that needs 2,450 mm and gets drawn at 2,520 mm may force a second plate.
- Order more than one. On quantities of five or more, nesting across the batch typically saves a few percent on material and noticeably more on setup.
Grade matters less than buyers expect. Moving from A36 to A572 Grade 50 buys yield strength at a modest premium, and sometimes pays for itself by letting you thin a section. The comparison is in steel grades for heavy weldments.
Finish: three systems, three price points
- Blast and shop primer: cheapest, appropriate for indoor equipment that will be top-coated later or never.
- Blast, primer and 2K industrial topcoat: the default for most exported equipment. Durable, and the colour is yours.
- Hot-dip galvanising: the most expensive by some margin, and it has design consequences: vent and drain holes, distortion risk on thin sections, and a size limit set by the kettle. Worth it outdoors or in wet process environments.
Full comparison in surface finish systems for structural steel.
Stress relief: small line, large consequence
Thermal stress relief itself is not expensive. What is expensive is discovering that your piece is longer than any available oven, which forces either a subcontract move with its own transport cost, or a switch to vibratory stress relief that your specification may not permit.
Ask about oven length early. The full picture is in stress relief for welded machine bases.
QC and documentation: buy what you will actually read
A mill cert, weld map and dimensional FAI are cheap and you should always have them. Volumetric NDT on every weld is neither cheap nor usually necessary. Specify NDT by joint, not by blanket statement, the reasoning is in NDT for heavy weldments.
Third-party inspection typically adds a fixed fee per visit. On a first order with a new shop it is money well spent. On the fourth repeat order, less so.
Freight: the line that ambushes people
Inland transport to port is proportional to how awkward the piece is, not how heavy. A frame that fits in a 40 ft flat rack ships at a predictable rate. A frame 20 cm too tall becomes out-of-gauge cargo and can cost several times as much.
This is why split-and-bolt frame design so often wins on landed cost even though it adds machining and assembly. The shipping side is covered in out-of-gauge and breakbulk shipping.
A worked example
A 6-tonne pump skid, S355, blasted and 2K painted, four machined mounting planes, single unit, FOB Ho Chi Minh City. Roughly:
- Plate and sections: 26%
- Fit-up and welding: 38%
- Stress relief: 4%
- Machining: 17%
- Finish: 7%
- QC and docs: 3%
- Packing and inland freight: 5%
Now change two things on the drawing: bring the four mounting planes onto two, and reduce a blanket 10 mm fillet to 8 mm where the calculation allows. Machining drops by roughly a third of its line and welding by about a fifth of its. That is close to a tenth off the whole job, from two decisions that cost nothing to make on paper.
That is the argument for sending a drawing before it is frozen. Send us what you have, even partial, and we will tell you which lines are carrying the cost, within 24 business hours.