My Shopping Cart
There are no items in your shopping cart.

Dynamic Control Plans Drive Continual Improvement (MFG361I)

Presented by: William A. Levinson, P.E.
(*) Single User Price. For multiple users please call 1-844-384-4744
Pre Recorded Webinar
60 minutes
Event Description
See the sneak peek before you buy!

Know How the Dynamic Control Plan Fits Into the Quality Management System to Support AQP

The control plan is an important part of advanced quality planning (AQP). It describes how, in terms of measurement procedures and equipment, the process is to measure and control critical to quality (CTQ) product characteristics. The control plan is in turn the result of the deliverables from the Failure Mode and Effects Analysis (FMEA), which is another element of AQP.

The FMEA and the control plan can also both be shown in context with quality function deployment (QFD) or the house of quality. The voice of the customer as reflected in QFD is an input that defines critical and significant product characteristics for the design FMEA. The design FMEA provides inputs to the process FMEA, whose outputs dictate the nature of the control plan.

The tabular format of the FMEA includes a row for each part (or process step), its potential failure mode (how it can fail) and effect (the consequences if it does), calculation of its risk priority number (RPN), existing controls, and action plans to reduce the RPN. The dynamic control plan is a logical extension of the FMEA with additional columns for control plan elements such as the measurement method (gage, sample size, and sample frequency), gage capability, process capability, and out of control action plan. This makes the FMEA and control plan into a single entity.

The result is a living document that addresses the risks (in terms of severity, likelihood of occurrence, and likelihood of non-detection) of non-conformance of CTQ characteristics. "Living document" means it is updated to reflect changes and improvements in the product and the process. The dynamic control plan also supports ISO 9001:2015's provisions for actions to address risks and opportunities, and also knowledge retention.

Join expert speaker William A. Levinson to learn how the dynamic control plan fits into the quality management system to support AQP, and also supports ISO 9001:2015's provisions for activities to address risks and opportunities, and knowledge retention. Know how to calculate the risk priority number (RPN), but also recognize that some RPNs are more equal than others. That is, a RPN of 100 with 10 severity, 2 occurrence, and 5 detection requires far more attention than an RPN of 200 with 2 severity, 10 occurrence, and 10 detection. Go beyond textbook FMEA by recognizing that risk involves not only the individual occurrence rating but also the number of parts or transactions involved; that is, the number of times we are exposed to the risk. Know how to combine the FMEA and control plan to get a living document, i.e. one that is updated in response to changes in the process, product, or customer needs, that focuses priorities on critical to quality features and ensures closed-loop actions to reduce risks and improve quality.

Session Highlights

  • Learn the role of the FMEA and the control plan in advanced quality planning (AQP)
  • Recognize critical characteristics (those that can impact safety or regulatory compliance) and significant characteristics (those that can impact conformance to customer requirements). These must be addressed by the control plan in terms of:
    • Measurement/inspection process including frequency and sample size
    • Measurement equipment (gages), which is in turn subject to calibration and measurement systems analysis (MSA) requirements.
  • Know the numerical rating systems for severity (S), likelihood of occurrence (O), and likelihood of detection (D), and how to multiply them to get a risk priority number (RPN).
    • Also recognize, however, that a severity of 9 or 10 puts the failure mode into the 'legal zone' regardless of RPN.
    • Recognize Army Techniques Publication (ATP) 5-19, "Risk Management," as a simpler alternative to FMEA when it is difficult to quantify the exact probability of occurrence and/or non-detection.
    • ATP 5-19 also recognizes that risk involves not only the individual chance of occurrence (p, nonconforming fraction as reflected by the FMEA occurrence rating), but also the number of exposures to it (n, the number of items exposed to the risk). It therefore offers a perspective that FMEA does not. Suppose for example that the probability that a nurse will connect an enteric feeding tube to an intravenous line is 1 in 1.5 million. This low probability would result in an FMEA occurrence rating of 1 (best possible), which would imply little need for action despite the maximum severity rating. When we recognize that nurses perform this task millions of times per year, however, multiple fatalities become almost certain. The same principle that risk is really np rather than just p applies to safety-related automobile components and many other products.
  • Know how to combine the FMEA and control plan into a dynamic control plan that addresses directly and explicitly the CTQ characteristics and their associated risks.

