We have a small mechanical assembly where every measured component passes its drawing inspection, but some completed units are too tight to move freely. Swapping one spacer sometimes fixes the problem, even though both spacers are within specification.
A simplified example
A housing has an internal width of 50.00 ± 0.10 mm. Three components sit side by side inside it, with widths of 20.00 ± 0.05 mm, 20.00 ± 0.05 mm, and 9.80 ± 0.05 mm. The nominal dimensions leave 0.20 mm of clearance, but that clearance clearly isn’t the same in every assembly.
How should I approach tolerance stack-up analysis for this situation? In particular, when is a worst-case calculation appropriate, and when can a statistical approach be justified?
I’d also like to understand how to specify minimum assembly clearance without unnecessarily tightening the tolerance on every component. What measurements would help identify which dimension needs better control?
