LDraw.org Discussion Forums

Hello,

I am working on a set of steam train driving wheels that I 3D print and I want to have LDraw parts to share. I know I can go the route of STL->Dat but the files are HUGE (10k+lines) and I would like to optimize a little better. 

So far my thought is to have a slice of the outer ring that is 1LDU thick but has the profile of the wheel then take and spin that 360 degrees. I can then create a center section of spokes and other features from there. The "problem" is that the largest size of steam driver I am looking to produce is 386 segments (123LDU diameter). Is that "too much" I see that there are other primitives that I can use to create some of the faces but I am worried that the resolution is not high enough for something so large. 

I have attached and example of this. 

I am really looking to make sure I am at least headed in the right direction and that I am not doing something needlessly over-complicated. 

I would think that the creation of a file would go something like this

• Create inner connection 
  
• Create rim(s) at sizes needed
  
• Create associated details
  
• Create the "final" wheel(s) which will have reference to the various pieces needed
  

Additionally, I have also developed a spreadsheet to calculate the ring positions and it would be easily adaptable to help fill out the circular primitives if that is helpful.

It's clear that you need (very) high resolution to 3-D print the wheel, and stl file is fine for that. But imho to make a LDraw part used for rendering and building instructions, 48-primives are really fine for a wheel that large.

Hello,

While these details are necessary for 3D print, doing this way the file will end up much too details.

We usually use a circle of 48 divisions for this size.

So as an approximation with existing Primitives that would end up as:

0 BFC INVERTNEXT
1 14 0 0 0 45.75 0 0 0 14.8 0 0 0 45.75 48\4-4cylo.dat
1 14 0 0 0 2.54167 0 0 0 1 0 0 0 2.54167 48\4-4rin18.dat
1 14 0 0 0 48.29173 0 0 0 1 0 0 0 48.29173 48\4-4edge.dat
1 14 0 12.8 0 1.17785 0 0 0 -12.8 0 0 0 1.17785 48\4-4con41.dat
1 14 0 12.8 0 49.46958 0 0 0 1 0 0 0 49.46958 48\4-4edge.dat

1 14 0 15.5 0 8.24493 0 0 0 -2.7 0 0 0 8.24493 48\1-4con6.dat
1 14 0 15.5 0 8.24493 0 0 0 -2.7 0 0 0 -8.24493 48\1-4con6.dat
1 14 0 15.5 0 -8.24493 0 0 0 -2.7 0 0 0 8.24493 48\1-4con6.dat
1 14 0 15.5 0 -8.24493 0 0 0 -2.7 0 0 0 -8.24493 48\1-4con6.dat

1 14 0 15.5 0 57.71451 0 0 0 2.74 0 0 0 57.71451 48\4-4cylo.dat


Gerald

(2021-05-13, 12:19)Gerald Lasser Wrote: [ -> ]Hello,

While these details are necessary for 3D print, doing this way the file will end up much too details.

We usually use a circle of 48 divisions for this size.

So as an approximation with existing Primitives that would end up as:

0 BFC INVERTNEXT
1 14 0 0 0 45.75 0 0 0 14.8 0 0 0 45.75 48\4-4cylo.dat
1 14 0 0 0 2.54167 0 0 0 1 0 0 0 2.54167 48\4-4rin18.dat
1 14 0 0 0 48.29173 0 0 0 1 0 0 0 48.29173 48\4-4edge.dat
1 14 0 12.8 0 1.17785 0 0 0 -12.8 0 0 0 1.17785 48\4-4con41.dat
1 14 0 12.8 0 49.46958 0 0 0 1 0 0 0 49.46958 48\4-4edge.dat

1 14 0 15.5 0 8.24493 0 0 0 -2.7 0 0 0 8.24493 48\1-4con6.dat
1 14 0 15.5 0 8.24493 0 0 0 -2.7 0 0 0 -8.24493 48\1-4con6.dat
1 14 0 15.5 0 -8.24493 0 0 0 -2.7 0 0 0 8.24493 48\1-4con6.dat
1 14 0 15.5 0 -8.24493 0 0 0 -2.7 0 0 0 -8.24493 48\1-4con6.dat

1 14 0 15.5 0 57.71451 0 0 0 2.74 0 0 0 57.71451 48\4-4cylo.dat


Gerald

Hey its a year later and I am finally getting back to this project. I'm slightly lost as to why n-fcon6 was used in the way it was. Is there some kind of calculation performed to show that I need to scale con6 by some amount to get the size I needed? 

Additionally if there is such a calculator or formula to determine what I need, then how do I used that to do the back of the wheel that is missing in the example you provided? 

Would I be better of creating 48 resolution primitives for my wheels and then just performing my original proposal?



As an aside to this, I am very interested in making some train track using a similar method where I can create the rail profiles needed and then calculate path points. When talking in circles that are in the 500mm radius range, is there a primitive set that is suggested or am I better off making segments that are 2LDU long and placing them all over?