Tolerance stack-up for multiple components to find interference risk

[RedCoat999]

I have two components (diameter feature) with a size and a runout tolerance defined on the prints for both features. I also have actual measured data for these features too.

My original hope was for these features to run in clearance 100% of the time. Since the 3 sigma actual measured dimensions show there can be interference, how can I determine the percent chance of this happening. For one particular feature and histogram normal distribution curve, I know how to use the "NORMSDIST" function in Excel to get the percent chance, but with multiple features and curves I don't understand how to put them together and get the percent chance of interference.
Thanks

 

[BillPSU]

Determine the percent of each part being in the interference zone and multiply.

 

[RedCoat999]

BillPSU, Thank you for responding.
I guess I am having a hard time determining what is interfering. For size and runout of one of the two parts, can I just take the mean size and add the mean runout, and add the size 1-sigma and runout 1-sigma together to get a normal distribution curve? Then I will only have two bell curves that will intersect, and I guess that will be easy to find out the percentages.

 

[monkeydog]

RedCoat999,
In your original post you said you hoped to have 100% clearance all the time. Was that based on your drawing requirements? Where are your parts relative to the drawing tolerance? It sounds like you have a "piston" and "cylinder" configuration. Is your drawing taking into account other features such as runout, concentricity, etc?

 

[RedCoat999]

To make a short story long, an error was made on some original parts where the print tolerances and runouts easily allow inteference, which I do not want to happen. The print has now been corrected for future components.

But what the components have been machined too is much tighter than the print. I had hoped that after considering 3-sigma variation on actual measured pieces for size and runout, that a clearance would still be found, but this is not the case. So I need to know the percent chance of interference, or near-interence, occuring.

 

[monkeydog]

RedCoat999,
I looks like you have all the data.
I think you need to plot out your data, consider the bell curve, and look at you averages, your sigma, and overlap your two components. Where the intersecting bell curves are will tell you the probability of interference. You will need to look up the statistics for your lot and sample size.