Search the Elsmar Cove!
**Search ALL of** with DuckDuckGo including content not in the forum - Search results with No ads.

The FMEA Mini-Series: Using an FMEA vs. SxO for Prioritizing

  • Thread starter kim.humphrey@at.siemens.c
  • Start date

Arun Kumar

Hi all,

My opinion is that the RPN indication doesnot make sure that your process is safe-guarded against potential failure modes with the controls that you identify. There is an inherent methodology called Process review that can be practiced for the same. For eg., consider the following option:
S=5*0=7* D=5 = 175. We come to the conclusion that an action has to be taken.
But consider the following:
S=9 * O=5 * D=1
This will give an RPN of 45, but when the problem occurs, it will have a higher impact ont he customer. Hence the analysis tool has to be changed. What we are discussing will not work if you are targetting improvement.

Arun Kumar


Captain Nice
Staff member
Just a note: RonSmith is unregistered. I don't know exactly what the occasion is, but today Ron has left 7 messages each of which has been a confirmation of omnex. I wouldn't question this particularly, but with no registration it seems like a lot of love for omnex. I do hope this is not SPAM.
Arun, I feel that the FMEA manual provides all the guidance necessary to identify and analyse potential problems.

Ron, I have tried many software programs; the only one that gets my vote is MS excel.
We tend to make things to difficult by jumping from new fad to new fad and the one thing I don't need is some overated software package making my life more difficult.


RPNis an indication of how the safe the process is against the potential failure modes. However to have a perfect FMEA the following needs to be considered.
01. Potential Failure causes shall be identified taking into consideration Man, Material, Macheines, Methods & Environment.
02. Process Controls need to be identified from the following:
a.Controls that can Prevent the Failure Mode.
b.Controls that the Prevent the Cause of the Failure Mode.
c. Controls that can detect the Failure Mode.
d. Controls that can detect the Potential Cause of the Failure Mode.

If a FMEA is prepared taking into consideration the above the resulting RPN will have a close link with the Process and the Failure Modes.

Considering your example 2 even thought the RPN is only 45, Still action is required since the Severity is 9 which is very high which can affect the safety of the user & can be against the govt. regulations.

[This message has been edited by Marc Smith (edited 18 April 2001).]

Al Dyer

If the severity is 9, occurrence is 1, and detection is 1, would you need to act on the situation.

I can't see why, severity is supposed to be determined by the customer and hypothetically out of your control. (internal or external customer doesn't matter)

As a supplier I would worry more about occurrence & detection than severity and RPN. O & D are within my control. This is not to say that RPN is discounted, if your O & D are high the actions would be commensurate with the severity level.



Paul Alexander

On the Severity issue - if it is high - could you not suggest a design change to lessen its impact?

Only a thought.......

Al Dyer

Originally posted by Paul Alexander:
On the Severity issue - if it is high - could you not suggest a design change to lessen its impact?

Only a thought.......
You sure could, it always helps to improve youself and help the customer!


Help Me

I would suggest that there is NO way to lower a severity rating with a design change.

For example, suppose the design in question is for a parachute. The severity of a failure to open would have to be a 10. How do you lower that severity?? A secondary parachute? That doesn't really lessen the severity, though. It just shifts it to the "n'th" redundant chute.

The fact of the matter is that severity of failure is not a function of the design. However, controlling the OCCURENCE and DETECTION RPN's is definitely facilitated by design alternatives, ie. design a chute with a 99.9999% successful deployment rating, and throw in a couple of redundant back-up chutes with equal success ratings, and your occurrence is extremely low.

But if the chutes don't open......

The severity rating cannot change. At least, I don't understand how.

What am I missing?
Top Bottom