Replacement of multi-position measuring devices and gages by multi-sensor coordinate measuring technology

by Dr. R. Christoph, D. Ferger and U. Lunze

Introduction

Many manufacturing processes are overseen with the aid of gauges and multi-position measuring devices. These systems are capable of reliably solving complex measuring tasks. They do however have the disadvantage of high costs for manufacturing, upkeep and the necessary calibration.

Multi-sensor coordinate measuring technology, together with the most up-to-date software packages, has now developed into an important tool for the efficient replacement of the aforementioned methods. Modern sensing procedures for the recording of the component geometry such as mechanically touching or optical touch-free measurement together with the appropriate software for control, image processing and evaluation open up new applications to coordinate measuring technology. Examples of this are shape measurement by means of coordinate measuring devices and the measuring of size, shape and position deviations of complex component geometries that are assembled from several individual surfaces or contours. The latter becomes necessary for individual deviations on account of the growing functionality of mechanical components with increasingly tight tolerance requirements. The functional interrelations between the individual characteristics must be taken into account during the determination of the size, shape and positional deviations, so that the prescribed tolerance can be completely utilized in the inspection process. Otherwise, faulty ratings of components cannot be ruled out.

Physical gauging

Fig.1 shows the drive plate of a vehicle clutch. Components like these are manufactured in large numbers and in multiple variations. The manufacturing is carried out through stamping by means of follow-on tools and hardening, whereby malformations can occur during the cooling down.

The function of the drive plate is influenced by the diameters of the bores, the width of the windows and their positions relative to one another. For example, for each of the four small bores the permissible position deviation is dependent on their diameter. Without inhibiting the function, the position deviation can also be greater in the case of a diameter with greater permissible maximum value than for a smaller diameter (Maximum Material Principle-DIN EN ISO 2692).

Therefore for a functional test, physical gages in the form of insertion gauges were used. Fig. 2 shows the insertion gage for the drive plate example. under consideration. The pins of the insertion gauges represent size and position of the appropriate elements of the geometrically ideal mating piece.

The handling of one of these insertion gages is simple. The work piece must fit onto the insertion gauge without the use of excessive force (Fig. 3). This leads to the quality statement "fits" or "does not fit" or "go / no-go". This physical gaging does not require any additional training of the workers who carry out this test as a self-test in production. It is robust, which is especially relevant to the manufacturing processes of stamping and hardening that are considered in this example.

Regardless, in this case physical gaging has considerable disadvantages. For each series of parts to be tested the gage must be individually manufactured. The wide range of parts to be tested makes multiple gages necessary. Since the test must take place both after the stamping and hardening, the number of gages is doubled. Along with the development and manufacturing costs for the gages, further expenditures are incurred for their storage and regular monitoring including the means of testing and gages necessary for this.

The physical gages only allow for a good/bad statement. For example in the case of statistical process monitoring, reacting to events is only possible if faulty parts have already been created. In this way they are only suitable for the achievement of a zero-fault-strategy in a limited way.

The statement of a gaging test is not in the final analysis 100% free from user influences (individually varying use of force in gauging).

Multi-position measuring device

Fig. 4 shows a furniture hardware component, with very high complexity.

Components like these are punched and shaped in many processing stages. They are likewise manufactured in large quantities and in many variations. Since these parts have already gone through several metal working processes, the measurements of the tolerance chain must be tested over and over. For solving these tasks, multi-position measuring devices are often used (Fig. 5).

By means of stringing together many inductive measuring probes even the most complex measuring tasks can be solved quickly and reliably. These measuring probes are linked in a complex structure and evaluated with computer technology.

The calibration of the system is carried out with so-called "master parts" (Fig. 6), which must be calibrated appropriately. Further on in the procedure, the parts are then laid into the "measuring appliance" and practically measured in comparison against the master part and evaluated in accordance with the deviations.

This has the advantage that the result is a direct deviation value and not just the "go / no-go" statement. Additionally, the measurement values are available electronically for adjusting the manufacturing process.

The costs for a multi-position measuring device including the required means of testing can very quickly reach amounts of tens of thousands of Euros. On account of the specialized structure, they can not be put to use universally and of course each multi-position measuring unit requires a master part appropriately calibrated for the task.

