Tackling Design and Manufacturing Issues at an Early Stage

Geometric inspection of small components with CT-Scanners and PolyWorks

A completely new way of thinking is needed for the inspection and functional analysis of small injection-molded components and assemblies. Today, non-conventional hardware and software tools are used in a unique combination, yielding better and easier-to understand results when dimensional inspection is done on small and complex geometric forms. This article describes how Bolton Works used the PolyWorks® suite to perform a virtual assembly using scanned components and CAD models, allowing specialists to tackle design and manufacturing problems at an early stage. Through a case study examining miniature valve and pump manufacturer Parker Hannifin’s Pneutronics Division, this article also explains how CT-scanners can be used to measure small features efficiently and perform diverse quality control tasks such as Geometric Dimensioning and Tolerancing.

The Challenge

With the advent of better design tools and materials, engineers seek to integrate multiple functions in injection-molded parts. By doing so, the overall cost, assembly time, and reliability of the assembly are improved. Unfortunately, “contact” measurement tools like calipers, gages, or CMMs (Coordinate Measurement Machines) have proven to be ineffective for measuring small components (< 25 mm) with feature sizes smaller than 1 mm.

Here are some of the reasons why:

  • Mechanical touch probes, due to their inertia, set limits to the speed at which they can operate and therefore limit the number of data points that are sampled.
  • Geometry created by using slides and inserts in the injection mold often creates undercuts, which cannot be reached by conventional tools in a single setup.
  • The size of a probe limits the size of features that can be measured.
  • A CMM needs to be programmed like a CNC machine tool to get the probe to the area of interest while avoiding the fixture which holds the part.
  • Gages, optical comparators, microscopes, etc. all work in 2 dimensions. These instruments work in planar sections and do not offer information about how the measured features relate to other features in 3D space, a major drawback whenever datum need to be used.

The Solution

Turning to CT-scanners

Pneutronics, a division of Parker Hannifin Corporation (Hollis, NH), engineers and manufactures miniature solenoid valves, miniature diaphragm pumps, and system solutions for critical fluidic applications. The Pneutronics valve product line includes miniature digital, proportional, and multi-media solenoid valves with orifice sizes from .003" to .250”, which are used in a range of medical technologies and analytical instrumentation.

In 2003, Pneutronics was searching for technology that would allow it to analyze a valve in a non-destructive manner. This product, XValve® is a two-position, three-way miniature pneumatic valve (24 mm x 8 mm x 9 mm) that consists of a glass-filled polymer valve housing, a stainless steel core, a stainless steel actuator, a plunger with an elastomeric seal, and a solenoid.
Finding a solution that could measure both the plastic and silicon rubber components accurately was an important requirement for deciding on the technology. Pneutronics development engineer, Drew Brenner, investigated many options before turning to Bolton Works and its CT-scanning service to generate a complete “digital” paper trail of the metrology process. 

CT-scanners (Computer Tomography) are used in the industrial and medical worlds to capture information from both the inside and the outside of an object under investigation. A CT-scanner consists of an X-ray source, a platform that holds the object, and a detector. A CT-scanner measures the attenuation of the X-ray beams passing through an object and creates a complete 3-dimensional volumetric dataset. The digitized model consists of very dense point clouds that can reach several million points.

To handle this great amount of information, Bolton Works turned to PolyWorks®, InnovMetric’s point cloud processing software solution. After several years of point cloud processing in the automotive and aeronautic industries, PolyWorks has proven to be the most robust software package capable of efficiently handling the large point clouds generated by CT-scanners.

Using CT-scan technology and PolyWorks, Pneutronics decreased product-development time, gained significant input, and streamlined their testing process.

The Inspection Process

1) Part digitizing 

In order to work with the mix of materials present in the X-Valve assembly, Bolton Works began by scanning the (blue) plastic housing separately, which resulted in a dataset consisting of 1500 images with a distance between the images of 20 μm. The total size of the image dataset produced was 1.8 gigabytes. From this stack of images, a 3D point cloud model consisting of roughly 6 million data points was created.


  • The benefits of sampling so many data points are significant, especially when compared to traditional measuring techniques.
  • With one CT-scan, Pneutronics obtained all the data needed to study the valve, eliminating time involved with additional scanning for new measurements.
  • Most importantly, the high number of data points greatly improved the precision and detail of the model, resulting in more accurate design data and a more complete product analysis.

