Questions on downstream model usage

[Shadowspawn]

Hey guys,
I need some feedback dealing with the business case for downstream usage of the models generated. This could easily get a bit long winded, so please bear with me. I need to get a better idea as to what the benefits "really" are for spending the time/energy in the design cycle to create a model that facilitates downstream usage (and what those uses are: vendor, drafting, manuf, etc) by incorporating sketches, wave linking, features, and so on the way that UGS intended versus very simplistic revolve/extrude type models created with 2d explicite curves (as in using the 2d crvs to revolve a tool solid, which is subtracted from the body to "create" a machined hole for an helicoil insert) that don't lend themselves to downstream usage. The basic arguement here is that it's just not worth the time/effort in our dept to utilize sketches, wave linking, UDF's, etc. as downstream usage is virtually non-existant, it takes "too long" to modify and update, too many updating errors are generated, too much complexity and associativity causes too many problems, etc, etc, etc.
As a little background, a lot of our product line is fairly unique and low count type stuff (aerospace gboxes and pumps)and there is very little customization of the toolset. In addition, the culture stems from a 2D toolset that tries to beat UG (a 3D toolset) into functioning like their "old" 2d toolset, (and methodologies reflect this).
So what do you think? Is it justifiable to the company to spend the time money/time/effort to update methodologies, incorp some of these "higher-end" functionalities UG offers (wave linking and hole features, not KF, smart models, or anything cool like that), customize the toolset (as in UFUNC and GRIP), and enforce these changes??? ...or is the reality in the real world that downstream usage of the models just isn't to the point where its cost justifiable to do so? (From what I can tell, there's very limited manufacturing usage of associated models and all vendors get simple parasolids to work with).
Thoughts, opinions, comments, and suggestions are welcome. And just so you know, I'm the guy advocating strict model checking prior to release, sketches, parametrics, wavelinking, mating conditions, toolset customization, editing of features versus deleting/recreating features, etc.

 

[looslib]

Start simple. Use the features that come with UG in a solid modeling environment. No shortcuts, just use it as designed, without reverting back to teh 2D way of doing things.

Use the basic shapes, blocks, cylinders, etc, along with sketches to create your initial starting block. Add features like bosses, pads, holes, chamfers and rounds to make the fnal part design. Do not use a cylinder and subtract to make a hole. The advantage of the hole feature is that it can be changed in size and moved in location. Try doing that with a subtracted cylinder!

I would put wave-linking bodies at a higher level of usage. Until you get the basic modeling techniques under control, stay away from the advanced stuff. There are alot of file interdependencies that get created that you really need to understand how they interact with the various models.


"Wildfires are dangerous, hard to control, and economically catastrophic."
"Fixed in the next release" should replace "Product First" as the PTC slogan.

Ben Loosli
CAD/CAM System Analyst
Ingersoll-Rand

 

[ewh]

Shadowspawn,
The arguement that "it's just not worth the time/effort... to modify and update, too many updating errors are generated, too much complexity and associativity causes too many problems, etc, etc, etc." seems weak to me. It may be true if poor practices are used.
With good practices and proper discipline, it is much more efficient to use the tools available to make a change once and check that it is carried through all levels that the part is used than it is to make the same change to multiple files.
Using revolved curves to create a hole is inefficient. To properly control those curves a sketch would be needed. As Ben stated, the hole function is a much better way to go. But a case can be made for using subtracted cylinders as holes. One of the great things about UG is that you can tailor it to suit your needs. We use subtracted extracted cylinders for holes in non-planer shapes (aircraft lofts). You can modify the size through expresions and relocate the parent tool cylinders as required. This is much simpler for us that creating and maintaining the large number of planes that would be necessary to use the hole function.
The important thing is to define your procedures and stick with them. If a new method is proposed, it can be evaluated and added to the procedures. This makes it much easier for someone unfamiliar with the part to modify it or create a similar part. It also keeps those updating errors at a minimum and eases the complexity of the part. While it may be a headache to implement the proper procedures, once everyone involved understands the reasoning behind them and the importance of following them, productivity will be greatly improved.
Just a few thoughts... I'm sure others will join in.

