If you are a product designer sending STEP files to a contract manufacturer, the most useful thing to understand is not any single design rule. It is the timeline your file enters the moment you hit send — and the exact point on that timeline where changing your design stops being a file edit and starts being a machining bill.
Here is the short version. After you send the file, your part moves through four stages: quote, tooling, trial shots, and production. Your ability to change the design cheaply drops off a cliff at one specific moment: when the mold steel is cut. Before that, a change costs you an afternoon in CAD. After it, a change means re-machining a tool, welding metal back into steel, or paying for a new mold.
This post walks through each stage, what the factory is actually looking at, and why the window to fix a geometry problem closes so hard.
What does the factory actually check when it quotes your part?
When you send a STEP file for a quote, the factory reads the geometry to price the job — not to review your design in depth.
To build a quote, a moldmaker pulls the overall dimensions (length, width, height), the part volume and projected area, the weight, the plastic family and specific grade, and that material's shrinkage rate. It also looks at your nominal wall thickness and whether the walls have basic draft, because both affect whether the part can be molded and ejected at all.
What the quote stage usually does not include is a deep review of tight tolerances, complex internal fillets, or every face of your geometry. Those get scrutinized after an order is placed, not before. And there is a structural reason for that: in the traditional workflow, most manufacturers will not provide detailed DFM feedback until you place an order. Some modern rapid-manufacturing platforms now include a free DFM review with the quote — flagging things like gate location, parting-line marks, ejector-pin marks, and predicted weld lines — but that is a newer convenience, not the industry default.
The practical takeaway: a clean quote is not a clean bill of health. It means the factory can price and mold the part, not that the part is free of the geometry problems that surface later.
How long does injection mold tooling take?
Tooling lead time is driven by the tool material and the production volume it is built for, and it ranges from about one week to two months.
- Rapid or prototype tooling (aluminum or soft steel like S50C): about 1 to 2 weeks, or roughly 10 to 14 business days. First samples can come as fast as 10 days. These tools are built for low volumes — on the order of 1,000 to 10,000 parts.
- Production tooling (hardened steels like P20, NAK80, or S136): about 2 to 4 weeks, up to 4 to 6 weeks for more complex molds, and as long as 60 days for large or intricate tools. These are built for 100,000-plus cycles.
Mold cost tracks the same tiers: roughly 2,000 to 5,000 US dollars for low-volume aluminum tools, 5,000 to 15,000 for mid-volume steel, and 15,000 to 50,000 or more for high-volume hardened tooling. (Exact figures vary by supplier, region, and part complexity — treat these as order-of-magnitude reference values, not a quote.)
The reason this matters for design changes is simple arithmetic: every week and every dollar in that tool is committed the moment the moldmaker starts cutting steel. That is the point on the timeline you want to reach with your geometry already sorted.
What are T1 and T2 samples, and what is being checked at each?
T1 and T2 are trial shots — the first physical parts pulled from a newly built mold — and they are where design problems finally become visible in the real world.
T1 is the first real proof. The moldmaker mounts the tool, shoots plastic, and checks all the drawing dimensions against the model, hunting for measurement mistakes and gross molding defects like flash (plastic seeping into the tool's seams) or short shots (the cavity not filling completely).
T2 comes after the tool is adjusted based on what T1 revealed. This round verifies the refined dimensions and looks hard at surface and cosmetic issues — sink marks, a wavy surface finish — plus functional testing and dimensional inspection against the specified tolerance grade (commercial or fine). For demanding parts, this stage can include an initial process-capability study (Cpk) or destructive structural testing.
If a T1 or T2 sample needs a change, you do not simply email a new STEP file. You submit an engineering change order, the tool is modified, and you shoot another round of samples. This is also why moldmakers recommend starting a new product with a single-cavity tool: it is cheaper and faster to build, precisely because your design may still change and require new tooling before you commit to a multi-cavity production mold.
This is the stage where the physics defects that are invisible in a CAD file — sink marks, weld lines, and the rest — finally show up on a real part. By then, the steel is already cut.
Why do design changes get so expensive once the mold is cut?
Because a mold is the inverse of your part, and there is a hard asymmetry between removing steel and adding it back.
A mold cavity is a negative of your plastic part: the empty space becomes plastic, and the solid steel becomes the part's holes, edges, and boundaries. That geometry creates a one-directional rule that toolmakers call steel-safe (or metal-safe).
