Skip to content

Dimension loop

Results

Nominal gap
—
Mean gap
—
Worst-case range
—
RSS range (±3σ)
—

Contribution to variation

    How to use it

    1. Pick the gap you care about and walk the loop from one side of it to the other.
    2. Set each dimension's direction: + if making it bigger makes the gap bigger,− if it makes the gap smaller.
    3. Enter tolerances as positive numbers. For 10 +0.2/−0, type 0.2 and 0.
    4. Add your minimum gap to see how many assemblies are expected to fail it.

    The method

    Each dimension is first re-centred to its mean size and a symmetric half-tolerancetit_i. The mean gap is the signed sum of the mean sizes:

    Gˉ=∑isi dˉi(4)\bar{G} = \sum_i s_i\, \bar{d}_i \tag{4}

    Worst-case adds the half-tolerances; RSS adds them as a root sum of squares:

    tWC=∑iti(5)t_{\text{WC}} = \sum_i t_i \tag{5}tRSS=∑iti2(7)t_{\text{RSS}} = \sqrt{\sum_i t_i^{2}} \tag{7}

    The share below your minimum gap uses the normal distribution withσG=tRSS/3\sigma_G = t_{\text{RSS}}/3. Equation numbers match the article Worst-case vs RSS: when each stack-up method lies to you, which explains every step and works through this page's default example.

    Assumptions and limits

    • 1D only: every dimension must be parallel to the gap.
    • Worst-case is guaranteed but often far wider than what production delivers.
    • RSS assumes independent, normally distributed, centred dimensions with tolerances at ±3σ. On short runs or with off-centre suppliers it is optimistic.