Who Should Attend

  • Quality management professionals
  • Engineers and technicians with quality management system responsibilities
  • Design engineers
  • People with responsibilities for ISO 9001:2015 or ISO/TS 16949 compliant systems
  • VP Engineering
  • Office Managers & Admin

Click here to read Transcript


Important Copyright Information


Dear Valued Customer:

We have received quite a number of questions regarding the permissibility of copying and distributing our print transcripts, as well as questions regarding the

conferencing of calls to other sites by our listeners. The main questions include:

May I photocopy my transcript packet and send it to others?

May I fax or email a copy of my transcript packet to one or more colleagues to use?

The quick answer to all of the above questions is “No.” If you or others you know are doing any of the above, you are infringing on the audioconferences’ copyright.

Probably you weren’t aware this is a legal infraction, but it is, and it is punishable by law through damage awards, etc.

To explore low-cost alternatives for legally having more than one copy of this transcript, please contact our Customer Relations Department at (800)223-8720. Our

representatives will be pleased to help while staying in compliance with copyright law.


Audio SolutionZ

P.S. We trust that our audience of professionals such as you will respect our legal copyright “on the honor system.” But if we learn of willful violations, we may have

to seek legal remedies.

P.P.S. If you know of any unauthorized copying or distribution of attendee packets or dial-in information, please let me know about it in confidence so that we may

follow up with the party and set them on the right path. (Your name will be kept confidential.)

Comment (A): I would like now to introduce your speaker for today, William A. Levinson.

William is the principal of Levinson Production – Productivity Systems, P.C. He is an ASQ Fellow, Certified Quality Engineer, Quality Auditor, Quality Manager, Reliability Engineer, and Six Sigma Black Belt.

He is also the author of several books on quality, productivity, and management, of which the most is “The Expanded and Annotated My Life and Work: Henry Ford's Universal Code for World-Class Success.”

William, welcome to the program. We are now ready to begin.

Comment (WL): Good afternoon and thank you for coming. We're going to cover “Dynamic
Control Plans”. The key learning objectives are as follows:

First, what is a dynamic control plan, role of the FMEA – Failure Mode Effects Analysis, and Advance Quality Planning or APQP which is the automotive version of letters required in automotive under ISO/TS 16949, how to prepare the FMEA, and then when you add the control plan to the FMEA that is combine them into essentially a single document, you get a dynamic control plan.

Okay. What is a dynamic control plan? The FMEA and the control plan are the key elements of advanced quality planning. And the handout reference you should have gotten has a reference, JD
Marhevko – “The Marriage of a FMEA and Control Plan”. And that's how she describes a dynamic control plan - or the title that presentation describe the concept perfectly. And it is available online for anybody who wants it.

The FMEA plays a central role in defining the control plan. So it's a logical extension to expand the FMEA and the control plan into a single document or spreadsheet. And what you get is a dynamic control plan.

Dynamic means changing, adapting, and evolving rather than static, also known as a living document. It's a closed loop improvement process rather than the once-and-done activity. Well, why – once we've done the FMEA and put together a control plan to go with it, why do we have to expand on it and continue to improve it?

Well, the issue is management of change, if you heard that, or MOC, Management of Change. It's a safety-related term from the chemical process industry. And it recognizes that any change in the process can create a potential hazard that could include a change in the process or even maintenance or repairs, anything that's different or unusual can create a problem.

The approach to management of change is similar to that of an FMEA or the Army's Risk Management Process which by the way is an outstanding reference. And the other nice thing about it is it’s a publication of the U.S. government which puts it into the public domain.

The figure on this page is from Army Techniques Publication 5-19, “Risk Management”. That's
also available online and as I said it's in the public domain. What you have here is a close loop improvement cycle. And then FMEA follows the same improvement cycle which isn’t surprising.

The Army, in fact, has a risk assessment worksheet that's very similar in concept to an FMEA but it's a lot simpler. Other Hazard Identification and Risk Assessment approaches also follow this improvement cycle, which also by the way is consistent with ISO 31001.

So ISO 31000 which is the Risk Management Standard from ISO, they're all very similar.

