Design for Inspection Part 2 | Thinking Outside of CAD for Successful Outcomes

As we reviewed in Part 1 of this series, by recognizing that GD&T is a language, we need to think carefully about how we communicate our design intent with multiple audiences.

Aaron Nelson
AC Mechanical Engineer & Technical Lead

“We can think about GD&T as ‘Design for Inspection’ (DFI), which helps us anticipate and specify how a part and its features are to be measured.

This process enables us to fine-tune our efforts to enable successful part inspection outcomes.”

To do this, we have to think outside of the “perfect” world of Computer-Aided Design (CAD), where everything is square and planar and lined up, because, in reality, parts rarely come out that way.

When we design a part in CAD, it’s easy to feel like we have controlled the part after just a few clicks and notations, but it’s an illusion if we don’t understand how the part is manufactured or measured in reality.

We must consider the material used, the physical outcome of the part, the tools, and the techniques that are used for measurement and inspection – all of the real-world impacts that help us anticipate how to use GD&T to better represent the physical part in our drawings.

So, let’s take a look at a few ways we can think outside of CAD to drive the successful inspection of our manufactured parts.

Anticipating the Misinterpretation of Our Design Intent

We all wish our designs were elegant enough that our intent is self-evident. Unfortunately, that’s not often the case. We further complicate this when we pass our design off to a drafter or another engineer to complete our drawing for us because they don’t know as much about our design and intent as we do.

The explicit nature of GD&T gives the impression that there is no room for interpretation, but there is. Like any language, there are assumptions, biases and expectations.

Remove ambiguity by communicating our design intent

In reality, our drawing is almost exclusively the only document we have that carries our design intent. It conveys our manufacturing and inspection instructions for measuring and qualifying the part. In fact, it is the sum of all the hard work put into designing the part. So, we should own the drawing and how we precisely communicate within it.

Yes, the CAD file will accompany the drawing, but it does not often include the type of information required by the inspector to determine our intentions.

Start by understanding the inspector and their approach

So, how can we improve the drawing to accurately describe our design intent? We can start by anticipating how our drawing might be interpreted (or misinterpreted) and then use GD&T to remove that ambiguity from our drawing.

It’s important to find out what the inspector understands about GD&T and what their approach will be to inspecting our parts.

Whenever possible, we should ask to watch the inspector review our drawing, in person or virtually. By watching them go through the process to do the inspection, we learn their level of understanding and what tools they will choose to do the inspection. This information is crucial to help us communicate with GD&T more precisely and ultimately drive a successful inspection outcome.

Pros and cons of different inspection types

Let’s consider some of the different inspection tools to understand how the pros/cons might affect the inspection of our parts.

Inspection Tool



CMM (coordinate measuring machine)

> Pro: Has many different configurations and capabilities

> Pro: Most common tool for inspection

> Con: Does not do well with parts that can deflect under the load of the probe

OMM (optical measuring machines)

> Pro: Very good for 2D parts with an edge

> Pro: Uses the contrast between the edge of a part and the light behind it

> Con: Does not do well when light cannot pass around an edge, such as when measuring the height of a screw boss

3D Scanner

> Pro: Has a variety of configurations, each with their own pros and cons

> Pro: Non-contact inspection is good for fragile parts

> Con: Can be time-consuming

> Con: Some scanners require a spray applied to the surface of the part

Hard surface tools

> Pro: Great for assembly line go/no-go inspection

> Con: Not flexible

> Con: Typically enables only a few measurements with each fixture

Anticipating the Physical Imperfections of Manufactured Parts

Technical drawings may look very straightforward to us. In CAD, everything has sharp corners and perfect circles. But even as we think about such a simple feature as a hole in a part, we see it can be measured in different ways.

For example, in theory, all we need to define a circle is three points, right?

This should mean that when the manufactured hole is inspected, we should be able to compare it to the circle diameter on the drawing. That’s our expectation, but it’s almost never the case. In fact, a manufactured hole is never perfectly round.

To illustrate this simple challenge, let’s look at a rough depiction of a manufactured hole, as it turned out in reality. With this image below, imagine we’ve zoomed in a lot to see the deformities.

If we simulate three points for inspection, we can see how it won’t actually represent the feature at all.

So, what can we do to anticipate a better representation of our circle in the drawing? What kind of data should we expect to get back? Well, it depends on the tools the vendor will use to measure this hole. To improve our drawing for inspection by a CMM, we should anticipate the need for more points.

For example, if we expect six points, the measured circle would then result in a better representation of the actual physical part. Keep in mind that the inspection tool (CMM or OMM) usually calculates the best fit circle for the points measured, whereas a pin gauge will measure the smallest included circle. However, we typically don’t specify the number of points – that’s left up to the inspector. There’s room for interpretation on the inspector’s end and our experience is that if there’s a shorter or faster way to do something, our inspector will probably do it. We need to keep that in mind.

The point is: even something as simple as a circle can have multiple layers of interpretation and meaning. It’s more complicated than it seems. But as long as we understand these assumptions and expectations, we can improve our part drawings.

So, to reduce ambiguity, we need to consider how the physical part is going to be formed and measured, and then improve our drawing to better represent the part. This translates to precise communication of our design with the inspector.

In this example of a circle representing the hole of our part, we could improve the drawing by:
> Using more points to better represent the physical part
> Adding a profile to control the form and position of the circle
> Specifying a pin gauge when we can’t anticipate how many points it takes, and the shape is crucial to the assembly

When our intentions are understood, we prevent ambiguity that can lead to misinterpretation and failed inspections as a consequence.

