Don is in charge of quality assurance at an OEM that manufactures automotive components.
Don has a problem: his company recently received a lucrative contract to produce 30,000 cam shafts for a new engine and Don has just been informed that the first shipment was rejected for being out of specification.
Don’s company has been using a combination of portable measuring tools to check parts on the production line, which had worked well enough until now. Unfortunately, the cam shafts required more measurements with a higher degree of accuracy than the portable tools could provide. Line operators were struggling to keep up with the pace, so they cut corners.
Mistakes made by personnel using measurement devices are a major source of error in metrology, but coordinate measuring machines (CMMs) can reduce it significantly.
The machines can be set up to fulfill repeated measuring tasks automatically without the need to reprogram them each time. They can also obviate the need to have a variety of measuring instruments for individual jobs, since a single CMM can be programmed to perform any number of measurement tasks.
All CMMs have three orthogonal axes (X, Y and Z) operating in a 3D coordinate system. Each axis has a scale that is used to indicate the system’s position or location in space.
The machines read inputs from a sensing device programmed by an operator or using computer numerical control (CNC).
Integrating a CMM into your production process—whether in a lab or on the shop floor—can make a significant difference to the quality of your product, but choosing the right CMM for your application is no easy task.
“It’s hard to choose the right machine for yourself,” said Benjamin Viering, senior director of marketing and business strategy at Carl Zeiss Industrial Metrology. “You might end up buying something that will not fulfill your specifications, or you’ll spend way too much money for a machine that’s more precise than you need.”
Today, the price range of CMMs varies from around $30K to over $1M, depending on the manufacturer’s specifications. Similarly, the cost of various probe/sensor configurations ranges from a few thousand dollars to tens of thousands.
This is why it’s worth seeking professional advice before making a final decision, but there are still some general pieces of advice that can help determine which CMM is best suited to your application.
1. Consider What and Where You’re Measuring
There are a number of ways to categorize coordinate measuring machines, but if you’re wondering which type is best for your application, ask the following question:
Do I want to bring my parts to the machine or do I want to bring a measuring device to my parts?
Comparison of part sizes measured on CMMs. (Left) A gear for a wind turbine. (Right) A gear for a watch. (Image courtesy of Carl Zeiss
The answer to this question will depend on what you’re measuring and where you’re measuring it. If you plan on bringing the part to the machine, then you’ll want a stationary CMM. If it’s the other way around, then you’ll want a portable unit (or a stationary one and a lot of muscle). In this article, we’ll focus exclusively on stationary CMMs.
There are four basic types of stationary CMM:
- Horizontal Arm
Bridge CMMs are the most common type of coordinate measuring machine.
In a moving bridge
, the measuring head determines values on the Z-axis by moving up and down on a bridge that spans the machine’s base. The head determines values for the X-axis by moving back and forth across the bridge. Values on the Y-axis are determined by moving the entire bridge over the granite base.
Schematics of moving (left) and fixed (right) bridge CMM architecture. (Image courtesy of Sandeep Y. Bagul.)
A fixed or stationary bridge CMM determines values on the Y-axis by moving the table, rather than the bridge. The added rigidity that comes from keeping the bridge immobile reduces measuring uncertainty, but because the table needs to be mobile its maximum load is more restricted.
At one time, bridge CMMs were difficult to load because they required workpieces to be mounted precisely on their granite base plates. However, these days the CMM coordinate system can be automatically aligned to the workpiece’s coordinate system, which is based on its CAD model.
This means that modern CMMs only need to align their coordinate system with that of the workpiece in order to ensure a good starting point, significantly decreasing the time they take to load.
However, the bridge itself still limits the accessibility of these CMMs—so depending on your accessibility requirements, a cantilever or horizontal arm CMM may be the better option.
differ from the bridge-style in that the measuring head is only attached on one side of a rigid base. Hence, they are restricted to smaller measuring ranges, since the projecting part of the mobile column must be relatively short in order to maintain its rigidity.
Schematic of cantilever CMM architecture. (Image courtesy of Sandeep Y. Bagul.)
However, the column’s large support base combined with its low weight enables cantilever CMMs to move very quickly. Their high level of accuracy and low measurement uncertainty are the reasons cantilever CMMs are used predominantly for measuring gauges and master parts.
Horizontal arm CMMs are less accurate than the other types because their design makes them more susceptible to deformation. However, it also makes them useful for measuring large workpieces or those with hard-to-reach features.
The ZEISS Pro Horizontal Arm CMM. (Image courtesy of Carl Zeiss.)
In automotive applications, for example, two arms can be paired to measure both sides of a vehicle’s frame simultaneously as it moves down the assembly line. This is done by synchronizing both machines to the same (automotive) coordinate system. Because of their design, the arms are able to reach features inside the vehicle that would be inaccessible to a bridge CMM.
