Multidirectional pattern...

[AdrianHomutescu]

Hi


I have a question regarding patterns... Pro/E allows patterns on a single direction or on two directions (first and second). Still, in the case of patterning a complex group, need can arise that more directions arenecessary to complete the model. What I tried was to model a thorn, rotate it around an axis to obtain a thorn corona, pattern the corona along a spiral, and then rotate the corona laden spiral about an axis. That would require a three direction pattern, and such model is possible and rational. Four, five and so on direction patterns can be imagined, still I could not find any resource on such a topic.


Does any one know how to pattern on more than two directions? I will appreciate greatly any suggestions from you out there.


Thanks for reading,


Adrian

 

[Dell_Boy]

You can pattern with up to four independant repeats

Create a two repeat pattern in the normal way.

Group the pattern, then pattern the group.

If you want to use the same dimension to repeat (double stepping), you either have to do it by making the first direction the same as the second at either the pattern or group pattern level.

The picture shows a relatively simple 4 repeat pattern of the bottom left squarewhich has been double stepped in both directions


View attachment 1323


however there was a small amount of skullduggery employed to get the x and y location dimensions stepping in boththe x and y directions.


Use of additional tactics such as duplicating entities in sketcher and feature/copy/dependant can also make it APPEAR you have up to six or eveneight repeats without much trouble.


Getting back to your original question though. From what I can understand of your requirements, you probably can do the 3-way pattern. Just make sure you do two repeats at either the Pattern or Group Pattern level


DB
Edited by: Dell_Boy

 

[AdrianHomutescu]

Thanks for the hint.


Still, it's not quite the thing. I need to pattern a larger group, made of one solid feture and several references. I want to pattern in three directions at the same time.


1. A thorn by an angle (forming a corona)


2. A point (corona center) that travels along a curve by its position relative to one curve end.


3. The3dcurve(defined by equations) thus filed with coronas about the COORD SYS created especially for it.


That is three completely different directions. More complex patterns can be imagined. Too bad Pro cannot manage them.


Adrian

 

[Dell_Boy]

Like I said it was a "simple pattern" purely to show the principle of a
quad repeating pattern. In this case the pattern only exists in four
degrees of freedom (x, y, length,width). You can easily extend it to
include others.



You can have more than two degrees of freedom in a single pattern but
you can't have more than two repeat values. Hence you combine an
ordinary pattern with a group pattern to get up to four repeat values.



It is perfectly possible to pattern a group consisting of one or more
base features and one or more child patterns. Just make sure the child
patterns have appropriate references.



For example a spiral staircase involves one repeat value with 2 degrees
of freedom z and θ. You then group the features making up
the staircase and use a group pattern by table to place several spiral
staircases thoughout the floors of a building (another 3 degrees of
freedom, 4 if you include start angle) and this total of six degrees of freedom has only 2 repeat values.



Another thing to remember with multiple reference patterns is that it is
generally more reliable to pattern the youngest grandchild FIRST



If you can achieve the patterns one at a time you should be able to
combine two of them into a two directional pattern. The patterning
along a curve is potentially the hardest to do.



Is there any chance of posting a picture of what you are trying to
model because it is highly likely that what you want is achievable.





DB


Edited by: Dell_Boy

 

[AdrianHomutescu]

OK.


You have the picture below, showing the model tree also.


The pattern was created on two directions. I need a third to wrap several such thorn spirals around an axis. That would be the angle of the CS1 Coord Sys. The thing is there is no third dimension provided... And Wildfire cannot pattern or group patterned groups.


So, the only thing I can do is assemble one spiral and then copy move it several times. I wanted to somehow perform the same operation in part.


Regards, Adrian.








View attachment 1327

 

[Dell_Boy]

I have had a little time to think about this now.



First thoughts are that I would probably pattern the thorn around the corona and group these features together.



Next I would create a group pattern of the coronas along the curve in
the first direction and rotate the curve in the second direction.
Provided you do not have unnecessary groups within the structure, in
principle this SHOULD work.



A workaround would be to use feature/copy/DEPENDANT but with a modified
angular dimension to rotate the thorns around one at a time. Break the
groups but not the dependancies. This then gives you the possibility of
using ref pattern to a two directional pattern of datum curves to
generate the field of coronas. Alternatively you may want to try
re-grouping dependant copied features and then patterning that group.





DB




Edited by: Dell_Boy

 

[Dell_Boy]

On second thoughts it is more complicated than that and the scheme that SHOULD have worked probably won't.



You are not just patterning features in multiple directions, you are
trying to pattern a group of a pattern of a group. Furthermore the
corona appears to need ref patterning to the datum curve as well
patterning along the datum curve.



To me this says that the pattern along the curve needs to be grouped
with the curve then patterned as a group. What may upset this is that
the corona itself is already a group.



I still think it should be possible to achieve your desired result, but
I will need to give it more thought and do some testing because in
reference patterning of groups the order of operations and the
references you choose can easily make the difference between success
and failure.



Interesting challenge





DB



Edited by: Dell_Boy

 

[AdrianHomutescu]

Not only interesting!


It is worth that such a question be addressed to PTC in order to perfect the following release of Wildfire with this option of patterning groups in "n" directions, not just two. They have the software engine already so all they have to do is perform minor changes to the interface! Rationally thinking, this must not be that difficult for them.


And provided you have proper computation capacity, incredibly complex patterns would be at hand for everyone!


So thanks again for your interest !


Adrian

 

[Dell_Boy]

I am not sure you why didn't model the corona as a part and then pattern it spatially in an assembly which would have been much easier to do and several orders of magnitude faster to regenerate,but that would have spoilt the challenge.

I am almost certain that it is possible to pattern in the manner you intended as I have just completed what I consider to be a reasonably analogous 2D pattern in about 10 minutes but as expected, the constructions, references and order of operations need to be well thought out

View attachment 1333
The core of the star is referenced to a single datum curve. The flares are patterned around the core of the star. The datum curve, core and flares and other necessary datums were grouped to form a single group as previously suggested.

This group was patterned along the datum curve in the first direction and the curve was patterned around the galaxy in the second direction, again as previously suggested, for a total of 357 features.

There is not much you can't do in Pro/E; you just have to know alternative methods of achieving the same results.


DB

Edited by: Dell_Boy

 

[Dell_Boy]

While I was at it I thought I would put in a fourth repeat value because that is the theoretical limit. While I can't see an obvious way of adding a third repeat to the "star" group it is possible to add a second repeat to the flares prior to grouping as shown below.


View attachment 1334

DB

Edited by: Dell_Boy

 

[AdrianHomutescu]

As I said, it's interesting to know how exactly can Pro/E be pushed further. It's not just how you do a certain model, it's how you harness the 3D modeling as a Pro!


Well, I did not model the corona itself as a whole, because I intended to have a variable number of rays/thorns. I also intended to do the same for the number of spirals... Too bad it wasn't possible in a similar way. because of software limitations. I did an assembly copy, as I wrote in my third post and as you suggested in your last. Still, it's not as editablwe as I intended it to be.


Check below! You might have become curious to know what I was up to!


All the best! Adrian

 

[Israr]

I would suggest you use MAYA for this kind of stuff.


I took 35 hours to model a bird in ProE and it took me only 4 hours to do the same in MAYA.


Israr

 

[AdrianHomutescu]

I might just do that if I had time to teach myself Maya...


Until then...

 

[Dell_Boy]

Just for completeness, I did manage to get a better, true fourth repeat into the pattern with a few extra degrees of freedom to boot.


View attachment 1335

DB

Edited by: Dell_Boy