Pro/E user attempting switch to UGNX 6

[fighterpilot]

After over 10 years as a designer and CAD admin for Pro/Engineer I need to learn UGNX 4. This is proving to be quite difficult. I'm sure I'll have more questions but I'll start with this one. I have a cylinder 10" in diameter, 8" long and want to create a turned down section 4" in diameter and 5" down the axis.

I want to revolve a rectangular section about the axis of the cylinder to create in Pro/E terms, a revolved cut. I can seem to get the rectangular section created in UG sketcher but now I'd like to apply a diameter dimension in my sketch (4") to what would be the OD of the revolved cut. In Pro/E sketcher I would just drop a centerline down the center of the part, align it to the existing cylinder axis then dimension the revolved cut by selecting the sketch line, centerline, and then the sketch line again. Viola, I have a diameter dimension. How do I accomplish a similar scheme in UG?

Next I would like to constrain the other section line to the projected edge of the main cylinder OD and then constrain the bottom edge of the sketch to the lower edge of the large cylinder. In the end I want two dimensions driving this cut, the 4" diameter and the 5" cut length.

How do I go about doing this in UG?

Thanks...

--
Fighter Pilot
Manufacturing Engineer

 

[NXMold]

"but what frustrates me is the attitude that common complaints (sketcher not recognizing off-plane curves) is not a problem. "

Yes, that is exactly my impression since I have started with NX. Frustrating, and ultimately dissapointing. NX cannot be all things to all people, but I don't think that statement addresses the issue.

 

[phillpd]

UGII_DRAFT_EXPRESSIONS_OK doesn't work too well when using the Master Model concept.

Slightly off topic I know, so I do apologise, but the one thing that gets me about Diameters (other than inside a Sketch that it is to be reolved) is that Info->Object and selecting a cylindrical face ONLY returns its Radius. Why ?

Specialty Engineered Automation (SEA)

http://www.sea4ug.com

a UGS Foundation Partner

 

[JohnRBaker]

Note that our strategic direction is to support PMI (Product and Manufacturing Information) and for NX 6.0.x (not sure exactly when it will be ready for prime-time) you will be able to explicitly make Sketch dimensions (parametric constraints) PMI objects which will be accessible from the 3D model views both in terms of seeing them as well as double-clicking on them to edit the sketch without having to actually opening the sketch or digging through the expression system to find the relevant sketch parameters. Now part of the intention of using PMI is to someday avoid having to create 'drawings', at least in terms of they being the only place where non-geometric specifications and annotation is placed to complete the documentation of a part design. However, if 'paper drawings' (or their electronic equivalents are needed) once views have been places on a drawing sheet, the PMI dimensions can be inherited to the drawing sheet and will be associatively linked back to the PMI data in the part model.

For a short description of what PMI is, go to:

http://en.wikipedia.org/wiki/Product_Manufacturing_Information

Also be aware that PMI has been adopted by and is being promoted by the major standards setting organizations, as noted in the article above. Therefore don't look at this as being something limited to only NX or even Siemens PLM products, although it is true that our company has been one of the leading champions of getting this adopted as a standard in the industry, and in that capacity has worked very closely with the standards setting organizations since the inception of this approach and it's suggestion that it be taken up as an industry supported standard.


John R. Baker, P.E.
Product 'Evangelist'
NX Design
Siemens PLM Software Inc.
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

 

[hudson888]

Bfleck,

You were right about the studio spline in the sketcher. I haven't had much call for it, so it doesn't actually bother me, though I can see some uses for it. After reading your later post I tried using it and moving the line only to find that the spline once created is a static entity. As far as I'm concerned this is almost totally useless, and does need looking into. I find that the same feature outside the sketcher works in the opposite manner and is for all intents an purposes superior. In summary I don't know why you'd use the version inside the sketcher as it stands when you can use the other one, but at the same time if they're going to have a studio spline inside the sketcher they might as well have one that works or not at all.

Yes as I said the expression based dimensioning isn't up to much, but the attribute based stuff for hole types and threaded holes isn't bad. As I also said I seldom find it too difficult or time consuming to add a few vertical and horizontal dimensions that would take care of most suitable parts. As the complexity increases surely you'd have trouble positioning a larger number of dimensions automatically, and may have to revert to more manual methods. I can only say that I don't miss what I haven't ever had, but I didn't mean to discourage your idea for what may be a good enhancement to NX in the future. Rather I should ask what does Pro-E do with models that don't lend themselves to any kind of automatic dimensioning, ie free-form or molded products with draft applied can be difficult to draft conventionally. My point being that on balance a lot of our work may not benefit from enhancing conventional drafting, perhaps there is something in the PMI package more suited.

