Some surfaces behaviour makes me confused

[Jacoolo]

ok, james, no problem


there are already some Curvature analisys so just unhide them

since yesterday, I have updated it a little, so now I could achieve what I wanted. However it was due the fact I change the sequence of quilts. When You open it You will see. Part is in WF 2.0 format.

Second - I change the way round is controled and know I am sure it has much more smooth transitions. Take a look

View attachment 3863

View attachment 3864

View attachment 3865
here is a part


2007-07-11_033200_mouse4.prt.zip

 

[Jacoolo]

still I am waiting fot Your(all) response considering the way of:

*checking curvature of sketch entities without converting them into spline
*adding\deleting controls points in boundary blend(till this time I found out that it is possible by direct change of sketch to spline which affects in deleting control points, or dividing entity in pieces what ends in adding points. But what about curves that come from Intersection?
*manipulating a "curve through pionts" in such a way to set the tangency for each of points(start and end) not only to one ref(curve or surface) but to two, i.e - in start point I want the curve to be tangent to two planes -surfaces, is it possible?

I am really counting on Your(all) response. If the answer is available somewhere in net, send a link. I find a time to read it))))

and today another problem with new case. I hope Bart from Design-engine won`t be angry if I use screens from de page to make models based on them(they look really chalenging for beginer as I am))))

View attachment 3868

View attachment 3869

View attachment 3870

and here is a problem

View attachment 3871
yeah, strange look. I bet it could be solved easy. I am sure I will dig it till I find it. But now, I do not know where the cause is? Both curves have positive curvature radius. I can not set the perpendicularity in second curve because the curve in second dir is not perpendicular to the plane. Sounds like maze? Indid it is))

 

[design-engine]

No.... never anger for me. I am here to help everyone. I like meeting people and if I can help then I made a new friend. Of course if I come to <?:namespace prefix = st1 ns = "urn:schemas-microsoft-comffice:smarttags" /><st1:country-region w:st="on"><st1lace w:st="on">Poland</st1lace></st1:country-region> I expect a couch and a beer! <?:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" />


I built my chair with just one surface in an effort to utilize and show strengths of ISDX over Pro/SURAFCE with a Curve On Surface example. Your model is broken up into multiple surface patches.


I am not sure what is your problem from looking at the screen shots... I think your big problem is your using 3 part boundaries. On Page 2 of this thread you are clearly using 3 part boundaries and getting the globing affect. This is bad practice because you are not able to get the tangencies the way you may desire.



Edited by: design-engine

 

[james.lynch]

*checking curvature of sketch entities without converting them into spline<?:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" />

Do you mean G2? If so, this cannot be achieved within sketcher, it can however be approximated with the use of the

 

[james.lynch]

I built my chair with just one surface in an effort to utilize and show strengths of ISDX over Pro/SURAFCE with a Curve On Surface example. Your model is broken up into multiple surface patches.

Hi Bart,


I'd be interested to see a curvature plot of your one surface chair - I would expect to find singularities at the top and bottom apexs. (unless you over built the surf and cut it back, esentially haveing little control over the edges...


James

 

[design-engine]

yes. over built.


James, I want to come to England again and speak at the Pro/E user conference there. Have you ever been to a Pro/E user conference? I did a bunch of classes in London and middle england in 02'. Triumph - British Aerospace - a few ID firms I want to do a track day on a race track or two in England.


About that model ... dont you still have the login and password to our tutorials?
Edited by: design-engine

 

[Jacoolo]

thanks for response

to Bart:

About coming to Poland and a beer - well for sure it worth. The Pro\E situation is unfortunately not so good. The use of this software is growing, but it has to take time to achieve the level of countries like Germany(strong Automotive industry). However, you can find in Poland good beer(Lech, Tyskie, Zywiec, Spiz in Wroclaw), and what I am really proud of - prettiest girls in the world. I should not say that(I am married and my wife could interprate that wrong, I hope she does not))) but it really is so. Come to Wroclaw. I will take You for a round in pubs in the marketplace(more than 600 pubs) and you will see what I mean.

