Home / Courses / Engineering Graphics and CAD / Lesson 18

Engineering Graphics and CAD · Lesson 18 of 35

CAD data, coordinate systems, and modelling strategy

Understand what a parametric CAD model *is* (features on a history tree driven by parameters) and plan before modelling.

Readiness check

Learning objectives

By the end of this lesson you can:

Describe parametric, feature-based, history modelling and how it differs from 2D CAD.
Identify the origin and the three default planes in a CAD part.
Explain the role of the coordinate system in model stability.
Plan a modelling strategy from a drawing before opening the software.
Justify a chosen base feature and order against likely design changes.

Check your starting point

Five to ten minutes.

In a CAD model, is the final shape stored directly, or is it built from a series of steps?
What do you think the "origin" of a CAD part is?
Before modelling a part, is it better to start drawing immediately or to plan the sequence first?

Interpretation.

Q1: It is built from a series of steps (features) that can be edited. This is the essence of parametric modelling.
Q2: A fixed reference point (0, 0, 0) with three planes, from which the model is built. Skill 18.2 covers it.
Q3: Plan first. A short plan prevents fragile models. Skill 18.4 shows why.

You need Parts I-III conceptually; CAD encodes the same geometry and definition you learned to read and dimension.

0 or 1 weak itemsContinue with this lesson.

2 weak itemsReview the learning objectives above, then return.

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

The core idea

What it is. Parametric, feature-based CAD builds a solid model as an ordered series of features (sketches turned into solids, then modifications), each driven by parameters (dimensions and relations) and recorded in a feature tree (the history of steps). Editing a parameter re-runs the steps and updates the model.

Why an engineer needs it. A parametric model is editable and reusable. Change a dimension and the whole part updates, drawings and all. This is why CAD replaced fixed 2D drawing for design: the model captures not just a shape but the recipe and intent behind it.

What problem it solves. It lets a designer change a part quickly and consistently, and it stores design intent (what should update together), which a static drawing cannot.

What goes wrong when it is ignored. Treating CAD like electronic paper (drawing lines with no structure) throws away its power. Modelling without a plan produces fragile models that break on the first edit (the subject of L23).

A simple mechanical example. To model the stepped block, you sketch a rectangle and extrude it (feature 1), then sketch the raised block on top and extrude that (feature 2). If the base width is a parameter, changing it updates the block and everything referenced to it. The model is a living recipe, not a frozen picture.

The stable foundation: origin and default planes. Every CAD part has an origin (the point 0, 0, 0) and three mutually perpendicular default planes (in Onshape: Top, Front, Right). These, and the coordinate system they define, are the most stable references in the model. Building from them (rather than from fragile model faces) makes a model robust, as later lessons show.

Modelling strategy. Before modelling, plan: which plane hosts the base feature, what that base feature is, and the order of subsequent features. A good plan chooses stable references and an order that survives the changes the part is likely to see.

Neutral terminology: Onshape Part Studio is the environment where parts are modelled (comparable to a part document in SolidWorks, a component/design in Fusion, NX, Creo, CATIA, or FreeCAD). The feature tree is called the feature list (Onshape), design tree (SolidWorks), or timeline (Fusion).

Part 4: Parametric CAD foundations.

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

The lesson map. CAD data, coordinate systems, and modelling strategy becomes manageable when you move through the four checks in order and verify each result before continuing.

The skills, taught in order

Skill 18.1 - Understand parametric, feature-based modelling

Concept. A model is an ordered set of parameter-driven features, not a static shape. Terminology. Parametric, feature, feature tree/history, parameter. Procedure. Think of each part as a sequence of features; expect to edit parameters and have the model update. Reasoning. Editability and reuse come from the feature structure. Failure mode. Treating CAD as static drawing. Check. State what happens when you change a driving parameter.

Skill 18.2 - Locate the origin and default planes

Concept. The origin and three default planes are the model's stable foundation. Terminology. Origin, default planes (Top, Front, Right), coordinate system. Procedure. Identify the origin and planes at the start; plan to build the base feature on a default plane through or near the origin. Reasoning. Default planes never move, so features built on them are stable. Failure mode. Ignoring the planes and building on arbitrary faces. Check. Name the three default planes.

