My normal distribution chart is flat. How do I fix it?

[landon.crews]

Hello and thank you for hearing my cry for help. I am working with an excel chart to show the process capability given a list of results for a specific test. I've attached my workbook because it'll be way easier to look at than to try and describe. I highlighted the problematic chart in yellow and also highlighed the tabs involved in yellow. If you see the other charts above the problematic one, the distribution looks correct. I tried to recreate the chart and I do get a normal distribution when I do this, but when I try to add the first error bar, the distribution goes flat. Can anyone tell me what I'm doing wrong?

[MrShorty]

My debugging steps:

1) There is not vertical axis to see what values the chart is plotting, so I first add the primary vertical axis to the chart. I see that it is showing numbers from 0 to 250.
2) Go to the "A852 - Results" tab and see what values are in column E. Column E from 3 to 2192 ranges from about 0 to about 0.02. (For comparison, Column E in the next series down ranges from 0 to values up to about 100).
3) The vertical bar series have a y value just less than 200.

I am not sure what you are expecting, but it looks like Excel is charting the requested values correctly. Charting values that range from 0 to 0.02 on a scale of 0 to 200 should look pretty flat at the bottom. I suspect that you have not correctly calculated or chosen the y values for the different series on this chart. Get more appropriate values for the y values and/or a more suitable range for the y axis max and min should fix the chart.

[landon.crews]

I am trying to chart the normal distribution of a set of values. The x-axis should contain the actual result and the y-axis should show the normal distribution function value. I don't understand why this method worked for the first two charts on 'Summary' tab. When I create a new scatter chart and select those two coloumns (Result / Distribution) a normal curve appears. So the values are fine. I guess I might not completely understand you. The problem starts when error bars are added.

-EDIT-

I understand what you were saying now. Changing the scale of the vertical axis allowed me to see the normal curve correctly. Thanks so much!

[MrShorty]

The point I am trying to make is that your standard deviation lines are plotting a y value (196) much larger than the normal curve is calculating (0.02). If I enter something like 0.02 into M85, the automatic axis scale shifts and the normal curve shows up on the chart as expected. I don't know what the purpose of the MAX() function in M85 is, so I hesitate to suggest this as a final solution, but it should show that the problem is because the value in M85 is so much larger than anything calculated by the normal distribution function that both sets of values cannot be effectively seen on the same chart. Basically, you need to either figure out what is wrong with the calculation in M85 that makes it so large, or figure out what is wrong with the normal distribution calculation in A852-Results that they are so small.

Edit: it looks like you figured something out while I was typing my response. If that solves your problem, be sure to mark the thread as solved (see thread tools near the top).