NX Solid modeling best practices 2

[TechnicalConsultant]

Hi all

We're up and running with NX4 and TCEng V9 at last! Just need to get everyone to start using 3D now.

I posted a thread a while back (thread561-145399) about best practices (specifically solid modeling) and so far haven't had much look in finding everything. I have documented what I believe to be correct but for all I know we could be doing things completely wrong.

If anyone has anything documented the could share it would be greatly appreciated (looslib, you previously commented you may have something I could have a look at?)

Many thanks in advance

 

[ewh]

Kirsten's suggtestion in the previous thread is a good place to start. There will be differing opinions depending on the industry and type of parts being created.

 

[TechnicalConsultant]

Kirstens suggestions were great and I have gone through all of those and tailored them to suit our needs. I didn't originally think it would be such a large amount of work but the more I look into it the bigger the task seems to be

 

[Kirsten]

Thank you, thank you! I thought you might be interested in what happened when the whole group tried to reach an agreement on the standards. It got a little heated. We still have not implemented them, because there's no one to enforce them.
* means mandatory:

Applies to all new models & files. Standards to be met prior to checking & drawing release
0. One primitive maximum. First feature located at 0, 0, 0.
No centered blocks—a corner must be at 0,0,0.
*1. No suppressed features
Exception: detailed threads may be suppressed
*2. No Out-of-date features
*3. No blanked objects
*4. Sketches fully constrained – except, for example, extruded text curves
*5. No “Fix” sketch constraints
*6. Proper layers
1 sketch per layer, generally
Categories must be assigned to every layer with an object on it
Category names for sketches must match sketch name
Master layers: Solids 4-100, all else 101 and up; 256 file notes
Drawing layers: Pattern 1; drafting annotation 3; components 4 & up; File notes 256
* 6a. All sketch names start with SK and have a descriptive name. Default name not allowed.
*7. Reference Sets set (solid, mating, drawing are minimum required ref. sets, each contains solids except for files with no geometry)
*8. Attributes assigned (must match between drawing & master)
*9. Item & file named properly
*10. Mating conditions must have a minimum of 2 constraints per condition.
Exceptions: models with no geometry
mate taper to taper constraint
Base component
*11. No dumb dimensions except “Text”
*12. All symbols & notes associated to views, “relative to view” is minimum origin type. Exception: Main block of notes
rev. blocks by main notes (preferred that rev. blocks are part of the notes)
13. No unparameterized features (applies to new models)
Exception: Vendor-supplied models
Models converted from SolidEdge
14. Scale expression preferably linked to drawing views & matching title block scale, unless scale is NONE
15. No faked parts lists (preferred)
*16. No suppressed components (assemblies)
Exception: Members of part family assemblies
17. Smart ID balloons when smart parts list is used
18. Smart callouts in drawing notes when smart parts list is used
19. For a part family that is saved to a new part number: family spreadsheet deleted, or all instances updated to new part number
*20. Correct assembly structure
*22. Component name matches item ID
*23. Non UG datasets that are linked to previous revisions must be cut from new revision
24. Photo materials assigned in master (optional)
25. Solid density assigned in master models
Exceptions: parts that will not be sold, like production fixtures
Assemblies
BEDocument types
Parts with no solids
26. Use expanded view for faking stuff--no lines/text that appear over a drawing view allowed on drawing sheet
28. NO interpart expressions
Exception: drafting expressions



30. Use feature ISET for arrays (recommended)
31. Assemble in order—create in order : For example, if you assemble washers to bolt holes, first assemble the washer at the beginning of the series, then the next closest, and so on. Don't assemble the 1st one, then the fourth, then the second. Do the same when making the holes: make the 1st, then 2nd.. (recommended)
32. Don't embed a feature in a feature: Don't make a cut to remove unwanted features. (recommended)
33. Links broken
*34. Family members updated
35. Title length per company standrd (total of “title” plus “title2” part attributes). Preferred that all characters be assigned in “title” attribute.
36. Do not revise a part family member. Saveas on a part family template to a new part number is not preferred.
*37. No unassociated OBJECTS (Dimensions or Leaders)
*38. File must be saved as follows:
Drawing File, Sheet 1 in drafting.
Master model, modeling, shaded, only solids visible except for file notes, fit to screen, unexploded, layer 256 (File Notes) visible, model oriented to show best view & detail

*39. No extra dwg sheets
Drawing only. Delete all unused sheets from drawing file.