NXMold,

What and why would you want to do with off plane sketch curves? It is a piece of software which probably treats a curve slightly off the plane in the same way as one that is sloping away at a 45 degree angle to the plane. Taking the later case as equivalent what outcome would you expect if you tried to constrain it using the sketcher. The problem that exists is there to be solved according to how you apply the tools to solve it. What solution could you reasonably expect the software to provide? What more appropriate response do other systems provide?

The only curves that I ever use in a sketcher are created withing the sketch and line exclusively on that plane. If I take in curves from the outside I can project them first where necessary. What else is there?

I'm not sure which way you were going with the post, whether you were taking somebody's earlier quote to agree of disagree? What I will say is that if you're posting to ask questions or solicit tips don't expect much help if you approach us with a negative attitude. The best way to overcome your frustration is to consider the proposition that other people are happily and successfully making use of this software, somehow there must be some knowledge worth seeking that might make that difference to you experience of it. So why cut yourself off from that experience by focusing on the differences between NX and whatever you have preferred in the past. If this is unfairly directed at you then I apologize, you may take it to be fairly directed at others posting here.


Phillip,

On information about radii.
I think that it has usually been this way, and that it has only become something that I noticed since posters like yourself raised it as an issue.

Interrogating an edge will return a radius and a diameter value. I know that isn't always particularly helpful in all cases, but it may be useful to keep in mind if you weren't aware if it in the past.

Analysis>Geometric properties will return the approximate radius of any surface adjacent to a point that you indicate. Since all faces aren't regular and cylindrical in the case of free-form surfaces for example the value won't always correspond to a diameter. I think that the information on a selected face works on more or less the same principle.

I think what John is referring to above is that this kind of information could and possibly even should be contained within the PMI. The idea there being that design intent and surface topology are treated as two separate things even if they co-exist in the one piece of geometry.

John,

You could comment on the above if I have put the wrong interpretation. I wasn't otherwise sure of where the PMI became relevant to posts made above.

Best Regards

Hudson

 

[JohnRBaker]

OK, a couple of things about PMI. To start with, NX has supported the creation of PMI objects (annotation, notes, finish symbols, dimensions, GD&T, etc.) for at least a couple of releases now. Now until now, those objects were similar to Drafting dimensions and annotation in that they are created by referencing existing aspects of the model (edges, faces, points, etc.) and like Drafting objects they will update when the model changes, but they will not 'drive' the model. However, the NX 6.0.x project I described will allow Sketch Dimensions to be 'exported' from a sketch and to be treated as PMI dimensions, only in this case, you will also be able to double click on that dimension and edit the value which will cause the sketch to update, all without having to open the sketch or even KNOW that the dimension came from a sketch. Of course, at the moment, only sketch dimensions exported to PMI will be able to drive the feature it references (but that will change in the future).

Now as to why I brought this up in the first place was relative to this idea that somehow it was both critical and strategic that NX somehow allow users to drive feature dimensions from a dimensioned view as contained in a traditional CAD drawing. While much of what we are doing with PMI will assist in the creation of drawings, the real consequence will be that drawing creation will become more automated and thus less critical in terms of being on the critical path in companies work-flows. In fact, many of our customers have indicated to us that their gaols are to move all drawings from the critical path and moving what drawings currently supplies to their work-flows closer to the data models themselves.

John R. Baker, P.E.
Product 'Evangelist'
NX Design
Siemens PLM Software Inc.
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

 

[hudson888]

Ahh! John,

The penny dropped!

Thanks muchly,

Hudson

 

[fighterpilot]

I've been reading the later posts and see this has turned into a discussion about the fundamentals of NX compared to other CAD tools. That's OK, becuase as I originally stated I'm moving from Pro/E over to NX4. As long as it remains civil I suppose the discussion will continue. Any comparison between the two from people as to how Pro/E accomplished something compared to how it's done in NX4 is beneficial.

