Home / Courses / Engineering Graphics and CAD / Lesson 24

Engineering Graphics and CAD · Lesson 24 of 35

Multi-feature mechanical parts and mating geometry

Model realistic multi-feature parts and the specific geometry where parts meet.

Readiness check

Learning objectives

By the end of this lesson you can:

Decompose a real part into an ordered set of features.
Model mating faces, bosses, and locating features deliberately.
Maintain robustness across many features.
Identify the interface geometry another part will touch.
Keep a clean, named feature tree for a complex part.

Check your starting point

Five to ten minutes.

When two parts bolt together, which surfaces actually touch, and why do they matter more than the rest?
If two parts must share a hole spacing, how could you make sure they always match?
What keeps a ten-feature part understandable to another engineer?

Interpretation.

Q1: The mating faces and the fastener holes; they matter because the assembly depends on them fitting. This lesson calls them interfaces.
Q2: Drive both from the same parameter (a shared variable), so a change updates both. Skill 24.2.
Q3: A clean, named feature tree and a deliberate order. Skill 24.4.

You need L18-L23 (robust single-part modelling).

0 or 1 weak itemsContinue with this lesson.

2 weak itemsReview Lesson 21, Lesson 23, then return.

3 or more weak itemsWork through the prerequisite examples before continuing.

The core idea

What it is. A multi-feature part is a realistic part built from many ordered features, some of which form interfaces: the faces, holes, and locating features where the part meets another part. This lesson is about modelling those parts, and especially their interfaces, deliberately.

Why an engineer needs it. Real parts are not single blocks; they are stacks of features, and the ones that matter most are the interfaces. A part can be beautifully modelled internally and still fail if its mating faces or fastener holes do not match the part it joins. Modelling interfaces deliberately, from stable references, is what makes assemblies work.

What problem it solves. It produces parts whose interfaces are correct, robust, and easy to keep matched to their mating parts.

What goes wrong when it is ignored. Interfaces modelled casually (wrong size, fragile references, not matched to the mating part) cause assembly failures, and a tangled unnamed tree makes a complex part unmaintainable.

A simple mechanical example. One jaw of a mechanical clamp has a mating face (where it grips), a locating pin hole (that aligns it to the other jaw), and a fastener boss (that takes the clamp screw). These three interfaces must match the other jaw. The rest of the jaw's shape matters less; the interfaces decide whether the clamp works.

Key ideas:

Interfaces first and stable: model the mating faces, holes, and locating features from stable references (L23), because other parts depend on them.
Shared parameters: where two parts share an interface dimension (a hole spacing, a bore size), drive both from one variable so they stay matched.
Clean, named tree: name interface features and keep the order legible, so the part stays maintainable.

Part 5: Parts, assemblies, and drawings.

Check: explain the decision in your own words before using a CAD command.

The lesson map. Multi-feature mechanical parts and mating geometry becomes manageable when you move through the four checks in order and verify each result before continuing.

The skills, taught in order

Skill 24.1 - Decompose a part into features

Concept. A real part is an ordered set of features; plan the decomposition before modelling. Terminology. Feature decomposition, base feature, interface feature. Procedure. List the part's features, identify which are interfaces, and order them stable-first (base, then interfaces, then detail, then finishing). Reasoning. A planned decomposition keeps a complex part robust and legible. Failure mode. Modelling features ad hoc with no plan. Check. For a clamp jaw, list its features and mark the interfaces.

Skill 24.2 - Model interfaces deliberately

Concept. Mating faces, locating features, and fastener holes are modelled from stable references and matched to the mating part. Terminology. Mating face, locating feature (boss, pin, pocket), fastener hole. Procedure. Model each interface from stable geometry, sized to match the mating part; where a dimension is shared, drive it from a common variable. Reasoning. Interfaces decide assembly success, so they deserve stable references and matched dimensions. Failure mode. Interfaces on fragile references or not matched to the mating part. Check. Model a locating pin hole from a stable datum, sized to the pin.

Skill 24.3 - Maintain robustness across many features

Concept. Robustness rules (L23) apply throughout a complex part. Terminology. Robust tree, stable reference, finishing-features-last. Procedure. Keep referencing stable geometry, ordering finishing features last, and naming key parameters, even as the feature count grows. Reasoning. Complexity magnifies fragility; discipline keeps the part editable. Failure mode. Letting later features reference fragile geometry as the part grows. Check. Confirm a mid-tree feature references stable geometry.

