Partimation

A real case

Re-open the process choice — before the geometry commits you

A housing drawn for 3-axis machining because the last generation was, with a forecast that just moved from 150 a year to 5,000. One upload screens six process families with the constraint behind each verdict, and solves the exact quantity where one process gets cheaper than another.

The situation

A concrete case

A 6082 aluminium lever housing, one STEP file, gate review a week on Friday. The part is drawn for 3-axis machining from bar — sharp internal corners, no draft anywhere, a boss that a die would have to core around — because the previous generation was machined from bar and nobody re-asked the question when this one started. At the last program review the forecast moved from 150 a year to 5,000. The geometry has not caught up with that number, and every week it stays as drawn the machining assumption gets harder to unpick: the corners, the draft and the boss are all an afternoon's work now and a redesign after release.

Step by step

How it plays out

01

Upload the model as drawn — there is no per-process version to make

One STL or STEP model goes in, exactly as it sits in CAD today. You do not redraw the part for casting to find out whether it casts, and you do not need the volume decided before you upload. Detected holes come back named with their own confidence, alongside the overall envelope and the part's face counts.

02

Read six verdicts and the constraint behind each one

Casting, molding, forging, machining, sheet metal and additive are screened together off that single upload, each coming back feasible, marginal or infeasible with the constraint that drove the verdict. A missing draft, an undercut a closed die cannot core, a wall the process cannot fill — each arrives named, which is the difference between a score you have to take on faith and a specific thing you can design against without ever having run the process yourself.

03

Read the crossover quantity instead of remembering it

Every feasible process carries a should-cost curve across quantities, and the exact quantity where one process gets cheaper than another is solved in closed form rather than eyeballed off two lines on a chart. At 150 a year you read the curve at 150. At 5,000 you read it at 5,000, and whether the crossover sits between those two numbers is the entire question the gate is actually asking.

04

Put the tool cost inside the curve, where it moves the answer

One block-and-machining tool-cost model prices dies, molds, patterns and core boxes, and amortizes that cost into the per-piece curve rather than leaving it beside the curve as a lump sum. Amortization is the thing that moves a crossover quantity, so a tool number sitting in a separate cell on a spreadsheet is exactly how the decision gets made wrong.

05

Take the ledger into the gate, not the total

Every cost line traces to a named rate, posture setting or engine rule instead of an opaque total, and every result carries a confidence value that states its basis — a published prior, shop-calibrated actuals, a deterministic rule, or a flag that it needs human review. Every panel mounts against the same job, so moving from a process verdict into its cost detail is a panel switch rather than another upload.

What you get

Where you end up

You leave with a verdict per process, the constraint that produced each one, and a curve you can read at both forecast quantities. At 150 a year the billet route may well still win, and now you have said so with a solved number rather than an inherited habit. At 5,000 the answer may be a different family altogether — and the geometry changes that go with it, the draft, the corner a die can fill, the boss moved off the pull direction, are an afternoon now rather than a redesign later. Where this stops matters as much as what it does. The screen ranks processes for this part at this quantity; it is not a tolerance stack-up, a joining study or a part-consolidation exercise, and none of those exist here. The assembly work that does ship is a costed BOM roll-up and a coarse advisory clearance screen, nothing further. Partimation is decision support, not a guarantee — confirm the fit with your supplier and your own engineering judgment before tooling is committed.

Already shipped

What backs this

One upload, any file

Built into PartDFM.

Cross-process fit screening with a named reason

Built into ProcessDFM.

Should-cost curves with an exact process-crossover quantity

Built into CostDFM.

Shared tool-cost model across dies, molds, patterns, core boxes

Built into ToolDFM.

Itemized cost ledger traced to a named rate or rule

Built into CostDFM.

Confidence bands with a named basis on every quote

Built into CostDFM.

Assembly BOM roll-up + interference screen

Built into AssemblyDFM.

Shared-job architecture — every specialist panel mounts against one jobId (DEMO_TRUTH_AUDIT.md, LIVE)

Built into ProcessDFM.

FAQ

Frequently asked questions

Does a part have to be redrawn for each process before it can be screened?

One model, exactly as drawn, is enough. Casting, molding, forging, machining, sheet metal and additive are screened together off a single upload, each returning feasible, marginal or infeasible with the constraint that drove the verdict — so a part drawn for machining from bar still gets a casting verdict, and the named reason behind it if that verdict is no.

What has to change on a machined part before it could be cast instead?

Draft, undercuts and wall thickness are the three that usually decide it, and the screen names which one it was rather than returning a score. An infeasible verdict arrives with the constraint attached, which is a specific thing to design against — open the corner, add the draft, move the boss off the pull direction — rather than a general warning that casting might be difficult.

Where does the exact crossover quantity between two processes come from?

Closed-form solution of the two should-cost curves, not a reading taken off a chart. Every feasible process carries a curve across quantities with tool cost amortized into the per-piece figure rather than sitting beside it as a lump sum, and the quantity where one curve overtakes another is solved directly — which is exactly why the tooling assumption moves the answer.

How do I defend a process change at a design gate?

Named constraints and an itemized ledger are what survive that room. Each process comes back feasible, marginal or infeasible with the constraint that drove the verdict, every cost line traces to a named rate, posture setting or engine rule instead of an opaque total, and every result carries a confidence value stating its basis — a published prior, shop-calibrated actuals, a deterministic rule, or a flag that it needs human review.

Where this fits

Explore further

Get started

See what this looks like on your own parts

Send a model or a drawing and book a walkthrough — or start free and run one through yourself.

Book a demo

Book a demo

Tell us about your shop and we'll set up a walkthrough on a real part.