-
The Case for Parametric
Most of my friends know that I’m a parametric evangelist, but it can be really difficult to explain why over the course of a conversation. I’ve used Rhino, Alias, SolidWorks, and CATIA V5 professionally, and once you learn to work with a truly robust solution like V5 or NX, it’s painful to go back. So my goal with this and subsequent posts on the topic will be to explain myself. Read on…
“I can work faster in Rhino” is a phrase I hear a lot. I don’t buy it. I’m personally a CATIA V5 or SolidWorks guy, but even if you opt for UG (NX) or ProE, the fundamental reality is that for truly efficient product design and engineering it’s a cinch that the proper use of a good parametric modeling solution will save you time and money. Yes, even with the seemingly-high prices of packages like V5 or NX, companies are learning that ultimately they’re a bargain when put alongside the downstream efficiencies gained.
High level reasoning:
1. Price
2. Speed / Efficiency
3. Control / Consistency“Parametric Modeling” is a type of CAD that uses dimensions, constraints, and functions to derive a 3D representation of a part. SolidWorks and Pro-Engineer are currently the most popular parametric modeling software packages in product manufacturing.
“Parametric” should not be confused with “historical” or “associative” modeling. While Parametric modeling is by definition both historical and associative, there are plenty of history-enabled modeling packages that by my definition do not qualify as parametric. Alias Studio Tools is an example. While Alias does have a very robust tool set for designers, and even many dimensionally-driven tools, the user can not easily create associations to drive every aspect of the finished design. This is a broad statement, and one not easily explained to the uninitiated, but I’ll have to save that one for a future write-up.
1. Price:
I’ve heard a lot of people say “Rhino is the best for us financially, if only because it’s so cheap!” It’s true that when compared with the up-front cost of CATIA V5 or Unigraphics, Rhino is practically free. But you really do get what you pay for.
Lets say that Carl costs the company $70,000 a year after overhead, benefits, et cetera. What if there were a magic tool that could make Carl work twice as fast? What would that be worth?
Well, considering that the alternative is to pay another Carl, it’s worth about $70,000 per year. I’ll make the case for speed later, but for now lets put it this way: if V5 can make Carl twice as fast as a guy using Rhino, then V5 is worth $60,000 per year. I can argue (and will later) that V5 can make a tight organization like B&D five or ten times more efficient overall than one trying to scrape by with Rhino.
2. Speed / Efficiency:
“Speed” is not to be confused with “efficiency.” Speed should also not be confused with “lurch.” I would argue that Rhino is an incredible interface for “lurch” modeling, in that you can build models very, very quickly, but when it comes time to make changes or add engineering features, you lose much more time than you’ve gained in the beginning. I know many designers who use Rhino for “sketch” models, and then hop into Solid Works or V5 to do the final surfacing. This is not a terrible approach, but it’s still redundant.
Ideation and Iteration > Design > Refinement / Engineering > Tooling Design
The Rhino mentality thinks only about the front half of the design cycle. Designers think that because they are able to do their job faster, they are making the process faster. This is a fundamentally flawed way of thinking. Just because the Ideation, Iteration, and Design phases happen very quickly doesn’t mean that the project will be proportionally shorter. In many cases quite the opposite occurs.
Instead of thinking of product design as something separate from engineering, the most effective way to shave time from the development cycle is to combine the two processes into one. By working collaboratively with Engineers, back-and-forth with complete mutual freedom to make changes and adjustments as necessary, the design cycle can be cut down dramatically. Using the same CAD system makes this process much, much simpler. And if you use your CAD system correctly, you can make it easy for engineers to make ID updates and vise-verse.
3. Control / Consistency:
With Rhino, it’s a given that your toolmaker is going to re-build your data before cutting steel. This is because as adjustments need to be made during the tooling process, he’ll need to be able to make said adjustments easily and accurately. Why trust an engineer to accurately recreate your design? Why not just build it the way that you want it, and then let them make exactly what you built? When designers can reign over their own surface models, they have better control over the final result, and can achieve greater levels of design consistency between products in a range.
Next time I’ll touch on why I choose V5 over other packages.
01/30/2007
Posted in Articles.