- Making the plastic part larger or thicker means cutting more steel away from the cavity. Removing steel — a bit more CNC milling or an EDM burn — is fast, cheap, and low-risk.
- Making the plastic part smaller, thinner, or filling in a feature means putting steel back into a tool that has already been machined. That requires welding metal into the mold and re-machining it, which is slow and expensive and can degrade the tool — and in some cases is not feasible at all, forcing a new mold.
This is exactly why experienced moldmakers deliberately machine tools slightly on the "safe" side of a dimension: so that any correction can be made by removing material rather than adding it back. A dimensional fix is possible in one direction and painful in the other. When you discover at T1 that a wall is too thin or a snap-fit hook won't eject, which direction the fix runs decides whether it costs an afternoon or a new tool.
The uncomfortable part for designers: the cost of a change roughly follows the timeline. A change caught in CAD costs an afternoon. A change caught at quote costs a re-quote. A change caught at T1 costs a tool modification and a fresh round of samples — days to weeks and real money. A change caught in production can cost a new mold.
When is the cheapest time to catch a design problem?
Before the file ever leaves your desk — specifically, the problems that are fully determined by your geometry.
Not every issue is catchable early. Sink marks, weld lines, and air traps depend on melt flow, gate placement, and cooling — the kind of physics that surfaces at T1 and genuinely needs a manufacturer's DFM review or flow simulation. But two of the most common tooling-killers are not physics problems at all. They are pure geometry:
- Draft — whether every vertical wall tapers enough to release from the steel.
- Undercuts — features that hook around the mold and prevent it from opening without side-actions.
Both are deterministic from the STEP file alone. They do not depend on material or gate location. They are exactly the kind of problem that is cheap to fix in CAD and expensive to fix in steel — a missed undercut can turn a straight-pull mold into a side-action tool, and a wall with zero draft can mean drag marks on every part in the run.
This is where checking your own geometry before you send the file pays for itself. Not to replace the manufacturer's review — to arrive at it having already cleared the problems that are plainly visible in the geometry, so the conversation is about the harder physics, not the basics.
Fabdose reads your STEP or STP file on your computer and locates, face by face, which walls are below the draft threshold and which features create undercuts — in about a minute. Your CAD geometry is processed locally on your own computer; the defect findings and their descriptions are processed by AI. It covers injection molding draft and undercut detection; it does not simulate the flow-dependent defects (sink, weld lines, air traps) that require a manufacturer's DFM review, and it is not a substitute for that review.
If you have a STEP file about to go out for a quote, the cheapest possible moment to catch a draft or undercut problem is right now — while a fix is still a file edit and not a welding job.
Check your design with Fabdose
FAQ
What happens after you send a STEP file to an injection molding factory?
The part moves through four stages: the factory quotes it from geometry and material, the mold is designed and cut, the factory runs trial shots (T1 and T2) to verify dimensions and surface quality, and then production begins. The key point is that changing the design cheaply gets much harder once the mold steel is cut — before that a change is a file edit, after it a change means re-machining steel, welding metal back into the tool, or cutting a new mold.
How long does it take to make an injection mold?
Rapid or prototype tooling in aluminum or soft steel typically takes 1 to 2 weeks, with first samples in as little as 10 days. Production tooling in hardened steel usually takes 2 to 4 weeks, up to 4 to 6 weeks — or as long as 60 days — for large or complex molds.
What is a T1 sample in injection molding?
A T1 sample is the first trial shot from a newly built mold. The moldmaker checks all drawing dimensions against the model, catches measurement mistakes, and looks for gross defects like flash or short shots. It is the first time your design exists as a real molded part rather than a CAD model.
What does "steel-safe" mean in injection molding?
Steel-safe (or metal-safe) is a moldmaking strategy where a tool is machined slightly on the conservative side of a dimension so that later corrections can be made by removing steel rather than adding it. Removing steel makes the plastic part larger and is cheap; adding steel back requires welding and re-machining and is expensive or sometimes impossible. It is why the direction of a design change decides how much it costs.
Fabdose is a desktop tool for checking STEP and STP files against injection molding design rules. It locates draft violations and undercuts face by face. Your CAD geometry is processed on your own computer; defect findings and descriptions are processed by AI.