Continue to Site

Welcome to MCAD Central

Join our MCAD Central community forums, the largest resource for MCAD (Mechanical Computer-Aided Design) professionals, including files, forums, jobs, articles, calendar, and more.

Register Log in

Good modeling practice.

Can someone else help me out in explaining this?



Let's say I decide to create a protrusion or a cut. I give the appropriate command (e.g., Insert > Protrusion > Extrude), select a sketch plane, direction of feature creation, and reference plane. That will put me into the Intent Manager, in which I will select sketch references, sketch geometric entities, dimension and constrain the sketch, and modify my sketch dimension values. This creates my sketch. I now have a sketch. It is part of my protrusion or cut. Therefore the protrusion / cut owns the sketch.



The purpose of inspection is to verify that your parts have been manufactured within designated tolerances. It sounds like your experience is mostly from working within a drawing-centric as opposed to a model-centric organization.



I create parts and assemblies that are solid models of the products I want to build. When I build my parts and assemblies, my design intent guides my choices. My design intent reflects the purpose of my models-- what I want them to do and be-- and the information I build within my models that governs how I want features to react parametrically when design changes occur.



For example, my part gets 2 inches shorter, so I change my main protrusion. By my choice of dimensioning schemes, parent-child relationships, relations, optimization study features, etc., subsequent features will change parametrically. Hopefully, if I've created my model correctly, the subsequent features in the part will update the way I planned them to and expected them to, without features in my model failing and sending me into Resolve Mode. If this is a part in an assembly, hopefully other parts with external references will also update without failing. It doesn't always happen that way, but that's the point of an associative parametric solid modeling tool. (Trust me on this, I used to work for PTC.)



The point of a 2-D production drawing is for me to communicate how I want a part to be built and inspected. Sometimes it is necessary for me to give different information for building and inspection than I used as my design intent for creating the solid model of my part or assembly. This often happens in complex surface design. This is why I'll have created dimensions and reference dimensions on a drawing. The shown dimensions reflect my design intent. For complex part and assembly design, it is sometimes necessary to create dimensions on the drawing that I do not have in my part features. Therefore, the production drawing does not always reflect the entirety of my design intent-- but again, there should be a huge amount of overlap.



David Martin

Torgon Industries
 
The purpose of inspection is to verify that your parts have been manufactured within designated tolerances.



Inspection is to help verify that the Engineer's design intent is met - checking tols is only one aspect - checking dimensions, material certs, notes etc are all parts too.



My design intent reflects the purpose of my models



Geewillikers Batman !! You're talking about the design intent of your models only !!!!!



What about design intent of your actual designs ???



I'm talking about the design intent of actual, real world parts - not the same - design intent as in how parts fit, how stack-up affects the part functionality etc - you know, the minor stuff !!



The point of a 2-D production drawing is for me to communicate how I want a part to be built and inspected.



Depends on the drawing, detailed part drawings are always inspection drawings. They're also legal documents used in the procurement of parts.



Trust me on this, I used to work for PTC. - some things just seem to say it all..
 
I'll avoid the baiting in the last post and get back to the original question regarding good design practices.



Many people think that good design concerns how quickly one can make a model or how simple a model can be (i.e., fewer features). Or how quickly / easily one can make a drawing. That may be the case in other CAD packages, but not Pro E.



Pro Engineer recognizes that the initial design of a part or an assembly is a small portion of a product's lifecycle. Engineers spend most of their time managing change and reacting to changes in a product's requirements.



Therefore, good design practices in Pro Engineer concern building additional intelligence in a model so that it's parametric-- making changes to one feature propagates changes to other features in a manner one expects and has planned for.



I stand by the practices I listed in the 6th reply in this thread. The list is by no means inclusive. Model Check looks for 800 different things in a model IIRC. If I gave some more time to think, I could probably come up with about another dozen guidelines pretty quickly. I did some consulting at Raytheon Tucson once and saw their Best Practices list. I wish I had written it down, because it was the best Best Practices list I've seen at any company. But they all reflect the same core idea: build your model so that it is as adaptable and as robust as possible when faced with design changes.



I've learned a few things from teaching and consulting in Pro Engineer. For one thing, every company does things differently. Don't assume because you have one set of standards at your company, that others adhere-- or should adhere-- to those same standards. Secondly, you can explain to someone what good design practices are and why they should follow them, but at the end of the day, it's an individual decision whether or not they choose to follow them, and arguing with them when their minds are already made up is fruitless.



David Martin

Torgon Industries
 
good design practices in Pro Engineer concern building additional intelligence in a model so that it's parametric-- making changes to one feature propagates changes to other features in a manner one expects and has planned for.



