Analyze needs then benchmark!

Ray Kurland

Description

SELECTING A SYSTEMSelecting an MCAD system:determine needs, then bench-mark!by Raymond KurlandPresident, TechniCom, Inc.(Excerpted from an article to be published in the March/April issue of SOLID Solutions Magazine)February, 1998similar future. Reusable designs, based on partsIntroductionusing variables and features will revolutionizeAs analysts and consultants following the Mechani- today’s design techniques and must be an integralcal CAD market we strive to understand how the latest part of every long term plan for new MDA sys-solid modeling software compares. The real question is tems.how to compare such complex software? Since we often Before we get started on the results, let’s go overadvise customers how to select a system that best fits what precedes technical comparisons.their needs, we decided to use a similar process in acontinuing benchmark to understand how today’s hotEvaluate the business needsmid range CAD systems compare. The process issimple, but deceptively hard to follow: first iden- Remaining competitive is the key to success intify the needs, then test alternatives for the best fit. today’s fast paced, ever-changing marketplace.Politics, unfortunately, often intervene. The meth- Evaluating what you need to keep your edge isodology outlined in this article helps make the constant, or should be. A wise business continuallyprocess more objective. examines its design-through-manufacturingTechniCom has worked for years in this arena ...

Sujets

• management

Informations

• Publié par: Ostlue
• Nombre de lectures: 25
• Langue: English

Extrait

SELECTING A SYSTEM

Selecting an MCAD system: determine needs, then bench- mark! by Raymond Kurland President, TechniCom, Inc. (Excerpted from an article to be published in the March/April issue of SOLID Solutions Magazine) February, 1998

Introduction As analysts and consultants following the Mechani- cal CAD market we strive to understand how the latest solid modeling software compares. The real question is how to compare such complex software? Since we often advise customers how to select a system that best fits their needs, we decided to use a similar process in a continuing benchmark to understand how today’s hot mid range CAD systems compare. The process is simple, but deceptively hard to follow: first iden- tify the needs, then test alternatives for the best fit. Politics, unfortunately, often intervene. The meth- odology outlined in this article helps make the process more objective. TechniCom has worked for years in this arena. To help users make these decisions we both publish reports and consult with users. The data in this article is extracted from those reports and our recent experiences. Two reports readers may find interesting and relevant are:a Modern Selecting Mechanical Design Automation System: A How-to Guide andMid Range MCAD Benchmark Report. For more information on these reports visit www.technicom.com. We’ll skip the usual hype about the need for solid modeling and assume that if you are reading this article you are already convinced of the need and advantages of solid modeling, as compared to other alternatives. We’ll also assume that you are not comparing solid modeling to surface or wire- frame modeling systems. If you are, these are called explicit systems. Such systems can no longer compete with variable driven design systems’ productivity and impact on future design changes. Like computer programmers, who envision a future where reusable code modules lead to rapid designs, so too can mechanical designers foresee a

TechniCom, Inc.

similar future. Reusable designs, based on parts using variables and features will revolutionize today’s design techniques and must be an integral part of every long term plan for new MDA sys- tems. Before we get started on the results, let’s go over what precedes technical comparisons.

Evaluate the business needs Remaining competitive is the key to success in today’s fast paced, ever-changing marketplace. Evaluating what you need to keep your edge is constant, or should be. A wise business continually examines its design-through-manufacturing processes and the associated tools to insure work- ing “smarter” to decrease costs while increasing profits. Critical areas of the business not meeting competitive challenges may be in part due to the tools being used for product development. If so, the process of evaluating and selecting the right tools can be a time consuming and costly endeavor. However, making the right decision is often crucial to your company’s ongoing success. A well chosen mechanical design automation system (MDA) facilitates successful product development by promoting a development process in which design, drafting, analysis, support and manufactur- ing departments work concurrently to bring a product to market. It simplifies communication between members of the product development team. It makes product iterations and revisions easy to handle. It has powerful modeling capabili- ties, yet is user-friendly and intuitive for the engineers using the software. In summary, it makes your design-through-manufacturing process quick and efficient, minimizing the time and cost re- quired to optimize a product design and bring the