Skill 24.4 - Keep a clean, named tree

Concept. Name interface features and keep the tree legible. Terminology. Feature naming, clean tree. Procedure. Rename important sketches and features (mating face, pin hole, screw boss) so their purpose is clear. Reasoning. Named features make a complex part maintainable and reviewable. Failure mode. A tree of unnamed default features nobody can follow. Check. Rename one interface feature to its purpose.

Worked example 1: model a clamp jaw with its interfaces

Problem. Model one jaw of a mechanical clamp (Project P5) with three interfaces: a flat gripping (mating) face, a locating pin hole of diameter 6, and a fastener boss with a tapped hole for an M6 clamp screw.

Planning. Base first, then the three interfaces from stable references, then finishing, with named features.

Solution.

Base. Model the jaw body as the base feature on a default plane, centred sensibly on the origin.
Mating face. Ensure the gripping face is a clean, flat face (a stable reference for the other jaw to seat against). Name it "mating face."
Locating pin hole. With the hole tool, cut a diameter-6 pin hole positioned from stable base faces (not from a fillet edge). Name it "pin hole." Size it for the locating pin.
Fastener boss. Add a boss and a tapped M6 hole for the clamp screw, positioned from stable references. Name it "screw boss."
Finishing. Add any fillets and chamfers last.
Tree. Base, mating face, pin hole, screw boss, finishing, all named.

Result. A clamp jaw with three deliberately modelled interfaces (mating face, diameter-6 pin hole, M6 screw boss), each from stable references and named, ready to mate to the other jaw.

Why the method works. Modelling interfaces from stable references and naming them makes the jaw robust and its mating features clear.

How to verify independently. Edit the jaw body size: the three interfaces, referenced to stable geometry, should stay correctly placed. If they move unexpectedly, a reference needs fixing.

Worked example 2: match the second jaw with a shared interface

Problem. Model the matching second jaw so its interfaces correspond to the first, and set up the shared pin-hole spacing so a change to it updates both jaws. The complication is keeping two parts matched through one parameter.

Planning. Mirror or re-model the jaw, then drive the shared interface dimension from a single variable used by both parts.

Solution.

Corresponding interfaces. The second jaw needs a mating face that meets the first jaw's, a pin hole that aligns with the first jaw's pin hole, and a boss/clearance for the same screw. Model these to correspond.
Shared dimension. The pin-hole spacing (the distance from a reference to the pin hole) must be identical on both jaws. Define it as a variable (say pinSpacing) and use that variable to drive the dimension in both parts.
Change test. Change pinSpacing once: both jaws' pin holes move together, staying aligned. Without the shared variable, you would have to edit each jaw and risk them drifting apart.
Clearance versus tapped. One jaw has the tapped M6 hole (the screw threads into it); the matching jaw has a clearance hole for the screw to pass. Model the correct one on each.
Result. Two jaws with corresponding, matched interfaces, and a shared variable keeping the pin spacing identical.

Comparison. Driving the shared dimension from one variable keeps the jaws matched automatically; dimensioning each jaw independently risks mismatch when one is edited. Shared parameters encode the intent that the interface must stay common.

Result. The second jaw's interfaces correspond to the first, and the pin-hole spacing is driven by a shared variable so a single edit updates both jaws.

Independent check. Change the shared variable and confirm both jaws update identically. Matched movement confirms the shared parameter is working.

Misconceptions and diagnostics

Misconception	Why it seems reasonable	Why it is wrong	Evidence that reveals it	Correction	Diagnostic question
"Interfaces are just faces like any other."	They look like ordinary faces.	Interfaces carry the assembly's function and tolerances; they need stable references and matching.	An assembly fails when a casually modelled interface does not match.	Model interfaces first, from stable references, matched to the mating part.	"Does another part depend on this face or hole?"
"Model each part's interface independently."	Parts are separate files.	Shared interface dimensions drift apart when edited independently.	One jaw's pin spacing changes; the other does not.	Drive shared dimensions from one variable.	"Is this dimension shared with a mating part?"
"Any order works for a complex part."	It is just more features.	Complexity magnifies fragility; interfaces should be stable and early.	A later edit breaks an interface built on fragile geometry.	Keep interfaces early and on stable references.	"Are the interfaces built on stable geometry?"