Okay. Any change in a cause and effect diagram category creates a risk. For example, if the job is operator dependent, manpower, changing the operator can create a risk, changes in the machine, changes in the material. There's an example in the handout notes called “Recognize Hazards to Recognize

And they give an example of a process that uses caustic soda, sodium hydroxide. And the danger that can happen from changing suppliers even if the – even if the specified concentration doesn't change and is deemed otherwise identical. If the supplier is using municipal water with trace chlorine in it, then you have trace chlorine in your caustic soda solution or probably in the form of sodium chloride by that time.

But that can eventually degrade stainless steel pipes which obviously - or it can degrade and
cause cracks in the stainless steel pipes but that really designed to stand up to chloride ions for long periods of time. So a trace in the material that you don't – you can't even think of or probably wouldn't think of ordinarily can cause very unpleasant and unforeseen consequences.

With regards to environment, Henry Ford cited a situation where there were – there were arguments and even threats of lawsuits between a customer and supplier because both of them had their inspectors in the same shop inspecting the parts. And the customer was accusing the supplier of trying to pass nonconforming parts. And the supplier was accusing the customer of trying to cheap them by calling parts nonconforming.

And it turned out that one of the inspectors was on the cold side of the shop and the other one was on the warm side of the shop. So because of thermal coefficient of expansion, their gages were in fact telling them exactly what they were claiming. Nobody was lying. It was just a question at the different environment. So anything – any change in the traditional cause and effect categories creates a risk and the same concept carries over into FMEA.

A startup in the chemical process industry and a new setup of a corresponding situation in discrete manufacturing also are changes. This is why chemical process accidents often happened on startup of a batch reaction. The handout cites the Process Safety Beacon with the case studies of industrial accidents. And in – when the chemical reactor was involved, it's almost always a batch reactor rather than the continuous flow reactor.

Similarly, where we do tightened inspection of the first parts from a new setup is a control and that FMEAs and control plan talk about controls of the tightened inspections then administrative control that addresses the risk from a new setup.

Okay. So the issue of management of change, with regard to quality as well as safety, requires
FMEAs and control plans to evolve and adapt accordingly.

Okay, next. Roles of FMEA and Control Plan in Advanced Quality Planning. The handout by the way – we're talking – going to talk about AQP and AQ – and Advance – APQP. There's a free reference on Production Part Approval Process from Cooper Industries and the reference is given in the handout.

That's nice to have it because the one from the Automotive Industry Action Group, I think it's like $150 unless your company is a member of that organization. They're excellent references. But actually, like the ISO Standards they're rather expensive. So hopefully that will be a useful reference.

Okay. The FMEA and the control plan were both part - parts of AQP and APQP. The latter one is on – focuses on automotive. And as I've mentioned, there's a reference available from the Automotive Industry Action Group.

Okay. AQP is a set of cross-functional activities – that's an extremely important concept that it is
cross-functional -- and it delivers quality plan that aligns the process, product, or service with customer requirements. The process is documented which makes it a quality record as far as ISO 9000 is concerned.

Applications. Design of new products that's where your design FMEA comes into play.
Development of new processes, that's where your process FMEA comes into play. Notice, again, the issue of management of change because it also applies for process or product changes that may affect safety or regulatory compliance and also transfer of tooling to new processes or plants. Again, anything new or different creates a risk of quality or other problems.

Here's a figure that explains it. This is the house of quality or Quality Function Deployment where we have the house of quality on top or the customer requirements are translated into product features and importance of those product features.

The importance of a product features, the criticality of the product features. Then provides feeds of the design FMEA. The output of the design FMEA identifies the critical and significant characteristics of the product that has to be accounted for.

And that influences the process FMEA because the corresponding process steps or operations
have to be controlled which, of course, is the focus in the control plan, to make sure that they are taking care in those critical and significant characteristics. When you put the process FMEA in the control plan side by side, you get a dynamic control plan - DCP.

Okay. What is an FMEA? It's a cross-functional planning activity that identifies potential failure modes. Basically, what can go wrong is the simple term for design or process. We're going to focus on processes because the process FMEA is what really dictates the control plan.

As I've mentioned earlier, the design FMEA indicates what we have to look at in the process FMEA that the deliverables of the process FMEA show how we have to prepare the control plan. This is mandatory in automotive applications. And actually many aspects are also mandatory in ISO 9000.

