In CMM inspection, great efforts are expended to maintain accuracy. The equipment is calibrated regularly, and the environment carefully controlled for consistent temperature and humidity. Measured points locations are meticulously planned, and analysis algorithms carefully selected for follow the rules of GD&T. Yet there is an issue that is often overlooked by designers that can defeat these efforts and compromise the relevance of the measurements – part flexibility and restraint. The result can be a lot of effort to get accurate measurements that are questioned because the part was measured in an unrealistic or irrelevant condition.
The Y14.5M-1994 and Y14.5-2009 dimensioning and tolerancing standards provide tools to specify the restraint condition that tolerances apply in. However, most designers do not make use of these tools nearly as much as is required. The default condition is that all dimensions and tolerances apply in a “free state” condition. This means that only the force of gravity is applied - the inspector is not allowed to apply restraining forces to deform the part for inspection.
For parts that are subject to free-state variation that is deemed to be excessive, the designer can treat the part as “nonrigid” and specify a restraint condition. This can be in the form of a note directly on the drawing, or a reference to a separate document. The restraint can be described in as much detail as required. Restraint notes are sometimes as simple as a statement such as “the part is clamped to a surface plate on datum feature A”. In other cases, the exact configuration of clamping locations, force or torque settings, and sequence of force application is specified. The objective is to enable a repeatable setup that mimics the restraint that the part will experience when assembled. In the absence of such a note, the inspector is not allowed to apply clamping forces. So it is important that designers specify restraint conditions when applicable – this technique is still underutilized. Otherwise, the inspector can be placed in an awkward situation in which the specification does not allow additional restraint, but the part has very little chance of conforming to the tolerances without it.
Rigid Or Non-Rigid?
With thin, floppy parts, it is obvious that the part has significant free state variation and must be treated as such. In other cases, however, the non-rigidity is more subtle. Machined castings are one example – these appear to be extremely rigid but may deform (bend and twist) slightly when released from the machining fixture. If inspected in the free state, the part may not conform to the tight form tolerances that are often specified on the machined features (and conformed when the part was in the fixture). So a restraint note is required in these situations as well, so that the part can be inspected while subject to the same loads that will be applied when it is bolted to the mating part.
Some companies create their own internal standards documents that go beyond or replace the requirements in Y14.5. One method is to allow a certain amount of force over a defined area or length. For example, a common specification is to allow a 5 pound force for each linear foot. While the definition itself is simple, this type of specification can be difficult to interpret and apply consistently. The distribution of the force is not well defined, especially on parts with irregular shapes.
The issue of part restraint also arises when datum targets are specified. These are often specified on thin, flexible parts. The usual intent is that an accurate fixture will be constructed with hardware such as tooling balls, blocks, and pins to act as datum target simulators. The part is intended to be restrained and deformed, to make contact with all of the simulators. Large numbers (more than a 3-2-1 configuration) of datum targets can be specified, that would overconstrain the system if the part was considered to be rigid.
When coordinate metrology equipment capable of measuring large parts became more popular, such as portable CMM’s and laser trackers, part of the selling point was that expensive fixtures were no longer necessary. Even complex datum target specifications could be inspected easily by directly probing the part surfaces, because the software could “best fit” the coordinate system alignment. It’s true that the software could easily create an alignment – it just wasn’t the same alignment that would have been created with a fixture. Worse still, the lack of an accurate holding fixture meant that the part was being inspected in a different stress state and shape. Inspectors were expected to use the 3D metrology system to reproduce the results obtained with a fixture, without being able to restrain the part in the same way. This led to disappointing results, and questioning of the accuracy of the metrology system or the competence of the programmer.
When a restraint condition is specified, this often makes the inspection more inconvenient or difficult. In addition to requiring special hardware, the requirement to restrain certain surfaces often limits access to those surfaces and others. It would be much more convenient if the part could be inspected in the free state, and then have the data compensated to reflect the proper restrained condition. This would be similar to the way that compensation can be made when measurements are taken at a non-standard temperature (although the mathematics would be much more complicated). Perhaps the point cloud gathered by the CMM or other 3D metrology device could be processed to “unbend” and “untwist” it, to simulate the deformation that would occur if the part was physically restrained. This would be a very useful technique, if it could be shown that the result was realistic and not just a manipulation to make the results appear to be better.
It is critically important that dimensional inspection be performed under conditions that are the same as, or at least similar to, the conditions that the part will experience when assembled. If this is not the case, then the relevance of the inspection results will suffer accordingly. Restraint condition is a significant issue with CMM inspection, as the CMM is capable of making high-accuracy measurements even on large parts. If the part is not restrained appropriately, the resulting errors can be quite large and significantly affect the validity of the measurements. So it is important to follow any restraint notes specified on the drawing, and to advocate for them to be specified where needed.