Geometric dimensioning and tolerancing (GD&T) is the only means we have to say precisely what we want so that others can do precisely what we say. To not use GD&T is to remain beholden to our own and others’ tribal understandings in manufacturing and inspection, and to suffer the consequences which unreliable or different tribal understandings impose on new product ramp-up cycles.
However, in order for GD&T to pay its dividends, it must be turned from a hate relationship into a love affair, and this can only happen if its use is made easier, which in turn can only happen if its implementation is largely automated. Is this possible?
What is GD&T?Computer-aided design (CAD) and GD&T aided design (GAD) are the two equally important halves of the science of machine part geometry specification. CAD provides the means to generate, manipulate and communicate the nominal geometry of machine parts.
GD&T, on the other hand, is a symbolic language which serves to research, refine and encode the functions of machine parts in order to guarantee assembly and operation prior to drawing release. CAD without GAD is very bad because it represents just half the story, and the lack of any GD&T is demonstrably responsible for bumpy, new product ramp-up cycles, unnecessarily costly manufacturing processes and questionable inspection results, all based on tribal understandings.
In short, GD&T is a highly sophisticated, encodable and decodable symbolic language for managing the risks of machine part design, manufacturing, inspection and assembly, without which there is no scientific basis for these processes.
Contributions to the Bottom LineProperly implemented, GD&T makes big contributions to corporate profitability. It does so by enabling design engineers to research, refine and guarantee the operational and assembly functions of parts before drawing release, instead of in the normal manner through costly trial and error on the assembly floor.
It also contributes to profitability by encoding requirements so clearly as to permit completely unambiguous decoding for reliable manufacturing and inspection process management. Without GD&T all tolerance stack-up analysis is merely a game; all manufacturing processes must rely on tribal understandings and hope, and all inspection is pure invention on the part of the inspector.
Nowhere is GD&T more important than in the newly disintegrated world of global manufacturing, where its symbolic language makes the user independent of local languages and tribal understandings, and most importantly, once again allows us to say so precisely what we know is worth doing, that others can do precisely what we say.
Furthermore, only GD&T eliminates remaining beholden to particular suppliers, because with GD&T their success is not due to their having finally understood what the engineer wanted in spite of his failure to say so, but is based on the drawing itself, which states those needs absolutely clearly to anyone-of course only to anyone who understands GD&T.
Finally, only GD&T makes it possible for coordinate metrology to function at all, and only GD&T forms a completely reliable basis for accepting or rejecting manufactured parts, which if acceptable, also are guaranteed to be assemblable and operational.
Who Needs GD&T?The GD&T user community is broadly based, and the needs of each group are different.
Level I users are technical specialists, including machinists, assembly leads, mechanical inspectors and machine part buyers who are primarily dedicated to decoding GD&T in order to implement design objectives. In order to succeed, they must come to respect GD&T, must be familiar with its most basic concepts, tools and rules, and must have ready access to computer based GD&T decoding systems-computer-aided manufacturing (CAM) and computer-aided inspection (CAI)-and to a live GD&T advisor team.
Level II users are engineers active in design, manufacturing and metrology, whose responsibilities include encoding and decoding GD&T to specify and manage engineering objectives. Design engineers use GD&T to research, refine and encode the functions of parts in order to guarantee assemblability and operation and to communicate objectives unambiguously to manufacturing and QA. Manufacturing and QA engineers decode the GD&T and turn it into reliable manufacturing and inspection processes.
In order to succeed, level II users must be familiar and comfortable with all the fundamental concepts, tools, rules, encoding and decoding processes. But they also must recognize that they may not be in full control. As a result, they too need access to computer based GD&T encoding and decoding systems-GAD and TSUPA-and to a live GD&T advisor team to ensure success.
Level III users, or GD&T advisors, are those team members in design, manufacturing and metrology who have come to thoroughly respect GD&T and have refined their knowledge to the point of being able to support others. They are the go-to people. In order to truly benefit the organization, GD&T advisors must be supported and encouraged by corporate management, must be required to participate in all critical engineering review meetings, must be required signatories for drawing releases, and should offer occasional in-house training sessions to further refine their own and others' knowledge.
Barriers to Effective ImplementationGD&T is widely used but also widely feared, and frequently raises supplier pricing instead of reducing it. Reasons for GD&T being widely feared include:
Its symbolic language is highly complex and requires deep understanding to be functionally encoded and correctly decoded.
If applied incorrectly, GD&T increases costs by misleading suppliers and can lead to the rejection of functional parts.
