3D modelling: time to sink or swim?

3D solid modelling is a good thing: we have known that for years. If you contact any of the solid modelling software vendors they will give you truck loads of case studies extolling its virtues and lauding its benefits. Here are a few to convert you to what, for many, is a ‘no brainer’. Mike Gascoyne, when he was chief designer at Jordan Grand Prix, said: “Previously, you would call a meeting to discuss say, the front suspension and you would spend the first hour of the meeting over the layout with someone telling you what they’d done. Now with 3D you walk around the design office and look over someone’s shoulder and there it is – you’re looking at it. And you can see immediately what he’s done.” Other critical areas for 3D are packaging (making everything fit in the smallest possible space), detailed design and avoiding mistakes. “It’s not that noticeable at the component level,” says Gascoyne. “It’s when you are working in large assemblies that you see ‘the lights come on’ for everyone.” Chris Tarr, senior design draftsman at Delkor, likewise says: “The speed at which changes can be made and 2D detail drawing views can be created is a major benefit to us in our design projects. With our 2D system it was too difficult and just took too long, so we restricted them to the absolute minimum, which could have a detrimental affect on the manufacturing process. “Accuracy and speed are the two main benefits [of 3D]. It considerably accelerates the design iteration process, so we have more time to explore alternatives. The 3D visualisation facilities help us to see more easily where changes are needed. We can ensure design integrity through structural and finite element analysis on parts by exporting models to third party analysis software. And we can be sure that the BoM (bill of materials) is accurate because it is calculated automatically from the 3D model. Finally, because it allows us to adopt a top down approach to modelling, the final model will be correct... all detail drawings are correct, which reduces errors down the line.” Then again, John Greenaway, NEC’s mechanical group leader, says NEC noticed time to toolmaking was considerably reduced on one of its products because of 3D, and that, most significantly, there were virtually no detailed queries from the toolmaker while making the tools. This represented a reduction of at least a factor of 20 over non-solid modelling. “The power of the model makes it all worthwhile,” says Greenaway. “If our solid modeller disappeared tomorrow we would be back to design cycles twice as long; we would lose our links to toolmaking and we’d effectively take a two or three year step backwards.” There you have it from the horses’ mouths. Measurable and tangible gains can be brought about by using high performance solid modelling software. Today, speed is crucial: the faster you can make the design process the sooner you can commence manufacture. Solid modelling enables you to freeze your designs sooner, bring your product to market faster, and earn profits sooner. And it provides benefits in other areas. Used effectively, it provides a flexible tool to help companies respond more quickly to their customers’ needs, improve their design and manufacturing productivity, and deliver higher quality products. Why isn’t everyone using it? There are lots of reasons – the one quoted most often is that solid modellers are not good at producing fully detailed drawings. They are right, but “cannot detail like AutoCAD”. Says David Parry, managing director of Nova Design: “At the end of the day we don’t make solid models, we make drawings. The model is only a means to an end.” Other objections include the price of training, the perception that solid modelling is difficult and that you need expensive operators. But CAD has always been partisan, fuelled by vendors’ marketing. “In reality a mixed design environment has been with us for years,” says Ian Dampney, managing director of Random Design. “Good engineering design needs a mixture of both 2D and 3D disciplines in order to function to its full potential.” And there is merit in the argument, ‘Just because contemporary hardware can cope and there are affordable, new solid modelling systems available, if 2D systems are serving you well and there are no compelling reasons to change, why bother?’ As Dampney says: “Unless you can assure people that the leap to 3D isn’t a binary state, very few will give up their 2D and take a leap in the dark. Most small companies are not prepared to take big capital risks, or for that matter, big technology risks.” Time to think again? Whether or not you go to 3D also depends largely on where you sit in your supply chain. There is no point in going 3D if the rest of your process to manufacture is not moving that way. Today, automotive OEMs insist on sending native model files to toolmakers, which generate drawings from the model file and cut from these. What’s wrong with that? The toolmaker keeps the business for the price of the software. True, he’s missing out on 70% of the functionality of the code, but to take full advantage of the software he would need to significantly change his process, which has cost implications. It’s not realistic to compare the validity of a 2D or a 3D definition of an object; these are just isolated elements in a total process. The only thing you can do with 2D is look at it – there is very little to pass on to the next step – but you have to be seeking gains elsewhere in the system, whether it’s checking fits or checking assemblies, before you commit to 3D. If you don’t look at the big picture there’s never been any sense in going to 3D. Fact: 3D only makes sense if the design/development stages and the software are linked from concept to machined prototype or finished product. The main challenge is thinking about your process to decide where and whether 2D or 3D is relevant. It’s less a question of how to turn your drafters into modellers, than how do you get your people to think about the whole process and choose the right discipline for each part. In 3D you have to plan your model construction looking much further forward, because it’s more difficult to change, even with current technology. But you handle this by assuming that what you’re doing now is a ‘good guess’. If you want to modify a 3D model, the simpler the geometry and 3D constraints, the quicker and easier it is to control the changes. Consequently, it’s more likely that you are prepared to change it, rather than start again from scratch or compromise. The efficiency you get from having a good forward view does ultimately produce a better product. This is more pronounced when your design environment is centred around ‘one-offs’. It becomes a little easier if you are doing family of parts work, or variations on what you did before. You have to make use of history trees: in some instances you plan the build of the history tree because you are looking forward to what you might want to treat as variables. Modern software is rich in tools that allow you to reach ‘constraint gridlock’ much too easily and much too quickly. Keep it simple and constrain as little as you have to. If you give ultimate rein to creative people, you get history trees that turn into forests – 3D has to be approached with discipline. Too many designers just use their creative side – the “I’m sitting in front of it, so I’ll fix it!” mentality. These kinds of designers should be encouraged to use a little more ‘left brain’, the logical, analytical, rational side. In a recent example requiring a model rebuild, a component history tree was reduced by 80%.