Published 05/13/2026

3 Principles to Achieve Test Cuts Worth Trusting

A test cut — justified by specific risk, designed on a clean canvas and used to confirm rather than analyze — can expose if uncommon machining traits fit your needs.

Basics

John Miller

President, Way Of The Mill

Source: Getty Images

Reader Question: When is a test cut justified in a machine buying decision, and how do you set one up so the result means something?

You’re evaluating a machine. The brochure looks right, the references check out, the math works on paper. Somebody suggests a test cut. Now the question isn’t whether to do the test cut; it’s whether it will tell you anything you don’t already know. A test cut takes time, money and the cooperation of a vendor who has every reason to make their machine look good. So, it should do real work in the decision, or it shouldn’t happen.

There are three principles worth working through: when a test cut is justified, how to set one up and what a test cut isn’t.

When a test cut is justified

In my view, test cuts are for specific risks, not generic capability. Frankly, most machine evaluations don’t justify one. A standard aluminum part at typical tolerances on a well-cared-for showroom machine is going to come out fine and you knew that before the trip. Therefore, a test cut is justified when the work is doing something a spec sheet can’t promise.

What that looks like in practice:

Specialty material to your designs or your customers
Finish requirements tighter than a typical surface callout
3D tolerances or blending across features
Metal-removal rates at the edge of the machine’s power claims

And most importantly, program enablers. This is where the machine being able to execute this one thing could eliminate some downstream operations entirely.

Big process swings belong on this list too. Moving a process from a five-axis vertical to a live-tool lathe with bar feeder isn’t a tweak; it’s a different way of making the part. The result of failure is owning a piece of equipment that doesn’t fit the work. A test cut on a swing like that should expose risk and benefit equally. You usually know the trade-offs going in, so design the cut to expose both sides of risk and reward. Also, stay open-minded that the grand idea may end up wrong. That’s why we test.

If the cut isn’t aimed at one of these risks, it’s a demo, not a test. Demos are fine, but they are typically generic, play to a machine’s strengths and are made to deliver a point. They shouldn’t carry the same weight in a decision.

How to test cut

A test cut should be designed into a single representative vessel such as a test plaque. Geometry that summarizes what you need to learn and ideally clamped simply in a vise. Workholding stays generic on purpose. If the part comes out poor, you want to know it was the machine, not the soft jaws or the fixture stack. Not to mention, vendors are also more willing to run an easy-to-set-up test rather than someone else’s production setup, which gets you cleaner cooperation and more honest results.

What goes on the plaque depends on the question, but typical features may include 3D surfaces with tight blends, dowel hole sizes and positions, tight corners, deep pockets that exercise motion and rigidity. Pick things that expose the risks you’ve identified and want answers to. Resist the wish list. Every feature you add is another variable to interpret, and a plaque that tries to test everything ends up telling you nothing clearly.

There are a few things worth supplying yourself for the sake of consistency:

Material. Bring your own stock. The cert and the lot are yours, the comparison stays clean across vendors, and you remove a variable that vendors have every reason to optimize.

Tools and toolholders. Same logic. If you're trying to compare machines, the cutter and toolholder shouldn’t be changing between platforms. Lock those down and let the iron do the talking. Be prepared to invest. A serious test cut isn’t free. They cost stock, tools, your time, sometimes paid engineering hours on the vendor’s side. If you’re really splitting hairs between platforms, the resources are part of the price of getting a real answer.

Parameters get handled differently. Give the vendor your starting numbers and tell them so. This is how we run this, or our best current guess. Then invite them to push past it. Set left and right bounds and let them cook. The real impact in a test cut often comes from an expert showing off the equipment to win the order, leaning into the tricks that make their machine special. You’ll see things you wouldn’t have programmed yourself, and that’s the point.

What a test cut isn’t

A test cut isn’t a simple engineering evaluation for generic expansion. Cycle time, throughput, capability on paper, ROI, fit in the shop. Those are desk exercises. Engineering judgment closes most of those loops, and they should be mostly closed before you ever schedule a cut.

What the test cut exists for is the last risk that math can’t evaluate — the one where you’ve done the analysis, the spec sheet says “yes,” the references say “yes,” and you still don’t know until you see it run. That’s the gap a test cut fills, and it’s a narrower gap than people give it credit for. Use the test cut for that, and the trip is justified. Use it as a substitute for the paperwork, and you’re letting a short demo carry weight it can’t carry. The math comes first, the test cut comes last and the order matters.

Remember that spec sheets are designed to put uncommon machines on common paper. A test cut — justified by specific risk, designed on a clean canvas and used to confirm rather than analyze — can expose if the uncommon traits fit your needs. Your work, on their iron, in front of you. Get all three right and the result is data you can act on.

Do you have a machining question? Ask the expert. John Miller leans on more than a decade of industry experience to answer machining questions from MMS readers. Submit your question online at mmsonline.com/MillersEdge.