40. File notes
Layer 256. Layer Category to include view name that notes are placed or be View- Dependent-Edited, View-to-model to allow it to show in all views.
The note should contain any non-typical model or drawing features or practices that had
to be used in the file or any information that would be considered helpful to the next
person to work on the file. The notes, when needed, should be simple and concise.

*41. No incomplete models allowed (missing features), unless a file note is created saying “This model is incomplete” and the note must be included in the solid, mating & drawing reference sets.

42. No MAX dimensions. Models and features shall be nominal

*43. Attributes in Title Block (not dumb text)

44. Correct Item Type in TeamCenter Eng.

 

[ewh]

Kirsten,
I can imagine how heated it got. I agree with the majority of points, but have to take exception with a couple of them (and am sure that I would have been quite red-faced at your meetings)-

0. ...First feature located at 0, 0, 0.
No centered blocks—a corner must be at 0,0,0.

-All of the aerospace companies that I have worked with use aircraft coordinates to locate the models, with the exception of parts used in multiple place (such as hardware). This does not mean that mating is not necessary, but it makes life much easier for the vendors if they don't have to "reconstruct" the aircraft to determine which relationships must be established/maintained, as well as for the customer when they pull our models into their assemblies. We manufacture composite parts (often with NO square edges or circular features), and it would be a HUGE headache if we had to relocate our part models to match the customers interface or mold models, etc.

*39. No extra dwg sheets
-We use simplified schematics to represent ply layups. These schematics are referenced to part views using detail views and are labeled as such. The actual detail view goes onto a "scratch" sheet where it will not be referenced.

While the rules listed are good, legitimate ones (depending on the industry), I would hesitate to establish so many. You have to factor in what it will cost to maintain these standards vs the benefits derived, and allow for future situations where they may become cumbersome or not apply at all. Saving a large assembly in shaded mode, for example, will cost you extra time (money) every time someone opens that file.

 

[JohnRBaker]

For what it's worth, I personally avoid the use of layers for several reasons, not the least of which is that they are a holdover from the 2D drafting days and were never intended to support modern modeling practices. I feel that Blank/Unblank (or Hide and Show, which will be the terminology used in NX 5) works and since it's binary in nature, it's much easier to understand and use. Besides, anyone with a rule that there are no blanked objects will find that some of the nice 'features' used in the navigators to control visibility of objects will find that they not be able to use them.

Also the rule about 'no suppressed components', well then you will not be able to use Arrangements since it depends on using 'Component Suppression' to control the contents of an Arrangement.

Anyway, these are just some personal observations and comments, but then I have the luxury of not working in a production environment and always knowing ahead of time, what's coming next

John R. Baker, P.E.
Product "Evangelist"
NX Product Line
UGS Corp
Cypress, CA

http://www.ugs.com

 

[TCMartist]

I find this post particularly interesting, thanks to noyce and all those who give their best practices.

I did the google search shadowspam suggested in the previous post and found several intersting things. one of them was refering to a book :

Practical Unigraphics NX3 Modeling, Drafting and Assemblies (Paperback)
by Stephen Samuel; Anuranjini Pragada; Gautam Baksi (Author)

does anyone had a look to it? I could not manage to find any developped comments or content on the web...

 

[TechnicalConsultant]

That book sound interesting, I'd be very interested to hear if anyone has seen it?

 

[Kirsten]

EWH,

I've always said that majority rule shouldn't be allowed when it comes to setting standards--some people just aren't qualified enough to vote, but that's what we did to arrive at these. Some of these rules were focused towards certain users' bad habits. And we have so many because everyone has a pet peeve.
We make Beverage & Oxygen components and (so far) seldom deal directly with the airframes, so don't use the aircraft coordinates. We have a couple of guys who place the components off in space and even at odd angles into the drawing file, and then ask for help when there drafting views are skewed. They've been using UG for 6 years minimum! (One of them draws his points using 4 lines, I found out yesterday, and he's one of our "better" modelers.) So that's where 0. ...First feature located at 0, 0, 0 came from.