Regarding drawings and not being able to show model dimensions. The inability to simply show feature creation dimesions is what really floored me about NX compared to Pro/E. If I, as a engineer/designer/model creator, am the one creating the parts and assemblies and I have built in a specific intent into the models then I really don't want to have my drafter, who in most cases was not me, to create dimensions for features using references I don't use in my models. If they were to dimension a feature from a surface that was not the same one I used to model the part then there is a possibility the actual machined part could behave differently than the model. At my previous company we probably spent more time up front than most companies in the early stages of product development figuring out how to manufacture something efficiently. We drove that intent into our model/assy creation. We also fully dimensioned, toleranced, annotated, and documented the model in the early stage as those tools were available in Pro/E long ago.

The "wall" between design and manufacturing was eliminated as soon as we started on Pro/E because we determined the overall design/manufacture/deliver process was shortened by doing so. However, the design process was a bit longer because of this collaboration. Initially management frowned upon this when we were not physically creating prototopyes on the shop floor as had been done in the past. When our first physical prototypes worked as they should on the first try, they were soon convinced.

Isn't the whole intent of CAD in general to shorten the ENTIRE design thru delivery cycle? In my opinion it is. If I want to drive intent into my models at the beginning stage because I know it saves me time on the back end then that's what I want to do. I'm not one to "cob" something together at the early stage just to have someone else come behind me and clean it up. That is not an efficient use of the tool.

At this point all I can do is ask questions of people who have used both tools and try to figure out the logic behind the new system. I'm not trying to slam either one, just make comparisions and figure out the best way to accomplish my tasks.

--
Fighter Pilot
Manufacturing Engineer

 

[NXMold]

I'm not trying to insult or alienate anyone here on the board, this is a fantastic resource. But objective facts are what they are, and from a couple encounters with GTAC plus a couple threads here, the recurring impression is that suggestions for change are unwelcome. I could feel the frowning glare through the phone from GTAC. Being that such sugestions arise from a real need to do work, either the user needs help understanding the software, finding a work around, or the software lacks something useful. I'm not the only one coming away disapointed with the impressions left, and to be frank my company wishes to find a suitable replacement for NX because we do not get the level of support availible from other packages with local independent resellers.

I really hope this observation does not impede the help I get here, you guys are making my NX transition much smoother.


"Interrogating an edge will return a radius and a diameter value. I know that isn't always particularly helpful in all cases, but it may be useful to keep in mind if you weren't aware if it in the past."

I didn't know that, it will help me in some cases. Too bad this is not availible in the measure radius dialog as well.

The PMI arrangement is interesting. We don't use drawings, machinists use a viewer and attach dimensions to the 3d model (parasolid exported from NX) in a way that sounds similar to what you are moving toward. I wonder if there will be (or is already?) a cheap NX viewer (or free, ideally) that would be apropriate for viewing and measuring NX models.

 

[hudson888]

Fighter,

Thanks for you well reasoned post. I'll put some perspective into what I mean by getting the most out of the system that may show that we may have different ways of going about things and perhaps that leads to different priorities, or stems from working in different types of industries.

My limited knowledge of Pro-E and other systems besides NX shows that many users are very well organized in their logical processes because it is within the scope of those systems the best way to work. When confronted with the flexibility of NX many find it almost anarchic and prefer to stick to the sketch based modeling that they're familiar with. If you were to tell most operators brought up in the NX way of thinking that they can't use some combination of feature based modeling, chunky solids, direct modeling and in short use it as a hybrid modeler in order to express their desire to model things the way that they think then they would reject the imposition of such limitations. A lot of this comes about because there are a preponderance of older Aerospace and Automotive users for whom the rationale for having UG latterly NX was based in the free form requirements for body, airframe or plastics design, none of which lend themselves readily to conventional drafting. For that type of application most of the time you wouldn't get the downstream benefit in terms of a payoff at the drafting stage regardless of how organized you tend to be. The real payoff is that CAM processes will be heavily relied upon in the manufacturing process and the future if not the present holds possibilities for applying tolerances and GD&T directly to the models so drawings may not even be required. We live in a world that insists that the model and not the drawing is the master.

A big part of the History of CAD and NX in particular was linked to these requirements, because both McDonnell Douglas and Dassault more or less simultaneously adopted 3D modeling technologies to develop new designs that benefited from the application of computing to their design process. Shortly after that they discovered that they were challenged to get this information to their suppliers and found solutions such as CAM combined with NC machining whereupon they made the extraordinarily astute decision get into the business of selling CAD to their own suppliers. Now that is just the opening chapter, perhaps not strictly the way the business arrangements were actually structured, but a fair accounting for what drove the decision making process that saw us progress towards the way we continue to use these technologies even today.

