Thread: What's a good cad program

Finite Element Analysis for Idiots is one of the worst things that ever happened in engineering. Only 30% off? Miraculously good! I've seen heaps of a lot worse, interpreted as gospel. How 'bout a symmetric structure with symmetric loads, and UNsymmetric response because of haywire boundary conditions improperly entered?

In the 1950s and 1960s, developers were throwing all sorts of shape functions and mesh points at methods of analysis. Some theories came from outer space, it seemed. I don't know how much of them ended up in commercial programs.

I suggest avoiding FEA, until you become familiar with its limitations. One of the best references is "Introduction to the Finite Element Method", by C.S. Desai - provides a step-by-step development of the theory. The first edition had a Fortran program included; probably in later editions too. Also study any tome on Theory of Elasticity, e.g. Timoshenko's, for enhanced understanding.

Joe

It doesn't take long for a thread on CAD programs to degenerate. Like the Ford vs. Holden debate, each has their followers and devotees and every CAD forum has at least one thread that has been heavily edited or closed due to rising tempers and flame wars.

I currently design cars and components for the automotive trade and I have several programs at my disposal including Catia, Solidworks, Inventor, Alibre, Solid Edge and AutoCAD.

I work almost exclusively in 3D, designing some highly complex components and I actually prefer to use AutoCAD when I can, although in truth, I use it in conjunction with several add-ons and complimentary programs.

I won't deride any CAD program; everyone has their own interpretation and method of using CAD and as with everything else in life, everyone has their preferences.

I would recommend anyone casually interested in CAD to download all the freely available programs and see which best suits you as you'll get nothing but conflicting opinion if asking which is 'the best' program to use.

I am well aware of fea limitations, infact I was a prominent and unpopular sceptic of them early on, but I have also seen modelling done by good people that yields good useful results. I don't use the things myself, but the tests were part of our evaluation process.

If you know something about fea you know there were standard tests developed to assess the results from various engines. The inventor results were alarming.

Solidworks comes with its own FEA program (Cosmos) and it, along with most FEA programs, are only as good as the operator. Ten different people using the same program to analyse the same component can easily arrive at ten totally different outcomes.

FEA was the 'in' word about ten years ago and was bantered around with other uber-hyped phrases like 'aircraft grade alloy' and 'hydrogen embritlement'. I know of few (individual) engineers who actually use FEA; it's usually the preserve of university engineering departments and high end companies who can afford to empirically test their products to back-up and thereby justify their FEA departments.

Originally Posted by colin2000

hi
i'm looking for a easy to use cad program to design my new workshop space so every thing i want will fit and also to design some furniture on. i want the program to be quick and easy to learn ( i want to play with wood not spend all my time learning a new program,) thanks Colin

Hi Colin

I used graph paper, scissors and cut-outs.

Cheers

Graeme

If you have a rectangular piece of steel of a fixed thickness and a hole in it and 2 fixed and 2 load points and a specificed mesh density your answer should be in a certain range.

Solidwork does not suck, but it's not as good as nx and is about the same $.

I work for a "working" engineering business and fea is used as part of the design validation process, probably for legal reasons as much as anything else. I've seen people with engineering degrees who can't do newtonian physics, and who came create fea results that defy newton, but then you can give a piece of clear blackwood and a table saw to some people and they will only make sawdust and firewood. It's a problem not restricted to fea.

Originally Posted by damian

...you can give a piece of clear blackwood and a table saw to some people and they will only make sawdust and firewood. It's a problem not restricted to fea.

How very true!

I accept that I was somewhat harsh about FEA for Idiots. Mostly because Desai himself was on my thesis committee at Virginia Tech in the 1970s, and Melosh (of Turner, Dill, Martin, and Melosh, 1956, a fundamental paper in the development of FEA) was a neighbour and informal mentor. I abandoned academia, quite simply because I ran out of money.

In my later career in heavy Civil and Military Structures, I found that while mathematics may be "exact," construction isn't. Key to successful design is the provision of "fault tolerance," i.e. designs that can be modified during construction and operation as needed to reduce drama. For example, in drawbridges we provide pockets in the counterweight to accept many, many balance blocks of about 75lbs (~35kg) each to adjust the CG of the leaf. (Think about wheel weights, but on a larger scale.) In colder climates, even that doesn't work; ice accumulation on the leaf must be jackhammered off to allow the bridge to open.

I don't think FEA is solely a legal requirement. In the early days, computing time was counted in hours or days, and some desperate measures were employed to achieve efficiency. In general, only three mesh refinements can suggest an asymptotic optimum farther out. This is where both CAD and FEA are best used in combination, as interim designs can be saved and further refined, e.g. for constructability.

Bottom line though:
1. Learn how to make sawdust and USEFUL firewood.
2. Draw cartoons on the wood, and cut. (It does grow on trees, y'know).
3. Try some free CAD programs to decide whether it's at all worthwhile. Like you said above, Woodwould.

Joe

Originally Posted by joe greiner

In my later career in heavy Civil and Military Structures, I found that while mathematics may be "exact," construction isn't.

Joe

One of the key facts I remember from high school is that physics isn't an exact science.

In maths, 9 means 9, no more and no less.
In physics, 9 means 8.5 or more, but less than 10.5. If one needs more precision than that, one needs better measurements.

Sometimes, better precision is implied by definition - 1 gram means .001 kilogramme, and an FPS measurement _might_ justify one additional (or one fewer!) decimal point when converted to one of the metric schemes.

Of course, the same rules of imprecision apply to most practical aspects of life, road speeds, cooking, and even woodworking. It's just that a lot of people don't get it.

ok then is .9' equal to 1 ? what is the square root of -1 ?

Wait till you do calculus where you have an infinite number of correct answers but not all answers are correct.

Then come with me into the world of quantum mechanics and the related maths. Wear a bag over your head so when it explodes we'll be able to find all the bits.

Originally Posted by damian

Wear a bag over your head so when it explodes we'll be able to find all the bits.

Good advice, but received too late. Exploded a long time ago, and about half the bits are missing. Might be able to collect some of the dust, if only I could find the dustpan and whisk broom.

Joe

You don't need a CAD programme, you need this..

http://www.grizzly.com/workshopplanner.aspx

It allows you to design on line and they supply all the representative objects that you will find in a workshop. No learning curve and it is simple.