Jewelry Prototyping: What A Sample Really Costs
Home
Yiwu, Zhejiang, China
Mon – Fri 9:00 AM – 6:00 PM · Sat 9:00 AM – 4:00 PM
[email protected]

Jewelry Prototyping: What A Sample Really Costs

Alan Xiao, Founder of Sourcing Pioneer
Alan Xiao
· Updated November 17, 2026 · 9 min read

Jewelry prototyping is the stage where a drawing becomes a metal object you can hold, weigh and argue about.

You are not buying a sample. You are buying information, and the price depends on how much of it you need before you commit to a run.

That is why two factories quote very different sample fees for the same pendant. One sells a printed pattern. The other sells a cast piece with the stones set and the finish done.

This article covers the four stages a prototype passes through, and where each one lies to you. It also covers what resolution really buys, and six lines that keep sample costs from becoming a subscription.

A red printed wax ring pattern beside a finished cast ring on a beige studio surface
Pattern on the left, metal on the right. Everything between them is the part buyers underestimate.

The Short Answer

A prototype exists to answer questions. Does the shape read at this size. Does the wall hold. Does the stone sit. Does the clasp close.

Each question costs a stage. Skipping a stage does not save the money, it moves it to the next one, usually with interest.

The cheapest sample is a printed pattern checked in the hand. The most expensive sample is a full cast piece in the final alloy with the final stones.

Most programs need both, in that order, plus one revision round that nobody planned for.

The engineering side of this is settled. A jeweller carves or models the shape, a printer or a mill produces a pattern, and a casting house turns the pattern into metal.

The commercial side is where projects go wrong. A pattern is not a product, and a photograph of a pattern is not evidence of anything except that a file opened.

The Four Stages

Every prototype passes through four stages, and each one can be approved or rejected on its own.

One, the drawing and the render. Cheap and fast, and it settles proportion. It cannot settle weight, balance or how a prong feels against skin.

Two, the printed pattern. A physical object you can hold, at roughly production size. It settles fit and silhouette, and it is where thin walls show up.

Three, the cast sample. The same shape in metal, finished and set. It settles weight, colour and the stone seats, and it is the first piece a customer would recognise.

Four, the pre-production piece. One unit from the actual production route, with the actual mould and the actual setting bench.

The gap between stage three and stage four is where most surprises live. A hand-finished sample can hide a process that does not repeat.

Four prototyping stages, from drawing and render to pre-production piece, each showing what it settles and what a redo costs
Four stages, and the cost of going back one. The ladder only works in order.

Approving stage three as if it were stage four is the commonest expensive mistake in this trade. The sample is usually better than the run, because it was made by the best bench on the floor.

There is a second trap in the same gap. A sample made by hand from a printed pattern proves the design, and it proves nothing about how the factory will produce it at volume.

So ask for the route, not just the object. The route is what you are actually buying when the run starts.

Wax, Resin, Or Mill

Three routes produce a pattern, and the choice is mostly about who accepts what downstream.

Printed wax. A light-based printer builds the shape in a wax-filled material, layer by layer. It is a direct pattern, and many casting houses treat it as the safest handoff.

Printed castable resin. The same kind of printer, a different material. The pattern then needs washing, a post-cure, and a burnout schedule the foundry has tested with that resin.

Milled wax or milled plastic. A cutting tool carves the shape out of a block instead of building it up. It suits simple geometry and leaves tool marks that have to be polished or avoided.

The decisive question is not which is better. It is which route your casting house will accept without changing its process.

Some houses work only in professional wax and decline resin or plastic patterns outright. An untested resin can leave residue in the flask, and that residue shows up as porosity in the metal.

RouteHow the pattern is madeMain limit
Printed waxLayered build in a wax materialAccepted by most casting houses
Castable resinLayered build, then wash and cureBurnout must be validated by the foundry
Milled waxCut from a blockSimple shapes only, tool access limits
Hand carved waxCut and built by a modellerSlowest, best for organic shapes

What 25 Microns Buys

Layer thickness is the number suppliers quote first, so it is worth knowing what it does and does not cover.

On current jewelry printers, the finest setting is often 25 microns. Other machines and materials run at 50 or 100 microns, according to the printer maker’s own material page.

A thinner layer gives a smoother curve and softer stair-steps on a dome or a bezel. That is a surface benefit, and it is real.

It does not fix a wall that is too thin, a cavity that cannot drain, or a sprue attached in the wrong place. Those are design and process questions.

It also does not decide how the metal comes out. A perfect pattern still goes through investment, burnout and a pour, and each of those steps can add a defect the printer never caused.

So read the resolution figure as one line in the spec. A supplier who quotes 25 microns and says nothing about burnout is answering the easy half of the question.

Where A Render Lies

A render is a promise made by a surface with no inside. Four things hide in it.

Thin walls. A shader draws a wall at any thickness. A casting needs enough metal to fill, and a wall under the practical minimum will come out short or porous.

Enclosed cavities. A hollow section looks efficient on screen. In the flask, the pattern has to burn out and the metal has to flow. Every enclosed space therefore needs a way in and a way out.

Support marks. A printed pattern is held up by supports that are cut off afterwards. On a polished face, those marks are the first thing a customer sees.

Finishing access. A tight inside curve can be printed and cast and then be impossible to polish, which turns a design detail into a rework line.

Four cards naming what a render cannot show: thin walls, enclosed cavities, support marks and finishing access
Four things a shaded picture cannot show, and four reasons to hold the printed part.

The fix is procedural, not technical. Review the printed pattern as an object, not as a picture, and check it with the bench that will finish it.

The Casting House Is Part Of The Spec

Most prototyping plans stop at the printer. The cast is where the plan meets reality.