But you know what the good news is that those whose requirements are more aligned with conventional or structural engineering shouldn't think that just because they co-exist with this bigger other reality that means a system like NX shouldn't fully meet their needs. So if you want to work differently by all means do, and in doing so by all means lobby for better features to suit your needs. The bugbear for me is that what seems to get most misunderstood is that flexibility affords you the choice not to use all the tools if you either don't want or need to, whereas inflexibility doesn't have that benefit.

I'm a bit concerned that some of this may sound inescapably elitist so I'd also offer the thought that I've had my fair share of involvement with the straightforward nuts and bolts side of things from time to time and I'd like to offer a though in that vein. My thought is that while you may achieve a time saving benefit from inheriting the sketch dimensions if you take the requirement to think out of the drafters task then the benefit of collaboration with that other individual, a second set of eyes over the design, may be largely lost. I'd hate to see anything that comes about of a modeling necessity translate into an incorrect drafting emphasis by default if large expensive parts or safety issues are involved with the design.

Best Regards

Hudson

 

[acciardi]

Fighterpilot:

I agree with you 100% on spending time upfront deciding how datum structures and geometric tolerances be used to embed as much intent into the part as possible. We spent some time training our users (back in '94) on the best modeling practices with Pro/E. The enforcement mechanism was abuse and humiliation of anyone foolish enough to submit a lousy model, which worked fine

This ended up having a huge payoff in the ability to pass parts between engineers (typically an issue in history-based systems). We regularly pass designs off, or more often, leverage other peoples' parts into a new design, so it is very important to follow a good practice. For instance, we insist that all features have a descriptive name, because there is nothing more frustrating that seeing a model tree with 200 features named "Extrude".

You will also find that it is really not necessary to use Layers to organize the model, even if it's a complex assembly. NX's Reference Sets are actually a very good way to set visibility of construction geometry of components in an assembly (there John, I said something nice). I suggest you make an effort to use this tool.

We have carried our part design philosophy forward into NX. We use no primitives and our models start with three primary datum planes. So far we have also eschewed direct face modeling. We attempt to always sketch on planes, not feature geometry, which as you know tends to reduce interdependancies between features.

Passing the model off to a draftsman (draftsperson?) does require you to spend some time with them and explain the design intent. Alternately, they can play the model back to understand how it was constructed. I tend to do my own drafting, so I know how the part was structured.

One thing that may help is to construct datum targets right on the model (using sketches with points or small circles in them). This tells the drafter at least where to dimension from. You can add the sketches that define the datum targets to the "MODEL" reference set so that they will appear in the drawing, then view-dependant edit them away in view in which you do not want to see them.

Hope this helps...

Ed

 

[hudson888]

Ed,

If I respond to this it is just to contrast not to criticize. I'll start with just mentioning the background that you mentioned in the first paragraph because the rest is your experience from your perspective and I've certainly had enough to say so I think what you said is all good and I have no problem with it at all. In fact I have no problem with anything you said we just differ in how we use CAD. The point that I would make is how what you reveal in the first paragraph differs from the way we work and the reasons why.

We would differ in how we go about setting up datum structures. This is because we work in absolute most of the time in a CAD sense. In the design sense the datum intent might be to achieve consistent panel gaps or other appearance driven concerns. For that we can set up the part datums as a means to an end and vary them throughout the design process.

Our company would take a very dim view of any enforcement method that involved humiliation and abuse. You can't even tell a dirty joke these days for fear of offending somebody, and getting sued for it. So we have a program to mete out the abuse for us, politely of course. We make good use of checkmate to enforce our modeling standards, we provided some tools to take the drudgery out of organizing some of the data, and we use layers when we need to. We use two standard reference sets one for the Solid Model and one for the Faceted. We do have reference sets in the assemblies but we manage them by means of automatic updates. We eschew any more than that because of the three assembly deep reference set problems that we experienced in the past.