Even though the arms are on opposite sides of the frame, their synchronization with the vehicle’s coordinate system enables the data they collect to be related to a single coordinate system. Moreover, the flexibility of the arms allows them to reach inside the vehicle frame to measure features that would be inaccessible to a bridge-style CMM.
The DEA DELTA SLANT gantry CMM. (Image courtesy of Hexagon.)
coordinate measuring machines are structurally similar to bridge CMMs, but tend to be much larger. In the gantry-style, the bridge is mounted on raised pillars, increasing overall measuring volume. Their structural similarity to bridge-style machines gives gantry CMMs a similarly high level of accuracy. This, in combination with their large measuring volume, is why gantry CMMs are particularly prevalent in the aerospace industry, where components can be large and yet still require a high degree of accuracy.
- Cantilever CMMs tend to be the most accurate but the least flexible
- Horizontal arm CMMs tend to be the most flexible but the least accurate
- Bridge and Gantry CMMs lie between these two extremes
The Quality Lab vs. The Shop Floor
The TIGO SF is an air-free, high-performance CMM for the shop floor. (Image courtesy of Hexagon.)
The last decade or so has seen a trend of CMMs migrating out of quality control labs and onto shop floors for what is variably called in-process or in-situ measurement.
According to Zvonimir Kotnik, director of business integration at Hexagon Manufacturing Intelligence, “CMMs have been used traditionally in quality labs because there you have a constant temperature—usually 20°C (68°F). In a quality lab you can measure a part under very controlled conditions.”
Unfortunately, measurements taken in a quality lab occur at the end of the manufacturing process. If a product is out of specification, there’s no way to tell where things went wrong in the operation.
“You want your metrology to be part of the manufacturing process so that you’re providing actionable information as it happens. That way you can adjust the process in almost real time, ensuring a much more consistent process. This reduces downtime and increase the overall quality of the product,” Kotnik added.
Moreover, maintaining a quality control lab with all the necessary climate conditioning is costly—so unless your applications requires very precise measurements, it may not be worth it.
“If you want really precise measurements you still need a measurement lab, but even then the part needs to be the same temperature as the lab, so you cannot bring a part right out of production—which might still be heated—because that will affect your measurements,” said Viering.
Ultimately, deciding which stationary CMM is best for your application depends on what you’re measuring and where you’re measuring it. See the table below for a summary of the above information.
Best for Measuring
Medium Components Requiring High Accuracy
Small Components Requiring Highest Accuracy
Large Components Requiring Low Accuracy
Large Components Requiring High Accuracy
2. Choose the Right Probe for the Job
A coordinate measuring machine is only as good as its probe(s).
Like CMMs, probes come in several types. The basic distinction is between contact probes, which measure workpieces by actually touching them, and non-contact probes, which employ lasers or machine vision. The former are more accurate, but the latter are faster to use.
There are also multi-sensor probes, which combine touch and optical scanning.
CMM Contact Probes
The most common contact probes fall into one of two categories:
- Touch Trigger Probes
- Analog Scanning Probes
Touch Trigger Probes
Cross-sectional view of a touch trigger probe. (Image courtesy of Heidenhain.)
consist of a stylus attached to a bearing plate connected to pressure sensors inside the probe housing. They generate an electrical signal each time they contact a point on the workpiece.
The first touch trigger probe was invented by Renishaw founder Sir David McMurty to solve a specific inspection requirement for the Olympus engines used on the Concorde.
The probe head is mounted at the end of one of a CMM’s moving axes. It can be rotated manually or automatically and accommodates a variety of stylus tips and attachments. Touch probes are versatile and flexible.
The incorporation of piezo-based sensors eliminated the effect of stylus bending and advancements in strain gauge technology ensured that the probes trigger at a constant force regardless of contact angle with the workpiece. This eliminates directional sensitivity, which gives these probes a submicron level of accuracy.
Analog Scanning Probes are also stylus-based and are used to measure contoured surfaces, such as sheet metal assemblies. Rather than touching individual points, the probe remains in contact with the workpiece as it is dragged across it, yielding analog measurements.
This offers dramatically increased levels of data acquisition. Continuous analog scanning (CAS) probes are based on continuous (rather than point-to-point) data acquisition. They are particularly useful for complex contoured shapes such as crankshafts, cams, turbine engine blades, prosthetics and automobile bodies.
Renishaw's SP600 analog scanning probe. (Image courtesy of Renishaw.)
