Managing tolerance stack-up in assemblies with multiple mating components

[indiancad]

When several machined components fit together in an assembly, even small dimensional variations can accumulate and affect the final fit. A drawing may show every individual component within its specified tolerance, yet the assembled parts may still create alignment problems or unwanted clearance.

This becomes particularly challenging when dimensions are controlled independently rather than from a common functional datum.

When checking tolerance stack-up in mechanical assemblies, do you prefer worst-case analysis, statistical tolerance analysis, or a direct measurement approach? How do you decide which method is appropriate for production?

[SLW210]

That somewhat falls on the engineer/designer doesn't it?

 

I follow (normally) GD&T and the use of LMC and MMC when necessary along with the tolerances per class of fit.

 

If properly designed with the appropriate  tolerances and machined within those tolerances, it should fit properly.

 

I usually draw objects that have tolerances at the median of the size if the tolerance is ± if it is -0 or +0 I draw it at that size in general, but really depends on the parts and the LMC/MMC.

 

Unless instructed differently, most items I just look in the books and go with snug fit or loose fit depending on the assembly.