1) Point of order - when you say design practices you are actually refering to modeling practices - 2 completely different things.



Design practices are things like recommended hole spacing, component fits, hole edge distances, explosion proof (in my business) requirements etc, etc



2) Modeling needs to be based on fit, form & function of the REAL PARTS and not on what's convenient to model.



3) Pro/E is just a tool to simulate real parts - preferably as close as possible.



I realize as a consultant that your deliverable is usually a Pro/E model but most users theirs is actual real life hardware.



4) Models and drawings should REPRESENT REAL parts as closely as possible - do you disagree ??



That includes tolerances, dimension schemes, datums, geometry etc, etc



5) every company does things differently. Yes but fit, form & function is forever.



We shouldn't be modeling just to facilitate Pro/E - classic case of the the tail wagging the dog. I agree robust models are important and desirable but (REAL) part functionality ALWAYS has to come first - think about it.



At the end of the day what's important is a functional, cost effective, reliable, producible, competitive, marketable product - not some nice fully parametric Pro/E models.
 
In my first five years of using Pro/E, I defined good modeling practice as ...so THAT's how I should modeled this!



Can't ever go wrong by modeling around ASME Y14.5 or equivalent..
 
When Y14.5 gets updated to include large assembly management, I'll agree with that... :)



How does Y14.5 affect large assembly management ?



Y14.5 is for Dimensioning and Tolerancing - large assemblies typically don't instruct many dimensions or tolerances :)
 
exactly...



My point was just that some people mistakenly limit the topic of modeling practices to part modeling practices.



-Brian
 
My point was just that some people mistakenly limit the topic of modeling practices to part modeling practices.



True and I'm glad you refer to modeling practices and not design intent !!!



After having said that the majority of issues are only relevant in part mode..
 
After having said that the majority of issues are only relevant in part mode.



Try telling that to companies who have designers who use external references improperly! Depending on the user and product, good design practices in assembly design can be just as important or more important than part design practices.



And for the sake of this discussion and the original question, I'm limiting my discussion to modeling practices. We'd be here all day if we discussed good product design practices (I'll try to be very precise in my terminology). The stories I could tell from my five years as a structural analyst at Unnamed Big Multi-National Defense Contractor...
 
Cccccccc........



Do you work on your job or you are so good that most of your job time you can surf the web? :)



PS I will agree with Brian



...so THAT's how I should modeled this!



He he he



Speling
 
Can't let Speling have the last word:



1) Design for product design intent first and foremost - after all this is why we, all our companies and all our clients exist in the first place..



2) Model around ASME Y14.5 or whatever standard you're required to conform to for dimensioning and tolerancing.



3) Use realistic tolerances.



Remember that a 0.014 diametral tolerance zone equates to +/- .005 in rectangular coordinates.



4) Remember that Engineering models and drawings are always inspection documents and as such are based on functionality.



5) Use GD&T - if you don't know it, learn it inside and out.



6) Make your models robust and in a logical manner:



Make a guess at your primary, secondary and tertiary

datums as early in your model creation as possible.

Remember that these are function based so they are

the Engineer's call - not manufacturing or inspection.

This will enable you to set up your dimensions and

minimize reworking your model later on.



Don't save with suppressed features.



Name features in a logical manner to assist in model

navigation.



Consider redefining features before creating additional

features.



Minimize geom checks.



Don't make rounds parent features if at all possible.



Put as much information into your model as possible -

gtols, set-up datums etc instead of drawings.



Avoid faking dimensions and tolerances in drawings -

as long as this practice persists there'll always be a need

for drawings.



If you specify any holes in your design, create them as

hole features and not cuts. One exception are

compound holes - these can only still be created as

revolved cuts (misfire may be different, anyone

know ??)



Use start-part features as your references as much as

possible - these are probably the most stable features in

any model.



When creating a start-parts don't just put in 3 datum-

planes, put in a center point, a csys, and 3 datum-axes

as well.



Always use standardized start parts.



If you're not on misfire, use make-datum to

avoid datum clutter and unnecessary feature

proliferation.
 
Here's what SolidWorks documentation says about sketch complexity:



View attachment 334



Note that even in this simple example the number of features is halved by using a complex sketch - imagine a 2-300 feature model ??



Although it's SW the same factors apply to Pro/E too.



Proeguru, your link is old (it refers to PT modeler !!) and out of date - most users I think understand the need to save frequently.
 
Although, you have to admit... if you were then told to go back and add draft to a part like that, you'd probably move the rounds to separate features that come after the draft features...



Or should we be building draft into sketches as well? :)



-Brian
 
You must log in or register to reply here.

Sponsor

Articles From 3DCAD World

Back
Top