Printable View
Quote:
Originally Posted by Allen123
Ok Allen, I finally had time to put this together. This is actually drawn in 3D.
While I can create a surface on this, it isn't right. It overhangs the boundaries. Let me know if you can figure it out!
BTW, I learned a good bit while working on this, and was actually able to get some of it to work. This is the next to the last part, but the real hard part is yet to come.
Thanks,
Don
Hi Don,
I skined this by dividing the top straight line by 2, then I could do a skin in 2 operations and connect the surfaces.
The geometry was not symetrical so I left 2 for you.
The one on the left is 2 seperate surfaces as I did the skin and is your drawing exactly. The split in the left half seems to be caused by a wierd "Back and forth" over the top of itself that "erase doubles" doesnt get rid of. Maybe somebody else with more experience knows what this is. (What program made the dwg? BC? The funky thing is in the bottom segment, bottom left corner fist segments right side)
The one on the right is stiched together and is the result of the "right half" skin, then mirrored to the left side. (This showed the asymetry).
Hope that helps.
Burr
Attachment 75405
Heres a better example Don:
http://www.i386group.com/bobhuh/surface/cnczone_dwg.htm
That's really cool! I was using the four edged surface function. Maybe that was the problem. I'll have to look at that geometry to see what's funky in the left side there.
Thanks!!!!
I've been following along and also looked at your file.Quote:
Originally Posted by DonW
There's some very short line segments at the end of the middle section chain that may get missed when chain selecting.
Or maybe they can be Optimised together?
moldmker
Optimized? Hmmm.... I guess I need to look into that and see what it's all about.
Meanwhile, here's another section of this mandolin top, the bottom half. When I use the "Surface 3 Edges" command, the centerline at the bottom gets slightly caved in. I assumed that it would make a smooth transition between the two broken curved splines, and the result would be a smooth looking dome. Instead, it seems to draw more strength from the horizontally oriented chain, and then drops into the vertical one.
I actually did an offset for the bottom of the re-curve, which sits slightly below the outside perimeter, around 5/8" from the edge. Then I created a surface of just the outer edge to the bottom of the recurve, and then from the bottom of the recurve through the rest of the geometry. I'd love to know what I'm doing wrong, or what I could do better.
I have to admit though, that what I thought was a weakness in BobCad's ability to do things, is as much if not more, my own weakness at not thinking things through to get the software to do what I want it to do...
They're really short because when I exploded the spline (imported from Illustrator as a .dwg file) it had a lot of short segments, and I had to split a few of those at the intersection points of bounding geometry (Break at Entity). It just worked out that way. I wish there was a way to re-combine geometry into a spline or entity of some sort... similar to a Boolean ad function for solids. Once combined, they act as a single entity. Just a thought for those BCC Engineers to think about.Quote:
Originally Posted by moldmker
I'm not sure I followed very well. I surface3d the geometry you posted and thought it went very well. The 3 edges was a good selection to use and its what I used since you had selected that method. I thought the surface was nice.Quote:
Originally Posted by DonW
Attachment 75449
WHen doing your chain selection, use the "zoom area" tool to zoom to the begining and end points of each section. As mold said earlier, it's easy to miss a little begining piece or end piece and end up with an odd surface.
Try lightening up the color of the surface, then rotate it around, and you'll see what I mean. Along the center, on the dome, right before it goes into the recurve, there's a slight dip in the dome, slightly concave. Are you using V22?
I'll post a pic of what happened when I did it later this evening.
Don
Quote:
Originally Posted by DonW
Hi Don,
I think I got what your saying now. (Last file was in V23. Switched back to V22. Thread title keeps throwing me off.)
You mean the lack of control of the skin operation between area 1 and area 2. So right now it follows more closly area 1's profile for the entire length of that curve, then dips all at once at the area 2 curve. Correct?
Attachment 75454
I need some time to look at this more closely, others may jump in here too, but in the mean time as food for thought, a great way to keep the control would be to use something more like cross sections.
Here is a file I did with cross sections. I used your geometry file and brought it into a surfacing package I'm more familiar with and created the surface i think you are after. I then brought that surface into BobCad and and created a series of lines above it. I projected those lines onto the surface then deleted the surface. This gave me my cross sections to get my skin from.
