Parametric Design - UG and think3 compared 4

[gciriani]

I've stumbled into think3 software (

http://www.think3.com),

and that company shows demos of parametric design, where the designer can easily modify the 3D. We use UG, and my chief designer tells me that UG can do the same. He also says that in our type of design (metallic watchbands, with up to 300 parts, of which 30 different part numbers), it does not make us gain time.

Do you concur with the two statements:
(1) think3 and UG are equally good for parametric design (ease and capability);
(2) parametric design does not make one gain time when developping a new product with 300 components (30 different part numbers)?

Regards, Giovanni Ciriani
Promotion Group +39-348-155-4029

 

[ewh]

1) It is difficult to compare the two based only on a web site. Only a "test drive" would allow one to answer either way.

2) I disagree. Parametric design can greatly enhance development vs non-parametric.

 

[cowski]

I have heard of think3 but never actually used it; I have been using UG since version 15 - I can't really give you an objective comparison of the 2. But I would like to make a few points:

1) If you are using UG you are already using parametric design - the question is "to what extent?". You would really have to go out of the way to entirely avoid parameters in UG (though I suppose it could be done). UG does have excellent parametric design capabilities. I would encourange you to get more familiar with the tools you already have before you decide to jump ship.

2) Never trust a demo. A demo has been finely tuned to a) work flawlessly and b) show off all the cool software features. Keep in mind you may not actually need the feature that made you say "wow", and said features may not work as advertised.

3) Carefully think through the pros and cons of switching design software. Here are a few cons: a) there will be lost productivity while the users get to know the new software b) do you reuse existing files or is every design brand new? since reuse is a good idea, what do you do with your existing UG files? if you are lucky there is a translator that (works flawlessly) to convert your parametric UG files into parametric think3 files - if not you can either remodel them in think3 (more wasted time) or import them in as step files (thus losing much of the parametric design). c) Do you use the 3d files just to make drawings or is there a downstream application that uses them (eg CAM tool paths)? will these downstream applications be able to use the new file format?

These are just a few issues I could think of off the top of my head, I'm sure there are more you must consider.

 

[cbrn]

Hi,
I used ThinkDesign and now also a bit of UG; though I can't be consideredan expert of neither of the two softwares, here are a few considerations:
1- be very careful when you consider shifting from one CAD environment to another: you will first have to face the "legacy" problem: how will you convert / reuse your files? UG and ThinkDesign really don't like each other very much... Then, the learning curve: how much time will it take to regain the same productivity you have nowadays?
2- both UG and TD are hybrid modelers, allowing to work parametrically or explicitely as needed, and by mixing wireframe, surface and solid geometry as well. TD suffered till v.8.4 of a general lack of robustness: because of that, you could easily "destroy" the history tree of your model and never be able to regenerate it correctly any more; Think3 guys seem to push a lot over their new FF functions (global shape modeling, GSM), which I agree are very powerful (though not as much as UG's ShapeStudio, mainly because they are a bit confused and confusing) but they desperately try to silence TD's weaknesses in the pure-solid parametric field. Maybe something changed with more recent versions, which I don't know. I don't know if it's your case, but whenever you need to drive geometry by math laws or expressions, UG is far superior to TD (though not as much as Pro/E, I think)
3- I do find that parametric modeling makes one save much time especially in the basic-design phase, which is crucial in the product's lifetime. But from that point of view, UG is already very strong, you would gain nothing or very little from a software shift.
4- if you use post-CAD facilities (CAM, FEM,...) then be careful with ThinkDesign: it's a "closed" system, no CAM preprocessor, no embedded FEM, etc... (as regards FEM, of course many of the FE packages are bi-directionally compatible with TD)

Only my 2-pence thoughts, of course...

Regards

 

[gciriani]

Thanks Cbrn,

I agree that migration from one CAD application to another would be a nightmare. I wanted to understand which software had better parametric design. And that you clarified it for me, thank you.

I'm interested by your further comment that: "parametric modeling makes one save much time especially in the basic-design phase". I'm not part of the design crew at my company, but my chief designer tells me that parametric would waste his time when he is pressed to make quick changes for the customer, or is pressed by deadlines. He asserts that parametric design takes a longer time to set up the initial design. I would appreciate the details of your rebuttal.

Regarding Think3, I'm attracted by the PLM tools that the company offers, but I wouldn't switch CAD for that.

