DFMEA Severity Ranking of EMC Non-compliance

Six Sigma – iSixSigma Forums General Forums Tools & Templates DFMEA Severity Ranking of EMC Non-compliance

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
  • #53952

    Henrik Orvin

    Hi all,

    I’d like to hear your opinion on how to rank the severity of effects like non-compliance to EMC standards. The issue I’m facing is that I can interpret the non-compliance to EMC standards in two ways:

    1. As non-compliance with government regulations, and thus it should be scored 9 or 10.
    2. I could argue that in most cases this would not be a problem, and the effect would be only be an annoyance for a few customers, which means a severity score of 2 or 3.

    If I use the first interpretation, there will be a lot of very high scoring failure modes in my FMEA, and I lose the prioritization, which is part of the reason for doing the FMEA in the first place. Therefore I tend to like the second interpretation. What do you think…? Would you include all kinds of standards that a product must fulfill in “government regulations” and score the severity 9 or 10…? Can we limit the definition of “government regulations”, e.g. only to something which has to do with safety…?



    Mike Carnell

    This is a little difficult to answer but let’s try. Basically the rankings in a FMEA is a Likert Scale of sorts. The ends are extremes and the middle is a minimally acceptable condition. 10 is not necessarily an optimal liket scale (normally 5 or 7).

    I am not sure why you consider non-compliance with government regulation as a non issue but I would assume that would constitute a “severe condition.” If you could get non-compliance and some sort of health risk – the combination would be the most severe.

    I haven’t had to play with this stuff in almost 2 decades but it used to be measured in terms of voltage. That makes it much easier 0 or some low level out is a 1 (least severe) and work your way up to whatever the maximum output would be – pull out all the shielding and see what it does (max severity).

    Just my opinion.



    @hepe – really comes down to what are the effects for non-compliance to the process/product being evaluated. If I’m evaluating a design, and EMC compliance will not signficantly affect the performance, then it might be low. If I’m looking at the organization, and EMC non-compliance will result in a fine or will fail certification (UL, CE or other) and thus will not be allowed for sale, then this could be very high.
    I’ve never been a fan of the 1-10 scale (as my good friend Mike identifies, this would be better at a 5 or 7 level Likert scale). I will often force the raters to use a 0, 1, 3, 9 rating scale – this forces the rating team to think is there an effect (then at least 1), is it a critical failure, then clearly a 9, if not, then is it more than just a nuisance, then a 3, otherwise, it’s a 1. Ultimately, you are looking to prioritize the identified risks. It doesn’t matter how many discrete buckets they get organized into, so long as they are prioritized.

Viewing 3 posts - 1 through 3 (of 3 total)

You must be logged in to reply to this topic.