Full Version: Polygonal Converter.

From: chrisd (CHRIS_DORDONI) [#20]
 29 May 2014
To: ALL

T-Splines, PowerSubdNurbs, and other "converters" that have been referenced here are quite limited in the number of polys they can process into a nurbs surface. These tools are intended to be able to manipulate a nurbs surface based on a polygonal "reference" surface. The "conversion" is a by product.

If anyone has found something that can handle millions of polys, please indicate what it is. In my experience, the only applications that can fit a nurbs surface to millions of polys are Geomagic and RapidForm.

I am particularly interested in Resurf3D with their RhinoResurf application, which looks like it can handle more polys. Unfortunately I have not been able to test it because the Rhino trial and the RhinoResurf trial don't play well together.

Reply


From: OSTexo [#21]
 29 May 2014
To: ALL

Hello,

The last scan mesh I imported into Rhino (a few days ago) was around half a million polys. Using a combination of T-splines and VSR Shape modeling I was able to take that very organic 20MB surface and end up with a watertight STEP file around 340K in size with deviation of less than .1mm, well within tolerance for the application. I would not recommend using the technique indicated earlier in this thread, that just leaves you with a jumbled mess of control points, highly unmanageable. Using a combination of submesh creation and proper T-splines and VSR surfacing tools will leave you with not only a much cleaner and lighter set of surfaces, it will allow you to modify those designs more easily.

Reply


From: chrisd (CHRIS_DORDONI) [#22]
 29 May 2014
To: ALL

OSTexo,

500,000 was the max number of polys for T-Splines if a recall correctly from a few years ago.

Has this limit been increased?

Some of the work I have been doing just can't be decimated down that far. I'm trying to find something I can afford that will do a reasonable job of fitting nurbs to a high poly model, and does not cost $15,000 or more.

Reply


From: OSTexo [#23]
 29 May 2014
To: ALL

Hello chrisd,

Don't try to convert your mesh to a T-spline. It depends on the scan, but I like to append faces and use the world retopology snap after laying out the general resurfacing plan. After some tweaking, adding and deleting edges and faces you are able to check your surface deviation within Rhino itself. In areas of detail you can go back and change density in specific areas, add creases, etc. In the end you'll be left with a much cleaner and lighter set of surfaces.

Reply


From: chrisd (CHRIS_DORDONI) [#24]
 29 May 2014
To: ALL

The thing that I have found is LARGE SURFACE VARIATION = LARGE NURBS MODEL. There does not seem to be any getting around this one yet. Tooling producers will cry if you tell them you want to give them a triangulated model, but the nurbs representation can be just as large, or even LARGER than the triangulated version. I fail to see in many situations where the nurbs is actually going to better to machine from where there is a lot of surface variation.

Reply


From: OSTexo [#25]
 30 May 2014
To: ALL

Hello,

I'm not sure I understand what you mean by large surface variation. I'd like to see a small example of what you're referring to.

Reply


From: Frenchy Pilou (PILOU) [#26]
 30 May 2014
To: ALL

Try Decimation Master of Zbrush before some exporting! ;)

Reply


From: chrisd (CHRIS_DORDONI) [#27]
 30 May 2014
To: ALL

hows this?

Image Attachments:
relief section.jpg 

Reply


From: Frenchy Pilou (PILOU) [#28]
 30 May 2014
To: ALL

ZSurf and you will obtain a Nurbs surface!

http://www.kf12.com/blogs/techno/2007/08/zsurf/

Reply


From: chrisd (CHRIS_DORDONI) [#29]
 30 May 2014
To: ALL

Hi Frenchy

Yes, ZSurf can do this for low levels of detail or debossing a logo, but really is not useful for more complex reliefs. It also does not import 16-bit greyscale files if I recall correctly.

Reply


From: OSTexo [#30]
 30 May 2014
To: ALL

Hello chrisd,

Have you considered modeling the item to a specific level of detail, have the molds produced, then sending your molds to be engraved and textured appropriately?

Reply


From: chrisd (CHRIS_DORDONI) [#31]
 30 May 2014
To: ALL

Osteo, I think I still do not understand what you mean.

If the mold was done in a 2 step process, would the last step of adding the fine detail need to be done from a nurbs model?

Reply


From: OSTexo [#32]
 30 May 2014
To: ALL

Hello chrisd,

The engraving and texturing maps can be created from various bitmap and vector formats, they don't need to be NURBs. I believe you can also use scan data as well. I'd investigate having the items modeled to within a manageable tolerance, have the molds milled conventionally, then have the NURBs data, maps and/or scan data sent along with the molds so the detail can be mapped accurately to your molds. The molds are then engraved by subtracting a relatively small amount of material for the finish and detail you desire. I know you can accomplish this chemically, through EDM and laser ablation.

Reply


From: BurrMan [#33]
 30 May 2014
To: ALL

Hi Chris,
"""""""""""" I fail to see in many situations where the nurbs is actually going to better to machine from where there is a lot of surface variation.""""""""""""

The NURBS will be preferred by your CAM system. With a NURBS model, you are allowing your toolpath generator to perform the path calcs and decisions with regard to the settings within the CAM system.

