Every quote has two numbers: the price and the lead time. Buyers fight hard over the first and mostly accept the second — which is why prototype schedules slip. The honest answer to "why does this take three weeks" is that only a fraction of that time is the machine cutting metal. Understanding the rest is how you compress a schedule without paying rush fees.
Where the calendar actually goes
A typical 20-piece aluminum prototype order breaks down roughly like this:
- Quote and DFM review: 1–2 days. Someone has to read the drawing, flag the impossible tolerance, ask about the finish, and price the material. This step compresses well if the drawing is complete; it stalls completely if the drawing is not.
- Material procurement: 0–5 days. 6061 plate is on the shelf at any distributor. 7075 in an odd thickness, 17-4PH in a specific condition, or a PEI/PSU resin for molding can add days. This is the most common invisible delay.
- Programming and fixturing: 0.5–2 days. CAM for a five-sided part with tight true position is real work, and soft jaws usually have to be machined before the first part exists.
- Machining: hours, not days. For most prototype quantities, spindle time is the smallest item on the list. Twenty brackets at 12 minutes each is four hours of cutting.
- Finishing: 3–7 days. Anodizing, powder coating, and passivation run in batches at outside vendors. Your parts wait for the next tank, not for your parts. Type III hard anodize and chem film are slower than Type II.
- Inspection and documentation: 0.5–1 day. First-article reports and CMM time are scheduled, not instant.
- Shipping: 1–5 days. International express is 2–4 days door to door; economy freight is cheaper and much slower.
The pattern is obvious once it is written down: finishing and material are the schedule, not machining. If your project is late, it is almost never because the spindle was busy.
The five things that actually stretch a schedule
1. Number of setups. A part machined from three sides needs three fixtures and three datum transfers. Each one adds setup time and a chance to lose tolerance. A design that can be done in two setups routinely quotes a week faster than the same part requiring five.
2. Tolerance distribution. Blanket ±0.01 mm on every dimension forces slow passes, in-process probing, and sometimes a second op that would otherwise be unnecessary. Specify tight tolerance only on the features that mate or locate — usually 5 to 10 dimensions out of 60.
3. Outsourced finishing. A single finish spec is one vendor queue. Three finishes on one part — anodize the body, chem film the lid, laser-mark the panel — is three queues that must be sequenced and cannot overlap.
4. Drawing ambiguity. "Finish: as required," a missing thread callout, or a model that disagrees with the drawing all trigger a clarification email, and the clock stops until it is answered. Ambiguity is the most expensive defect in the whole process because it costs calendar time, not just money.
5. Quantity surprises. Ten parts and 100 parts are different jobs. Small quantities favor machining; larger quantities can justify a die casting tool, which is four to eight weeks for the tool alone before the first good part exists. Choosing the wrong side of that line is the most expensive scheduling mistake in hardware.
How to compress the calendar
Send a complete package the first time. A STEP model, a 2D drawing with tolerances and finish, the material grade with temper, and a quantity. Where a decision is genuinely open, say so explicitly rather than leaving a field blank — a supplier can quote an option, but not a void.
Use the material the shop stocks. If 6061-T6 meets the requirement, do not specify 6082 or 7075 for a marginal strength gain. Material substitutions are the fastest legitimate schedule win available.
Consolidate finishes. One finish for the whole assembly, applied after all parts are made, beats three specialized finishes that have to be coordinated.
Ask for a partial shipment. If the anodized cosmetic covers are the long pole, have the functional brackets ship first so assembly and testing can start. Many shops will do this at no extra cost if you ask at the quoting stage rather than after the parts are made.
Design for fewer setups from the start. Chamfers instead of sharp corners that need EDM, one datum scheme, holes that can be drilled from the open side. These are free at the CAD stage and expensive later.
Where the time really cannot be removed
Some steps are physics. Anodize needs time in the tank; a die casting tool needs to be cut, spotted, and sampled; a plastic injection mold needs steel, a trial, and often a revision. Anyone promising to compress these to zero is either rushing a step or planning to rework later.
The useful question is not "can you go faster" but "which step is the constraint." Ask a supplier to show the schedule broken down by step — quote, material, machining, finish, inspection, ship. A shop that can produce that breakdown is usually a shop that has already optimized it. A shop that answers "about three weeks" without a breakdown is quoting a habit.
For most prototype programs, the realistic floor for machined metal with a cosmetic finish is 7–10 working days, and 5–7 days for unfinished parts in stock material. If your program needs less than that, the answer is usually to split the build: machine the critical geometry first, finish it later, and start testing on the as-machined parts. Reviewed end to end, that is the only compression that does not cost quality.