The coded material is from the Automotive Industry Action Group, FMEA manual, identifying assess in terms of severity and likelihood, the ways in which your product or process might fail.

Now, the Army Risk Management process also does that. They also use a severity and a frequency of occurrence rating. They don't use a detection rating which makes it a lot simpler. But basically, they uses the same approach. It's a logical, rational approach. And many organizations have different ways of doing it.

Okay. Once you've identified the potential risks or hazards, identify actions and we call this controls for process – process controls – to eliminate or mitigate the failure modes in question. And finally, document it. In other words, it becomes - everything becomes quality record under ISO 9001.

Okay. The code is again from the Automotive Industry Action Group. FMEAs also apply to internal as well as external customers. Note the tie-in with Quality Function Deployment, which I showed on the previous slides.

Okay. The FMEA deliverables include requirements for controls of critical to quality characteristics. That's where we get the control plan for the realization process. And the dynamic control Plan is simply going to combine the process FMEA and the control plan into a single entity.

Dynamic means it's updated. It's adapting. In other words, it's a living document that drives continuous improvement. And as I've mentioned earlier, management of change, anytime we do something new or different, changing suppliers, changing the design, changing the environment or anything else creates an opportunity for trouble. And it might be necessary to update the FMEA and the control plan accordingly.

And addition of course, even if there aren’t any changes, if we find something we've overlooked or a way of suppressing a risk that we haven't thought of before, we would update the FMEA and the control plan to reflect the new – the new control and the new situation.

Types of FMEAs. We have – I've mentioned, design then process FMEA. And again, the latter is
going to drive the nature of the control plan. And actually we put side by side with the control plan to get a dynamic control plan.

The design FMEA, I'm not going to spend a lot of time on this because our focus is going to be on the process FMEA which is part of the dynamic control plan. But the importance of the design FMEA cannot be – cannot be over emphasized because we have to build quality into the product rather than inspecting it into the product.

And not only that, there's a limit on how much quality we can build into a badly designed product. You talk about process capability, making sure the process is capable of producing the product to how much variation is there in the process relative to the specification limit. If the product is badly designed, there's a limit as to how much quality we can even build into a badly designed product.

The phrase, “lipstick on the pig” comes to mind. State of the art process can do only so much for a product that's not designed for manufacturer. It simply not going to work.

I do remember from the semiconductor industry, this is probably about 25 or 30 years ago, we
were etching holes in a dielectric layer to allow electrical connections – three dimensional electrical connections.

And one of the design had their holes so close together that the action which was hydrofluoric
acid, it also etches sideways as it etches down and that was etching the holes into each other, that the holes would end up -- that were supposed to be side by side – end up having channels etched between them which would lead to a short circuit.

That was an example of something that was not manufacturable by the existing technology and that's just one example. The handout gives examples of extremely wasteful processes that whether, even if the product can be manufactured with a high level of quality, it still waste an enormous amount of the material.

This is from an American Society for Quality presentation from 1992. The reference is in the
handout. It's not available online though. It's just from my notes. And in the event, they have a study by
Ricoh Copier that set the design improvement or design – a positive design change have 100 to 1 pay off. A process improvement has 10 to 1 pay off. And a - usually its corrective action during the manufacturing stage has one to one pay off which underscores the need for design for manufacture. We can design not only quality into the product but also cost effectiveness into the product.

And Henry Ford said of this - this is what Henry Ford talked about, essentially designed for manufacturing and designed for assembly except they didn't call it that then. But this was more than – more than 90 years ago, 1922, that would be about what, 90 – yeah, 94 years ago when he wrote this and obviously, design for manufacturing and design for assembly.

DFMEA also supports the ISO 9000 (9001?) provision for Design and Development. There's a
similar one in the 2015 version of the standard. It just had a different numbering. And of course, it's required by ISO/TS 16949 in automotive.

The DFMEA Deliverables. And this is where it connects to the PFMEA, the Process Failure Mode Effects Analysis. Identifies the critical and significant characteristics. Critical, this is - the definitions are from Dean Stamatis book on “Advanced Quality Planning” which is shown in the handout.