January 17th, 2008 at 9:13 pm
I couldn’t agree more about your comments on Rhino. Rhino was the first 3D CAD software that I learned at CCAD and fell in love with it immediately. It allowed me to build beautiful models, which could never be made. About two years ago, when entering the “real world” of design, I quickly realized that this was not an option. Breaking-up with Rhino wasn’t easy but I just couldn’t see the relationship going anywhere. Currently, Solidworks is the only 3D CAD software that I use. Its capability to go back and make revisions is absolutely essential.
August 12th, 2008 at 4:32 am
I found your site on technorati and read a few of your other posts. Keep up the good work. I just added your RSS feed to my Google News Reader. Looking forward to reading more from you down the road!
September 13th, 2008 at 4:51 am
Hi Adam
Your analysis of the cost/benefit situation is correct and your explaination of this is spot-on. One thing I don’t quite agree with is “Parametric modeling is by definition both historical and associative”.
Parametric modeling doesn’t imply history, it only means that you can parametrically adjust the model. I think a ore meaningful distinction is between explicit and implicit associativity.
I am a Rhino fan, I admit. I have done some projects with rhino as cad tool, and was able to produce boat hulls and decks directly from the Rhino files without problems. Adjustments did take some time, but I can’t really see how doing It in Catia would make a huge difference. The modeling principles and surface structure would be nearly identical. Only in Catia, i would have to go back in the history tree once in a while.
Using Rhino save-as and undo and Ctrl+C has until now worked kinda like “poor-mans history. Things are changing now as the Grashopper plugin out.
I have used Catia and Solidworks earlier, and I think they are both impressive. For me, Solidworks just doesn’t do surface modeling good enough. It too often ends up with ugly patches, wrinkles, and tiny holes in the model.
Catia seems to me basically like a “solidworks with enhanced surfacing“. I mean the interface is nearly identical (only more stuff in it), but the surface tools of Catia seem ok.
I don’t know what’s the best option. What do you think would be most sensible for a small Industrial design company to invest in: Solidworks+Rhino or Catia?
Regards
September 13th, 2008 at 6:57 am
Hi Fredrik:
You make a good point, particularly in light of newer “direct-modeling” style applications such as SpaceClaim and the new toolboxes in Unigraphics and CATIA. It seems that the latest trends are moving toward this more direct approach, which is arguably “parametric” in the strict sense of the word without being “historical”.
Rhino is very popular among boat designers, and I’m not familiar with the process of designing a boat enough to make a judgment as to its suitability to the purpose.
I have friends who are hard-core Rhino junkies, and have developed working styles and methods around the particular toolset available in the program. Far be it from me to say that these folks aren’t productive; I’ve seen Rhino at work in the hands of a real Rhino Pro, and it can be impressive.
Where I differ with you is the statement that “Catia seems…like a ’solidworks with enhanced surfacing.’” I would use that characterization to describe Unigraphics NX, as the NX system is much more similar to SolidWorks than Catia. [In fact, SolidWorks runs on the NX kernel.]
Catia has a totally different way of dealing with history and associativity than Solidworks has, and in my opinion is much, much more powerful in this regard. Creating complex design geometry for an intricate product with lots of components and interfaces–such as a power tool–is where Catia really shines.
As for your final question–my recommendation for a small firm–I would probably recommend SW, Rhino, and SpaceClaim. I’ve been playing with SpaceClaim at work, and even though it’s in its infancy as a program, its integration with Rhino is already a huge boon.
If I, personally, were to go it alone, I might think more seriously about Catia. But that is mostly because I’m so used to it, and can work most powerfully with it. But if you’re starting fresh, or already have an affinity toward Rhino or SW, the benefit of Catia in the context of a small firm may be less pronounced. If you were working in a larger organization, however, a more integrated large-scale system like Catia or NX begins to make more sense.
Best of luck!
Adam
September 13th, 2008 at 7:13 am
Thanks for your insights and advice on Catia.
I actually used it only briefly during studies, but i remember it as a one of the coolest programs.
