DFMEA Detection Rating

Quick reference

• 00:04 Hello, I'm Ray Sheen.
• 00:06 Well, it's now time to move to the third of scoring criteria for
• 00:09 the Design FMEA, and that is scoring detection.
• 00:15 The detection ratings are just a bit weird for the Design FMEA.
• 00:19 You are evaluating the probability,
• 00:21 that the failure would be detected during the design process.
• 00:25 The assumption being, that detected at that time,
• 00:27 a design change would be made to eliminate or control the failure mode.
• 00:32 What is a bit odd about the scoring, is that a high score is for low detection.
• 00:37 And a low score is for high detection.
• 00:40 for this reason, a better name for
• 00:42 this rating would probably have been Undetectability, not detection.
• 00:47 When doing this rating, you should first describe the detection method used in
• 00:51 the development process, to find this type of failure.
• 00:54 List that on the FMEA form, and then score it.
• 00:57 And one more time, I normally score the severity, occurrence and
• 01:01 detection for one failure mode, then go to the next failure mode.
• 01:05 I don't score all the severities, and then all the occurrences, and
• 01:09 then all the detections.
• 01:11 I work across the form horizontally.
• 01:14 Let's consider data sources.
• 01:16 Your primary source of information for this rating will be your design and
• 01:20 development process.
• 01:21 By that I mean analysis, testing, verification and validation.
• 01:25 That is done along with the design control aspects of the process.
• 01:29 An area I also like to include is the design configuration warning and
• 01:33 cautions, that are included in the design documentation.
• 01:36 If your organization does not have a standard process for
• 01:38 all of these, then use the project plan for the product development.
• 01:42 This area of rating can also consider customer experience with detection.
• 01:47 This would include the results of customer installation or application, and
• 01:52 the customer returns and complaint history.
• 01:54 For new products, you may wanna include a customer focus group.
• 01:57 They can help to make sure your development process is robust, and
• 02:01 considers all the customer uses and applications when doing your testing and
• 02:05 analysis.
• 02:06 This area will focus heavily on test data, but don't overlook the power
• 02:10 of a good simulation or analysis of your product or system.
• 02:14 Okay, let's look at the ratings.
• 02:16 In this case, we are detecting the failure mode.
• 02:19 The rating of 10 is the worst for detection.
• 02:22 In fact, it is the inability to detect the existence of the failure mode.
• 02:26 Use this score if there is no design control for this element, or
• 02:29 if the ability to either detect the cause of the failure, or
• 02:33 the failure itself can only be done by the customer once they are trying to
• 02:37 use the product, and they have been negatively impacted by that effect.
• 02:43 Now, let's look at the rating of nine.
• 02:45 This is used when there's a very remote chance of detecting the cause or failure.
• 02:49 Either the design controls like testing and analysis for
• 02:52 this condition are very weak, or the tests and analysis that are used
• 02:56 don't correlate with the actual experience or conditions at the customer.
• 03:00 This is also the score if the failure mode cannot be detected by you,
• 03:03 the manufacturer or supplier of the product, but the customer
• 03:06 is able to detect it immediately before they feel the effect of the failure.
• 03:11 The score of 8 is for the remote chance of detecting the failure or cause.
• 03:15 So in this case, you could catch the failure in a verification or
• 03:18 validation testing but only in testing that occurs after the design is
• 03:22 complete and frozen, which means it might easily slip through.
• 03:26 You can also use this score if the failure can be detected internally, but
• 03:30 only by random testing.
• 03:33 Usually, that means some type of random distructive testing of the product.
• 03:37 The score of 7 is for a very low chance of detection.
• 03:40 In this case, you're still relying on verification and
• 03:43 validation testing after design freeze.
• 03:46 But the testing does go all the way to failure.
• 03:49 So you can see if there is any design margin or
• 03:52 if the product is very fragile, with respect to this failure mode.
• 03:56 Next is the low chance of detection which gets a score of 6.
• 03:59 There is a fair amount of testing that is occurring with
