Eric N. Davis, PMP
A Mistake Proofing Methodology
Error and Mistake Proofing Definitions
An error proofing or mistake proofing device is any mechanism that either prevents a mistake from being made or makes the mistake obvious at a glance.
Poka-yoke is also a common term used for error and mistake proofing. This is a Japanese term (poka - inadvertent mistake and yoke - prevent). The term and process was invented by Shigeo Shingo in the early 1960’s.
Some organizations define Error and Mistake Proofing separately. Typically, in this context error proofing is a method used to identify potential process errors and either design them out of a product or process, or eliminate the possibility that the error could produce a defect. Mistake proofing is an approach used to develop methods to clearly identify errors that may occur and prevent them from becoming nonconformances.
What Brought Mistake Proofing to Industry?
Error proofing was brought into industry largely through the work of one man – Shigeo Shingo who was a industrial consultant for Toyota in 1961. On a visit to the Yamada Electric plant, Shingo used a dish to fix a re-occurring assembly problem.
From this experience, Shingo developed a system of poka-yoke and formalized Zero Defect Quality Control. He brought his techniques to the United States in the mid-1980’s.
To Err is Human
When performing a routine task the human brain’s default mode of operation is pattern recognition and autopilot execution.
If the pattern is familiar, a behavior that has been successful in the past is “launched.” It is only when feedback suggests that things are not going as planned that more in-depth thought is called up.
What is a Mistake?
A mistake occurs when:
A required action is not performed or performed incorrectly.
A prohibited action is executed.
Information essential for an action is misinterpreted.
Products with zero defects are possible
Disciplinary action does not guarantee good parts. (operator vigilance)
Process control does not guarantee good parts. A Cp>2 process with just 8 steps would result in 240 PPM. (Variance)
Mistake proofing can produce parts with no errors.
Mistakes - the dominant source of defects
Each type of undetected and uncorrected mistake occurs rarely, about 1 in 10,000 to 1 in 100,000 manufacturing operations.
Although rare, omission errors alone will result in defect rates as high as 100 ppm .
Complexity - the root cause of defects
The more complex the assembly, the more likely it is that mistakes will be made.
Product complexity is also the root of excessive variation (the things that use SPC, etc. to control)
General Principles for Mistake Proofing
As defined by Shingo, there are three types of mistake proofing inspection:
Judgment (sorting defects out)
Informative (inspection soon after completing a task)
Source (before the fact)
Source inspection is the most powerful technique – it happens before the error and the inspection can improve the process and prevent defects. Because source inspection is the most powerful, putting error proofing measures into the design is the most effective way to prevent errors.
Successive (informative) checks are nearly as effective as source inspection. Checks must take place at the nearest downstream point possible.
“Inspecting in” quality is not an effective quality management approach. Judgment inspection is therefor the least preferred.
Some good tools to use...
- Validation of equipment upon receipt using a Gage R&R type approach. This assures that the equipment does what it was designed to do.
- Daily checks to make sure that the mistake proofing is still working.
Some good tools to use...
- DFMEA, PFMEA, TGR/TGW/ DFM/DFA
- DFMEA - best opportunity to get at the design and see where error proofing is needed.
- PFMEA - best opportunity to look at the manufacturing process for weakness that mistake proofing can address
- Things Gone Right / Things Gone Wrong - best place to find past error and fix them for the future.
- Design for Manufacturing/Assembly -best chance to reduce the complexity of the assembly
- Manufacturing Flow Charts - best place to see possible interruptions to flow that could lead to parts packed/labeled improperly.
Some good tools to use...
Grouts Poke-Yoke Page - The best place on the world wide web for information on mistake proofing.http://www.campbell.berry.edu/faculty/jgrout/pokayoke.shtml
- Mistakes happen: To prevent mistakes, reduce complexity and 100% inspect the wherever mistakes could occur.
Error Proofing is our first line of defense. Mistake Proofing is our second.
All processes are susceptible to mistakes so look at every step of every process for opportunities to improve.
A Mistake Proofing Methodology
Sources of Errors
Errors in manufacture often generate defects in products. These defects can come from errors in design of the product or process. Alternatively, the defects can come from mistakes made during the manufacture of the product.