The critical and significant characteristics are the critical to quality characteristics that has to be
controlled, which is the function of a control plan, to ensure – to ensure quality. Critical and significant characteristics also come from process FMEAs and other sources. Again, this is from the Stamatis’ reference identified by special symbols especially in automotive applications. And again, this dictate the control plan which - or the dynamic control plan as we're going to see.

We're going to refer to them generically as critical to quality. Basically, any characteristics in which a nonconformance fails to meet the customer's needs, so then encompasses critical and significant characteristics.

We're going now to go to the process FMEA which evaluates the process operations and defines the required controls. When we merged the process FMEA into the control plan, we get a dynamic control plan. The process FMEA ties in with the process flowchart.

Now, the flowchart is of course, a basic quality tool. And actually a lot of people now use flowcharts to support their ISO 9001 documentation because ISO 9001 takes a process perspective of a business system rather than looking, for example, the individual or department activities, which is how the older versions of ISO worked.

ISO 9001, 2000, and subsequent versions want to look at your process from a process
perspective. So there's a process flowcharts often part of the quality manual or part of the supporting documentation for the quality manual.

So the Process FMEA ties in directly with your flowchart and it addresses the process activities
that affect critical or significant characteristics. It can also be cause – be called an anticipatory cause and effect diagram. In other words, cause and effect diagram is usually done after the fact when a defect has been discovered or some other problem has been discovered, use the cause and effect diagram for brainstorming to find out where it might have come from.

A process FMEA also can use the cause and effect diagram categories. But to anticipate and
prevent the trouble rather than trying to react to it after it has happened. Begin with a flowchart risk assessment of the process. Again, we're using the process flowchart which a basic quality management tool.

Otherwise, it's similar in concept to DFMEA. And the two are generally prepared in the same manner. Again, we're going to identify potential, essentially, what can go wrong? How could the process failed to meet process requirements or design intent?

Okay. Action plans are required for high Risk Priority Numbers. RPN stands for Risk Priority Numbers. And also for high-severity failure modes. We're going to see by the way the – all these ratings, severity, occurrence, and detection are on 1 to 10 scale with 1 being the best and 10 being worst. So a higher – you multiply them together to get a Risk Priority Number. And a higher Risk Priority Number means simply that there's more risk.

A couple of things to point out. The occurrence rating of 4 or more indicates a non-capable process. If use the occurrence rating as recommended by the Automotive Industry Action Group, namely the chance of occurrence or essentially defects for million opportunities, anything within the occurrence rating of 4 or more indicates a process with a capability of less than 1.33 which is generally defined as a non-capable process, 1.33 is generally the come off for a process being considered capable.

And also the RPN is not everything. That – the reference gave by JD Marhevko – The Marriage of a FMEA and control plan points that out very explicitly. The Risk Priority Number is not everything. In other words, you can't just rank the Risk Priority Numbers to determine which of your problem has the highest priority.

As a simple example, if it has a catastrophic severity rating, in other words, it endangers human life or safety, or compliance with the government regulation that requires attention regardless of the Risk Priority Number.

Let’s say, you have a 10 severity rating and maybe in occurrence of two - and a detection of 2, you have an RPN of 40 which doesn't really seem all that high, but because of the 10 severity, it requires attention regardless of how good the occurrence and detection ratings are.

Also, that occurrence rating can be deceptive which is surprising because FMEA has been around for a long time. But what they don't take into account and what the Army's Risk Management process does is that the risk is not just the individual probability of occurrence. It's also the frequency with which were exposed to the risk.

So if we're mass producing something or performing the same job very large number of times, our risk is going to increase regardless of the individual chance of occurrence simply because we're exposed to the risk so many times. More on that in a couple of slides.

Okay. FMEA and the Control Plan. The critical to quality characteristics require design and
process controls and some of these are listed. Notice, by the way, the tie-in to other aspects of the quality system like machine and process capability.

Process capability studies are basic and often required gages and gage capability, Statistical Process Control. By the way, those latter two gage capability and the SPC, the Automotive Industry
Action Group has manual for both those which are in fact required for many automotive applications.

Other aspects, again, we gone to that management of change issue where it says, FMEA should be performed for all new parts, changed parts, and parts carried over into new applications. In other words, anything new or different which in a safety concept that are management of change is the same concept.