I did it this way for speed and precision to what you were after with limited information, and to illustrate cross sections for you. If you can design the sections, you can have more finite control of the skin. (Again, this is V22)
I will see if I can figure out skins a bit later as this may be the way you need to do it and I also need more info in this area.
Talk to you later,
Burr
Hey Burr, Yeah, I think you've got it. Attached is the file of what I was able to do going at it in a slightly different way. You'll see that I did the edge/recurve sections on the bottom half separately from the dome part. Where it gets funky is along the center line. This file is saved so that it's upside down and the highlights show the inward bump at the center down at the bottom. I'm just not sure why it's doing that. By isolating the edge/recurve section, it seems to hold the recurve area at a consistant distance from the edge. Without doing that, it wants to stray from the example of the chain. Notice on yours, the middle area between 1 and 2 moves closer to the edge. What I'm trying to achieve is a uniform recurve around the perimeter. In case I'm confusing you with the term "recurve", it's just the are that dips downward and then back up again, going from concave at the edge, to convex and doming upward to the middle. The recurve is that transition area.
I don't know. It seems there are no easy and clear-cut answers.
Hi Don,
Yeah I hear you now. Using cross section can give you exact control of the surface your going for. Trying to draw those cross sections is your challenge. In the file I posted last, you can delete the surface and view the curves. They were just created from a surface generated in the manner you posted before so the results are the same.
Bobcad is limited in its surface creation tools and if you need a more powerful toolset for this you may need to look for other packages that are more specific to this task to acheive this. Higher end packages can give more control over where your surface goes and how the edges blend through.
I see the result your after but might be beyond me with the curves supplied.
Maybe Allen has a good way to proceed.
Bueller.....Bueller?
Yeah...that's what I suspected. And we haven't even got to the hard part yet - the scroll area...
Sorry, I haven't gotten back to you. It looks like a few guys chimed in and got you some answers.
I have been at the solidworks show in Orlando. What an awesome show! I'll take a look at your file latter today.
Hey Burr,
So what's out there that can draw stuff like this? I know Rhino probably can do it, and I can get an educational price on it, but it's an entirely different way of thinking when using that software... Anything out there reasonably priced (read: cheap, poor-boy hobbyist)?
http://i564.photobucket.com/albums/s...juice/mk34.jpg
The video on how to draw this surface is a good one! Good Job,
A few things to consider.
#1 instead of divide, you should break the geometry.
- This breaks all the geometry where it intersect and really that is what you want.
#2 You don't need to create two surfaces for this shape. You can skin all 5 paths.
- use the same selection method and order. The only difference is with the skin you can use more than 4 chains.
* When you created the sample surface then did a mirror you could see the other side did not match up proper. That's because the part seems not to be symmetrical. This is why I used all the chains and did one skin.
Allen, thanks a bunch. Any idea what's causing that line? I've been through the geometry, and I can't find any doubles or overlaps etc.
Not sure what is causing that line, but I don't feel it will be an issue when it comes to cutting the part.Quote:
Originally Posted by DonW
I take another look
Quote:
Originally Posted by DonW
Yeah Rhino kindof specializes in "Surface" creation. When trying to reverse engineer stuff and match exactly a shape, having a richer toolset is necessary. No hobbiest packages will do it well. BobCad can make that surface but you have to draw the curves, and when "Matching", it becomes very difficult. The higher end packages can provide additional "Paths and rails" to use with the surfacing that remain manipulatable with a history update of the surface so you can tweak and move and match.
The learning curve will be much higher.
DonW,
You can try Alibre. Free version is called Alibre Express. It a parametric modeler with a rich tool set. The tutorials are great and I found it easy to learn. Be prepared to spend about 40 hours learning it. But the things you learn will pertain to just about any solid modeler. Export to STEP and import into BobCAD from STEP.
With what you are doing, I'd say you need Alibre Express. It's free and you will have a blast just creating things in it. You probably won't need any of the Alibre Standard or Professional features, as you are not going to have to use a bunch of part assemblies.
Steve