We have been through processes that involved best practices and found that what it usually involved was a generalized list of things not to do, which eventually got chipped away at by exception when no other solution could be found. We have since moved on to shift the emphasis to guidelines that talk about making the data maintainable, so they're all things that you should do, with only a few that you shouldn't. If you also want your designers to be creative and flexible recognize that telling them what not to do doesn't work. The smart ones reject any dumbing down mentality and leave for better paid jobs with your training under their belts. What we say is use the tools that best fit your needs, try to stay parametric, and document your data with naming and collecting it so that others will be able to follow your construction method. Using all the tools holds no fear for us and we have Mentors in the office who get help the newbies when they get stuck. The expectation is to generally raise the standards of expertise and experience to a fairly high level. We also accept that there are final models and studies and that the needs differ between the two, i.e. most initial layouts or studies need not be maintainable.

On straightforward engineering projects I always do my own drafting and I often enjoy it; money for jam . On automotive or aerospace projects never. I leave the GD&T to the experts and (try to) value their opinion. It is just that kind of different discipline I think.

In terms of the terminology that you were struggling with for a concept akin to copying other peoples' designs we tend to use the term "Benchmarking" as code for that. As in we benchmarked several designs and found this one mirrored our requirements, ergo we pinched it .

Where this leaves us is with two sets of requirements for different kinds of industries and applications. We also have different backgrounds in terms of how we learned CAD. To me NX is a toolkit which we paid for; I know how I want to use it and I post here because I'd like others to get the same satisfaction from their jobs. I don't think I could ever recommend changing it in ways that suit any narrow application of design when that would curtail the flexibility needed to make a broader range of possibilities happen.

I think John has a few tricks up his sleeve in NX-6 (no pressure...), that will hopefully make us all think differently about how we tackle some of these tasks. What I have seen and he has mentioned thus far makes me realize even hope for some better possibilities to integrate the design data into engineering processes that really by and large bypassed conventional drafting years ago.

Regards

Hudson

 

[acciardi]

Hudson:

Our industry backgrounds are somewhat different - as I mentioned earlier, there is virtually no surfacing used in any of our design work.

One of the ways we evolved our current parametric bias goes back to the early nineties when we first started using Pro/E. Before that, Unigraphics was the standard in our industry (minicomputers and disk/tape storage). We used UGII from about 1980 to about 1994. We dabbled in UniSolids a bit (you may remember this was an option in UGII v8. It was great for computing mass properties, but drafting had not quite caught up with the solid modeling capability, so all production work was still done in wireframe.

We had a pretty good idea of how long and what resources it took to develop a product, so when we switched over to Pro/E in 1994 (I believe UG was at V10 around then) we were shocked at how much more productive we had become. Design teams literally went from 6 engineers and 4 designers to 4-5 engineers doing all their own drafting. We have been at these staffing levels ever since. Our management of course still bitches about how long it takes to do things

SO we have tended to stick with what worked so successfully for us. Change does come slowly, as evidenced by how long it took for the market to embrace solid modeling. We are probably due for another major shift, as the basic technology as defined by PTC in 1986 is now 22 years old.

I will be very interested in seeing what NX6 has to offer.

Ed

 

[JohnRBaker]

...there is virtually no surfacing used in any of our design work.

Then you may be right when you say...

We are probably due for another major shift,...

...since I think you will find that some of the new capabilities that we will soon be talking about as the NX 6 launch gets underway (April 22nd is the day) will be something that you will want to pay close attention to. Now I warn you that your parametric background (this is NOT a reference to Pro/E, but rather the idea of creating everything with sketches and parametric features) may cause you to react with some skepticism toward some of the claims and even demonstrations that we will be offering, but as you alluded to above, there is a 'shift' coming in how models are going to be modified and updated in the future.

Anyway, keep your eyes, ears and your mind open ;-)

John R. Baker, P.E.
Product 'Evangelist'
NX Design
Siemens PLM Software Inc.
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

 

[acciardi]

John - you're such a tease. But you sound genuinely stoked, so I'm eagerly awaiting the rollout. Being basically a lazy ne'er-do-well bastard, I am always looking for ways to make my life easier.

It's been my observation that we've pretty much hit the limits of what can be done in feature-based parametric systems. UIs are improving all the time, but all the waste in the workflow has been squeezed out. CPUs have reduced regen time from minutes to just a few seconds, but you still have to struggle with part history and the challenges therin. IMHO, Pro/E did it to perfection in their 2001 release. Everything since then has been designed to help new users get up to speed more quickly, but there's really been nothing 'new'.

Here's to paradigms!

Ed

 

[JohnRBaker]

...but you still have to struggle with part history and the challenges therin.

Hold that thought please.

John R. Baker, P.E.
Product 'Evangelist'
NX Design
Siemens PLM Software Inc.
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/