Okay. It's cross-functional. In other words, it's not on one side where it's not done, for example, by a design engineer or quality engineer. You involve everybody relevant. For example, the manufacturing people who actually have to make the product has to be involved with it as well. It's a living document. As I've mentioned before, should be updated as necessary. And QFD, as we showed earlier, fits into the – fits into the requirements.

Okay. What does an FMEA do? It identifies proactively potential trouble sources and critical to
quality characteristics and that's where we get the control plan. Prioritize is, again, subject to those caveats I've mentioned, Risk Priority Number is not everything. And finally, removal or mitigation of the trouble sources.

It prioritizes the failure modes according to the severity of the failure, likelihood of the failure,
that's the occurrence rating, and the chance of detecting or mitigating the failure. And the product of those three ratings is, again, 1 is best and 10 is worst. You multiply them together to get the Risk Priority Number. And again, we're going to see RPN is not everything.

The next step is define, implement, and make sure they work – the action plans to remove or mitigate the failure modes. That's simply close loop corrective action system or close loop proactive system where you identify an improvement or, in this case, the control implement it. And finally, make sure it works, verify that it works that's part of any close loop improvement cycle, to remove or mitigate the failure modes.

And the action plans will then be reflected in the control plan. So when you put the process FMEA side by side with the control plan, this is – it's inevitable that the action plans will be reflected in the control plan.

Okay. Understand mechanism and mode. The failure mechanism is what causes the failure mode which is the subject as an FMEA. The mode then has a consequence with regard to a product or service failure.

For example, the GM ignitions switch, I think they have a weak spring in the ignition switches. I remember, the incident being described where you have a heavy keychain, it would actually turn off the – just turn the car off which was the consequence. Obviously, a 10 level severity because that creates the chance of an accident. So that would be the failure mode of the mechanism involve the weakness of the spring in the ignition switch.

Okay. The mechanism is why does it fail? And the – for example, cyclic stress, fracture due to a stress concentrator, hydrogen embrittlement, I think I've mentioned the chloride ion causing cracks in stainless steel pipes would be the mechanism. And the mode is what happen is - what happens when it fails. Resulting failure is the failure mode.

Next. How to prepare the FMEA? Okay. Again, the process flowchart which is a basic quality tool starting point and it consist of a columnar table as shown below. It has 13 columns. And the headings of those columns are from the Automotive Industry Action Group’s FMEA manual.

And this is where we're going to get to the dynamic control plan. The control plan also has a number of columns. And we're simply going to add those columns to the FMEA shown here, to get a dynamic control plan. This is where it's helpful to have a spreadsheet because the resulting document is probably going to be horizontally too long really to put on a standard sheet of paper.

The Army, by the way, in the reference has given in the handout, it's called the Deliberate Risk Assessment Worksheet, uses a similar but much less elaborate approach. And again, this is public domain, it’s a publication of the U.S. Government. Work is worst looking at though because even though it has a fewer columns, you'll see almost a one-to-one relationship between some of the Army's categories and what you see in the FMEA.
These are the column numbers item one as the item, part, or function. Two, is the potential failure mode. And it is how could the part fail to meet the design effects or the process requirements? In other words, what happens when it fails?

Now, the potential effect is the consequence of the failure mode. For example, ignition system shuts down (cargos) of electrical power and goes off - goes off the road before the driver can recover. So that's the consequence of the failure mode. And that's, by the way, where you design the severity rating.

Severity rating is the seriousness of the effect on the 1 to 10 scale. Usually reserved for effects that endanger human life or noncompliance with a government regulation. In other words, if the failure result in the chemical release or chemical spill that would also have a 10 severity; 8, the item is inoperable.

So for example, I've mentioned the GM ignitions switch problem. That would have a 10 severity
because it endangers human life. If a car simply won't start, that's an 8 severity. Obviously, the customer is going to be highly dissatisfied so that the car won't start, it's pretty hard to have a car crash.

Now, that severity quick off to a nine or 10 though if the vehicle is something like a fire engine,
or a police car, or an ambulance where if it doesn't start somebody could die as a result that normally a vehicle not starting would be an 8. However, lower severities can still result in customer dissatisfaction.
Obviously, the car won't start, the customer is going to be pretty unhappy.

