Let’s start with a familiar scenario.

You’ve got a simple machined plate. Nothing fancy—just a mounting surface. You call out a tight flatness tolerance because, well, it “should be flat.”

The part comes back late. Expensive. Maybe even rejected. And now everyone’s asking the same question: “Did we actually need that flatness callout?” 

Most of the time, the answer is no.

Flatness is one of the most misunderstood tools in GD&T. It’s easy to apply, easy to overdo, and surprisingly expensive when you get it wrong.

This issue is about getting it right—so your drawings reflect real function, not accidental constraints.

What Flatness actually controls (and what it doesn't)

Flatness is a form tolerance. It doesn’t care where your surface is. It doesn’t care what it’s aligned to. It only cares about one thing: how much a surface deviates from a perfectly flat condition.

Technically, you’re defining a tolerance zone of two parallel planes, separated by your tolerance value. The entire surface has to sit between them.

That’s it.

No datums. No orientation. No location. Just shape.

And that’s where things go sideways.

Because in real assemblies, surfaces don’t exist in isolation—they exist relative to other things. GD&T as a whole is designed to communicate those relationships clearly.

Flatness doesn’t.

Where flatness goes wrong

Flatness gets used as a catch-all because it feels safe and precise. But most of the time, it’s solving the wrong problem. 

These are the most common flatness mistakes (apologies in advance - you're going to start seeing these everywhere):

• Flatness on large, non-critical surfaces (adds cost without improving function)
• Flatness where parallelism is needed (miscommunicates assembly requirements)
• Over-tight flatness on thin parts (leads to warping after machining) 
• Using flatness to "clean up" loose designs (masks root cause tolerance issues)

These issues are best understood within the design, manufacturing, and inspection domains. 

 →  For engineering

1. Flatness vs. Parallelism

One of the biggest culprits on the engineering front is when flatness is used instead of parallelism. Flatness controls surface form, while parallelism controls orientation relative to a datum.

If your surface needs to sit squarely on another part, flatness alone won’t guarantee that. You can have a perfectly flat surface that ends up tilted in the assembly.

That’s how you end up with parts that pass inspection and still fail in real life.

2. Flatness vs. Profile

Another engineering infraction is when flatness is chosen over profile.  

Flatness assumes the surface should be perfectly planar.

But what if your geometry is intentional—curved, blended, or functional?

That’s where profile comes in. It controls deviation from a theoretical shape, not just flatness.

Using flatness in those cases doesn’t just miss the point—it can make parts impossible to manufacture correctly.

 →  For manufacturing

Manufacturing is where the proverbial wheels fall off the design wagon fast when we're talking about misapplied flatness. 

Machining processes don’t produce flat surfaces automatically—they produce surfaces with patterns, distortion, and residual stress. A few realities to keep in mind:

1. Machining processes have natural limits

• Face milling can produce good flatness—but depends heavily on toolpath strategy and setup

• Grinding improves flatness, but adds time and cost

• Lapping gets you extreme flatness… and extreme pricing

The tighter the tolerance, the more likely you’re forcing an expensive process change.

2. Parts move (way more than you think)

Even if a part is machined flat in the fixture, it may not stay that way.

• Residual stress releases after machining

• Thin parts deform when unclamped

• Thermal shifts can introduce variation

And softer materials? Even harder to control, because they flex during cutting.

3. Tooling isn't perfect

Tool wear, cutter deflection, and machine accuracy all stack up.

Modern CNC machines rely on compensation to hit tolerances precisely—but that compensation has limits.

Flatness tolerances that don’t account for these realities often result in scrap.

 →  For inspection

Flatness isn’t just hard to make—it can be hard to measure.

Common inspection approaches:

• Surface plate + indicator sweep

• CMM measurement with point sampling

• High-density scanning (less common, more advanced)

CMM inspection, for example, captures multiple points to evaluate the surface condition. But here’s the catch: you’re never measuring the entire surface.

So your accuracy depends on sampling strategy, measurement setup, and constraint assumptions.

If those aren’t defined, your tolerance isn’t fully defined either.

When and how to actually use Flatness

Flatness is incredibly useful—when it’s tied to a real failure mode. Good examples include sealing surfaces (e.g., gaskets, O-rings), load distribution faces (e.g., bearing pads, clamping surfaces), thermal interfaces, and optical or mechanical contact surfaces. 

In each case, the question is simple: what breaks if this surface isn't flat? 

• If the answer is leakage → flatness makes sense

• If the answer is misalignment → you probably need orientation control

• If the answer is “nothing critical” → don’t call it out

Flatness should always be a response to a specific risk, not a general preference.

A simple framework goes a long way: 

• Start with function (i.e., what does the surface actually do?)
• Identify failure mode (e.g., leakage, misalignment, load instability, etc.)
• Choose the right control
  • Form → flatness
  • Orientation → parallelism
  • Complex geometry → profile
• Match manufacturing capability (don’t specify what your process can’t reasonably produce)
• Think through inspection
  If it can’t be measured reliably, it can’t be enforced

Final thoughts

Here’s the rule most engineers learn the hard way: tighter tolerance = disproportionately higher cost

Not because machinists don’t care—but because:

• Feeds and speeds have to slow down
• Additional setups may be required
• Secondary operations (like grinding) get introduced
• Inspection becomes more complex and time-consuming .

These factors have an exponential impact on cost, and flatness is one of the fastest ways to trigger that curve.

Flatness doesn’t locate a surface in space; it just controls its shape. That distinction matters in tolerance stacks.

You can have a flat surface in the wrong place, and at the wrong angle, and the part can still technically “pass.”

That’s how seemingly precise drawings create assembly problems.

Before adding a flatness callout, ask:

• Is there a functional reason this surface must be flat?

• Am I actually trying to control orientation instead?

• Do I know how this will be manufactured?

• Do I know how this will be inspected?

If any of those answers are unclear, the tolerance probably is, too.

And if you made it this far, here are some more resources that may help on the datum and technical drawing side of things: 

• Anatomy of an Engineering Drawing
• 10 GD&T Best Practices from our Engineering Team
• Tolerances Cheat Sheet 📄
• Technical  Drawing Checklist 📑

Was this email forwarded to you? Sign up here to receive our monthly newsletter. 

We created Tolerance Stack to help engineers who design with microns in mind and understand that manufacturability is just as critical as innovation.

Whether you're optimizing a part for 5-axis milling, selecting materials for a Class III medical device, or navigating the complexities of GD&T, Tolerance Stack delivers the insights you need to make confident, production-ready decisions.