 

[MSPBenson]

I agree with ewh. Using UG without using some of the features like the "Hole" feature is like buying a car and only driving it in first gear.
Some of the features like sketch and hole etc are (once you get used to them) much quicker and more efficient to produce models with. The complexity issues are normally down to the skill of the CAD users.
Errors when updating once you understand what they say are normally very easy to rectify and with a little thought can be avoided in the first place.
You can still keep models simple and use sketches instead of curve strings. A sketch can be attached to a Fixed datum plane and will act like any other string of curves however the profile can be modified with greater ease


Things to bear in mind if you do implement a change in your usage of UG are:
Don't overload the users with too many new features. It has to be a slow process.
Make sure everyone is brought up to the same level with each stage of training. For example if only half your guys know how to use sketch then not everyone will be able to modify each others models efficiently.
Accept that the change will reduce output in the short term but increase it in the long.


I've been through this in a company with 250 UG seats and what ewh suggests about procedures and standards is critical if your user base is large.

In short get your money's worth out of UG.

Hope that's of help,



Mark Benson
CAD Support Engineer

 

[cowski]

To address the original question:

...better idea as to what the benefits "really" are for spending the time/energy in the design cycle to create a model that facilitates downstream usage...

All depends on what is downstream from you. I think you answered this for yourself when you said:

(From what I can tell, there's very limited manufacturing usage of associated models and all vendors get simple parasolids to work with)

To address the underlying question of How can UG make the designers' lives easier? Come with me on a little trip.

Welcome to Company A
Company A produces gearboxes customized for their customers' needs. Through years of experience designing these gearboxes, they have worked out a wonderful system. 90% of their orders can be boiled down to 10 independent parameters that customers can change (shaft placement and rotation, expected loads and speeds, etc etc), and in turn these parameters can drive dozens or hundreds of other parameters. Through the years the designers have created seed parts that incorporate these relationships; a few worked great from the start, others needed some refinement. Today a designer enters in the parameters, hits a button, and wanders away to find a necessary coffee refill. The macros jump into action calculating the gear ratio, calculating the minimum required shaft size, rounding the shaft size up to the next stock size, calling out the correct bearings, etc, finding the correct seed parts, and updating expressions; now interpart expressions and wave linking take over to make sure all parts work together correctly, an exploded assembly view is generated and plotted, part prints are generated and dimensioned based on sketches from the seed files, a bill of material is generated and formatted correctly, and all necessary information is entered into the process control system. On the way back from coffee break the designer stops by the plotter and picks up the drawings. Congratulations, what used to take 2 designers a week or more, one designer has accomplished over a coffee break!
<cue the ukelele music and tropical sunset>
Welcome to CAD nirvana (population: too few, if any).

Welcome to company B
Company B makes consumer coffee makers. The design department receives specs with items like small footprint, modern look, 8 cup capacity, and $50 price point. After years of design experience they have a very good grasp of how to turn water into coffee. Where they spend most of their time is on the look and feel of the product. They start their design by using a previous project with as much in common as the proposed project as possible. Because the part count, feature set and styling are never quite the same it has been found that wave linking and interpart expressions can actually create more work than they save. It has also been found that the styling aspects are easier to work with when created with plain curves as opposed to sketches. The designers tweak, fiddle, and edit until they run out of time and are forced to shove what they have out the door and call it done.
<cue the performance reveiw>
Welcome to CAD reality (population: too many).

There are a few points I would like to make
1) Certain products/industries get more benefit from certain tools in CAD
2) For those in CAD nirvana; it has been a long, hard journey - it didn't happen overnight
3) If you reach CAD nirvana (some of us never will); enjoy it while you can, some stay longer than others, but no one is a permanent resident
4) All those cool features that sound great on paper can be counter productive for your particular situation (and even if they do eventually lead to CAD nirvana for you they can often seem like a roadblock, especially when first learning them)
5) Even at company A, an oddball order will come through every once in a while that just doesn't fit nicely in the system.

Those of us responding to your post will naturally be inclined to recommend to you what has worked for us - keep in mind it may or may not work for you. If a certain tool follows along with what you are designing, great; if you have to force a tool to do what you want with workarounds and such, you may want to take a step back and look at the big picture (and the help files!).

Sounds like you made a big step recently moving from 2D to 3D and a lot of the "old guard" is resisting. Take Ben's advice and learn the basic tools, then go slowly from there. And if I may add a bit of advice, be careful about talking bad about the old system (especially if any in the old guard helped make the system), people get strangely attached to methods, and if you attack it you are by proxy attacking the people attached to it. Tread lightly and you'll catch more flies with honey, as they say.

 

[motorsportsdesign]

Design for downstream usage has much more to do with the manufacturabilty of the shapes designed than the features used to form them.

An effective way to get a design group to get behind advanced use of 3D is to have an under-rated designer do a couple of pilot projects. If he succeeds, then the others won't have any arguments about difficulty.

Jonathan T. Schmidt

http://www.schmidtmotorworks.com