And for a process FMEA, you have to consider process safety and continuity of operations.
Again, anything that the affects workplace safety would be a 9 or 10, of course.

Examples of severity, obviously a plane – if it can cause a plane crash, that's obviously a 10. There have been numerous incidents I know involving American Airlines. And I think Northwest where passengers were stranded on the runway for more than three hours. That would be a 9. More than three hours is not in compliance with the government regulation, the airline can be fine for it.

Also, anything that involves vehicle with the flashing red lights is a 9 or 10 severity by definition. And there have been situations where this – where the airline stranding people on the runway and that would not come out of the airplane is resulted in medical alerts – medical situations as required responses from ambulances, so that's obviously a 9 severity.

Brakes or ignitions switch failure – that ignitions switch will drive and that would mean the car
won't start, I mean actually while driving that will be a 9 or 10 severity.

Fuel pump failure, that actually happened to me. The car won't work. Then again, that could be higher because of my experience that happened in an intersection. And my car wouldn't start in the middle of an intersection so that could have created a hazard other than the car not going.

A squeaking part could be a – it will annoy the customer but the car still works so that might only
be a 2 to 4 severity.
The potential cause or mechanism. The mechanism is what causes the failure to occur. Likelihood of occurrence on a 1 to 10 scale in the handout, gives the Automotive Industry Action Group guidelines. For example, 4 severity is 1 in 2000 which by the way is a non-capable process. For process to be capable, the occurrence rating has to be 3 or less.

Here is the recommended occurrence ratings. Yeah, 1 in 15,000 means a process capability of 1.33 or more. And 1 in 1.5 million that is a Five Sigma process. It means the process capability is more than 1.67.

The occurrence rating is, however, not all-inclusive. And here's what the Army's Risk
Management document has to say about it. It says, “The probability is frequent – as defined and it cites specifically exposure. It's the frequency and the length of time were exposed to the hazard or hazards.”

And for manufacturing operation, if we're making a lot of parts like millions of parts even if the
individual chance of failure is very low, our risk can become very high if the part is say, critical to life or safety.

I phrase that as “Risk = NP” where N is the number of opportunities and P is the probability of occurrence rather than just P which is reflected by the FMEA occurrence rating. That NP, by the way, is the mean as the binomial distribution, if you're familiar with statistics. The expected number of nonconformances from the binomial distribution of number of parts, times the individual chance of being nonconforming.
This is a good example for the want of a nail. And the Marhevko reference also cites this.

Here we have a horse shoe where there's eight nails per shoe and you have four shoes per horse, and you N the horses in the regiment. So that means there's a lot of horse shoe nails a seemingly humble part. If you have eight per shoe, four per horse, and you have N the horses in the regiment, how many horses they have in the cavalry regiment?

So I think it was some – it was certainly in the 100. So you're looking at the lot of opportunities
for failure. And even if in the individual probability that the nail is effective or nonconforming, you multiply that by the number of opportunities for failure, you can have a real problem.

Now, the only thing on our side is that there's eight nails per shoes. So the failure of one nail will
probably not lose the shoe. It will increase the likelihood because it increases the stress on the other nails that's what’s called the shared parallel redundant system.
Now, of course, instead of horse nails today, we have fasteners. And I know there is a big issue with counterfeit fasteners. For example, SAE, if they want a SAE grade 8 which is the highest grade, there were certain unscrupulous suppliers who would label lower grade fasteners is grade 8 which of course creates an enormous risk.

Here's how the Army addresses that issue of Risk = NP, rather than just P. This is the Army's
Risk Assessment Matrix.
Now, notice the upfront similarity to an FMEA because you have a severity rating. The rose or the severity ratings ranging from catastrophic to negligible. And you have an occurrence ratings where it says, hazards probability ranges from unlikely as being the best, frequent as being the worst.

But here's what the Army's Risk Assessment Matrix in the FMEA per company. The – suppose
the – you have an occurrence rating of 1 for an FMEA. Say, 1 in 1.5 million is a – 1 in 1.5 million is the criterion for that. That would go in the unlikely column.

And let’s say, if the failure is catastrophic but the probability is unlikely, you would call that a
moderate risk. Note that can't be low because it's catastrophic. Automatically requires attention just like a severity rating of 9 or 10 requires attention.

