Thousands of holes

I have no idea of the physical specification of your system, but I once tried to pattern only a couple hundred holes with my system (1GB RAM,

3.0 GHz CPU, NVIDIA Quadro 1300 FX), it was unusable. The only thing I can think of that _may_ help is to enable the "Geometry Pattern" check box in the pattern property manager.

I wouldn't put much hope in it though...you are asking a lot IMHO.

Dave

Personally, I would do one of the following, rather than make immense patterns:

- create a sketch of the area which has holes in it and cross hatch it, use a note to specify the holes and pattern

- split out the face or make a trimmed offset surface and put a color or texture on the surface that looks like holes if possible.

- if you really "have to" have the holes, make configurations with and without. You wont want to be the one that makes a 2d drawing of a part with holes like that.

- there is a new function in SW06 which will fill an area with a pattern. Pretty slick for irregularly shaped patterns.

If all you need is the appearance of holes, you might apply a texture to it. Or, what about a sketch - it might not carry the overhead. I don't know - never tried it.

WT

MG,

You might want to try creating a sketch containing the hole centerpoints.

This sketch could be used to pattern your holes. For a lightweight version of your model, simply suppress the pattern, and show the sketch.

Hope this helps.

CG

If you have to make the holes for some reason you might want to run the Patbench benchmark to show that there are limitations.

Here is my secret.

We make circular patterns of 10K holes or more on single parts. My situation is a round part that I model one pie shape, which I try to keep holes in this pie section to MG,

This would be quite interesting to benchmark - Paul???

WT

I have been contemplating something like this. Since memory runs out during the generattion of a pattern and not when the pattern is finished it may be that some scratch memory is used and released. What has been discussed here suggests that there is a way to greatly reduce the amount of scratch memory used. If I can figure a way to do this I'll roll it into a benchmark if only to show SW it is possible.

Well, what about just creating some parts with the different methods and look at the rebuild statistics. It wouldn't be quite as automatic, but would give a comparison of the methods.

WT

That is what I was contemplating.

First brute force, then pizza slice.

I could never find it but there was somebody who posted problems with PC board layouts with thousands of vias. I did build a test for that. Looks kind of cool when done.