• 04:02 respect to this failure mode and cause.
• 04:04 There is verification and validation testing that includes degradation testing.
• 04:08 This will enable modeling of the design to determine if it is robust with respect to
• 04:12 this failure mode, or the customer may install the product and
• 04:15 conduct their own testing, as part of the final acceptance of the product system.
• 04:19 That would obviously be much more realistic than internal testing, but
• 04:23 it's normally used only for custom built products.
• 04:26 This is also the score to use when there's a very good testing program, but
• 04:29 some of the customer requirements are not known, and therefore cannot be tested
• 04:34 or if different customers react differently to the failure.
• 04:37 The score of 5 is labeled the moderate chance of detection.
• 04:41 In this case, the validation tests are conducted prior to the design freeze,
• 04:45 providing time for the design team to make changes to the product,
• 04:49 to reduce the cause or the impact of the failure.
• 04:52 However, the nature of the testing is just to make sure everything passes the spec,
• 04:56 not to determine how much design margin exists.
• 04:59 The testing may also be done by a customer testing prototypes of the product, and
• 05:03 providing feedback before the design is completed.
• 05:07 Double 4, gives us a moderately high chance of detecting the failure
• 05:10 during development, and eliminating or reducing it.
• 05:13 This score is used when the validation tests are done prior to design freeze, and
• 05:17 the tests are run to the point of failure to determine the fragility,
• 05:20 and allow the team to add design margin.
• 05:23 This is also used if the failure mode is a well known condition that is tested and
• 05:28 evaluated.
• 05:29 During development as a stand-alone part of the development process.
• 05:33 Level 3, is a high chance of detection.
• 05:36 The validation tests are again done prior to design freeze.
• 05:39 And this time, they are done to measure degradation.
• 05:42 That way, the design team can predict when the design begins to become fragile and
• 05:46 can design a way from that point.
• 05:48 Level 2 has a very high chance of detection.
• 05:51 It's used when the development process includes design analysis and
• 05:55 detection controls embedded in the development process have demonstrated,
• 05:59 that they have a high detection capability.
• 06:02 This takes time and a number of development processes to achieve.
• 06:06 Once in a row is not good enough.
• 06:08 You have an excellent simulation such as a digital twin, and
• 06:12 it has been highly correlated with actual field data.
• 06:15 You can use this to test the failure condition and rate a score of two.
• 06:20 Finally, to get a score of 1,
• 06:21 the failure mode must be fully prevented through the product design.
• 06:25 This would mean applying Poka Yoke techniques that prevent the failures from
• 06:29 ever occurring.
• 06:31 So let's continue with our pen example.
• 06:33 We're analyzing the ball of a Ball Point Pen.
• 06:35 And as we considered the development process and use, the assembly process is
• 06:40 analyzing and testing prior to design freeze to determine any assembly problems.
• 06:45 So it is essentially testing to failure and a rating of 4.
• 06:49 The testing of the ink drying out is also done prior to design freeze, but
• 06:54 that test is just a test against the ink specification for pass or fail rating.
• 06:58 So it is a score of 5.
• 07:00 The corrosion test for the ball is done prior to the design freeze also.
• 07:04 But some of the customer application environments are not well understood, so
• 07:08 there are likely some requirements that are not known, we give this a 6.
• 07:12 With respect to the failure of damage to the ball due to dropping or abusing it,
• 07:16 this tested before design freeze and tested to failure, so the score is 4.
• 07:21 Finally, the testing for dirt or other substances that prevent the ball from
• 07:25 rolling freely, is tested with our standard test.
• 07:28 But the standard test does not have some of the unique customer
• 07:31 environment contaminations, because they are not known.
• 07:34 So, this failure mode also gets a score of 6.
• 07:39 The detection analysis score is a score of your development process' ability,
• 07:44 to find and predict whether the problem is likely to occur at some point in time
• 07:49 during the product lifecycle.

Lesson notes are only available for subscribers.