Sourcing machined parts in difficult materials

Why identical drawings return different prices

Send one titanium housing drawing to five shops and the spread can be four to one. Buyers reasonably conclude someone is gouging them. Usually nobody is, and the spread comes from five specific places.

Material buying position. On these alloys raw material is often 30–60 % of the part price. A shop with titanium bar already on the shelf quotes from that cost; a shop buying one length from a distributor pays a premium and a cut charge. On nickel alloys, mill-direct against one-off distributor pricing can differ by 40 % on its own.

Whether they have done it before. A shop that runs Inconel weekly has proven data and knows its cycle time within 10 %. A shop that has not either adds a risk premium or — worse — does not, wins the job, and discovers the real cycle time halfway through. The cheap quote is sometimes the opening position of a change order conversation.

What they assumed about the ambiguity. This is the biggest source, and the only one entirely within your control. If the drawing says “Inconel 718” without a condition, one shop quotes solution-treated and one quotes aged. If inspection is unstated, one quotes a certificate of conformance and one quotes full AS9102 first article inspection. Those two assumptions alone can double a price.

The process route they chose. Machine soft and heat treat, or buy treated and machine hard. Both can be right, and they land 40 % apart.

Overhead and how badly they want work. The part everyone assumes explains the whole spread. It usually explains the least of it.

The practical reading rule: discard the outlier low, not the outlier high, and ask every shop to state its assumptions. A quote reading “assumes ISO 2768-m general, solution treated bar, CoC only, EXW” is comparable. One reading “$310” is not.

The package that gets quoted this week

A machine shop that cuts titanium and Inconel is not short of work. When an RFQ lands, someone spends about ninety seconds deciding whether to quote it now, later, or not at all. Quoting a nickel alloy part properly takes an engineer two to four hours, and nobody spends that on a drawing that raises more questions than it answers.

What gets checked in those ninety seconds:

1. A 3D model, not just a PDF. STEP AP214 or AP242 plus the dimensioned PDF. Without a model someone has to rebuild your geometry before estimating.
2. Material by grade and condition, with the standard.
3. Tolerances that mean something — a general callout plus tightening only what mates, seals or locates.
4. Quantity now and per year. One-off, 50 and 500/year are three different manufacturing problems.
5. Surface finish by surface, not a global callout.
6. Heat treatment and coating with the sequence, and who supplies them.
7. Inspection and documentation level, decided before the quote.
8. A date, stated as a date. “ASAP” is not information.
9. Incoterm and destination. EXW Vietnam and DDP Ohio are different numbers.

Fifteen extra minutes assembling that routinely takes a week off the quote, and removes the two or three rounds of clarification that otherwise eat the same week.

Documentation is where quotes quietly double

Decide the level deliberately, because “aerospace quality” is not a specification.

Requirement	What it proves
Certificate of Conformance	Someone will state in writing that parts meet the drawing
EN 10204 2.2	Producer’s routine test data, not your batch
EN 10204 3.1	The actual heat your material came from, certified by the mill’s independent inspection
EN 10204 3.2	The same, counter-signed by a third party
AS9102 FAI	A dimensional proof, not a material one — every characteristic ballooned and measured

A mill certificate tells you the metal is what the drawing says. An FAI tells you the part is the shape the drawing says. Buyers sometimes request one while meaning the other.

The expensive combination is a complex part with sixty characteristics and a requirement to measure all of them on every piece. Mark the critical characteristics and let the rest be sampled.

The lead time is usually the metal

Buyers chase machine shops about schedule when the constraint is upstream. A titanium bracket might take four hours of spindle time and eight weeks to start.

• Distributor stock, days. Common grades in common sizes.
• Distributor network, two to six weeks. The size exists, but not locally.
• Mill order, eight to twenty weeks. Large diameters, ELI grades, specific conditions, approved-source requirements, refractory metals, near-net forgings.

Ask one question with the RFQ — is the material in stock, in the network, or a mill order? — and insist that quotes state the material lead time separately from the machining lead time. “Twelve weeks” is not useful. “Ten weeks for material, then two to machine” tells you exactly where to apply pressure.

Two decisions that cost more than they look

A blanket surface finish. “All surfaces Ra 0.8 µm unless otherwise noted” on a prismatic housing applies a fine finish to every face including the ones nobody touches. Write it the other way round: a coarse general callout, fine finishes marked where they are needed.

Specifying the machine instead of the requirement. Demanding five-axis when the part does not need it narrows your supplier pool and raises the price. Give the model and mark the relationships that genuinely matter between features on different faces; let the shop choose the setup plan, then ask how many setups they quoted.

Reading in this guide