Partimation

A real case

Foundry should-cost: every line of a casting price, and its basis

Twelve grey-iron manifolds off printed sand, and the customer's engineer wants the price broken down before he approves a pattern. Here is what a foundry can show him: the route including degating, fettling and queue time, a ledger where every line names its rate or rule, and a confidence basis per number.

The situation

A concrete case

Twelve castings. EN-GJL-250 grey iron manifold, a validation build for an OEM that intends to order a few thousand a year afterwards, wanted in five weeks. No pattern exists and nobody is cutting one for twelve pieces. The buyer's engineer has asked for the price broken down before he will take the production tooling to his own approvals, and what he means by that is not "how much" — he already has three numbers from three foundries and no way to tell which of them is built the same way. What he is asking is what the price is made of. Most quotes cannot answer that, which is why the cheapest one usually wins and then goes wrong later.

Step by step

How it plays out

01

Build the whole route, including the parts estimates leave off

The route comes out in full — tooling, forming, degating and fettling, heat treatment, machining, finishing, and the inter-department handling and queue time an estimate normally leaves out entirely — then gets ranked cheapest, quickest and fewest-steps, with each step naming the rule that put it there. Every step is marked in-house or priced with your accepted vendor's markup, so the machining and the coating you send out are on the same sheet as the melt, at the markup you actually pay, instead of arriving later as the reason the job lost money.

02

Read the ledger, line by line, on your own rates

Every cost line traces back to a named rate, a posture setting or a rule in the engine — never an opaque total. The rates underneath are the ones you set once in Rates Studio: your melt cost, your moulding line, your fettling labour, your heat-treat, your handling. A customer's engineer reading that ledger is reading your foundry's cost structure with names on it, which is a different conversation from defending a lump sum he assumes is padded.

03

Price twelve pieces the way twelve pieces actually get made

Printed sand molds and cores have their own should-cost path — consumables and printer hours, no pattern tooling — rather than being a cut-pattern quote with a discount applied to it. Then the should-cost curve runs at every quantity, with the exact quantity where one method becomes cheaper than another solved in closed form, so the validation run and the production run he is really asking about are two readings off one curve rather than two unrelated quotes.

04

State how sure each number is, and on what

Every result carries a confidence value that names its basis: a published prior, a calibration built from your own recorded actuals, a deterministic rule, or a flag that it needs a human to look at it. On a manifold nobody has poured before, some lines will rest on a published prior and the quote says so instead of hiding it. That is the line that survives contact with a buyer: you are showing him where your own number is soft, before he finds it.

05

Correct the lines the job proves wrong

When the twelve are poured, whoever knows what actually happened enters the correction on the cost line itself — the fettling that took twice as long, the yield that came in under the quote — and that feeds a calibration record, so the confidence bands tighten around your foundry's own performance rather than a published average. Corrections are typed in by a person on the line they belong to; that is the loop, and it is worth knowing exactly where it starts before you build a process around it.

What you get

Where you end up

What goes back to the buyer's engineer is not a total. It is a route with degating, fettling, heat treat and queue time on it, a ledger where every line names the rate or rule that produced it, an in-house or vendor marking against each step, and a confidence basis stated per number including the ones resting on a published prior. He can hold his other two quotes against it and see which of them left the handling out. You can hold the same document up in five weeks when the production quantity comes back and read the crossover off the curve rather than re-quoting from scratch. Partimation is decision support, not a guarantee — every result names the rule behind it, and the reason that matters here is that a number you can take apart is the only kind a customer can be argued out of rejecting.

Already shipped

What backs this

Full manufacturing route builder, ranked

Built into RouteDFM.

In-house vs. brokered-vendor step pricing

Built into RouteDFM.

Itemized cost ledger traced to a named rate or rule

Built into CostDFM.

Shop-editable rates & cost model (Rates Studio)

Built into CostDFM.

Sand casting + printed-sand molds

Built into CostDFM.

Should-cost curves with an exact process-crossover quantity

Built into ProcessDFM.

Confidence bands with a named basis on every quote

Built into CostDFM.

Correction-feedback loop calibrates future quotes

Built into CostDFM.

FAQ

Frequently asked questions

A buyer wants our casting price broken down — what can a foundry actually show them?

The route in full, with each step naming the rule that put it there: tooling, forming, degating and fettling, heat treatment, machining, finishing, and the inter-department handling and queue time most estimates leave off. Underneath it sits a ledger where every cost line traces to a named rate, a posture setting or a rule in the engine, and each step is marked in-house or priced at your accepted vendor's markup. A buyer holding three quotes built three different ways can finally see which one left the handling out.

How is a twelve-off casting run priced when there is no pattern yet?

Printed sand molds and cores carry their own should-cost path — consumables and printer hours, no pattern tooling — rather than being priced as a cut-pattern job with a discount on it. The should-cost curve then runs at every quantity with the exact crossover solved in closed form, so a validation run of twelve and the production quantity behind it are two readings off one curve instead of two unrelated quotes.

How sure is the number on a casting we have never poured before?

Every result carries a confidence value that states which of four things it rests on: a published prior, a calibration built from your own recorded actuals, a deterministic rule, or a flag that it needs a human to look at it. On a part with no history at your foundry, some lines will rest on a published prior and the quote says so rather than hiding it — and the needs-human-review flag is where the last half hour before the quote goes out belongs.

Can our own melt, fettling and heat-treat rates sit under the number a customer sees?

Machine rates, labour rates and the cost model are set once in Rates Studio, and every quote across every process pulls from them — melt, moulding line, fettling, heat treat, handling. The ledger a customer reads is therefore built on your foundry's own rates rather than an industry average with a correction applied by hand afterwards, and each line still names the rate or rule it came from.

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.