What is Tolerance ? & Why its needed?
Method used in engineering and
manufacturing to manage and control the
What is Tolerance Stack Up? cumulative variations in dimensions that occur
when assembling multiple components
To Ensure proper Assembly & Functioning,
INTRODUCTION Why its used? considering each minor variations in each
individual parts
Misalignments and Poor Fits
Increased Scrap and Rework
Reduced Product Performance
What will happen if its not used?
Assembly Challenges
Increased Maintenance Needs
Difficulty in Interchangeability
The primary concern is making sure
BASIC EXAMPLE all four boxes fit into the channel in
a satisfactory manner
1 Define the Study
0 mm < Gap < 10 mm
2 Determine Inputs & Outputs
Generate Transfer Function by just
3 Generate Transfer Function
walking thru Dimensions
STEPS TO EXCECUTE TOLERANCE
STACK DESIGN Output ( Gap ) = +V-Z-Y-X-W
TOLERANCE
STACK DESIGN
Establish Nominal value and Typically its best to use Tolerances that are on
4 tolerance for each Input
the drawings and suppliers / Manufactures
agreed upon
Use WC, RSS or Monte Carlo These are multiple ways to Analysis Tolerance
5 Analysis to predic variation Design Study
Compare the results of stack against For Current Example Criteria is
Requirement 0 mm < Gap < 10 mm
6 Interpret & Act on Result
If the criterion is not met, changes have to be
made to the manufacturing processes, the
design, or the specification limits.
1 Estimates Actual Min & Max for Output
2 Quick & Straightforward analysis
3 Used when there are less inputs ( 2 or 3)
Min Gap (50-2)-(10+1)-(5+1)-(20+1)-(10+1) -1
Worst Case Analysis 4 Revisiting Example
Max Gap (50+2)-(10-1)-(5-1)-(20-1)-(10-1) 11
Specified tolerances are not able to meet
Criteria
Change the output Requirements
Tighten Tolerances of all Inputs
Interpretation Multiple Ways to resolve it
Change Nominal values of inputs
Combination of changes listed above
V= 50 +/- 1
Proposal-1 Min Gap (50-1)-(10+1)-(5+1)-(20+1)-(10+1) 0
Max Gap (50+1)-(10-1)-(5-1)-(20-1)-(10-1) 10
COMMONLY USED METHODS 1 Root Sum Square
RSS Analysis
RSS (Output Target Value) = 50-10-20-5-10 = 5
2 RSS ( Output Tolerances) =
m
“ ” is the target mean value of each input
"T” is the drawing tolerance of each input
Based around randomized numbers:
1
Assume input variation & predict output
Monte Carlo Analysis 2 Takes actual production inputs
3 Best if actual data available
4 Requires specialized softwares
Comparison Comparison
Pros Cons Use When..
Accurate for <=3 inputs Can require tolerances to be too small to Need to be 100% sure you are covered for
Worst Case meet design objective, especially with >3 worst case scenario
inputs
Widely used and understood, less Actual inputs may not be normally distributed You have large confidence in normal
RSS
conservative than worst case, easy to interpret with centered mean distribution
Most accurate since it uses actual data from Need to obtain actual data Actual data can be obtained
Monte Carlo
production Need to use special software