Practice ladder

Level A - Recognition

Task. On six part pairs, identify the interface faces and features on each. Deliverable. Six annotated pairs. Success criteria. Interfaces correctly identified in at least five. Answer guidance. Interfaces are where the parts touch or fasten. Common errors. Missing a locating feature. Difficulty. Low.

Level B - Guided application

Task. Model a scaffolded multi-feature part, with prompts, naming its interface features. Deliverable. The modelled part with named interfaces. Success criteria. Interfaces from stable references; features named. Answer guidance. Base, interfaces, detail, finishing. Common errors. Unnamed features. Difficulty. Medium.

Level C - Independent application

Task. Model a clamp component (Project P5, part 1) independently with its interfaces. Deliverable. The modelled component. Success criteria. Correct interfaces, stable references, clean named tree. Answer guidance. Identify interfaces first, model them deliberately. Common errors. Interfaces on fragile references. Difficulty. Medium.

Level D - Transfer and design

Task. Decompose a supplied part into features and model it with named interfaces and one shared variable for an interface it holds in common with a mating part. Deliverable. The model plus a note on the shared variable. Success criteria. Sensible decomposition; interfaces stable and named; shared variable drives the common dimension. Answer guidance. Find the dimension shared with the mating part and make it a variable. Common errors. Not sharing the common dimension. Difficulty. Medium to high.

Working with AI, and proving it yourself

Use AI as a tutor

Useful AI support:

Ask it to help identify interfaces on a described part.
Ask it to suggest a feature decomposition and critique it for robustness.
Ask it to explain shared variables across parts.

Limits:

A text assistant cannot see your tree or references.
It may miss which faces are true interfaces.

Verify AI output against: the interface-first principle, stable references (L23), and the shared-variable approach for common dimensions.

Prove it yourself

A plausible but incorrect AI answer, and how to catch it. You ask, "The two clamp jaws share a pin spacing. Should I just type the same number into each jaw?" and the assistant replies: "Yes, type the same value in both; that keeps them equal."

This is fragile. Detect it with the shared-parameter principle: typing the same number twice means the two are not linked, so editing one later leaves the other behind. The evidence is the change test: change one jaw's spacing and the other does not follow. Correct conclusion: drive both from a single shared variable so a change updates both jaws together.

Retrieval and spaced review

What is an interface on a part?
Why model interfaces from stable references and early?
How do you keep a shared interface dimension matched across two parts?
What keeps a complex tree maintainable?
Which parts of a jaw are its interfaces?
Why do interfaces matter more than incidental faces?
Cumulative (L23): How do the robustness rules from L23 apply to interfaces?
Reconstruction task: From memory, list the clamp jaw's three interfaces.

Answers. 1: the faces, holes, and locating features where the part meets another part. 2: because other parts depend on them, so they must be stable and correct. 3: drive both from one shared variable. 4: a clean, named tree and a deliberate order. 5: the mating (gripping) face, the locating pin hole, and the fastener boss/hole. 6: they decide whether the assembly fits and functions. 7: interfaces must reference stable geometry and be built early so they survive edits.

Suggested review intervals. 1 day, 3 days, 7 days.

Reference mapping and next step

Read further

Onshape docs
Giesecke ch.13.

Standards details must be checked against the current official edition used by your institution or employer.

Finish the lesson

You can now: decompose a part into features; model interfaces deliberately from stable references; maintain robustness across many features; share interface dimensions; and keep a clean named tree.

Self-assessment checklist.

I identify interfaces before modelling.
I build interfaces early, on stable references.
I drive shared dimensions from one variable.
I keep a clean, named feature tree.
My interfaces survive edits to the part body.

Next lesson: L25 - Assemblies: structure and mates. Why it follows: with parts whose interfaces are modelled deliberately, you can now bring them together into an assembly, connecting those interfaces with mates that reflect how the parts really join.

Required files or submissions: submit your Level C clamp component (Project P5, part 1). Optional extension: model the second clamp jaw and link its pin spacing to the first with a shared variable.