There are other lower level clearance KPIV that can be added to the list, but the above testing would cover more than 95% of the expected variation in Chirp. The real issue however is that any group of parts that are 4 sigma from mean are not behaving as expected. I would expect to see a variety of special and common cause signals in that FA, but we need to get the data to determine the corrective action against that yield loss. The drivers for those individual heads coming in at a very high clearance are clearly not affecting the heart of the distribution given the expected yield loss is less than 0.1%.
Essentially, what it comes down to is that the LCO team is asking for a high temp spec to bake additional robustness into the drive, the closure action at the HGA / component level to that request is to put the 4 sigma power spec in and then work the FA to reduce the yield loss associated with the spec request. . Does that make sense?