Methods for Effective ImplementationCorporate implementation success is critically dependent on top management support. Top management must therefore be aware of the financial benefits of GD&T and must be committed to ensuring the benefits are achieved.
The benefits of GD&T are best achieved by:
Well-managed training at every level.
The establishment and proactive maintenance of capable GD&T advisory teams.
The acquisition and implementation of CAD, CAI and CAM systems which include capable semi-automatic, smart GD&T encoding and decoding engines, which support designers with rule based, process driven processes and explanations.
Where are these engines? With the exception of just a few players, the world is still waiting.
The acquisition and implementation of smart GD&T decoding engines in CAM and CAI, which automatically convert GD&T encoded CAD models into rule-based manufacturing and inspection processes and eliminate dependence on interpreting GD&T. Where are these engines? There are some very capable, fully automatic coordinate metrology processing software systems already on the market, but they depend on good GD&T in the first place, which is sadly a great rarity for the time being.
Implementation should be undertaken in measured steps, division by division or product line by product line, and time should be taken to assess methods and success at each step. In the long run, successful implementation depends substantially on management support, on building and maintaining an accomplished and empowered GD&T advisor team, and on the availability of smart GD&T encoding and decoding software tools.
Implementation CostsThe greatest potential out-of-pocket cost of an implementation effort is the cost of failure because failure means continuing to be saddled with not using, or misusing GD&T. It is therefore recommended to proceed carefully with a well-considered plan.
The next greatest cost is the cost of time lost on the job during training. This cost can be minimized by selecting the most effective training possible and combining it with on-the-job applications.
The smallest costs are those associated with capable training providers and with reliable as well as smart GD&T based CAD, CAM and CAI software to the extent they are or will become available.
Concluding ThoughtsHaving discovered the power of GD&T, companies are more and more inclined to use it, but its benefits are always impacted by the difficulties of implementing and maintaining it, including the pervasive idea that it is open to interpretation. In the end, only large scale rule based automation of the encoding and decoding processes can guarantee success, but if users fail to bang on the doors of software providers to request such tools, they will never appear. It’s time to start banging. Q
Tech TipsGD&T is a highly sophisticated, encodable and decodable symbolic language for managing the risks of machine part design, manufacturing, inspection and assembly.
Properly implemented, GD&T enables design engineers to research, refine and guarantee the operational and assembly functions of parts before drawing release.
Only GD&T makes it possible for coordinate metrology to function at all.
A GD&T ExampleHere is a simple example of how GD&T can be used to encode the functions of each feature of a part in an absolutely reliable way which could be largely automated in CAD.
The functional objective of the short shaft shown in the drawing is to be attached by a flathead screw to a rectangular bar in such a way as to ensure that it clears a hole in a mating plate. The assembly process in which the threads of the screw engage the threads in the shaft and the head of the screw engages the countersink in the mating bar, reliably achieves this goal in a simple self-centering operation. In order to guarantee that the shaft clears the bore in the mating plate, one will need to control its size, its perpendicularity to the mating planar surface at the threaded end and its coaxiality relative to the threaded bore. The code shown performs these functions.
The planar surface at the left hand end of the shaft is the most important feature because the forces applied to it will cause it to constrain the shaft’s two most important degrees of freedom, namely pitch and yaw. This is encoded by identifying it as datum feature A, and by controlling the only error that could arise, namely flatness.
The next most important feature is the threaded bore because it serves to constrain the next most important degrees of freedom during the assembly process, namely two translational degrees. This is encoded by identifying it as datum feature B, and by controlling everything that could go wrong, namely the size and form of the threads using a thread specification, and its perpendicularity relative to datum feature A using the perpendicularity tool with a tolerance zone size modifier (S) to encode its aiming or self-centering function.
Finally, the requirement that the surface of the shaft be perpendicular to A and coaxial with B is encoded using the position tool which is referenced first to datum feature A, to constrain pitch and yaw, and then to datum feature B to constrain two degrees of translational freedom.
Furthermore, the tolerance zone size modifier (M) following the tolerance value encodes the clearance function, whereas the tolerance zone mobility modifier (S) following datum feature label B encodes the fact that threaded datum features always produce stable datum reference frames.
These steps could be largely automated in a smart GD&T encoding engine.
AcronymsCAD, computer-aided design
CAI, computer-aided inspection
CAM, computer-aided manufacturing
GD&T, geometric dimensioning and tolerancing
GAD, GD&T aided-design
TSUPA, tolerance stack- up analysis