Ask three engineering firms how long the mechanical design of a new product takes and you will get three answers: six weeks, four months, and "it depends." All three can be honest. Mechanical design is not one activity. It is a chain of activities whose total length is set by how many moving parts the product has, how tight the fits are, how many prototype rounds the design needs before it behaves, and how quickly someone on the client side answers questions. Projects rarely slip because CAD modeling is slow. They slip because the schedule counted only the modeling.
This article breaks a mechanical design program into the phases that actually consume calendar time, gives realistic ranges for each, and points at the places where weeks quietly disappear.
What Mechanical Design Actually Includes
When a founder says "design the mechanics," they usually picture a 3D model. The model is maybe a third of the work. A complete mechanical design phase produces:
- A mechanical architecture: how the product splits into parts, which parts move, where the loads go, and how it is assembled and serviced.
- Detailed 3D CAD of every custom part, plus correctly represented off-the-shelf components.
- Material and process selection for each part, because a part designed for injection molding is not the same part designed for machining.
- Analysis where it matters: stress, deflection, thermal, and tolerance stack-up across the assembly.
- A drawing and release package that a supplier can quote and build from, with dimensions, tolerances, finishes, and inspection callouts.
Skip any of these and the calendar time does not vanish. It reappears later as a prototype that does not close, a quote that comes back triple, or a first article that gets rejected.
A Realistic Phase Breakdown
The ranges below cover a moderately complex consumer or light industrial product: an enclosure with a mechanism inside, a motor or two, a PCB, and a user interface. Simple products land at the low end, machines with several subsystems run past the high end.
| Phase | Typical duration | Main deliverable | What stalls it |
|---|---|---|---|
| Requirements and architecture | 1-2 weeks | Layout sketches, part breakdown, interface list | Requirements that are still opinions |
| Concept mechanisms | 2-4 weeks | Two or three working mechanical approaches | Waiting on a decision between them |
| Detailed CAD | 3-8 weeks | Full 3D assembly, every custom part modeled | Electronics outline not frozen |
| Analysis and simulation | 1-3 weeks, often parallel | FEA results, stack-up study, thermal check | Load cases nobody defined |
| Prototype round | 3-6 weeks each | Physical parts, assembly, test report | Shipping, and finding test time |
| Drawings and release | 1-3 weeks | Manufacturing data package | Late tolerance decisions |
| DFM revisions after quoting | 2-4 weeks | Revised parts that suppliers will actually build | Supplier response times |
Add those up and a single-pass program is roughly 12 to 20 weeks. Add a second prototype round, which most products need, and 16 to 26 weeks is the honest number. That is mechanical design alone, running alongside the industrial design work and electronics rather than after it.
Complexity Tiers That Change the Answer
Static parts only
A bracket, a housing for a board, a mount, a rugged case. No mechanism, no moving interfaces. Detailed design is often 2 to 4 weeks and the whole program can close in 6 to 9 weeks including one printed prototype. Cost typically lands in the $6,000 to $18,000 range for the mechanical scope.
One mechanism
A hinge that has to survive tens of thousands of cycles, a latch, a dispensing head, a sliding drawer. Now the design has cycle life to prove, which means test fixtures and a real test plan. Plan 12 to 20 weeks and two prototype rounds. Anything with repeated motion pulls in fatigue behavior over the product's service life, which cannot be settled with a single printed sample.
Multi-subsystem machines
A benchtop instrument, a piece of production equipment, a motorized consumer appliance with a pump and a heater. Several mechanisms interact, the frame carries real loads, and safety standards apply. Six to twelve months is normal, and the schedule is dominated by integration testing, not by drawing parts.
Where the Weeks Actually Go
Across projects, the same handful of items eat the schedule:
- Unfrozen inputs. The single largest cause. If the PCB outline, the battery, the motor, or the display can still change, every downstream part is provisional. Freezing the mechanical interfaces early is worth more than any CAD productivity trick.
- Decision latency. An engineer sends three concepts and waits nine days for a reply. That nine days is on the schedule even though nobody worked. Weekly decision meetings with a named decision-maker routinely save a month.
- Prototype logistics. Printing takes days; shipping, customs on overseas parts, and finding a week when the client can actually run the test take longer than the fabrication.
- Discovering the requirement late. A drop height, a washdown requirement, an IP rating, a certification. Each one arriving in week ten costs more than it would have in week one.
- Skipping the release package. Sending STEP files without drawings feels fast, then produces parts that are dimensionally legal and functionally wrong. A complete manufacturing drawing is a schedule tool, not paperwork.
How Many Prototype Rounds to Plan For
Two is the realistic default for a product with a mechanism. The first round proves the concept and finds the three things nobody anticipated. The second round validates the fixes and gets close to production intent. A third round appears when a certification test fails, when the client changes the industrial design after seeing the first physical model, or when a supplier asks for changes during quoting.
Simulation reduces rounds but does not eliminate them. FEA work that pays for itself is usually the analysis that lets you skip a round of "is this rib thick enough," not the analysis that replaces physical testing of an assembly.
Compressing the Schedule Without Wrecking It
Legitimate accelerators exist. Running mechanical and electronics design concurrently with a written interface agreement between them saves weeks. Ordering long-lead off-the-shelf parts during concept rather than after CAD freeze removes a dead month. Designing the first prototype in printable geometry while keeping the production process in mind lets you test function early without redesigning later. And front-loading design for manufacturing decisions means the supplier review at the end confirms rather than rewrites.
The illegitimate accelerators are equally well known: cutting the analysis, skipping the tolerance study, releasing without drawings, and running one prototype round instead of two. These do not shorten the program. They move the delay past the point where it is cheap to fix, which is exactly the pattern behind most design errors that only surface in production.
What to Ask Before You Sign
A credible mechanical schedule names the phases, states how many prototype rounds are included, lists the inputs the engineer needs from you and when, and identifies the assumptions that would change the number. A schedule that is one line saying "mechanical design: 8 weeks" is not a schedule; it is a hope. Ask what happens to the timeline if the second prototype fails, and listen for a specific answer.
Projects House plans mechanical programs in phases with named gates, so you know before the work starts where the decision points are and what each one costs in calendar time. If you have a product that needs a realistic mechanical schedule rather than an optimistic one, send the details through our contact form and we will walk you through the phases that apply to your design.