No extra drawing sheets is because we have automated pdf creation, or will have, if we ever get an upgrade.

John,
I believe devoutly in blanking rather than layers! Unfortunately, the guy here who has a lot of pull with the boss and who hates datum planes is the reason for the layer, sketch naming and layer category rules. He tells me I'm too progressive ("The defaults were always good enough for me. Why do you have to change things?"). My boss tells everyone I focus too much on being good at using UG, because "CAD is only a tool." He (the boss) could "use a pencil & napkin and still get the job done."
Sure it's a tool, but I still want to be the best tool-user I can be! It's the main tool we use for our jobs!

John,
Is component suppression in arrangements available in NX2? If so, we didn't know about it.

 

[ewh]

Kirsten,
While I understand John's point about layers, I still find layer segregation with clear catagories is easier for the average user than blanking, but that could just be me. To me, blanking is binary, either on or off, and layers give more control.
I also understand about auto pdf creation. Instead of pdf's, we use cgm's, manually created. I think it is likely that we will start using .jt or acrobat 3d before long, though. Drawing importance is slowly fading with MBD.

TCMartist, noyce,
I have a copy of the book you are refering to, and while it is an excellent primer, it does not go into any depth. It covers what basic UGS classes do and not much more.

 

[JohnRBaker]

Kristen,

My response to people when it comes the 'defaults' is that if we didn't want you to change them, we would never have given you the option, we would have just hard-coded them and told you "TOUGH!".

As for the issue with respect to 'suppression' of Components, that enhancement was added to Arrangements in NX 3. In NX 2, where we introduced Arrangements, you could only use them to define alternate positioning of Components. In NX 3 we included the ability to also alter the 'content' of an arrangement and the 'mechanism' we decided to use was 'suppression' because it was already there and we had given the user ways to control it (with Expressions) so we used the same mechanism and just linked it to a particular Arrangement. Note that part of the rational for adding these NX 2 and NX 3 capabilities to NX was to provide support for workflows originally developed in Ideas using what was referred to as "Configurations" (we changed the name to reduce confusions with TeamCenter 'Configurations'). This also allowed us, when migrating Ideas models to NX, to convert those 'Configurations' into 'Arrangements' and preserve the design intent of the Ideas users.

ewh,

The direction that we are going is toward a more comprehensive 'grouping' sort of approach, using a 'folder' metaphor. We introduced this concept in both the Sketcher and Assemblies in NX 4. As far as the Sketcher is concerned, you can segregate curves into different folders and act on them as a 'group' such as selection or Blanking (in NX 5 this will be changed to "Hide" and "Show"). Also there is the concept of the 'active' folder (group) which means that any new Sketch curves created would automatically go into the 'active' folder.

With respect to Assemblies, what we did was convert what we used to call 'filters' into folders (groups) for both consistency and because we think it's a better understood concept. Again, we are providing options like Blanking and so on as well as 'rules' (which are the old 'filter' settings) so that we can again automate certain behaviors. Long term the idea is to eventually provide a comprehensive scheme using this folder metaphor with the ability to act on all members in a folder as well as define rules, like an active folder or one where I automatically put all objects of a certain kind, perhaps like sheet bodies or solids or dimensions or whatever you want. At that point, we could just 'convert' layers into 256 folders named '1' thru '256' and then let the user rename them, define rules, add to or remove from them, etc. The long term value is that we can apply common actions to both objects and folders of objects, such as visibility ("Hide" & "Show"), selectability, automated behavior, etc.

Anyway, just some comments about why we're going the way we're going and why some of us feel the way we do about current 'features' of the system.

John R. Baker, P.E.
Product "Evangelist"
NX Product Line
UGS Corp
Cypress, CA

http://www.ugs.com