Control charts are a great tool to monitor your processes overtime. This way you can easily see variation. Control charts are a great tool that you can use to determine if your process is under statistical control, the level of variation inherent in the process, and point you in the direction of the nature of the variation (common cause or special cause).

When Do You Use Control Charts?

Generally a control part in a DMAIC project is used in the control phase to help lock in the gains that you made and automate an alarm system to let you know if the process is misbehaving. However, if a process has existing data, you could use the same tools and techniques to illustrate the level (or lack) of control in the current state system. And of course the findings from analysis on a control chart could be a launching point for improvement initiatives.

A control chart is an extension of a run chart. The control chart includes everything a run chart does but adds upper control limits and lower control limits at a distance of 3 Standard Deviations away from the process mean. This shows process capability and helps you monitor a process to see if it is within acceptable parameters or not.

There are multiple kinds of control charts. You need to choose the right one depending upon the kinds of data sets you are mapping and other conditions. The kind of chart you use will affect the calculations of control limits you place in the chart.

Practical Thoughts Around Control Charts

A colleague once labeled the Upper & Lower control limits for a process he was responsible for as the “Time to update the resume lines” because if the process got out of control, he might be out of a job!

Use a control chart to distinguish between common cause and special cause variation in a new process. Or use it to determine how much common cause variation exists.

Control Chart Tips

  • Specification lines should NEVER be included on a control chart
  • You should gather data for a control chart in the order of production
  • The ease of data collection is not a major consideration.
  • It is more important to collect data that relates to a critical product or process parameter.
  • Never gather data from inspection records, because it is too late — the cause for a point out of control, shift, or trend is lost because it happened hours earlier.
  • Put at least 6 points in the range of a control chart to ensure adequate discrimination.
  • A control chart can be used to identify the following assignable causes
    •  Shifts
    • Trends
    • A point outside control limits
    • NOT anything to do with specification limits.

What Kind of a Control Chart to Use?

control chart decision tree
control chart decision tree


control chart decision tree 2
control chart decision tree 2
charting calculations
charting calculations


Control Chart vs a Run Chart

A run chart can reveal shifts and trends, but not points out of control (A run chart does not have control limits; therefore, it cannot detect out of control conditions.) You can turn a run chart into a control chart by adding upper and lower control limits.

Control Limits

Control limits are the voice of the process (different from specification limits, which are the voice of the customer.) They show what the process is doing and act as a guide for what it should be doing. Control limits also indicate that a process event or measurement is likely to fall within that limit

Control limits are calculated by:

  1. Estimating the standard deviation, σ, of the sample data
  2. Multiplying that number by three
  3. Adding (3 x σ to the average) for the UCL and subtracting (3 x σ from the average) for the LCL

Mathematically, the calculation of control limits looks like:

control limit calculation
control limit calculation

A Control Chart Indicates a Process is Out of Control When:

The following are indicators of out of control conditions on a control chart:

  • Six consecutive points, increasing or decreasing.
  • Fourteen consecutive points that alternate up and down.
  • One or more points outside the control limits.

Control Charts: Usage & Terms

Trend: seven points in a row in either an upward or downward direction.

Shift: Seven points in a row either above or below the center line

Extreme shift: Seven points in a row that are beyond the control limits, not a trend

Trend: Seven points in a row in either an upward or downward direction

Drift: A process is expected to drift, for no particular reason, about 1.5 standard deviations over the long run

Control Chart Videos

Good ASQ Video here.

How to Create a Control Charts Using MiniTab

Other Notes About Control Charts

ASQ Six Sigma Black Belt Control Chart Questions

Question: The purpose of using control charts is to

(A) determine if the process is performing within specifications
(B) evaluate process performance over time
(C) determine how to recreate the process
(D) detect the causes of nonconformities

Answer: (B) Control charts are used to evaluate the performance of a process overtime. It is irrelevant how the specifications are set. A control chart may tell you if non-onformaties are present, but it will not tell you what they are without root cause analysis.


Question: For a process, X BAR BAR = 35.0 and σ = 5.0. If the subgroup size is n=5, what is the value for the upper control limit for the process?

(A) 37.24

(B) 37.89

(C) 41.71

(D) 52.50

Answer: (C) 41.71.   UCL = mean + 3*σ / n^(1/2) = X BAR BAR + 3*5 / 5^0.5 = 35 + 15/2.2361 = 41.71


ASQ Six Sigma Green Belt Control Chart Questions

Which of the following control charts is used to monitor discrete data?

  1. p
  2. I & mR
  3. X Bar
  4. X Bar – R


Answer: p charts are used to monitor discrete data. See the control chart matrix in on this page. Also, review attribute charts.

I & MR charts and X Bar charts are for continuous data and When you have subgroups of size = 1. You use the ImR (XmR) chart only when logistical reasons prevent you from having larger subgroups or when there is no reasonable basis for rational subgroups.

Use X Bar R Control Charts when you have small amounts of constant, continuous data and when you can rationally collect measurements in subgroups of generally between two and 10 observations.



Comments (19)

UCL = mean + 3*σ / n^(1/2)

Where did you get this formula? I could not find it anywhere in the Villanova SSBB information.

Amy, one of my misgivings about the Villanova program is that they test you on items not in their materials. They do list many books and outside references they want you to be familiar with. Their stance is that their exam is not a comprehensive test of their material, but rather an assessment of the industry’s material. I don’t recall which book, but I have a list of my references here.

Hey I have a doubt. In the question B) the answer for UCL is calculated to be 41.71. However with Mean 35 and S.D of 5, the value of Mean+3Sigma = 50. Hence, if the samples follow a normal distribution, they will fall outside the control limit of 41.71.
In SPC the control limits are assigned such that the variation falls within the limits. I dont seem to understand the logic behind this calculation. Usually the formula used would be X-DoubleBar + A2Rbar

Hi Narayanan, We cover question breakdown and approach in the Guided option of my Pass Your Six Sigma Green Belt study guide course. There you can post questions and discuss solution sets with experts. Join up and add this to the discussion!

The control chart decision tree given above confuses sample size with subgroup size, please correct me if I am wrong.

Good afternoon Ted, great article!! I represent Thomas Pyzdek of the Pyzdek institute and would love to link up and discuss your article!

Dear Sir
I am very much impressed to learn such kind of lesson what you shared. Really its amazing !!!

With Regards
Nur Mohammed Munshi

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.