A casting house runs a fixed sequence: sprue the pattern, invest it in a flask, burn out the pattern, melt and pour the metal.

Each step has settings, and those settings belong to the foundry, not to the file. That is why “castable” on a material label is a claim about the material, not about your flask.

Ask a foundry three questions before the first print. Which pattern materials do you accept. Which burnout schedule applies to each. And what geometry will you refuse.

The answers change the design. A house that accepts wax only rules out a resin workflow. A house with a long burnout suits one pattern material and not another.

It is cheaper to learn these answers in a phone call than in a failed flask of twelve rings.

What Drives The Cost

Sample fees are quoted as one number and built from six.

CAD hours. The modelling time, and the revisions to it. A complex pavé setting can cost more in modelling than in metal.

Pattern production. Printer time, material, supports, wash and cure. Small parts, low cost. A tall sculptural piece, much more.

Failed patterns. Prints that come out wrong, castings that come out porous, and the parts remade because a sprue left a mark. These are real cost lines, and good suppliers track them. A note on pattern routes puts failed prints and finishing labour in the same budget.

Metal. Priced by weight in the alloy chosen, plus the casting yield you do not get back. A prototype in final alloy costs more than one in brass.

Bench time. Filing, polishing, setting and finishing. This is where a sample becomes expensive, because it is the least automatable part of the chain.

The mould, if you want more than one. A rubber mould for repeat casting is a separate buy, and it only pays back across a run.

Cost lineWhat it coversWhen it grows
CAD hoursModelling and its revisionsComplex settings, late changes
PatternPrint, supports, wash, cureTall parts, fine detail
Failed patternsRemakes and rejectsUntested route or geometry
MetalAlloy by weight, plus yield lossGold and platinum samples
Bench timeFiling, polishing, settingEvery extra revision
Rubber mouldRepeat casting after approvalAny run beyond one piece

When a quote looks too low, one of these six is missing. Usually it is the bench time, which the factory absorbs once and stops absorbing on the second revision.

Revision Rounds: Who Owns What

Revisions are cheap or expensive depending on what has already been built.

A change in CAD is the cheapest event in the project. A reprint adds material and machine time. A re-cast adds a flask, metal and a full bench pass.

So the sequence matters. Collect every comment before the first print, and every comment again before the first cast.

Two rounds are normal on a new design: one after the render and one after the cast. Anything beyond that is usually a specification that was not written down.

Write the ownership into the order. Who pays for the second print. Who pays if the cast does not match the approved pattern. Who decides when the shape is closed.

A supplier who will not cap revisions will not cap the invoice either.

Approving A Sample

An approval is a decision with evidence behind it, so the evidence needs a list.

Dimensions. Measure the three or four features that decide fit, and compare them with the drawing rather than with the render.

Weight. Weigh the cast piece. Weight is the fastest check on wall thickness, and it is the number a customer feels.

The thinnest wall. Cut or measure at the narrowest point. That is where a casting fails first.

The stone seats. Check that each stone sits level and that the prongs cover the girdle evenly. A seat that is off by a little at sample stage is off by more in a run.

The surface. Look at the polished face under a loupe for support marks and porosity. Both survive polishing and both are reasons for a rework claim.

The signature block. Photograph the approved sample from four angles, date it, and attach it to the order. The photograph becomes the reference when the run arrives.

Six Lines For The Order Sheet

These six lines fit on one page and turn a sample fee into a scope.

One, the stage. Render, printed pattern, cast sample or pre-production piece, priced as separate deliverables.

Two, the material for each stage. Pattern material, and the alloy the cast sample is made in.

Three, the files. Format, revision number, and who owns the CAD after the project ends.

Four, the route. Pattern material plus the casting house that will accept it, named.

Five, the revisions. How many rounds are included, and the price of one more.

Six, the acceptance checks. The four measurements, the weight, and who signs the approval.

Put all six into the specification page the bench works from before the first print. If the factory side is still open, how the factory is chosen decides more about sample cost than any negotiation.

The alloy matters for the cast stage, so it helps to know what 14K actually contains before you approve a metal.

When the run arrives, a check before the run ships compares the goods with the photograph you signed.

Six numbered clauses for a jewelry prototyping order, from the stage and material to acceptance checks
Six clauses. The first decides what you are buying, the last decides who signs for it.
SCOPE THE SAMPLE

Send the drawing and the sample quote. I will check which stage you are paying for, and whether the casting house accepts that pattern. I will also count the revisions really included, and name the checks that should decide the approval.

Ask For A Sample Review
How long does jewelry prototyping take?

It depends on the stage. A render is days. A printed pattern is days to a week. A cast sample with setting and finishing adds bench time, and a revision restarts part of it.

Do I need a printed pattern if I have a good render?

Yes, for anything worn. A render cannot tell you how the piece balances, how a prong feels, or whether the wall is thick enough to cast.

Is castable resin as good as wax?

It can be, and only if the foundry has validated its burnout for that resin. Some houses accept wax only, which decides the route for you.

What does resolution actually change?

Surface smoothness on curves, mostly. It does not fix thin walls, enclosed cavities or a sprue in the wrong place.

Why is my sample better than the production run?

Samples are usually made by the most skilled bench on the floor, by hand. The run is made by the production route, so approve the route, not only the sample.

Who owns the CAD file?

Whoever the order says. It is worth settling in writing, because the file is the cheapest thing to reuse and the hardest to replace.

Questions & Comments

We read every one and reply within 24 hours

If a sample quote lands on your desk, send it over. I will name the stage and the revisions hiding inside the fee before you approve the print.

Subscribe
Notify of
guest
0 Comments
Most Voted
Newest Oldest
0
Would love your thoughts, please comment.x
()
x