"Drawing on Experience," Professionally Speaking, Nov. 2002

PROFESSIONALLY SPEAKING

drawing on experience

CAD rendering is pretty, but the engineer needs to be in the picture.

by Robert O. Woods

The boundary between engineering and the preparation of CAD drawings and solid models is fenced with serious barriers to communication.

The engineer is caught in the middle between a manager— who probably doesn't have detailed knowledge of the computer-aided design process but is expected to judge the product—and the CAD practitioner—whose specialty is sufficiently arcane that the engineer is pressed to leave details to that expert's judgment.

The engineer as a professional and the quality of the job both suffer. This mismatch can be treated only by frequent and in-depth conversations among the three parties: the engineer, the CAD jockey, and the manager or client.

In the Mechanical Engineering Design supplement in March, Jean Thilmany explored the idea that the advent of computer-aided design has served to blur the distinction between engineering and what used to be drafting. That was a very perceptive article.

Given the present state of software development, it takes a substantial investment in time and training for anyone to become proficient with a solids modeling system. This gives the operator every right to regard himself as a professional, but it doesn't make him an engineer.

During earlier days, any engineer could take a pencil and annotate a drawing, thus intruding into the province of the draftsman. It is not possible to do anything similar with solids modeling. Unless he has invested time in studying it, an engineer will not be very clear as to just what is involved in creating a solid model.

But engineering judgment must go into model preparation. The pinch comes when it is necessary to exercise judgment in establishing the boundary conditions to put on a model. In some cases, the distinctions are subtle enough that whoever is setting up the model doesn't even realize he is making assumptions.

Early mechanical perspective aided drawing: This woodcut by Albrecht Drüer depicts an artist who is in the process of drawing a lute.

Photo courtesy of the Metropolitan Museum of Art.

Unless the engineer is involved at every step, serious mistakes can be made because of honest ignorance. In one case in my experience, where it was necessary to judge how much of a cross-section was actually being developed when computing section properties, perfectly plausible but naive assumptions led to wildly erroneous conclusions. The moment of inertia calculation was thrown off by a factor of more than four because the CAD expert assumed that a bolted flange would provide enough shear in the plane of separation to cause two components to act as a monolithic beam. This looked perfectly good on the model.

In another case, the drafter assumed that a pair of components that were to be match reamed and pinned could be disassembled and a part replaced by simply bolting it on. Neither of these cases would look bad to anyone without engineering experience and a verbal explanation of what the drafter had done.

We can hope that solids modeling will become accessible enough that engineers will no longer have to depend on
other parties to model their designs. Certainly, two-dimensional CAD systems have become sufficiently user-friendly that any engineer willing to invest a little time can work with them. Solids modeling is moving in that direction.

Intrusion of technology has happened before. At one time, a typewriter and trained typist were expensive enough that it was necessary to keep them occupied by establishing a typing pool. Now everyone can do his own typing. When solids modeling technology becomes accessible enough, it will be unnecessary to have people who do nothing but create models.

Until we reach this stage, a real communication problem exists, which can be solved only by long and frequent discussions aimed at establishing some level of mutual understanding among the engineer, the CAD jockey, and whoever judges the finished product.

Robert O. Woods is an ASME Fellow who retired from Sandia National Laboratories, where he worked primarily in flight instrumentation design.