Skill 18.3 - Respect the coordinate system for stability

Concept. Anchoring geometry to the coordinate system makes edits predictable. Terminology. Coordinate system, anchor, stability. Procedure. Place the base feature symmetrically about the origin or planes where sensible, so the model has a stable, meaningful zero. Reasoning. A meaningful origin simplifies symmetry, mirroring, and later references. Failure mode. Floating the part far from the origin with no relationship to the planes. Check. Explain one benefit of centering a symmetric part on the origin.

Skill 18.4 - Plan the base feature and order

Concept. Decide the base plane, base feature, and feature order before modelling. Terminology. Base feature, modelling order, strategy. Procedure. From the drawing, pick the plane and base feature that best represent the part, then list the order of subsequent features, choosing stable references. Reasoning. A deliberate plan avoids fragile dependencies and rework. Failure mode. Modelling ad hoc and discovering late that the structure is fragile. Check. Write a two-line plan for a simple part (base plane, base feature, order).

Worked example 1: a modelling plan for the stepped block

Problem. Write a modelling plan for the stepped block (base 60 by 40 by 20; raised block 30 by 40 by 20 centred on top) before modelling it.

Planning. Choose a base plane and base feature, then order the features and note the driving parameters.

Solution.

Base plane. Use the Top default plane for the base footprint, so the part builds upward in a natural orientation, centred on the origin for symmetry.
Base feature. Sketch the 60 by 40 base rectangle centred on the origin and extrude it 20 upward. This is feature 1.
Second feature. On the top face (or better, using a dimension from the base), sketch the 30 by 40 raised block centred across the width and extrude it 20. This is feature 2.
Driving parameters. Base width 60, base depth 40, base height 20, block width 30, block height 20. The block is centred, so its position is driven by symmetry, not a separate dimension.
Order rationale. Base first (it is the largest, most stable feature); block second (it depends on the base). Centering on the origin makes symmetry and any future mirroring easy.

Result. A two-feature plan: extrude the centred base (feature 1), extrude the centred raised block (feature 2), with symmetry driving the block's position.

Why the method works. Planning the base, order, and parameters first gives a stable, symmetric model that is easy to edit.

How to verify independently. Check that changing the base width would keep the block centred (because it is driven by symmetry, not a fixed offset). If so, the plan is robust to that edit.

Worked example 2: two strategies for an L-bracket

Problem. Compare two modelling strategies for an L-bracket (two perpendicular arms). Strategy A sketches the full L profile on one plane and extrudes it. Strategy B models one arm, then adds the second arm as a separate feature on a model face. Predict which survives edits better. The complication is that the two strategies differ in robustness, not in final shape.

Planning. Consider how each responds to likely edits (changing an arm length or thickness).

Solution.

Strategy A (L profile extruded). Sketch the whole L shape (both arms) on the Front plane and extrude it to the bracket width. Both arms share one sketch, driven by a few parameters.
Strategy B (arm plus arm). Model the first arm as a box, then sketch and extrude the second arm on a face of the first. The second arm depends on the first arm's face.
Edit test, arm length. In Strategy A, changing an arm length edits one sketch dimension and the L updates cleanly. In Strategy B, changing the first arm may move the face the second arm was built on, which can shift or break the second arm.
Edit test, thickness. In Strategy A, thickness is the single extrude depth; it updates everywhere. In Strategy B, the two arms may have independent thicknesses that must be kept equal manually.
Verdict. Strategy A (single profile) is more robust: both arms come from one intentional sketch, so edits stay consistent. Strategy B couples the second arm to a model face, a fragile reference (the subject of L23).

Comparison. Both produce the same L-bracket, but Strategy A survives edits with a single, intentional sketch, while Strategy B risks breaking when the base arm changes. Prefer the strategy that builds from stable references and one clear intent.

Result. Strategy A (extrude the full L profile from one sketch on a default plane) is the more robust plan; Strategy B (second arm on a model face) is more fragile.

Independent check. Predict the effect of doubling an arm length in each strategy. Strategy A updates cleanly; Strategy B may shift or fail the dependent arm. That difference confirms the verdict.

Misconceptions and diagnostics

Misconception	Why it seems reasonable	Why it is wrong	Evidence that reveals it	Correction	Diagnostic question
"CAD is just drawing on a computer."	It looks like drawing.	CAD builds an editable, parameter-driven solid, not static lines.	Editing a parameter updates the whole model.	Model with features and parameters, not static lines.	"Can I change one dimension and have the model update?"
"The origin and planes do not matter."	The part looks fine anywhere.	Default planes are the stable references that keep edits predictable.	A part floated far from the origin is hard to mirror or reference.	Build the base on a default plane, centred where sensible.	"Is the base built on a default plane near the origin?"
"Just start modelling and figure it out."	It feels faster.	Ad hoc modelling produces fragile trees that break on edits.	The model fails when an early feature changes.	Plan base, order, and references first.	"Do I have a base-feature and order plan?"