Page 1

SELECTING A SYSTEM

product to market.business goals? Can you augment your How to determine which MDA system has thecurrent tools to accomplish your goals? capabilities required by your organization? ThereOr, do you need to look at a com- are a multitude of mechanical design products onpletely new set of tools to accomplish the market, all claiming to be the best system foryour goals? your needs. It can be difficult to cut through the·Select a system, if you need a new set advertising hype and industry jargon and decideof MDA tools. This step of the process what a system really can do for your company.is time consuming and is described in detail in the how-to-guide referenced previously. Steps to select MDA software The following are some simple steps you can Using business needs to select a follow to walk your company through the process of selecting MDA software. system ·Determine those areas which are There remains a great deal of controversy critical to the success of your business. regarding whether to run a benchmark or not. For example, quality of product, Frankly we see little alternative. Who would buy a individual/company productivity, car without first test driving it? Actually I did once time-to-market, and so on. and spenta great deal of time, effort, and money ·Review your design-through-manufac- fixing the problems. Either benchmark the software turing process, keeping in mind your or take the time and effort to train on a few systems company’s areas of focus. How does and install them at your facility to evaluate how your current design-through-manufac- well they satisfy your needs. Either way let’s call it turing process work? How do you a benchmark. The following are the general steps want your design-through-manufac- to follow in selecting an MDA system. They will turing process to work? lead you through a detailed, step-by-step process ·Identify the differences between how to assist you in designing and conducting your you want your design-through- own MDA evaluation. manufacturing process to work Create a listing of requirements for the compared with how it currently works. MDA system The changes should improve on the Generate an initial list of MDA software key weaknesses of your business. products that appear to meet the needs Review the best practices of the of your company. industry to make sure you will be on a Construct a matrix, listing requirements in par or better than your key competi- the left column and MDA systems in a tors. Consider using an outside con- row across the top. sultant to expand your thinking in this Based on the results of the completed area. matrix (via marketing and industry ·Analyze how you can implement the literature, vendor responses, and product new design-through-manufacturing demonstrations), you should be able to processes to meet your planned narrow down your MDA system list to process changes. What processes need two or three candidates. to be automated? Do some of your Design and conduct a meaningful software design-through-manufacturing pro- evaluation (benchmark). cesses need to be changed? Can you Based on the results of this benchmark and automate some processes to accom- pricing, make a decision for the MDA plish your business goals? This analy- system that best meets your company’s sis allows you to identify important business goals. tasks, strategic areas of focus, and deliverables in your design-through- Designing a thorough and meaningful evalua- manufacturing process so that you tion of competing MDA systems is a challenging may develop a list of required MDA and time-consuming task. As many users quickly system features and capabilities. discover, benchmarks present their own set of ·Evaluate the results. Using your difficulties. They must be hard enough to test for current tools, can you implement an the desired results, while at the same time easy MDA system to accomplish the enough to complete in the allotted time. In addi- changes necessary to meet your

TechniCom, Inc.

Page 2

SELECTING A SYSTEM

tion, a rigorous set of procedures must be put in place to evaluate the performance of the software, often observed at different times and by different personnel. By following these steps, you can have confidence that your software evaluation process is modern, comprehensive and unbiased and there- fore will produce valid results, ensuring that you are selecting a system that truly addresses your company’s needs. Diving immediately into an evaluation of MDA software and not following these steps, will probably result in the selection and implementation of a system that fails to gain the maximum advantage of today’s (and future) tech- nology.

Designing the benchmark to make the evaluation Fortunately we’ve done this explor- atory work and decided that we need to H2 benchmark mid range systems to see if they meet our requirements. Within the past two years the MCAD market experienced explosive growth fueled by mid-range (main- stream) software, with the entries generally offer- ing variable driven solid modeling, drafting, and assembly modeling for less than $7,000 U.S. dollars. Such systems seem to promise a lot— cheaper variable driven solid modeling software that costs less than high-end systems, fully inte- grated drafting and assembly modeling, using less expensive, Windows-based hardware and easier to use systems. Software in this price range usually runs on Windows NT and all use commercial modeling kernels developed by third-party ven- dors. Most are designed to be Windows centric, relying on Windows for their menuing system and making extensive use of OLE. For an industry review we need to simulate our needs. The benchmark described below success- fully meets most of the above issues and continu- ally evolves as our needs and the software changes. The information in this article is based on the data in our latest benchmark report, published in January, 1998.

