The MEboost tolerance calculator allows you to set a dimension's tolerance limits based on a standard deviation range. This allows you to estimate rejection rate while balancing the necessary tolerance for functionality.

Prior to using the tool you need to have an understanding of the manufacturing process used to create the part. In order to determine rejection rate, a probability distribution of the dimension must be specified.

Consider a part whose nominal dimension is 1.000 and the standard deviation of the manufacturing process is .005. The probability density function is shown below. We can specify the tolerance limits at +/-3 standard deviations and determine the percentage of parts rejected for being out of tolerance.

There is often a trade-off between tolerances and reject rate. If the tolerance limits are too tight, reject rate increases and therefore increases cost. If reject rate is too high, a costlier manufacturing method may have to be used. If the tolerance limits are too loose, performance or fit may suffer.

## The MEboost Tolerance Calculator

MEboost has a tolerance calculator that allows you to specify a range of standard deviations and it will calculate tolerance limits and their associated reject rate. To open the tolerance calculator, click on the *Tolerance Calculator* button in the MEboost ribbon.

The tolerance calculator form will appear. Using the example part, we select a normal distribution with a mean of 1 and standard deviation of .005. We will use a range of 6 standard deviations (+/- 3σ).

Clicking the *Calculate* button will show the following results.

For tolerance limits at +/-3σ, we can expect a reject rate of 0.27%. If we increase the sigma range, we could lower the reject rate. A smaller sigma range would result in a higher reject rate.

## Other Distributions

MEboost allows for the following probability distributions:

- Normal
- Uniform
- Triangular
- Trapezoidal
- PERT

The normal distribution is generally used when there is data available on the manufacturing process and is normally distributed. The uniform distribution might be used to model machine tool wear. The remaining distributions are generally used when you can only estimate the parameters of the distribution.

MEboost is an Excel add-in that includes a suite of mechanical engineering tools for analysis and charting.

Excel is a registered trademark of Microsoft Corporation. Used with permission from Microsoft.