However, if you have a couple of million chances for occurrence, your hazard probability goes up to frequent and then you have an extremely high risk. So you also have to account for – not only for the individual probability of occurrence but the number of times you're exposed to it.

Implications. Okay. Suppose 1 in 1000 of our product is defective, well, that's 100% poor quality to each customer involved. I think there is a local consumer watchdog article about some chavvy problems like I think a steering wheel problem and then electrical system problem. And GM answered,
“Well, we think that's only like 1 in 2000 cars might have that.”

Well. First of all, for the customers affected, that's 100% poor quality. And also because of the safety implications that simply not - unacceptable situation. When you have hundreds of thousands of that model on the road, the chance is something going wrong in the safety related context is a near certainty.

Controls should be what the Army called engineering controls rather than administrative controls
or what Henry Ford called “can't rather than don't”. And the definition of that is given in the handout. In other words, a mechanical control or other control that makes the defect or failure impossible is the only way to insure that you won't have trouble when you’re – when you’re doing the job thousands or millions of time.

And next, we have the current design controls or process controls. Again, remember, engineering controls versus worker vigilance. Shigeo Shingo uses the phrase “worker vigilance” many times in discussing quality related situations. Where he said, “Trouble was prevented through worker vigilance.” That's an immediate – that’s what the Army called an administrative control.

For example, a work instruction and then inspection instruction are saying be careful of this,
make sure of that. And, again, if you're doing the job thousands of times or millions of times, sooner or later, somebody – vigilance is going to fail. So anytime Shigeo Shingo talk about worker vigilance, that was an immediate warning sign that your control might be inadequate.

The Cooper Industries AQP reference, this is in the handout note says, “Controls must not be excessively dependent on visual inspection and should use a preventive techniques or what the Army called “engineering controls” and what Henry Ford called “can't rather than don't” whenever possible.

The detection rating is the chance of the existing controls will detect the problem before it causes
trouble, again, 1 to 10 scale. The detection rating is given by the EIG or highly subjective.

Now, here are some examples. For a detection rating of 1, the best possible self-check systems and Henry Ford used these things right and left. Essentially, 100% checking of parts with mechanical automated gages.

It wasn't the manual inspection. It was done automatically. They catch and reject the nonconforming work so they can see a certainty of detection. And jidoka or autonomation stops the process when an abnormal situation occurs.

Actually, Toyoda, now of course, it's spelled as Toyota, he had a power loom and the problem with looms was sometimes the thread would break and then it would weave nonconforming cause until the defect was finally caught. So he designed loom to detect a broken thread and stop automatically. That's an example of autonomation. And the handout and the reference as from additional information including a book by Shigeo Shingo on self-check systems and/or error-proofing.

Examples of detection. I've mentioned the automatic go/no-go gages. The Army would call those engineering controls. As Ford complain it was difficult. It's not impossible to detect casting defects. They didn't have a nondestructive test – testing techniques we have today. So that would have a very high detection rating. And of course, higher is worst....

Order Now To Read Full Version Along With Conference Material
About Our Speaker(s)

William A. Levinson | Quality and Productivity Management SpeakerWilliam A. Levinson P.E.
William A. Levinson, P.E., is the principal of Levinson Productivity Systems, P.C. He is an ASQ Fellow, Certified Quality Engineer, Quality Auditor, Quality Manager, Reliability Engineer, and Six Sigma Black Belt. He is also the author of several books on quality, productivity, and management, of which the most recent ... More info

Recommend this event to a colleague
  • Your Name:
  • Your E-mail:
  • Colleague's Name:
  • Colleague's E-mail:
    Event Title: Dynamic Control Plans Drive Continual Improvement
    Presenter(s): William A. Levinson, P.E.

Our Accreditation Partners Protection Status

Dear Valued Customers,

We regret to announce that ProEdTech LLC and all its affiliate brands will cease operations on April 1, 2019.

We are no longer able to fulfill online orders. We will fullfill all DVD and book orders already placed.

Customers of canceled webinars and subscription products may request a refund at (800) 223-8720 or You must do so by April 1, 2019.

Thank you for your business and loyalty over the years. We sincerely apologize for any inconvenience caused.

Best regards,
The ProEdTech Team