There are two types of CAS systems:
- Closed-Loop Systems: the probe automatically detects changes in the surface direction of the workpiece and adjusts itself to maintain contact
- Useful when digitizing unknown complex shapes
- Open-Loop Systems: the probe is driven along a path using dimensional information from a data file
- Useful for high-speed data-gathering on parts with geometry that is well defined by surface points and vectors or by CAD data
One advantage of analog scanning probes over touch trigger probes is that the former acquires 10-50 times more data than the latter in the same amount of time. More data means more confidence, which may be needed if there are large gaps between data points using point-to-point probing techniques.
A second advantage that comes with using an analog scanning probe is that it can also be used as a touch trigger probe, giving users greater flexibility. Operators can pick and choose which features to touch quickly and which to devote more time to if a particular feature is critical, for example.
CMM Non-Contact Probes
Non-contact probes are best for parts that are either more complex, smaller, high-precision or easily deformable. They are either laser-based or vision-based.
Nikon's InSight L100 CMM laser scanner. (Image courtesy of Nikon.)
work like touch trigger probes but use a concentrated beam of light instead of a stylus. The beam acts as an optical switch so that when it is projected onto the part, the position can then be read by triangulation through a lens in the probe receptor.
This is similar to the technique used by surveyors to find a position or location with bearings from two fixed points that stand a known distance apart.
are particularly useful for very small parts, like microprocessors. Rather than measuring the parts themselves, an archetype is electronically digitized to generate accurate dimensions for future workpieces.
Renishaw's REVO vision measurement probe. (Image courtesy of Renishaw.)
A high-definition camera then generates numerous measurement points in one frame, allowing features to be measured in comparison with the electronic model by counting pixels. Unlike other probes, which require recalibration, a vision system’s lens only needs to be calibrated once.
The main advantage of non-contact probes is that they enable users to collect data from a larger surface area in a shorter amount of time compared to contact probes. However, the trade-off is that they are less accurate.
3. Consider Your CMM Software Needs
A CMM’s suitability for an application depends on more than just its structure and its probe; it also depends on software.
Although there is a standard CMM programming language called the dimensional measurement interface standard (DMIS), it’s not used by every CMM manufacturer, though most at least support it.
DMIS doesn’t contain all the functionality required for all measurement tasks. Its limitations have led manufacturers to create their own unique flavors of DMIS for CMM programming and execution.
CMM Software and Measurement Uncertainty
Measurement uncertainty means that it’s impossible to know a part’s true value. We can only get close to the true value and this is why measurement uncertainty is so important in metrology.
There are several ways to determine measurement uncertainty for CMMs. It can be determined experimentally by taking many measurements on a given machine, theoretically by modelling the machine or statistically via, for example, a Monte Carlo simulation.
Determining measurement uncertainty in a CMM using Monte Carlo simulation. (Image courtesy of Zeiss
The more accuracy required by a process, the more measurement uncertainty matters. This isn’t just a matter of the size of your parts, but also the size of your part features. Engine cylinders and injection nozzles can be quite large, especially in aerospace applications, but they also tend to have important features that require a high degree of accuracy.
As Kotnik puts it, “It’s not just how the data’s collected or how fast it’s collected, but how confident you are in the data you’re getting. For Hexagon, it’s all about speed plus confidence. If you don’t have confidence in your measuring system, you can’t have confidence in your products.”
CMM CAD/CAM Compatibility
The last 15 years have seen a significant push toward model-based definition (MBD), sometimes called digital product definition (DPD). More and more manufacturers are trading master parts for master CAD models with embedded tolerancing, which makes compatibility between CMM and CAD software a crucial issue.
Renishaw's MODUS 2 intelligent CAD data extraction. (Image courtesy of Renishaw
Kotnik puts it this way: “As an engineer, I want to make sure that each operation using my model is using the same consistent data, otherwise, the saying is: random procedures create random jobs.”
“CMM software has to be able to use the native CAD model information, which means if someone is designing a part in Unigraphics vs Pro/ENGINEER vs SOLIDWORKS, the CMM software needs to be able to read that data,” he concluded.
If CAD compatibility is a major concern, look for CMMs with a direct CAD interface (DCI), which gives operators the ability to use CAD data without translating it into another programming language like DMIS. Not all CMMs have DCI; it depends on software.
Choosing the Right CMM for Your Application
In the end, there isn’t one single factor that will determine which CMM is best for your application.
Many factors need to be considered, including CMM type, probe type, software and of course cost.
However, on that last point, Kotnik offers the following piece of advice:
“Do not buy on price alone. By the time you get a low-cost system up and running you’ve spent as much as you’ve saved (or maybe more) to get it to work. Price is important, but there are many other factors that determine the best product for the job.”
By now, you should have a better idea of what those factors are, but to reiterate the starting point of this article: it’s always worth seeking professional advice before purchasing a CMM.
Follow Ian Wright on Twitter