With a poly model, you will be forcing the toolpath calcs, tolerances, moves etc, in conjunction with the CAM settings, which is less desirable and can be problematic when you are doing precision work (especially molds)

Frenchy mentioned reducing the poly count, which I do a lot. The programs today are getting really good at this, but there are some considerations (mostly with molds)

Is DOES morph the intent (but so does the NURBS conversion), so if you are doing 2 part work, you will want to re-create the second/split half from the morphed result. The "decimation" can be done to allow for tolerances specked on the part.

I use 3dcoat for poly retopo's or "decimation".....

Reply


From: chrisd (CHRIS_DORDONI) [#34]
 30 May 2014
To: ALL

While I understand the relation of surface finish and precision to the nurbs surface, I have explained to numerous vendors that I am not interested in high polish or extremely high precision surfaces.

The fact is most cam systems even now compute a toolpath based on a tesselated model derived from a nurbs model. Therefore, even though I may give a vendor a nurbs model, their machine is cutting it as G1 or straight line moves. So when they complain that they can't machine from from a triangulated model, that points out me they actually don't understand their own machining process very well.

Its quite funny actually when I asked to see sample G-code from one shop, to clarify this, they refused.

What I believe is happening is that most machine shops assume high precision work, and they won't stop and switch gears for projects that might be a little different from what they usually do.

Reply


From: BurrMan [#35]
 30 May 2014
To: ALL

"""""""""The fact is most cam systems even now compute a toolpath based on a tesselated model derived from a nurbs model. Therefore, even though I may give a vendor a nurbs model, their machine is cutting it as G1 or straight line moves. So when they complain that they can't machine from from a triangulated model, that points out me they actually don't understand their own machining process very well.""""""""""

Well, yes to what you said, but:

Forcing the CAM program to generate it's path based off of the given little flat lines and patches and hundreds of thousands of edges (each requiring the calc to start and stop THERE as opposed to where IT DECIDES) is problematic for CAM programs and post processing into a proper toolpath. Some of your vendors may have just decided that the cost of problems that can arise because of that is not something they want to keep dealing with.

It's far better for the CAM program to be allowed to generate it's own little sliced up model based on the tolerances supplied within the system. Mixing those tolerances with the "forced dice" of a polygon model is hard at best.

Many CAM programs handle toolpath differently. Some will calc off of the "display mesh" created (way junk). These are the apps that will put a toolpath on a multi million poly model "really fast"... Some will just give you a set, hard coded tolerance which can burn past any of the really hard calcs. Most of the higher end stuff, will be looking at the surfaces supplied.....

CAM'ing a multi million poly model out when the specs are a little looser like you mention, is not super difficult, but it requires an understanding of the nature of the model. Not a lot of CAM shops are very versed in the underlying structure of the models they are working with. They are Machinists....

Reply


From: OSTexo [#36]
 30 May 2014
To: ALL

Hello,

Trying to calculate an optimized toolpath through even 100000 points seems to be some insane traveling salesman problem, let alone millions of points. Perhaps the reason nobody wants to touch the job is because they realize that computationally it is not practical, and there are already well established alternative procedures that will provide an acceptable result.

Reply


From: chrisd (CHRIS_DORDONI) [#37]
 30 May 2014
To: ALL

"Perhaps the reason nobody wants to touch the job is because they realize that computationally it is not practical"

That statement is simply not true.

If I had a mediocre CNC mill with a $2000 -$3000 CADCAM program like DeskProto or VisualMill (MasterCAM, SurfCAM, or Delcam not needed), I could cut this in steel myself. So the problem is not the file. And its not computation power.

I CNC milled files with more than a million triangles 15 years ago using DeskProto. Its ludicrous for a machine shop to say it cannot be done now.

Reply


From: BurrMan [#38]
 30 May 2014
To: ALL

You have to read the entire sentence:

"""""""""Perhaps the reason nobody wants to touch the job is because they realize that computationally it is not practical, and there are already well established alternative procedures that will provide an acceptable result.""""""""""""


And then:

""""""""""Its ludicrous for a machine shop to say it cannot be done now.""""""""""

Remember to note the difference between "can not" and "will not"....

Could probably find a thousand hobbiests to cut stl's.... However, it will be different for guy's running mastercam on $100,000 machines, with employee's doing the machine work (and even the CAM work)....

Reply


From: OSTexo [#39]
 30 May 2014
To: ALL

Hello chrisd,

Apparently you misunderstood my first statement which made my following statement unclear, I'm talking about calculating the optimal tool path from each point on each line of each triangle in relation to its neighbors. I'm not sure how you could say that my computationally impractical statement was not true. I didn't say it was impossible, but it is highly inefficient, and unnecessary given the other options that are available. Did the CNC mill each one of those million plus triangles poly by poly fifteen years ago? Of course not, tool paths were created from an input surface. But I guess that's the point, start with a lightweight surface set and add detail where needed through further secondary finishing processes that are appropriate for the desired output.

Reply


Show messages:  1-19  20-39  40-59  60-79  80-99  …

Reply to All

Back to thread list | Login

© 2017 Project Beehive Forum