2) Comparison to CAD

After scanning and generating the 3D model, the point cloud data-set was aligned with the 3D CAD model using a “best fit” method. The software compares every point on the point cloud to the CAD surface information, records deviations, and displays these in a color map.

3) Geometric Dimensioning and Tolerancing

The PolyWorks software allowed Pneutronics to perform geometric dimensioning and tolerancing (GD&T) and efficiently develop a mechanical drawing standard by defining a datum-based reference coordinate system for the assembly. It uses the 3D CAD model and automatically extracts primitives, such as planes, cylinders, and cones, from the point cloud. To define the datum planes for the pneumatic valve, the plane in the 3D CAD model was identified. PolyWorks then extracted the corresponding plane from the point cloud by using points surrounding this CAD feature within an 0.25 mm distance. It further excluded points of which the normals were not within a 20 degree tolerance of band of the surface normal. Through this filtering process, out of the millions of scanned data points, a correct datum was constructed. The initial best fit therefore is only used to align the points close enough for the software to find the points forming the datums and other features.
The drawing shows how the extracted datums, cylinders, planes, etc. relate to each other. The drawing specifies the tolerances with maximum/minimum material conditions. These relationships are fed into the PolyWorks inspection software. This is essential, because if the drawing specifies that a cylinder should be positioned to datum A, B, and C, the software should virtually align the point cloud accordingly. After extraction of the desired features and comparison to what had been defined on the drawing, the information is output in Excel format for documentation purposes.

4) Virtual assembly

Pneutronics wanted to enhance specific performance characteristics of the X-valve and needed an accurate GD&T report highlighting the current state of critical dimensions. Brenner decided to evaluate the interaction of the stainless steel and rubber components with the plastic housing. The valve assembly consists of components with different materials, including plastic, rubber, stainless steel, and copper wiring. All these materials have a different density, which means that different energy levels had to be used by the CT scan for each component. It was not feasible to segment out the materials for individual scanning and then try to reconstruct them into an accurate model. It was decided to employ another approach to visualize the assembly.

The cylindrical stainless steel parts were measured with conventional tools such as comparators and calipers. After confirmation that they were within tolerance, the part was virtually assembled in PolyWorks using the CAD models of the stainless cylinders and the rubber valve and then overlaying them in 3D with the scanned data of the housing.

The virtual assembly validated the design geometrically and confirmed that the valve should close as intended. As the design was now validated, the focus of the investigation went to the assembly process.

A CT scanner with a higher voltage was used to visualize only the metal components. The scans revealed that when the actuator was pressed in the housing, it could move out of alignment. A revision of the assembly process was therefore called for.

The geometric inspection and virtual assembly were instrumental in understanding the component interaction and allowing Pneutronics to concentrate on the assembly process rather than the design process to enhance the valve.

Automation and Reports

PolyWorks offers a powerful scripting language that enables specialists to automate entire inspection tasks. All tasks previously discussed in this article can be automated and executed in one click of a mouse. It is just a matter of loading the next point cloud and having the results output to an Excel spreadsheet or uploaded to a Web server in HTML format. This level of automation is especially helpful when inspecting multi-cavity tools. Bolton Works has been using PolyWorks since 2003 to create inspection reports automatically from CT-scan data.

The Benefits

As seen by the results at Pneutronics, CT scanning proved to be a cost-effective way of implementing a viable inspection process of miniature valves. The use of traditional “contact” measurement tools had proven to be difficult. The use of a CT-scanner combined with the right software tools has enabled to establish a complete digital inspection process at Pneutronics, applicable at different phases of the manufacturing process (design, prototype, production, and assembly). This inspection method proved to be very effective in validating parts through global analyses, GD&T measurements, and virtual assembly.
PolyWorks has proven to be the most appropriate tool for handling CT-scanner data for various reasons:

  • PolyWorks very efficiently handles the large datasets (2 GB) produced by CT-scanners.
  • PolyWorks offers instant global analysis through data-to-CAD comparisons.
  • PolyWorks has a unique embedded GD&T engine to verify that the parts comply to the latest (ASME Y14.5M-1994) standards.
  • PolyWorks enables users to “virtually assemble” scanned data with CAD models.
  • PolyWorks offers a powerful scripting language to automate inspection processes.

CT scanners require a substantial capital investment. By providing CT scanning as a service, Bolton Works made the technology affordable and still allows its customers to do their analysis in house, based on the point cloud and the help of the PolyWorks software solution.