Tolerance stack up types

[umatrix]

I’m almost always asked if I can perform a tolerance stack up analysis, and the answer is yes, a basic linear 1D, stack up.

I have yet to see a 2D, 3D or a radial stack up calculation. Is it just a case of these being more complicated and difficult so designers rely on the 1D stack up making these other types rare ?

How are the 2D, 3D and radial designs checked for tolerance?

This is a good article on the subject, but also wanted to hear from others engineers.

https://www.enventive.com/tolerance...erences-between-1d-2d-and-3d-tolerance-stacks

Thanks

 

[Belanger]

The reason you don't hear much about 2D or 3D stacks is because the great majority of stacks in the real world are 1D. Even most "radial stacks" are 1D in nature, and longitudinal or axial stacks are also 1D.
But as you suspected, 2D and 3D stacks are more complicated (that's not a reason to avoid them, though).

Attached is a screen shot from a presentation that shows a sample 2D stack. To solve for the unknown "z" requires some trigonometry.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[3DDave]

That only shows how to calculate the remaining thickness. The tolerance analysis is a bit more complicated. Even more if one desires to include the distribution of variability in the manufacture of the part.

I found it was far better to design part definitions that directly controlled the variability that is important to be controlled and apportion the allowed variability based on manufacturing capability to tolerances on the dimensions of the participating features. Some times it required redesign or requiring assembly line adjustment to make up for poor manufacturing capability. One group decided to skip all that and used grinders at final assembly.

An important consideration is part stress and strain that result from part variability - good luck ever finding any report tying imperfect geometry that is accepted by drawing tolerances and a stress analysis or other performance analysis.

I read further on that Enventive website - Hi guys! We bought 8 workstations worth of Cognition's Mechanical Advantage because our engineering department head demanded that "engineers don't make drawings!" Then he got us Applicon Bravo! which crashed and burned. (not literally, just their part of software market share.) I did like the MA sketcher for displaying unconstrained degrees of freedom - made the transition to Pro/E much easier and the Ultrix stations from DEC had a C compiler that I found really handy.

 

[Windback]

1D and 2D stacks are the most common ones however, there are specific problems that requires to perform 3D stacks.

Stack ups are usually made for room temperature but for the applications requires function in a wide temperature range, thermal effects on the geometry should be utilized in the stack up as well. I call them "hot stack" or "cold stack". (very creative huh )

If you are ending up with 2D 3D stacks usually, that could be an indicator of the dimensioning on the drawing might be set in a problematic way as-well.

 

[umatrix]

Found a good book titled MECHANICAL TOLERANCE
STACKUP AND ANALYSIS by Fischer with a good description of 1, 2 and 3D stack up approach.


"As all tolerance analyses and stackups are truly three-dimensional, the problem
solver is forced to frame the problem in such a manner as to facilitate a one dimensional
solution. Simplification and idealization of the problem are required."

 

[flash3780]

When dealing with multi-dimensional stackups, I typically try to stack the individual dimensions independently.

However, there can sometimes be interaction between them.

For example, you may need to consider multiple dimensions if you were stacking the gap between components installed into a circumferential slot in the rim of a disk.

In that case you'll need to determine the stackup for the radius of the components, and apply that to the stackup in the circumferential direction.

-Chris