Coordinate measuring machine technology

Therefore, to increase the economic viability of the testing procedures, alternatives had to be found. For this the following criteria were important:

• simultaneous evaluation of size, shape and position
• joint evaluation of all actual contours to the target contours
• automatic measuring procedure
• automatic interpretation and evaluation of the parts quality
• robustness in the workshop environment (temperature, dirt, vibration, operators)
• a measurement range for small parts with dimensions of a few millimeters up to parts with maximum lengths of 1,000 mm
• universal use for other measuring tasks
• SPC

For both applications, the requirements are best fulfilled by the use of a multi-sensor coordinate measuring machine with CNC operation, i.g.ScopeCheck (Fig. 7) or VideoCheck (Fig. 8).

The contours of the work piece are recorded by a high-resolution camera in the transmitted and reflected light modes and digitized (or analog by sensors Fig. 9). Digital filter techniques ensure the necessary robustness of the measurement so that dirt particles and other interferences do not influence the measurement result. These devices can be equipped with further sensors of coordinate measuring technology (mechanical touch and scanning probes, laser scanners, and others) and additional measuring axes (rotary and rotary tilt axes).

For CNC-programming of the measuring routine there are three possibilities (Fig.10):

• at the machine in the learning operation (teach-in) on the basis of the work piece
• at the machine in the CAD-online mode on the basis on the nominal geometry, which has been transferred from a CAD-system
• in the CAD-offline mode, at the CAD-workstation and transfer of the measuring program through the DMIS-interface (DMIS-Dimensional Measuring Interface Specification)

The measuring results can be represented graphically or numerically and output stored, and condensed for further interpretations (statistical process control).

Measuring on a multi-sensor coordinate measuring machine

Modern multi-sensor co-ordinates measuring machines cover a broad spectrum of complex tasks. The qualification of the machine operator stretches from the low-trained worker who only occasionally determines a few measurements, up to the specialist who processes quite difficult measuring tasks, making use of all the technological possibilities. The very different work practices are supported in the best possible way by the structure of the machine operation software. For example the measuring machines have several access leveles, which correspond to the varying qualification levels of the operators. The software has a modular structure, which suits ergonomic requirements by making available to each user exactly the functions and tools that they are trained for and require.

The "intelligence" of the software will then take over i.g. the selecting the parameters necessary for measurement (setting of a window), the geometric element to be measured (line, circle) or of the logic algorithms necessary for the determination of distances and angles.

Thanks to the user-friendly operating philosophy, measuring is as easy as reading a construction plan.

The machine can also be used by operators who do not know the test procedure in detail. For these users, the software makes it possible simply to select the part number and start an automatic program. Alternatively, this can be done by scanning a bar code.

By clicking on a particular feature, it is possible for an operator to test from which geometric elements the feature is constructed. Continued clicking will lead in steps to the individual measurement operation and its technology parameters (sensor system and lighting). Parallel to the feature tree, the appropriate characteristics, geometric elements, and measurement results will also be shown in the graphic representation of the measurement procedure and in the numerical measurement report. Computation operations can either be programmed in the feature tree or in the graphic view.

ToleranceFit - gages on the multi-sensor coordinate measuring machine

For a gage test of the components, a simultaneous observation of the functional scanned actual contours together with the nominal contours and tolerances is now possible. This is carried out by means of the software Tolerance-Fit (Fig. 14). Proceeding from the nominal geometry, which can be made available by the CAD-system, the interactive determination of the tolerance limits for the individual contours is carried out. The measuring of the components is carried out by scanning the contour by means of a camera or with mechanical probe system.

Further analysis is now carried out in two steps:

• Rough alignment
• Fitting into the tolerance zone (ToleranceFit)

For the rough alignment the user has two possibilities:

1. Interactively on the computer screen by dragging the actual values with the mouse or with manual input of values
2. By means of a software option, which realizes the rough alignment by fitting, in of the nominal and actual contours.

What is new here is the type of evaluation. The fitting-in is no longer only carried out on the nominal contour i.e. by the method of the smallest square or of the minimization of the largest deviation with subsequent tolerance comparison, rather it is directly fitted into the tolerance zone. Fig. 14 demonstrates this principal difference of the fitting-in strategies.

After a ToleranceFit fitting either all actual points are lying within the largest possible clearance in the tolerance zone or the unavoidable tolerance excesses are as small as possible.