Considering stuff surf related - this is what You mean by 3 part boundries(see below)

View attachment 3874

if so, you have right, I use it widely

this is what I wanted to achieve yesterday with my last problem

View attachment 3875


I am not sure what You mean by globing affect(effect)? But I realize that this 3 part boundries disturb the grid flow between patches(quilts)

View attachment 3876

However in this moment I can not imagine how could I do that without 3 part boundries.

*to James

"Do you mean G2? If so,.." yeap this is what I meant. I missed "curvature" constraint in Pro\E sketcher, or maybe it is hided?

"This is where the care and attention and forward planning comes in." - nothing to add. I think this is main thing to be learned in Pro\e surfacing. Not a tools - thyer relatively easy to use. But way how to use them, a way how to predicts the effect of such a use.

*"To create a curve which is tangent to two planar surfaces, intersect them and make you curve tangent to the intersect." - it looks like power of curve through points is limited.

Finaly this is what I achieved yesterday with the seat.

View attachment 3877

one can say there are no changes. Well they are. For me transitions are much more smooth now. I`ve done this by using conic arc in important - main curves. They do really nice effects.

 

[james.lynch]

yes. over built.


James, I want to come to England again and speak at the Pro/E user conference there. Have you ever been to a Pro/E user conference?

I was at Nashville and Orlando a few years ago, I'll go again next year!

About that model ... dont you still have the login and password to our tutorials?

just tried the login yo ugave me again, no longer works tho'...

James

 

[design-engine]

I saw the Moto GP races today near Berlin.TV not in person. I want to go to Europe again.Hats off to girls in Poland!


http://www.proetools.com/tutor l: student p:sushiwabi


I am working on Pro/CABLE tuotorals and Maya tutorials this month.


I meet you James in Tenn.at the International confrence. You had the battle ship and I was buying beers for the Ford Pro/E guys.


Anyone up for it, Ill be in Atlanta August 4th and 5th Road Atlanta Raceway (trackday) camping. Great for the kids too!
Edited by: design-engine

 

[Jacoolo]

Bart, sending the link with login and password you increased rating for this topic. Now is more than 1000) )))
Luckly You did not put sentence - email me to get a password. In this case it would be tons of thread like - me too... ))))))))))))

well, honestly, I check this link and the page, and I can`t wait till I have time to play with some of those examples

Now I am dealing with bike saddle, and it is more challenging than the previous examples I did - mouse surf, and seat

View attachment 3882

btw

James - what about my model? Was is it worth for You to check that out? I still do no know how to figure out the smooth transition between round and back surf?

View attachment 3883

What is the weahter in Ireland. Here in Poland after two weeks of rain we got temp up to 36C. I hope it is gonna be the same in August when I starting my holiday))

 

[Jacoolo]

yeap, I reached that point. Finally.

This model is based on Logitech mouse I have in the office. It does not fit realistic well, but,...hm, it looks simillar and that is what I wanted to achieve.

View attachment 3894

View attachment 3895

View attachment 3896

View attachment 3897

Yeah, I know, I missed some details, but it was not a goal to include all of them. In fact, I have not still reached the point I had wanted to with grid. I used 4 part boundary, instead of 3, but it is not smooth enough for me. I demand to much from myself )))) In genral the model was rebiuld from scratch from that one I attached few days ago.

 

[james.lynch]

I meet you James in Tenn.at the International confrence. You had the battle ship and I was buying beers for the Ford Pro/E guys

shur don't I remember ya well! but since you were asking was I ever at one I wasn;t sure if you remerbered! thanks for the updated login.. some nice models there.. I must spend a few hours on them next week when I get a bit of time!

James

 

[james.lynch]

Hi Muadib. the model looks great! CONGRATS!!

you are definatley progressing!
My company actually design and create most of the Logitech products and all in ProE.

Mice really are quite challenging! especially the higher end ones. we actually physically scuplt them from clay or modelling foam first, then cast them in a hard resin, then hoan the surfaces/flow/design of the model all by hand first. we then digitise this model and build our ProE surfaces over the scan (generally to with 0.1mm). it takes a bit of time (maybe 50 hours to capture a mouse of similar styling to to what you have show) to create a full master model which is good enough for production with all gaps and B-surfaces included. A more complex mouse master model (say 1500 features) may be twice that.. I'm working on one at the moment.. some really nice surfaces!! keep an eye out on the shelves in FY'08!

Keep up the good work!

