Parametric Modeling Techniques for the Team Player (MCAD Modeling Column)1 May, 2007 By: IDSA ,Mike Hudspeth
Think about how your design will be used by others.
You're sitting in traffic, biding your time and waiting for that treasured moment when you'll be able to pass the jerk in front of you who, for some inexplicable reason, wants to drive at 10 miles an hour. That's when you see it—that opening in the long line of cars that's just big enough to allow you to get around him. You gun your engine and start to swerve into the other lane—as you catch a glimpse of something big and dark in your mirror. You hastily swerve back to where you were barely in time to avoid colliding with the guy in the other lane who took advantage of the same opening you saw. "What an idiot!" you exclaim. "Don't you know how to drive?" But the same quirk of human nature that convinces us that we are really good drivers does the same for those who cut us off on the highway. They think they're pretty good drivers too—and that we are the idiots.
We in the parametric 3D modeling community know the frustration of having to use models that somebody else made. The other guy never seems to model things in a way that makes sense. Nothing is built the way it should be—or maybe I should say, as we would build it. We are, after all, good modelers—better than most. And we are always concerned with the inefficiency of the models we use. That's why it bugs us when we see things done that could've been done differently. We're inconvenienced when we have to figure out how squeeze the necessary functionality out of a model. It takes time that we are ill prepared to sacrifice to fix things. Most often, we end up just getting the job done without really fixing things. We add to the chaos with our jury-rigged repairs. Woe to whoever has to change the model again in the future!
The above scenarios are basically identical. If we're ever to rise above mediocrity, we're going to have to realize that none of us is perfect. We all have plenty of room for improvement. And we need to banish the "You're not the boss of me" attitude.
Figure 1. Nowadays you can send a 3D virtual model to the machine shop, and it can run tool paths and cut the model with no other references.
We must remember when we model that there's always going to be a downstream use. Other people are going to have to use our models. To be honest, there was a time when it was perfectly acceptable—even preferable—to throw a finished (or more often semi-finished) design over the wall for downstream applications. But those days are quickly coming to an end. My first position in this industry was as a drafter. Over the years, I grew to see that my position was ever-so-slowly being phased out. I saw the handwriting on the wall. The major 3D modeling vendors were aiming at nothing short of eliminating dedicated drafters by essentially automating the process. I transitioned to 3D modeling and industrial design.
Figure 2. The Autoconstrain feature most often will merely assign dimensions so everything is tied down in 3D space. Unfortunately, software is notoriously bad at assigning meaning—that's what you are supposed to do.
Today, engineers and designers can take their models and send them to be cut directly. It's amazing. Tooling without drawings (figure 1)! I never would have imagined it, but it will come to pass eventually. Until then, however, you need to model with drafting needs in mind. I always recommend that designers constrain their sketches as they'd like to see them in a drawing. Capture the important relationships first (figures 2 and 3) and then take care of the rest. I know some 3D modelers have an Autoconstrain function, but how often does it really guess the correct relationships? My experience says rarely. Most companies still use dedicated drafters or accommodate them because they know how to show things most designers don't want to bother with. Designers are interested in the design; drafters are more interested in documenting it. Lend them a hand and make their job easier.
Figure 3. If you intend a relationship between two features, constrain them that way. Don't just concentrate on fully defining everything willy-nilly. More often than not, that approach will come back to bite you in the end.
You also must accommodate tooling. You can send a 3D model to your machine shop and those crazy wizards—and I say that with the utmost respect and admiration—can cut parts or make tools directly from the 3D geometry. But even the most mild-mannered machinist will cuss you out if you hand him or her a model that has no draft (or reliefs or whatever it is in the process you use that makes the part, well, makeable). I have seen many designers send things to the shop that won't work. I've also seen designs fall apart because the designer waited until the model was finished to include draft or tooling considerations (figure 4). If it can't be made, it won't. Always model with tooling in mind.
Figure 4. An injection-molded part draft can make or break a design, especially in a parametric modeling system. Most often the software will have trouble overlapping things or eliminating faces or edges, so add draft as early as you can.
Men and Brethren, What Can We Do?
So you have started modeling with all of these things in mind. It's not you—it's still the other guys. How do we make them listen to reason? If you talk to them and they tell you what they think of your suggestions, then you have little choice. At this point I'm going to say the "S" word. If you want to avoid massive modeling chaos you need to establish standards. (Insert sharp inhalation here.) I know, no one likes to hear that, but neither do they like to have to work on junk. If you don't have the authority to mandate modeling standards within your organization, then establish them for yourself—and follow them—always. As I've told my children, lead by example. When people see that working on your files is easier, they will be happier. And trust me, happy people listen a whole lot better than unhappy ones. Write your standards down and make copies for anyone who is interested. Yeah, it's totally unofficial, but those are the kinds of rules that people would rather follow. Everyone hates having rules forced upon them.
Of course, some organizations mandate rules (kinda sounds redundant, doesn't it?). If that is your organization, then find out the procedure for creating them. Make sure you ask as many people as you can for input so people can't accuse you of trying to ramrod your own preferences. Get buy in and, again, put everything in writing.
So Happy Together
What the big guys in modeling don't want to tell us is that history-based modelers are prone to all of the aforementioned problems. Nonhistory-based modelers really couldn't be bothered by any of it. They have no parameters, no sketches and no histories to keep straight. It's all just geometry. I'm not really saying it's a superior modeling method, just that it has certain advantages. But that's for another column. In any case, if we model with the rest of our team in mind, we will have a much happier product stream. And happier is better.
Mike Hudspeth, IDSA, is an industrial designer, artist and author based in St. Louis, Missouri.