In Fig.14 b) it can clearly be seen how the deviations in the zoomed area become smaller by tolerance band matching than is the case with matching to the target contour in Fig. 14 a). In the left-hand (unzoomed) view of the part, the deviations are of course made larger by the tolerance band matching but remain within the tolerance limits.

The amount of deviation from the nominal is also report as numerical values.  These values, along with their positions, can serve as a basis for statistical process control. Unlike in the case of physical gauging, this enables intervention into the manufacturing process before scrap occurs. Additional fitting strategies can also be used to control the axial displacement and rotation between the actual and nominal contours.

Filtering options to eliminate the influence of dust, debris, or minor burrs are available. There are also options for controlling the graphic output display.

 

Measurement of furniture fittings integrated into the production

The VideoCheck gantry type coordinate measuring machines are placed close to the production area and use video sensor systems, tactile 3D-probes, and laser sensors corresponding to the specific measuring tasks. Right angle mirror attachments are also, at times, placed on the optics to be able to use the advantages of rapid optical measurement technology in other planes.

The main emphasis is placed on the optical measurements since considerable savings in measuring time can be achieved. For example at the moment 70 - 80% of the required characteristics are being optically measured.  However, around 50% of the measuring time is consumed by the,tactile measurements.  In other words, 20% of the tasks are taking up 50% of the time on account of tactile probing. The measuring machines are used in a 3-shift operation. The worker loads the machine with the appropriate fixtures and workpieces and then simply starts the measuring programs. To eliminate operator error, the manufacturing order is read by barcode reader and then the appropriate partial program is automatically started. Subsequent to the measurement, the measuring data is transferred fully automatically to an SPC system, evaluated, and if necessary, corrective measures are introduced automatically to the appropriate machining centers.

Gaging of clutch parts

The coordinate measuring machine VideoCheck IP is installed close to the production line and is enclosed within a housing to avoid or lessen disruptive influences from the production environment (temperature, dirt, etc.). The worker responsible for the particular process step independently tests a component (or a small random sample) with the Video-Check in conjunction with ToleranceFit at fixed intervals.

For rough alignment, a T square on the measuring table of the device is used and the worker begins the existing measurement program (Fig. 18).

Necessary instructions and special features with reference to the setup of the are conveyed to the operator on the computer screen. Measurement and evaluation are then immediately carried out automatically. The result of the BestFit is displayed graphically and is stored on a file server for documentation of the process quality and for further analyses.

The task of operating the machine are, in this way, reduced to a minimum for the manufacturing workers and are similar to the correct handing of a physical gage. In exceptional situations (for example, recognizable outliers on the scanned contour) employees of the quality control department are called over.
While all of the physical gages for each individual part and feature had to be on hand, now the measurement is carried out with a single measuring machine. Even though the machine is located in the production area, the production workers can sometimes make long trips between the manufacturing cell and the testing area. This expenditure is however more than offset by the lower frequency of testing, which becomes possible as a result of the considerably higher information content of the measurement.

Comparison of the systems

(↑=Advantage, →=Neutral , ↓=Disadvantage)

	Multi-position measuring installation / Gage		Multi-sensor coordinate measuring machine
↑	Complexity / handling	→	Complexity of the installations constant
↓	Flexibility with regard to parts changes low	↑	Flexibility with regard to parts changes high
↓	Set-up times high	↑	Set-up times low
↓	Test cost per part high	↑	Test cost per part low- as universal
↓	Manufacture of master part and calibration	↑	Master parts eliminated
↓	Feedback expensive	↑	Feedback provided at low cost
↓	Manufacturing cost of the installation high	↑	Manufacturing cost constant (series devices)
↓	EDP compatibility low	↑	EDP compatibility high

Multi-sensor coordinates measuring machines make it possible to measure even complex parts in one measuring program since all the sensors are working in the same coordinate system.

Existing measuring programs can be easily modified to accommodate workpiece design changes or families of parts. The set-up times and testing costs per part are considerably reduced since in principle all parts can be measured on one machine and each part does not require special measuring appliances or gages. The manufacture of calibration parts and master parts is eliminated.

Regularly reoccurring costs for calibration and maintenance (remanufacturing, means of testing, administration, etc.) are also eliminated. The compliance of the measuring results to traceable standards is implicitly given.

These investment amounts very quickly justify the use of the appropriate multi-sensor machine technology wherever there is a large variety of parts.