Using a coordinate measuring machine (CMM) involves securing your part on a granite table, calibrating the probe against a reference sphere, then running a measurement program that moves the probe across the workpiece to capture precise dimensional data. The process has a learning curve, but the basic workflow is consistent across most bridge-style CMMs regardless of manufacturer.
How a CMM Is Built
A CMM has three core components: a physical structure that moves along three axes, a probing system, and a computer with measurement software. The most common design is the bridge type, where a gantry with two legs rides along a heavy granite table. One leg follows a guide rail on the side of the table while the opposite leg glides along the granite surface itself. This bridge movement creates the X and Y axes. A vertical quill or spindle moves up and down through the center of the bridge carriage, creating the Z axis. The probe sits at the bottom of that quill.
The granite table serves as both the foundation and the reference surface. Granite is used because it’s extremely stable, resists thermal expansion, and can be ground to a very flat finish. Most CMMs use air bearings between the moving components and the granite, meaning the bridge floats on a thin cushion of compressed air for nearly frictionless motion.
Choosing the Right Probe
CMM probes fall into two categories: contact and non-contact. Your choice depends on what you’re measuring and how fast you need results.
- Touch-trigger probes take a single measurement each time the ball tip contacts the part surface. They’re the most common type and deliver the highest accuracy. You’ll use these for discrete features like hole diameters, distances between surfaces, and point-to-point dimensions.
- Scanning probes also make physical contact, but instead of touching and retracting, they drag the ball tip continuously across the surface, recording a stream of data points. This makes them ideal for inspecting complex contours, checking overall form, or verifying tight-tolerance parts where gaps between discrete points could miss defects.
- Optical and laser probes never touch the workpiece. Optical probes use cameras, while laser line probes use a beam to capture coordinates through triangulation. Both collect data quickly and won’t risk damaging delicate surfaces. The trade-off is reduced accuracy compared to contact methods, along with higher equipment cost.
For most shop-floor inspection work, a touch-trigger probe with interchangeable styli handles the majority of tasks. Scanning probes earn their place when you’re checking freeform surfaces or running high-volume production parts where throughput matters.
Calibrating the Probe
Every stylus tip needs to be qualified before you can trust your measurements. This step determines the effective radius of the ball tip, the length and orientation of the stylus, and any bending errors in the probing system. You do this by measuring a qualification sphere, a highly precise reference ball mounted on the CMM table with a known, calibrated diameter.
The process works like this: you enter the calibrated diameter of the qualification sphere into the CMM software, then the machine takes several contact points around the sphere’s surface, typically along cardinal directions (top, sides, front, back). The software compares its calculated sphere size against the known diameter you entered. The difference tells the software exactly how large the stylus tip really is and where its center sits relative to the machine’s coordinate system.
You need to recalibrate whenever you change a stylus, rotate the probe to a new angle, or at the start of a shift if your shop’s quality procedures require it. Skipping this step is the fastest way to introduce measurement error that won’t show up until parts fail downstream inspection.
Fixturing the Part
Before measuring anything, your workpiece needs to be stable on the granite table. Any movement during the measurement routine, even a tiny shift, corrupts your data. CMM fixtures range from simple clamps and V-blocks to modular plate systems with magnetic or threaded mounting points.
Start by thinking about which features the probe needs to access. The fixture has to hold the part firmly without blocking the surfaces you’re measuring. If you’re inspecting a cylindrical part, a V-block cradle keeps it stable while leaving the diameter exposed. For sheet metal or complex castings, modular fixture kits let you build custom support arrangements using standoffs, clamps, and locating pins.
Align the part consistently. If you’re running the same inspection program on multiple parts, each one should sit in the same orientation and position on the table. Consistent fixturing means your measurement routine doesn’t need manual adjustment between parts, which saves time and eliminates a source of operator error.
Setting Up the Measurement Program
CMM software is where you define what to measure and how the probe should move. Most modern packages let you import a CAD model of your part directly. Once the model is loaded, you name each feature you want to inspect: a bore diameter, a slot width, the flatness of a surface, the distance between two holes.
For each feature, you set the target coordinates, the measurement type (length, diameter, radius, angle, and so on), and the number, order, and location of touch points. A circle measurement might require a minimum of three probe hits, but you’ll get better accuracy with more. A flatness check might need a grid of points across the surface.
Many software platforms use a graphical, object-based interface rather than requiring you to write code. You click on a feature in the CAD model, select the measurement type from a menu, and the software generates the probe path. For more complex routines, experienced programmers can edit the underlying code to add conditional logic, loops, or custom reporting.
If you don’t have a CAD model, you can teach the program manually. This means jogging the CMM to each feature using a joystick or handbox, taking preliminary probe hits, and recording those moves as steps in a program. The software saves the sequence so you can replay it on subsequent parts.
Running the Measurement
With calibration done, the part fixtured, and the program loaded, you’re ready to run. Most programs start with an alignment routine where the probe touches a few datum features on the part, establishing where the part actually sits relative to the machine’s coordinate system. This alignment links the physical part to the CAD model or nominal dimensions in your program.
A typical alignment uses a plane (three or more points on a flat surface), a line (two points along an edge or axis), and a single origin point. This locks all six degrees of freedom so the software knows exactly how the part is oriented. If your alignment is sloppy, every measurement that follows inherits that error.
Once aligned, the program runs automatically. The bridge moves, the quill drops, the probe contacts the part at each programmed location, and the software records the coordinates. On a simple prismatic part, a full inspection might take a few minutes. Complex aerospace or medical components with dozens of features can take considerably longer.
Watch the first run carefully. Make sure the probe isn’t colliding with fixtures, approaching features from awkward angles, or missing the part entirely due to alignment issues. Most software lets you set safe travel planes and clearance heights to prevent crashes, but these need to be verified the first time through.
Reading the Results
After the routine completes, the software generates a report comparing measured values against your nominal dimensions and tolerances. Each feature shows the actual measurement, the deviation from nominal, and a pass/fail flag based on your tolerance window. Graphical reports can overlay measured data on the CAD model, color-coding features green for in-tolerance and red for out-of-tolerance, making it easy to spot problems at a glance.
Pay attention to the deviation values, not just pass/fail. A dimension that passes but consistently sits near the edge of its tolerance band may signal tool wear or process drift that will eventually produce rejects. Tracking these trends over multiple parts turns your CMM from a go/no-go gauge into a process monitoring tool.
Keeping the Machine Accurate
A CMM is only as good as its maintenance. The granite table surface needs regular cleaning to keep the air bearings functioning properly. Most shops use 90% or stronger isopropyl alcohol on both the granite and the guide rails. Avoid cleaners containing surfactants, as they leave a slippery film that builds up over time and interferes with the air bearing function. If you need to remove grease or sticky residue, clean it off with a degreaser first, then follow up with isopropyl alcohol to remove any residue the degreaser leaves behind.
Beyond cleaning, keep the CMM environment stable. Temperature swings cause both the machine and the workpiece to expand or contract, introducing measurement error. Most CMM rooms are climate-controlled to 20°C (68°F). Keep the door closed, avoid placing the machine near heat sources, and let parts acclimate to room temperature before measuring them. A warm part straight off a milling machine will measure differently than the same part at room temperature.