Architectural modeling, precision, floating point, make2d, tolerances, etc

Can someone please clarify what degree of precision can be “guaranteed”, and when tolerances come into play?

What is absolute tolerance / relative tolerance, angular tolerance, how do these relate to floating point accuracy issues, and so on?

If I model something far away from the origin, in inches or feet, rotate it to some 100th of a degree, extrudes, trim, etc… then I orient this closer to the origin and Make2d, should I expect that the output is going to get progressively worse with each operation? Then if I’m using the make2d output for creating 2d drawings and dimensions… how close are the tolerances? Can we be assured that the output is within .01", ready for fabrication after all these? I’m finding that this is a bit confusing, especially when boolean operations don’t work on simple shapes.

Any enlightenment you guys are willing to offer would be much appreciated.

Well, first of all you should NOT model FAR away from the origin. All accuracy suffers. Floating point math issues.:football:

that’s one of my questions. Say I was modeling a space elevator for fabrication with components +/- 1 mm. I would obviously have to model far away from the origin in order to get everything to scale… would I be able to do this in Rhino?

Hi @ConcernedCad,

Is this helpful?

https://www.rhino3d.com/accuracy

– Dale

Also see https://wiki.mcneel.com/rhino/faqtolerances

In general trying to model a very large distance from the origin can cause problems and should be avoided.

Why would you have to model very far from the origin. Model the mechanisms at either end of the cable using local origins near each mechanism.

Actually that was very helpful. Thank you.