Do you know if your process is healthy? Are you using SPC Control Charts to track and monitor the behavior of your processes?
Every manufacturing process generates a lot of numbers. From equipment and materials to specifications and tolerances, numbers are generated from every action performed by workers and equipment on the shop floor. These numbers are incredibly useful to understand key information about your process.
But many of the values within the data can seem random. So how can we decipher the data and use these number values to our advantage? How can we discern valuable statistical information over time from seemingly random data points?
The answer is with Statistical Process Control (SPC) and control charts.
Random data points are in many ways a sign of a good healthy process. We just want these “random” data points to be within acceptable tolerances. With this in mind, let’s explore what SPC Control Charts are and how to use them to maintain the health of your processes.
Why Use an SPC Control Chart?
When reviewing the health and effectiveness of your operation, tools like interactive SPC Control Charts are extremely effective. They help manufacturers monitor process behaviors over time and discover trends that could negatively affect the process and product.
For any manufacturing process to succeed, it needs two crucial behaviors to be considered healthy:
Stability: Process is accurate and variations remain within specifications and tolerances.
Capability: Process can be sustained and consistently creates value.
Natural variations occur within any process, but we need these variations to stay within a certain range. In other words, the process needs to be stable. If the data points are unstable, then the risk of creating defects or deviating from customer requirements is more than likely.
Instability leads to a process that is incapable of producing value for the customer or sustain profitability. Why assemble 1000 products if only half will be good enough for the customer?
An example of a cause of an unhealthy variation is a loose part within a machine. As time passes and the machine is used, the part will progressively loosen. This will affect the tolerances and the product will progressively deviate from the required specification range.
To observe the stability and capability of your processes, SPC Control Charts are visual and interactive tools that enable you to gain a key perspective on your process’s natural behavior. This is done by tracking a specific action or value over time.
But the power of an SPC Control Chart lies in its ability to gauge and predict future behavior based on past data. Trends, repetitions, or patterns within specifications can all be indicators of instability. If the behavior is leading towards (or already outside) warning and specification limits, then adjustments will need to be made.
Example #1: Roll the Dice!
A simplified example of this is with a 6 sided die.
There are 6 numbers that this die should present. If you roll the die one hundred times, you can expect a fairly even distribution between all the numbers. The natural variation and healthy behavior would give you a mean average value of 3.5.
But if at one point you roll a 7, you know something has gone wrong. The die has given a value outside of the specification limits. Similarly, if you throw the die and get 4 more often than not, then you know the die may be unevenly weighted.
It is the same with tracking the specifications of a manufactured product. If the product created gives a value outside of specifications (in our case, anything outside of 1-6), then it is a defect. And trends that show an uneven distribution also indicate an unhealthy process.
But with the use of SPC Control Charts that display accurate history and real-time data, you can quickly identify problems, review trends, predict future events, and find solutions. Recording key manufacturing data onto an SPC Control Chart greatly minimizes the risk of costly defects and machine breakdowns.
Reading your SPC control chart is very simple. The left axis is the recorded values and the bottom axis is the time or number of productions. In the case below, for Vacuum Mixer #1, values were recorded with each batch.
As we’ve discussed, an SPC Control Chart is a graph used to study and visualize how a process progresses over time. The control chart also indicates the warning limits and specification limits of the process.
Warning Limits:
Determines and defines the stability of the process. Is the process under control? Data points should remain within these limits as much as possible. If data points progressively or consistently deviate outside of the upper warning limits (UWL) and lower warning limits (LWL), then there is a trend leading to potential defects.
Specification Limits:
Acceptable parameters that meet tolerances, production goals, & customer requirements. Any deviation outside of the upper specification limits (USL) and lower specification limits (LSL) means there was an occurrence that was a defect.
With these limits in place, effectively monitoring if a process is in control or outside of the limits is made easy for you and your team. Plus, by reviewing the data over time, trends are much more easily identified before they become a problem.
7 Out of Control Variations
To help you identify the different types of variations that are markers of an unhealthy process, here are 7 out of control conditions or data variations. Many of these are still within the upper and lower limits of the SPC control chart but we’ll see how these can be dangerous to your process.