TechniCom benchmark descrip- tion Consisting of seven test units totaling 64 steps, we exercise key modeling aspects of the systems (including solid modeling, drafting, and assembly modeling), not FEA, or manufacturing. Nor did we test the ability to handle large concurrent engineer- ing, to exchange data with other systems, or to

TechniCom, Inc.

retrieve, access and share data. The vendors supply the operators, assuring us that we had skilled personnel “driving” the software. They received most of the tests in advance. Test 1 evaluates the ability to create 3D objects from profiles; shelling of solids; creation of draft angles; filleting and chamfering of pockets; and the creation and retention of rib features and feature

H1

H3

V1 V3 V2

patterns. We start with the following sketch:

Then we extrude the profile, add fillets, shell the solid, and add ribs. Test 2 tests for the associativity of the solid model with drafting, including two way associativity, cross sections, offset sections and isometric views. We build this part:

Then we construct an associative drawing from the solid, and change both the model and the drawing to test two way associa- tivity. Test 3 tests the ability to alter the features of the solid model in rather complex ways and correctly update the model and the associated drawings. We start with the Tee block from Test 2. After many alterations Page 3

SELECTING A SYSTEM

and measurements, the result is this:

Shelling then further tests the modeling capabilities. Test 4 builds parts for the assembly test of Test #5. We also test the ability to correctly perform complex blends. We wind up with this:

Test 5 evaluates assembly management functions: assembling parts, building and altering parametric relationships between parts in the assembly, and interpart rela- tions that drive the assembly from part parameters. We use the previous part to build this assembly:

TechniCom, Inc.

Then we alter both part and assembly parameters that drive the final assembly. Test 6 develops a more complex assembly where parts of the assembly relate to other parts for their position and parameter values. Here is an example:

Finally, Test 7 examines the ability of the systems to develop complex revolved surfaces, to add features to the surface and to shell the result.

The solid modelers we tested in Version 4 of this report include: ·Mechanical Desktop V1.2 ·MicroStation Modeler V5.5 ·PT/Modeler V2 ·Solid Edge V2 ·SolidWorks 97 ·Helix V4

What were we able to conclude in gen- eral? For the dollars, these systems offer very robust modeling and are improving fast enough to make the more traditional CAD vendors sweat. All of the vendors tested can essentially complete our entire benchmark, some more successfully than others. Each has their individual functional strengths and weaknesses.

Page 4

SELECTING A SYSTEM

While we did not notice a dramatic ease of use paradigm shift, we found a GUI nudge. Users like Windows centric systems and already are voting with their dollars. While the intelligent interface of SolidWorks comes close to the mark, none of these systems attack the real issues of mass appeal: ease of use, ease of trainability and common data exchange. They are all still too hard to use, requir- ing significant training and knowledge of the intricate sequences of events needed to produce workable parts. Mid-priced systems have jolted the MCAD marketplace. Last year at this time we predicted that the rapid advances in mid-range MCAD software would cause the high-end market to drastically change. Judging by all the acquisition activity we were certainly right. By coupling this power of desktop supercomputers with advanced function, mid-range software will continue to cause wrenching changes in the landscape of the MCAD industry. Furthermore, our prediction that by the end of 1997 half of the new MCAD seats going into

small to medium sized installations would be mid- range has been proved. Now if only they were even easier to use. . . .

About the author Raymond Kurland leads TechniCom’s mechani- cal CAD/CAM program. He consults with both users and vendors and regularly publishes in- depth reports and a newsletter on CAD/CAM products, their pricing and market positioning, both tactical and strategic. Prior to founding TechniCom in 1989, he held management positions at IBM and Dassault Systems. He holds Bachelors and Masters degrees in engineering from Rutgers and New York Universities.

TechniCom, Inc.Clifton, NJ 07013 USA66 Mt. Prospect Avenue [phone]FAX: [phone] TCI [e-mail] http://www.technicom.com

TechniCom, Inc.

Page 5