Simple dimensions can be interpreted in multiple ways

To further illustrate this point, let’s look at a relatively simple way of depicting a dimension. What possible interpretations could there be that require us to be careful about how we communicate using GD&T?

With this drawing, we might be tempted to think to ourselves: ““I just want to measure the location of the hole. So, in the drawing, I’m going to put a dimension to measure three holes. I don’t want to bother with creating a datum and coordinate system – that takes too much time.”

Sounds easy. The thing is, in CAD, everything is perfect: it’s square, and planar, and lined up. It might seem obvious to the designer that this dimension would fully constrain the holes, but as we’re going to see, there are multiple interpretations of this dimension. This drawing doesn’t control form, only position. And it’s not clear what this dimension represents.

Without a GD&T reference frame, or coordinate system, because the part probably won’t turn out perfectly perpendicular, the dimension could mean this:

Or, perhaps this is how our drawing is interpreted?

Or maybe the interpretation could be to measure from either of these corners?

The problem with a single point-to-point measurement is that we’re not being clear about how we intend it to be measured.

But that’s the beauty of GD&T: it formalizes a common set of terms and provides a language to convey meaning so we can prevent misinterpretation.

Why can’t I just use direct dimensions and +/- tolerances? Because it doesn’t control the form and is unclear about what the dimension represents.

Tip: Anticipate how tools are used to measure part features

As we’ve just seen, we might think our dimensioning in CAD is clear, but it can differ from the real-world measuring scenario. When a physical part is made, we can expect it’s not going to be perfectly planar or parallel, etc.

This also means that we need to think about the tools that will be used to measure the part features and add specifications to our drawings that help the inspector take an accurate measurement of the imperfect part.

For example, with calipers we may need to:
> Anticipate the best plane for taking a measurement of an imperfect part
> Add a dimension or tolerance that controls the form and position of the part

Tip: Specify measurement from a particular point of origin

By using the datum symbol (circle with crosshairs), we can specify that the measurement is taken from a particular origin, and then use a standard tolerance limit of the ± (plus or minus) symbol. Even this small notation change can prevent the kind of ambiguity that avoids misinterpration of our design intent to improve our communication and the inspection outcome.

Specifying the measuring technique

Understand tooling and how measurements will be performed

GD&T was designed for hard tooling

Decades ago, hard tooling and basic measuring tools were the standard. Height gauges, dial indicators, and other hard tools are still a way we can confirm the continuous measurement of a part, but hard tools don’t give us discrete measurements. “Continuous” means that the measurement can be made simultaneously over the entire surface under inspection.

For example, when using an inspection surface plate, we have the advantage of a continuous measurement reference plane.

But for most product development engineering, we don’t often see parts measured with hard tooling. We may use some go/no-go fixturing, but that’s not really GD&T because there’s no numerical measurement being performed.

Understand the advantages and limitations of measuring techniques

Most current tooling relies on CMMs to sample points to define a plane. However, that mathematically created plane doesn’t actually represent the physical part and can’t be confirmed as easily as with hard tooling. For example, a CMM cannot inspect a hole all the way around simultaneously as a pin gauge could because a CMM only samples points – it can’t sample continuously.

OMMs can pick discrete points faster than a CMM, but they also can’t do all points continuously. An OMM might “see” the whole feature at the same time, but it only picks discrete points. Even when you dial in the number of points to sample, such as finding the hole then measuring 100 points or 1000 points, it’s still not a continuous measurement.

By understanding the equipment that will be used by the inspector, we can tune our design and notation to work with their tooling, whether continuous or discrete.

A CMM may probe this surface that’s supposed to be planar, and it picks some points to create a best fit plane.

Looking at this best fit plane, we notice that some of the points of the part lie above the line and some lie below the line. But there’s no way to have the CMM just measure the high spots (or low spots, for that matter). It’s an important distinction, because when we mate something to this part, we can’t mate anything to the low points in the valleys. When we mate a part, it’s likely going to pick up only the high points.

In this example, we see that using a hard tool only picks up the high points. The bottom surface of this hard tool also does not match the best fit plane of the CMM – they’re not the same. However, because the mating part will be like the hard tool (we’re only going to hit the high spots), there’s a huge advantage to using hard tooling. It more accurately represents the mating interface, however, it often takes longer and is more expensive.

As we can see, using GD&T doesn’t mean it represents how the parts will be assembled – it’s just an approximation. What we really need is to understand what the inspection will be so we can anticipate those best fit circles and best fit planes. By doing that, we can adjust our drawings to the reality of the way parts are inspected and how they interface with other parts.

When best fit algorithms are used, it no longer simulates actual mating conditions, or hard tool fixtures.

Tip: Request the CMM hit map to anticipate the inspection

Sometimes we run into difficulties with a vendor’s CMM program. One solution is to ask for their CMM hit map to understand the planned CMM hits.

For example, this CMM hit map shows the locations where the CMM will touch the part represented by the black dots.

When we can’t be there with the inspector, this still gives us a virtual way to see what they’re planning and how they’ve programmed the inspection. From this information, we can improve how our part is dimensioned for those hits or ask the inspector to change the location of the hits.

Anticipating How Parts are Fixtured During Inspections

In addition to understanding the inspector’s measurement techniques, we also need to anticipate how some parts will need to be fixtured.

When we identify that a part requires an inspection fixture, we can plan for its inspection by understanding how the part is used in its assembly and what features will be inspected. This insight enables us to define our datum for part inspection. We’ll explore more about fixturing and defining datums in the next part in this series, so stay tuned!

Coming Soon in Part 3: Creating Better Datum Coordinate Systems

Read More