Summary of responses to Merging very complex assembly question

From: Steve Stewart (steve@protogenic.com)
Date: Thu Dec 18 1997 - 01:46:40 EET


Hello, all -

On Thu, 11 Dec 1997, I wrote:
-------------------------------------------------------------------------------
>We have a very complex assembly in Pro. You can think of it as a fan with
>about 80 curved, lofted blades. Each blade has, as its root, a barb-like
>affair which slides (apparently line-to-line) into the hub. Similar
>attachment to an external ring.

>So, we have about 100 solids which, when made into a single STL file of
>acceptable quality gives us a 20 Mb file. The size of the file shouldn't
>really be a problem - we've done big files before. However, this beast
>took 45 hours to slice on a 250 MHz Indigo2 with 128 Mb of RAM. Okay, we
>ate that. Now, however, we are getting an error on our SLA350 which,
>according to 3D Systems, means that features are too close together for the
>software to distinguish between them.

>We theorized that, as we are drawing shells in the .STL file which are
>immediately next to each other, if we could turn the 100 solids into one
>solid, we could eliminate this problem. I haven't got a huge amount of
>experience in Pro Assembly, but I found how I could merge parts, thinking
>this would produce the single-solid part I desire. Unfortunately, it
>appears that, even though I can pick all of the blades and tell it to merge
>them to the hub as the reference part, Pro does each merge separately,
>stopping and asking us for a "done" before it goes on to the next merge.

>The question: is there a way to tell Pro to merge all of the solids in an
>assembly into one part, in one single step?

-------------------------------------------------------------------------------

There were many responses.

The winning solution for us came from Chuck Alexander at Solid Concepts Inc.:

If I understand your problem correctly, you are getting an error, in build
on the SLA-350, because you have multiple shells in a single stl file which
have coincident geometry and that you are looking for an quick way to
boolean union the components together.

If that's the case, then I can't answer your question directly, but I may
have an alternative.

You could combine shells selectively into multiple stl files. For instance,
every other vane could be put in 1 stl file, the rest of the vanes in a 2nd,
hub and other geometry in a 3rd and so on. You would just have to make sure
that you didn't add shells, in a single file, which have coincident
geometry. 3D's software would then handle the separate stl's as independent
parts and you shouldn't get the error when building. Also you may see some
decrease in slice time, because most of the time spent in slicing large
amounts of triangles is spent in sorting the data, and you will now have
multiple, smaller, stl files.

This was an elegantly simple solution and solved all of our problems in a
matter of minutes! Thanks, Chuck!

-------------------------------------------------------------------------------

Another great idea if you wanted to truly do the merge in Pro came from
Stefan Knoch at German Aerospace Center:

there is a way to merge all parts of an assambley to one singel part: assemble
one additional part (with only one feature: a coordinate system to make your
constrains for the assembley process). Select this part to perform the Merge
preocess to and select all the parts of your assembley as reference parts for
the Merge process.I did this several times and it works well. The only
disadvantage of this funktion is, you will loos all your color settings of the
single parts, means your "merge-part" will appear in white.
Hope it helps.

We didn't try it, but the Pro/E gods in the offices above ours (CAD
Potential) thought it was the best idea.

-------------------------------------------------------------------------------

Other ideas included:

Running all of the .stl files separately, which would have been prohibitive
in terms of labor.

Merging all of the files in SolidView (might work, but we were getting some
pretty huge files as it was).

Creating all of the blades as a single part, merging that with the hubs. I
didn't want to do this, as the assembly came to us from a customer and I
didn't want to risk misinterpreting.

Creating a Part in Assy Mode using Copy Surf, Quilting each Copy Surf from
each blade and the hub, then creating a Solid from Quilt to obtain your
single Part Protrusion. I'm pretty sure I don't have the horsepower to do
that.

Pulling up the blade part and use the mirror feature command to create a
continiously bladed hub that exists as a single part. Again, I'm pretty
sure I don't have the horsepower to do that. Also, it would be a mirror,
so the blades would go the wrong way, no?

You could translate the files to UG via STEP (the file sizes would be a lot
smaller). Once in UG you could unite all of the solids together -UG does
booleans much quicker than Pro- remove any parameters (keeping file sizes
small) and be done with it.

Using surfaces to trim off the "non-blade" portion of the blade, making
copies of each blade. This one was, I felt, too likely for me to make a
mistake in the customer's design intent.

Thanks a million to all of you! This was a great learning experience.

Regards,

\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_
 \_ Steve Stewart Phone: 303-252-0212 \_
  \_ protogenic, inc. FAX: 303-252-0223 \_
   \_ 1490 w. 121 ave., suite 101 WebSite:www.protogenic.com \_
    \_ westminster, co 80234 e-mail:steve@protogenic.com \_
     \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_



This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:40:56 EEST