“Current Process Controls” and the “Detection Rank” assigned to them

[José Pedro Espinoza]

Here is a question about the “Current Process Controls” and the “Detection Rank” assigned to them that has been bother me for a while, to illustrate my concern let say that we are doing a FMEA about the process of how to make a “cup of coffee with sugar”, this process have several “Operations”: 10 – Heat the water, 20 – Pour the water on the cup, 30 – Add 30 grams of sugar, 40 – Add 30 grams of instant coffee, etc.

If we are doing the FMEA for the operation #30 –“Add 30 grams of sugar”, we can say that one of the Requirements of the operation is “to put 30 grams of sugar”, and then one “Failure Mode” would be “Add more than 30 grams of sugar”, one of the “Effects” of this “Failure Mode” can be “coffee too sweet” and just to keep it short assume that one of the “Causes” of the “Failure Mode” is “Lack of measuring device on the operation”, and if the process has an inspection operation that check the “sweetness” of the coffee, I say that this operation is checking the “Effect” (sweetness) not the “Cause” (lack of measuring device) neither the “Failure Mode” (Add more than 30 grams of sugar),so the detection rank must be “10” because it is testing the “Effect” not the cause or the failure mode. Am I right?

[BradM]

Can anyone provide some assistance to José?

[Marc]

One more bump, just in case someone can help with this.

[Jim Wynne]

In Reply to Parent Post by José Pedro Espinoza

Here is a question about the “Current Process Controls” and the “Detection Rank” assigned to them that has been bother me for a while, to illustrate my concern let say that we are doing a FMEA about the process of how to make a “cup of coffee with sugar”, this process have several “Operations”: 10 – Heat the water, 20 – Pour the water on the cup, 30 – Add 30 grams of sugar, 40 – Add 30 grams of instant coffee, etc.

If we are doing the FMEA for the operation #30 –“Add 30 grams of sugar”, we can say that one of the Requirements of the operation is “to put 30 grams of sugar”, and then one “Failure Mode” would be “Add more than 30 grams of sugar”, one of the “Effects” of this “Failure Mode” can be “coffee too sweet” and just to keep it short assume that one of the “Causes” of the “Failure Mode” is “Lack of measuring device on the operation”, and if the process has an inspection operation that check the “sweetness” of the coffee, I say that this operation is checking the “Effect” (sweetness) not the “Cause” (lack of measuring device) neither the “Failure Mode” (Add more than 30 grams of sugar),so the detection rank must be “10” because it is testing the “Effect” not the cause or the failure mode. Am I right?

I'm always a little leery of examples that deal with widgets or cups of coffee, but I'll take a stab. What if you did it this way:

1. Operation = Add 30 grams of sugar
2. Potential Failure mode= No measuring device
3. Potential Causes: (A) Scoop missing; (B) Plan didn't account for measurement device

Now you can deal with the actual causes; detection can be related to whether or not the scoop is present before starting to make coffee, and occurrence to the likelihood of the scoop being missing.

[tymer5]

It has been awhile since I have posted here due to a very busy schedule, but I believe I can answer José’s question. Here is how I do process FMEAs and how I train to do them. It is very mechanical and you only have to follow this step by step.

Whenever you have to do a process FMEA it is important to get the right level. Think work instruction level and value add. What do I do that adds value.

Now map your process. Yep that's right you have to create a process map. Not the ideal process map but what is actually being done.

Now for each step in your process map create a SIPOC (Supplier, Input, Process, Output, Customer). Don't skip this step! It is a huge time saver if you do it. You are now ready to work the FMEA.

Put your process step (P from SIPOC) into the step column of the FMEA.

Put your Outcomes (O from SIPOC) into the Process outcome/Requirements of the FMEA.

Consider your outcome and your failure modes that come from it and write them down in your FMEA. (Not at all, too little, too slow, too much, too fast, meets the requirement intermittently, unevenly or radically.)

Consider each customer and what impact it will have on them if you fail to meet the requirement in the manner stated in the failure Mode. Place the effect in the FMEA and rank it with the appropriate severity value per AIAG standard. Consider each high severity effect and determine if there are any recommended actions you can take to eliminate the Failure mode or mitigate the effect experienced by the customer. Write those RAs in the RA section of the FMEA.

Now consider each input and how it can go wrong. These are your causes. So from the example given the sugar is an input and you list how it will go wrong. (salt not sugar, sugar is old and hard, wrong sugar available (brown, white, powder). The measuring cup is also an input, and you want to list all of the ways it can go wrong as well. (Dirty, wet, hole in it, wrong size....) Rank each cause with its appropriate occurrence value. Consider the criticality of each Failure Mode, Cause combination and create any recommended actions. You are ready to move on to the controls.

Consider each failure mode and determine what controls you have in place to prevent or detect it. Consider each cause and determine what controls you have in place to prevent or detect it. Rank each detection control appropriately. Consider if there are any more RAs you can or should take regarding the controls, and do them. Your FMEA is now mature, and you only need to follow up on your RAs to ensure they are completed, and to create your Control Plan from the controls listed in your FMEA.

This method makes your FMEA very mechanical method makes your FMEA a simple matter of execution. It eliminates team conflicts and covers each process evenly and consistently. It avoids the common confusion of "is this input a requirement?" You will also see that a little pre-work up front of a process map and a SIPOC is a real time saver and creates consistency of approach.

I hope you find this of benefit.

Best of luck!

Tymer5

[José Pedro Espinoza]

Thank for your answer, but my concern still is with the ranking of the controls, becuase if they are detecting the defect (coffee too sweet) not the cause of the failure mode or the failure mode itself, what would be the correct ranking? (I'm still think it must be 10, but I would like to hear others opinions)

[Jim Wynne]

In Reply to Parent Post by José Pedro Espinoza

Thank for your answer, but my concern still is with the ranking of the controls, becuase if they are detecting the defect (coffee too sweet) not the cause of the failure mode or the failure mode itself, what would be the correct ranking? (I'm still think it must be 10, but I would like to hear others opinions)

The problem is that the failure mode (Add > 30 grams sugar) and the effect (coffee too sweet) are just different descriptions of the same problem. How did you determine that 30 grams was the correct amount? Wouldn't that determination mean that adding more than 30 grams would make the coffee too sweet, just as adding less would mean the coffee isn't sweet enough? Detecting excessive sweetness is detecting the failure mode (as you've described it) isn't it?

As far as the detection rating is concerned, the question should be, "What is the likelihood that the product (the finished cup of coffee) will be consumed without detecting the failure mode?" It depends. Some people might be more likely to detect an excessive amount of sugar than others. Who can tell the difference (in tasting the coffee) between 30 grams and 30.1 grams? It's not a failure until (A) someone complains and (B) you can demonstrate that more than 30 grams of sugar was used.

[Neil Dickie]

When assessing the prevention & detection controls; on the first pass through we score based on the potential cause of the failure. The idea being that if we can eliminate the cause or improve its' detection then we have improved our "right first time" and quality costs. Where we have not been able to reduce the risk at this level and we have a high detection score (9-10) then we look at the failure mode to ensure we can detect this to protect our customer. This is always a test at the end of each process step (like the tea tester for sweetness), but this doesn't prevent us from making the error in the first place. This score then goes on the FMEA at this stage.

Hope this helps