In our lexicology, we term the design based control of errors as error proofing. Specifically, error proofing is a method used to identify potential process errors and either design them out of the product or process or eliminate the possibility that the error would produce a defect.
By the same lexicological approach, we term the component of errors made by deviation from the designated process, mistake proofing. Mistake proofing is an approach used to develop methods to clearly identify errors that may occur and prevent them from becoming defects.
For now, we will ignore the very important design process except to make a couple of points. The most powerful way to eliminate defects in products is to make sure that products are simply and elegantly designed so that manufacturing defects are virtually impossible. Good design is the first and best line of defense against product defects. That said, modifying design of the part or of the process is the preferred way to eliminate errors. Using the very best mistake proofing devices cannot provide perfect protect against defects. Error proofing the design can. When ever defects are determined that can be eliminated by either error proofing or by mistake proofing, error proofing is the choice to make.
Looking For Mistakes
In order to prevent mistakes, it is essential to determine when and what mistakes might be made. As this is a cause and effect problem, a standard cause and effect (fishbone) diagram is a very effective model for determining possible errors. The causes are errors. The effects are defects.
A mistake proofing fishbone diagram has two weaknesses that you should be aware of; while it helps us to remember that mistakes come from a number of different locations, it subtlety undermines the importance of human error. Make no mistake, humans are the number one source of errors.
The other consideration is that since the fishbone is used to analyze the designed product and process, it sometimes doesn't provide the best insight into changes in part or methods that might correct though error proofing problems that you are trying to solve by mistake proofing.
Still, if you break down a process into the constituents of man, machine, materials, and methods you have a good start at identifying sources of mistakes.
Methods: Types of errors in methods that may cause defects include non-compliance to the operator instruction sheet, a poor job instruction sheet, and improper job sequence.
Materials: Types of errors in materials that may cause defects are possibly out-of-specification materials, previously damaged materials, and wrong materials being used in the manufacture of a part.
Machine/Process: One type of error in machine/process may involve a lack of maintenance. That would cause a defect such as an improper adjustment or a dirty proximity sensor switch.
Human Error: Under the component of human error, some errors that may occur are inadvertent mistakes, resulting in errors or defects.
Effective Mistake Proofing Countermeasures
You can avoid defects in parts by checking part or process as the process is unfolding, or by inspecting the product as or after the process has been executed. In either case, the work has to be interrupted by warning or by control/halt of the process.
Checking can be broken down in the following way: feature checking, count checking, or motion-step.
Feature checking is a method to check the physical characteristics of a part or process to differentiate it from some standard. Screw gun torque, presence of a boss or fastener, and time to HACS a part would all be things that could be checked as features. The color, shape, and dimensions of a part would also be features.
Count checking is verification of a predetermined number of tasks against some standard. The number of screws attached at a workstation would be an example.
Motion-Step checking is verification that the required sequence of events occur in the required order compared to some standard. A device that checks to make sure that glove box screws are always attached in the same order each time would be a motion-step check.
Inspection can be broken down into judgment, informative, and source inspection.
In judgment inspection, some review of the product is condutcted, typically by person who is not directly attached to the production of the part. Judgment inspection is typically the end-of-line or before shipment 100% containment. One of the important characteristics of judgment inspection is that provides little if any feedback.
Informative inspection forms a feedback loop, revealing where the mistake comes. There are two types of Informative Inspection: Successive and Self-Inspection.
Where Mistake Proofing Works Well
1.Manual operations where work vigilance is needed.
2.Where mispositioning can occur.
3.Where adjustment is required.
4.Where teams need common-sense tools and not another buzz-word.
5.Where SPC is difficult to apply or apparently ineffective.
6.Where attributes not measurements are important.
7.Where training cost and employee turnover are high.
8.Where mixed model production occurs.
9.Where customers make mistakes and blame the service provider.
10.Where special causes can reoccur.
11.Where external failure costs dramatically exceed internal failure costs.
Where Mistake Proofing Does NOT Work Well
2.Product rate is very fast.
3.Shifts occur more rapidly than they can be responded to.
4.Self-checks when control charts are used effectively.
Error Proofing Components
- Feature Check
- Count Check
- Motion-step Check
- Judgment Inspection
- Informative Inspection
- Source Inspection