I like your site by the way, nice projects and interesting content! Just bumped into it by googling. Actually it’s funny you mention Spaceclaim, iv’e just got the Rhino version for our office. It’s good for editing Rhino-files, adding mechanical details, moving faces etc. I am pretty impressed by what it can do.
Regards,
Fredrik
September 13th, 2008 at 9:21 am
I was a little quick there, to compare Solidworks with Catia. I understand that the underlying kernels differ quite a bit.
Anyhow, it will be interesting to see how the “new” direct editing approach develops. I think they will work alongside the history based approach, not really replace it.
Both answer to different needs: History gives structure and “scope-control”, while direct editing gives more speed. Feature recognition will probably also come stronger in the future, and speed up things like modifying files from other systems.
A history based design is good when you know what you are going to build in advance, right? But history can be inhibiting because fundamental design changes require a complete rebuild.
Not always though, history can also be a great way to explore different “prototypes”. It depends on the kind of task you have at hand I think.
I run a small firm doing product design. We are fresh in the game, and don’t have buckets of money. But i see your point on efficiency and high-end parametric software. For me Solidworks is high-end enough for now. Havent bought it yet, but we are certainly concidering it. Maby someday we will be sitting on Catia workstations :) For now i will keep on with Rhino, which i find very nice, and hopefully Spaceclaim.
Anyways, once again great site! Nice work, keep on making stuff.
Ps. sorry for double posting. You can put this into the previous post if you like.
-Fredrik
September 13th, 2008 at 1:08 pm
No problem, Fredrik. Thanks for the comments!
I’m currently working on a series of V5 tutorials that should help give an overview of how it works, and possibly how it differs from other packages.
I think that the idea that history is a hindrance to iterative design is misguided. I feel much more comfortable creating a wide range of quick concepts in a parametric app than in Rhino, but that’s largely because I’ve learned to “think parametric”. Rhino nuts will say that Rhino “works the way you think,” where as parametric apps make you “think the way the software thinks.” I think that’s a fallacy; Rhino guys “think like Rhino” because they’re used to it, just like I think like Catia. It’s purely a matter of time at the wheel!
Best of luck to you, Fredrik. Let me know if I can ever be of service.
Adam
October 15th, 2008 at 12:09 pm
Nice work very impressive projects.
I am interested to know how much you use Solidworks versus Catia? I am a Solidworks user doing some minor projects for inventers and have run into some walls recently attempting complex surfaces, organic shapes. So I am wondering if it is time to learn Alias or Rhino. Though I still feel there is lots to learn in Solidworks still.
My main concern is working on a project with two softwares at once. It seems dicey to be importing Rhino models into SW and editing them in context in SW.
Wondered if you or others could advise about the necessity to move beyond Solidworks and working on a model that was created by multiple softwares.
Thanks
October 15th, 2008 at 2:37 pm
Hi Rich.
You mention a couple of things here.
1. SolidWorks has increasingly good surfacing capabilities. The vast majority of surfacing needs can be met from within SolidWorks for the generation of final geometry. If you’re wanting to interactively push/pull surfaces in and out organically, you’re going to need to move to a better surfacing package like Catia, UG, Rhino, or Alias. I personally don’t work that way; I create constrained surfaces from curves and parameters. That being the case, SolidWorks is an okay solution (though Catia and UG are much better).
2. As for the necessity of working between CAD packages, it’s an issue that the whole industry is wrestling with at the moment. SolidWorks does read Rhino files and vise-verse, but I haven’t used this functionality. I typically will be happy to use SolidWorks geometry in Rhino, but never the other way ’round: I would never create a surface in Rhino and expect to use it in SolidWorks. The software may allow you to do it, but I wouldn’t recommend it. If you’re not careful, Rhino will make very bad geometry that will cause all kinds of problems in SolidWorks.
That said, in a multi-CAD environment, Rhino becomes the obvious center hub. It’s great for organizing CAD from other packages, which can then be exported piece-meal for editing elsewhere. Rhino/Spaceclaim is a good combo for this, but SolidWorks->Rhino works well too.
I can say that I’ve spent a good bit of time and energy this year in trying to work Rhino into my process, but ultimately I still avoid doing serious work in it; I much prefer the flexibility that parametric gives me.
Good luck!