WAH! Thanks for taking the time to point these errors out to me Michael.

If you don't mind fixing it, I don't mind the helping hand! LOL

C.

Hey Chris, don't take this the wrong way, you should try and fix this yourself for the experience.
Some might pick up nurbs modeling very quickly, but it takes longer and more experience to know why something isn't working and the best strategy to fix it. Micheal has already pin pointed a lot of areas so you have a good head start.
It will make things easier and quicker in the long run, if you get stuck the forum is always here to help :)

Cheers and good luck.

Hi Chris, I've attached a finished up version, I think this should be ready to send out as an STL now.

One of the key things that you want to look at when you're preparing something for STL is the object type indicator - that's the label in the upper right area of the window which shows some information about the selected objects.

You want to periodically check that indicator to see if the selected object is a solid or not, and also how many different objects are selected. Do a Ctrl+A to select everything and then look over here:




Your goal for making an STL is to make that readout show that everything that was selected by Ctrl+A was just a single solid. In your previous model, if you do that you can see it says "49 objects", which means you've got a bunch of different pieces and some are solids and some are open surfaces.

So to clean things up, I started by selecting some of the large top pieces that were solids and using boolean union to combine them together into larger pieces.

Then another key thing is make sure to hide pieces if they are getting in your way, otherwise it can be difficult to see if there are holes or things when you have other pieces sitting right on top of what you are trying to work on.

Once I got all the big pieces booleaned together into one large unit, I hid that and then started to look at the bottom parts, those were the areas that needed some planes to be drawn in and joined in to finish off some parts. Then that was followed with some more boolean unions between parts to keep on combining parts together.


But try to keep track of that type indicator, that will help you to know if the thing you are currently working on is a finished and closed up solid or if it still needs some pieces to be joined into it.

I hope this helps!

- Michael

If you create an object only for a photo is very different than if you create it for a 3D printer or other machine...
Sometime it is not the same way to work.
Don't change your design in order to resolve the errors.
Try, try and try again.
Good work

Fil

Hi all... So this model got bounced again by the RP house for being "non manifold", or not water tight. Very odd!

All I did was scale the model that Michael fixed by .044 to give me the correct model size, and then output the STL with approx. 400k polygons.

Any thoughts?

Cheers,

Chris

Hi Chris, what meshing parameters were you using to generate the STL file?

I did actually test generating an STL file from the model I posted, and it read into Rhino and verified that it read as a completely closed mesh.

But maybe some particular meshing parameters that you are using is creating some problem in the mesh, it could be that you are getting some areas with too many vertices being packed together, making the RP house's analysis tools think that they are all lumped together instead of being distinct. If that's the case then you may need to make a somewhat slightly looser mesh instead of an overly dense one.

It's hard to know exactly what you are running into though without some very specific information - please post your scaled model file and also give me the exact parameters that you are using for producing the STL file, that way I can generate the same STL file over here and examine it closely to see what might be going wrong.

- Michael

Hi Chris, it's also possible that you have created some details in this part which are just too small in relation to the resolution that the RP process can actually handle, like for instance here:






That's a little tiny fillet piece with a radius of 0.06, and that's actually from your previous model before any scaling down.

If you're then scaling down by a factor of 0.044, that's going to make that particular feature have a radius of 0.00264 , which is extremely small, probably quite a long ways beneath the actual resolution of what their RP process can produce.

Then since it is a curved surface it will get a bunch of smaller polygons generated in that area as well.


Such small features may be considered by their analysis software as being essentially collapsed together, which can confuse the analysis of the volume. If that's the case then it could be possible for them to adjust the tolerance being used in their analysis software to reduce it so that those things won't be considered to be lumped together. But it is also possible that you may need to eliminate these tiny features or possibly not use a tight angle value when meshing, to avoid this kind of problem of generating micro polygons that are smaller than what the RP house is expecting.

- Michael

Hi Michael. I guess this is why modeling full size and scaling down for output is a bad idea. LOL

This has been quite the learning experience.

C.

Hi Chris, well it could be possible to adjust some mesh settings in order to avoid creating too many little teeny-tiny polygons.

If you can post your scaled down model, and also let me know what mesh settings you were using for the STL file that was not accepted, that could help with that.

It may be possible to solve it by loosening the meshing angle somewhat, or possibly use the "Avoid smaller than" setting to make small things get a rougher meshing on them.


But if I could have a starting point with the exact model and settings that you used, that would let me take a look at the output and see if there is maybe some other problem with the mesh apart from that.

- Michael

Hi Michael. Hope this helps.

I created the STL with a setting of "6", to give approx. 400k polygons.

C.

Hi Chris, try sending them the attached STL file, it seems to check out fine.

To generate this file, I saved to STL and for the meshing parameters I used:

Angle = 6
Avoid smaller than = 0.001


The "Avoid smaller than" parameter tells the mesher that any curved surface under 0.001 units in size should get more roughly meshed, it helps to avoid generating a zillion little teeny tiny bits here.


This STL file only has 38,442 polygons in it, so using that parameter eliminates a whole lot of tiny pieces in this case.


You may want to download the "MiniMagics" STL viewer program from here: http://www.materialise.com/MiniMagics - that's something that you can use to check the STL output and see if it can find any problems in it.


Ideally you will want to avoid making curved features that are so very small though, when you are planning on doing a 3D print of your model. That's also just modeling time that is wasted because you're going way under the resolution of what the rapid prototyping machines can actually generate.


But I hope this STL attached here will work for you!

- Michael

Thank you Michael! The model has been accepted and is now scheduled for RP printing. Finger crossed, I will post photos when the model arrives.

Thanks again for all your help. Should this work as planned, I will be using Moi for some production level RP work.

Cheers,

Chris