My first post here, but have been reading for some time. I would really appreciate some advise on how to solve the problem I'm experiencing.
After 3 weeks of trying I'm about ready to give up - although that is not really an option with a paying customer waiting for my work.
Here is what I am trying to do:
I want to make a skeleton model of the top surface of a part. (I'm sorry, I am not allowed to tell you what the function of the part is)
The top surface is very complex. The main features are:
- a hexagonal array of cuts - array has almost 10000 copies
- almost as much protrusions ("burls")
- a (sort of) spiral shaped extrusion
- some features on the edge.
The picture 1-model_OK_hexsize-40.jpg gives an idea. The only problem is: the hexsize in it is 40 mm; I want it to be 3 mm. This means a lot more pattern instances. ProE seems to have a promblem with that.
I tried several approaches.
created several surfaces, merged them into a quilt
- hexagon quilt is created in a separate skeleton model.
- Burl quilts is created in a separate skeleton model.
- A third skeleton model uses external copy geometry features to import both surfaces
- All surfaces are merged into 1 quilt
The advantage over A: I can have several computers do the work simultaniously since creating such a complex quilt takes many hours.
- I started with a solid disc
- made a hex cut, patterned the hexagons
- built all the other features as solid extrudes
- started with a 2d disc
- cut out all hexagons
- added the spriral
- thickened into a solid
- did the same with the burls
This is the model I attached to this post
In each case, I started with a limited number of hex cuts, either by increasing their size or by limiting their number in the pattern (added a second contour inside the pattern area).
Each method worked.
I then modified the pattern to the correct number and size.
Several things then went wrong:
- all of a sudden, some features start to fail. Quilts won't merge anymore, cuts won't regenerate (2-failed_burl_hexsize-35.jpg).
- some instances of a pattern show up at completely the wrong location (3-hex_pattern_problem1_hexsize-3.jpg).
- the amount of time it takes to regenerate increases to "rediculous" imho. Two days, just to make the hex cuts in a 2d plane!
Maybe some system info would help:
Dell precision T3500, X5650 Xeon 6 core 2,67 GHz, 6 GB Ram, 64 Bit, Win 7, Wildfire 5
Does anyone understand why this is not working?
Or have any tips/tricks/ a better way to model this?
Did you say 10,000 cuts and 10,000 protrusions? That's a LOT of geometry! You just may be stretching the software beyond its capability. I had a similar problem several years ago trying to create a huge 3D repeating scaffold network, and had to give up. If not a problem with the geometry engine itself, it could be as simple as over-filled tracking registers or something of that nature. I don't know. It would be best if you could get some insight from someone at PTC. Have you contacted their support people? I don't know for sure, but I also suspect that the Merge approach just adds to the difficulty for the software; that's a lot of complex serial solving time. Without knowing more about your challenge part, I can't tell, but is there some way to simplify its representation?
Cant really help you directly but have a look at these tutorials, they might help
You may also need to look at the ProEs accuracy setting
Thank for all the replies guys.
Yes, I tried geometry patterns, this was in the attempt where I used several parts which were joined through external copy geometries.But this did not solve the problem.
Also tried reducing the accuracy.
I have contacted our supplier for support but they could not really help me. Their suggestion was to make an assembly; this really sounds like a desperate last resort kind of solution, which I would like to avoid.
Why would ProE have a problem with large patterns? Is there a memory limit? I noticed that of the 6 GB RAM, only 4.5 GB are used.
i use proe on a i-mac and i've open your part and no problem (memory and feature) . my relativ accuracy is 0.01.
my mac is a intel core duo e3835 3.06GHZ
This is an interesting result!
Just to be sure, did you set the hex pattern to the correct size?
To do this, you need to change the relation "zeshoek_grootte = 60" into "zeshoek_grootte = 3"
No i have only change to 35 like you exemple. 3 to small 10000 instance it is impossible. How is the goal of that 's ?
i think i've find the problem. Change the deth of the first repetition and the second (use trough next and flip the direction).
Thanks for your help Pierre!
I tried your suggestion but I think I don't understand what you meant.
I redefined feature #13 (first repetition of Extrude 10). Changed depth to "through next". This resulted in a hole through the entire part. I then flipped direction, but there is no material on that side of the feature.
Part of the problem I see is the order of the features. Based on what I understand you are trying to do the BURL and related featres should come before the PALM features. As you change the hexagon size there are BURL locations that end up in the PALM area and end up being inside solid geometry which may be causing the failures. Drag the PALM feature so it is the last feature before the publish geometry and see what you get. You may also want to try regenerating the patterns one at a time. There were some other things in the references for the features that I would change also. The last suggestion would be if you are trying to create a skeleton I wouldn't add all this detail.