Tolerance Stackup

1D linear tolerance stackup with worst-case, RSS, and modified RSS analysis. Sensitivity ranking shows where your tolerance budget is going.

Unit System
Optional — set both to enable pass/fail

Trace a path from one side of the gap to the other through each part. + dimensions increase the gap, dimensions decrease it.

Label Nominal Upper Lower Dir

Enter at least 2 dimensions to see results

Understanding Tolerance Stackups
When to use each method
Worst-Case Every dimension at its worst extreme simultaneously. Guarantees 100% of assemblies fit. Use for safety-critical assemblies, small runs (<50 units), or when rework cost is high.
RSS (Root Sum Square) Statistical method assuming normal distributions. Covers ~99.7% of assemblies (3σ). Use for production quantities where occasional out-of-spec assemblies can be sorted or reworked.
Modified RSS (Bender) RSS with a safety multiplier (k). Accounts for non-ideal distributions, tool wear drift, and process variation that isn't perfectly normal. k=1.5 is the standard starting point. Increase k for less-controlled processes.
Setting up a dimension chain
  1. Identify the gap or clearance you need to control.
  2. Trace the shortest path through the assembly from one side of the gap to the other.
  3. Each part or feature boundary you cross is a dimension in the chain.
  4. Direction is + if the dimension adds to the gap, if it subtracts.
  5. The total should equal the gap when all dimensions are at nominal.
Common mistakes
  • Missing a contributor (thermal expansion, coating thickness, weld distortion).
  • Wrong direction sign — flip + and − and your result inverts.
  • Using worst-case tolerances on production runs. Over-constrains the design, drives cost up.
  • Forgetting that geometric tolerances (position, flatness, etc.) also contribute — use the bonus tolerance calculator to get the effective tolerance, then add it as a row here.
  • Confusing size tolerance with geometric tolerance — they stack independently.
Worked example: pin-in-housing stackup

Three dimensions: housing bore, spacer, and pin OD.

DimNominalUpperLowerDir
Housing bore1.0000″+0.0050+0.0000+1
Spacer0.2500″+0.0020−0.0020+1
Pin OD1.2480″+0.0000−0.0030−1

Nominal gap:

1.0000 + 0.2500 − 1.2480 = 0.0020″

Worst-case:

Max: 1.0050 + 0.2520 − 1.2450 = 0.0120″
Min: 1.0000 + 0.2480 − 1.2480 = 0.0000″ (zero clearance)

RSS (3σ):

Mean of each dimension (midpoint of tolerance band):
 Housing: 1.0000 + (0.0050 + 0.0000)/2 = 1.0025
 Spacer: 0.2500 + (0.0020 + (−0.0020))/2 = 0.2500
 Pin: 1.2480 + (0.0000 + (−0.0030))/2 = 1.2465
Mean gap: 1.0025 + 0.2500 − 1.2465 = 0.0060″
Half-tolerance of each dimension:
 Housing: (0.0050 − 0.0000)/2 = 0.0025
 Spacer: (0.0020 − (−0.0020))/2 = 0.0020
 Pin: (0.0000 − (−0.0030))/2 = 0.0015
RSS tol = √(0.0025² + 0.0020² + 0.0015²) = √0.0000125 = 0.00354″
RSS max: 0.0060 + 0.00354 = 0.00954″
RSS min: 0.0060 − 0.00354 = 0.00246″

The worst-case analysis shows zero clearance is possible — every part could be at the edge of its tolerance. RSS says that's statistically unlikely and predicts a minimum gap of 0.00246″ at 3σ. Which method you trust depends on your production volume, process control, and consequences of interference.

All three methods assume 1D linear chains. For 2D/3D stackups with angular contributors or form tolerances, use Monte Carlo simulation.