Regards, Giovanni Ciriani
Promotion Group +39-348-155-4029

 

[cbrn]

Hi,
well, parametric design effectively needs more time to "set it up" because you have to think much more about the "best way to do it": even when you model explicitely, you have more than one way to achieve what you need, and these ways are multiplied when using parametric design, because for any topology you have several param patterns possible; among them, you have to select the one that will make your model "most clever". Two things can happen when trying to modify a parametric model:
- the modification is made possible by the param pattern -> the modification is EXTREMELY fast (only time to modify param values and regenerate - in this case, most of the time it's better to turn on "delayed regeneration" because otherwise modified values could collide with not-yet-modified ones). This is the kind of modification I usually perform. This occurs especially when new models have to be created from similar ones.
- the modification is not possible with the current param pattern -> in this case, probably re-creating the param pattern is longer than re-creating the geometry from scratch, though it's not a rule of thumb; by using the tools of UG it is hardly possible that you would be pushed into an "impasse". The only case where param re-creation is VERY hard is when you have to operate on a model made by someone else (thus, with a logic you really don't know...).

Regards!

 

[gciriani]

Hi,
very interesting what you say. But if

"param re-creation is VERY hard ... when you have to operate on a model made by someone else"

this goes counter the goals of team work, process independent from the operator etc. Then I have to make sure to standardize the rules for parametric modeling.

Regards, Giovanni Ciriani
Promotion Group +39-348-155-4029

 

[cbrn]

Hello,
yes, the assertion seems a bit contradictory... In fact, I meant that things get difficult when you know absolutely nothing about the logic, the intent, with which the model was created. In a team (if it's really a team and not a collection of independent workers each of whom is responsible at 100% for a different design), it's unlikely so: people talk to each other, they exchange opinions and "how-to's"; here at my workplace, when I have to modify a part created by other designers and I don't understand a logic, simply I ask for info/advice.
Anyway, probably parametric modeling is more difficult to "understand", and I think that's what your chief designer refers to when he says that param design is not so efficient: when you must finish a part modification "before yesterday", and the base part geometry is sufficiently simple or easy to understand, taking time to understand a param pattern may take longer than explicitely modify the part geometry; but this is true for a single modification pass: when you have understood a param pattern, you are ready to modify it as many times as you want, and in the case of 3, 4, ..., 10 modifications, the total time needed will be surely less than if you were to modify explicitely each time!...
That's why I believe that param modeling is efficient for the first phases of a project development, when the topology (the general geometric shape, the final intent, etc) is assessed but you have to make repeated adjustments or changes in order to "enginneerize" the whole. Probably, for the very beginning (pure research, conceptual design), explicit modeling is faster.

Regards!

 

[HellBent]

Again...if you're designing in UG you're already "parametric". And as far as PLM goes....look into UG's Teamcenter if you need a PLM solution.

 

[gciriani]

HellBent, thanks for the suggestion. I'm told though that think3 is the leader in PLM. Have you had any direct experience in comparing UG's PLM vs other companies' PLMs?

Regards, Giovanni Ciriani
Promotion Group +39-348-155-4029

 

[cowski]

Who the PLM market leader is depends heavily on who you ask. Out of curiosity, I did a quick google search on "PLM market leader" and from the first page of results I learned that PTC is in the "leader quadrant", UGS "leads in market presence", and MatrixOne is the "PLM mindshare leader". I'll leave it up to you to cut through the marketing hoohah.

 

[J-C]

I would think that think3's PLM is a bit lower end, while UG PLM is more high end. Depends on what you need. High end usually means harder to setup and longer rollouts but greater capability, while low end is the opposite.

Jason

 

[HellBent]

Cowski is correct...there are about a million PLM solutions out there. Trick is to find the one that best meets the needs of your organization. I've only used Teamcenter and just like UG it can be overwhelming at times, but will do just about anything. It also has modules specifically designed for interfacing with manufacturing, production, and various accounting softwares. It can also be heavily customized if necessary. I also have a good friend who works for PTC's PLM division. They specialize in custom written PLM solutions to fit your company, regardless of current softwares. I would consider their products to be very "high end" and very customer specific. As far as which one to choose...I would recommend looking into the one that is designed to work with your CAD software first. If it can do everything you need you could avoid a lot of the problems that come along with customization and the continuous traslation of data.

Take care...