James

 

[Jacoolo]

thanks james

Yeap, it would be nice to have one of Logitech mouse(G7 or MX518). I have A4 Tech X718, it is optical. Well it is not a best of the best, but it has possibility to change DPI for X and Y axis, and this is what I wanted for first. I play Counter-strike once or twice a week with guys from previous job, and this mouse makes playing better. But, with laser mouse, with better software(in X718 you can only go down with DPI to 200, and max to 600 I suppose), the results in game would be even better.

The procedure You use to develop mouse sounds impresive. However I tought all is done from start in Pro\E, with reverse enginnering. 1500 feats is a lot for such small size device as mouse housing is. I work now with back housing of 46" LCD TV. With almost all neccesary rounds and drafts I have about 2000 feats. But this model is more than 1 m long, and 80 cm height.

I think I buy new mouse when will be buying new PC. It means - not sa fast My kids soon will be more than 6 years old, and I am considering to buy Nintendo WI or X-box to play with them. So new PC has a little chances to be bought.

 

[james.lynch]

thanks james

The procedure You use to develop mouse sounds impresive. However I tought all is done from start in Pro\E, with reverse enginnering. 1500 feats is a lot for such small size device as mouse housing is. I work now with back housing of 46" LCD TV. With almost all neccesary rounds and drafts I have about 2000 feats. But this model is more than 1 m long, and 80 cm height.

To capture a form accurately surely size doesn't matter? be it 100mm or 100 inch or the size of a car or whatever? Surely it's down to the complexity of the surface and how well you want to capture it?

 

[Jacoolo]

you are propably right, if small feats are needed then it is possible that a total number of them is big, even if total size is small.



It is just hard to imagine why is there so much feats needed? I suppose
the reason is because there is a lot of small surfaces patches(quilts).
In fact I have not yet played with scan tools, and reverse modeling
approach with Pro\e



Could You tell me where most of those feats come into? Drafts, rounds, cavities?

 

[james.lynch]

well, the thing is that every round is built, every draft is swept, every curve is constructed and very little is a resultant!


It's very simple to create a shape which looks like it's supposed to, but to actually create something as it is actually intended to be is considerably more difficult!


I probably create things in a slightly feature intensive way but the result is a VERY robust model which is easily modifable and easy to navigate, regen times also are not very long, abuot a min fro top to bottom, but the thing is that generally I'm only working on a set of features... so no problem there!

 

[Jacoolo]

james how model could be ease to modify if it consists more 1000 feats?

I am sure that there is a moment, maybe it comes in the middle of model tree, that the changes to core feats from the beggining of tree, are not possible, or hard to provide - just two cents from my life.

Or you mean changes as only modifing the dims value?
sweeps for drafts - that sounds really impresive.

 

[james.lynch]

well the tooling on a complex mouse is really quite dificult and often simply cannot be drafted usind the draft tool. every blended surface is then builtoff swept drafted surfces.


Withcare and good reference control, very rarely do I say I cannot (in fact never ) modify that. it takes quite a bit of forward planning but it can be done!

 

[Jacoolo]

Is it widely based on external references? Do you divide the geometry on different models each one for another scope - one for outer geom, one for inner, another for cores?

Once, in previous job, had been new project started, and I was asked to guide designer with most efficient and robust way to deal with it. It was new valve. Big housing(as for air brake valves), with the need to have all in it(drafts, rounds, till this time it was up to supplier). So I decided to devide it in small models, like I mentioned earlier. Unfortunately I couldn`t see the final result. I quit that job, but I was informed it was very good approach for design, bad one for FEM. The body was solidified in special model, and one has no control on drafts and rounds. Guys from MES wasn`t happy. Pro\M always failed with such big and complicated model. Hence, it took quite a bit time to prepare model for FEM calcultions. I run some investigation before that to put rounds and drafts in seperate model and then somehow merge it with main geom, to suppress such features easily. In fact I never went so far with it to have intresting and valuable results. Just idea to play with.

So, I imagine that it also could be with such a model as yours, when there is a need to simplify geom for calculations. But I doubt You run FEM, rather plastic anlaysis(PartAdvisor?). I am also curious how You avoid problems with merge and normals, which change the dir for quilt if You change the sequence or surf in merge inside.