This is where I got involved. As a memeber of the test engineering team, I was responsible for building a piece of test equipment that would explain this odd failure issue. As luck would have it, this piece of equipment was put into service on Christmas Eve and the entire manufacturing line was off until after the new year.

Seizing the opportunity, I spent the next week testing and evaluating variations in Magnetic Overwrite Modulation (yah, it's technical) on dozens of drives until I was able to not only prove the problem but repeate it (get drives to fail on purpose).

Reporting my findings up the corporate chain lead to a meeting I'll never forget... Reminiscent of the Star Chamber, I entered a room with a huge conference table. I was ask to present my findings and I did so. I was then ask to leave. The head of Quality Control supported my findings but...

There was a small group of test technicians actually responsible for product testing and they disputed my findings based on their manual tests.

Why am I getting so technical and yet not technical enough for the computer geeks reading this? Here's the punchline...

This panel of Vice Presidents voted that there was in fact no problem! I was instructed to go back to doing my regular job, mind my own business and never speak of this again.

Sometimes the truth will out...

In a process like this, there is always one product that is defined as the perfect product. At this Bernoulli company, we called this the "Gold Standard." It was stored in a safe and if everything seems off, a manufacturer looks to this to determine what the calibration is supposed to be.

The problem was the Gold Standard failed it's own calibration, even the drive that all other drives were judged against has failed.