Cyclical Effect: The control chart is repeating the same sequence. If the process is repeating the same sequence over and over, then there could be something wrong with the data measurements or an undesirable consistent effect.
Points Outside of Control Limits: The process is creating defects or unwanted variations. This means that a defect was created. Now, you can remove the undesired product and investigate why the defect occurred.
Lack of Variability: Though the data points are close to the middle, there is not an even distribution of values. Going back to our 6 sided die example, if you consistently rolled a 3, 4, or 5, you’d suspect that something was wrong with the die.
Run: Consecutive values are occurring above or below the target average. Similar to our last one, the values are consistent in an undesirable way.
Excessive Variability: Values are regularity at the fringes and rarely in the center. Something within the process is causing values to occur at the extremities of the specification limits. This means that there is a loss of accuracy.
Trend: Data points are forming a trend pointing to eventually exceeding specification limits. Something within the process is pushing the data points in a downward or upward trend. Perform a cause analysis before the process exceeds limitations.
Alternating Values: Values frequently alternate between a few data points. This may seem like a well-distributed set of data points but it is again only hitting a few numbers. Like a metronome, it’s consistently alternating, which is a sign of a problem.
The Data Needs to be Random
This may seem counterintuitive but the best thing for your process is for your data points to be random within the control limits. All of the above examples of “out of control variations” have an element of overly consistent behavior.
How your control chart is deciphered will heavily depend on your production environment. You may have some consistencies that you have accounted for. Perhaps there is a natural cycle to your process that is needed or benign to the quality of your product. But in any case, unless for factors already accounted for, your process should produce completely random data within the specification limits.
A key marker of healthy statistical process control is that the process is not producing any defects while being unaffected by any undesirable factors.
Statistical Process Control Charts with Work Instruction Software
Did you know VKS has an SPC feature? You can either use it for specific work instructions or a general process across multiple guidebooks. It is a simple and effective method to empower employees with procedural knowledge and accurately monitor the health of your processes.
While following their instructions, workers are prompted to enter key data about the process. This data is then recorded over time in the control chart. Valuable information is extrapolated to fix problems and continuously improve.
Through the use of the rule engine within VKS, SPC control charts will also send notifications to operators and management if readings are out of ‘warning limits’ or (worse still) out of ‘specification limits. This enables the appropriate people to receive real-time data and respond accordingly in the least amount of time.
The interactive SPC Control Chart is easily accessible within the instructional guidebook and under the reports tab. You can glide your mouse over the data points and the chart will give you the information from that specific point in time.
Example #2: Humidity in Controlled Environment
Let's say you operate a pharmaceutical lab testing the effectiveness of live cultures in beer yeast. Keeping the environment at an optimal humidity level is crucial.
At the beginning of every procedure, your employees are required to take a reading of the humidity to verify that it is within standards. Once this information is verified and recorded, they can begin work with the yeast.
Each time an operator records the humidity data, the information is compiled into the SPC Control Chart within VKS. With an accurate history of variations in humidity, you and your team can see if there are trends or unseen factors that could negatively affect the testing of the yeast cultures.
When reviewing the chart you see a trend of higher humidity at the beginning of the week and lower humidity at the end of the week. The data points are still within specifications but, oddly, this cycle has been repeating itself over the past few weeks.
You investigate this issue and find out that there are different teams of people working at the two separate times in the week. And there is a slight variation between how each team checks the humidity. This is throwing off the humidity records and giving an inaccurate reading.
By using an SPC Control Chart, you have been able to find the issue and subsequently remedy the error by implementing a standardized method with your work instruction software. Now the humidity levels readings are where they should be.
Is Your Process Under Control?
For added information, take a look at this video to see a real-life use of SPC Control Charts with work instruction software for manufacturing chemical process control.
The power of data is irrefutable. Knowing more about your process enables you to make better decisions and even predict future events. By implementing work instruction software with SPC Control Charts, you have greater control over your process, your products, and your quality.
The more you know, the healthier your process will be.