Practice ladder

Level A - Recognition

Task. In five CAD screenshots, identify the origin, the three default planes, and the base feature. Deliverable. Five labelled screenshots. Success criteria. Origin, planes, and base feature identified in at least four. Answer guidance. The base feature is the first solid in the tree. Common errors. Confusing a default plane with a model face. Difficulty. Low.

Level B - Guided application

Task. Write a base-feature-and-order plan for a given part, with prompts for plane and base choice. Deliverable. A short written plan. Success criteria. Sensible base plane, base feature, order, and driving parameters. Answer guidance. Largest/most stable feature first; build on a default plane. Common errors. Choosing a fragile base. Difficulty. Medium.

Level C - Independent application

Task. Plan the modelling order for a supplied multi-feature part, with no prompts. Deliverable. A written strategy. Success criteria. Stable base; logical order; references chosen for robustness. Answer guidance. Anticipate which references stay stable. Common errors. Building later features on fragile faces. Difficulty. Medium.

Level D - Transfer and design

Task. For one part, write two different strategies and argue which better survives a stated set of likely edits. Deliverable. Two plans plus a comparison. Success criteria. Correct prediction of which plan is robust, with reasons tied to references and order. Answer guidance. Test each plan against the specific edits. Common errors. Comparing only the final shape, not robustness. Difficulty. Medium to high.

Working with AI, and proving it yourself

Use AI as a tutor

Useful AI support:

Ask it to explain parametric modelling and confirm against this lesson.
Ask it to suggest a modelling order for a part, then critique it for robustness.
Ask it to list the default planes in your CAD tool.

Limits:

A text assistant cannot see your part or tree.
It may propose an order that builds on fragile faces.

Verify AI output against: the stable-reference principle (default planes over model faces) and the edit test (does the plan survive likely changes?).

Prove it yourself

A plausible but incorrect AI answer, and how to catch it. You ask, "What is the fastest way to model this part?" and the assistant replies: "Just extrude each feature off whatever face is nearest; do not worry about planes or order."

This produces a fragile model. Detect it with the strategy principle: building off arbitrary faces creates parent-child chains that break on edits (L23). The evidence is the edit test: change an early feature and dependent features shift or fail. Correct conclusion: plan a base on a stable default plane and choose references and order deliberately, even if it feels slightly slower at first.

Retrieval and spaced review

What is a parametric, feature-based model made of?
What are the origin and three default planes for?
Why build the base feature on a default plane?
What does a modelling plan decide before you start?
How does parametric CAD differ from 2D CAD?
What happens when you edit a driving parameter?
Cumulative (L12): How does choosing stable references in CAD echo choosing functional datums in dimensioning?
Reconstruction task: From memory, write the two-feature plan for the stepped block.

Answers. 1: an ordered set of parameter-driven features recorded in a feature tree. 2: they are the stable reference point and planes the model is built from. 3: default planes never move, so features built on them stay stable. 4: the base plane, the base feature, and the feature order (and driving parameters). 5: it builds editable solids from features, not static lines. 6: the model re-runs its features and updates. 7: both anchor the definition to stable, functional references so changes behave predictably.

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

Reference mapping and next step

Read further

Onshape docs (Part Studios, planes)
Giesecke ch.5 (geometry).

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

Finish the lesson

You can now: describe parametric feature-based modelling; identify the origin and default planes; use the coordinate system for stability; and plan a base feature and order before modelling.

Self-assessment checklist.

I can explain what a feature tree stores.
I can find the origin and default planes.
I plan a base feature and order before modelling.
I build on default planes where sensible.
I can compare two strategies for robustness.

Next lesson: L19 - Sketch planes, reference geometry, and sketch entities. Why it follows: with a modelling plan in hand, the first concrete step is the sketch. L19 covers choosing a sketch plane and drawing clean sketch entities, the foundation every feature is built on.

Required files or submissions: submit your Level C modelling strategy. Optional extension: open Onshape, create a Part Studio, and identify the origin and the three default planes before modelling anything.