In 22947, I said "...the coil driver sent a bad status..."  This is too broad and points fingers at the Coil Driver.  While the zero bit may have originated at the coil driver, as we believe that occurrance will latch the state it is very possible the bit dropping to zero may have originated down stream.  There is the binary interface chassis bundling all the 9pin channels into db37s and then the Bin I/O card in the I/O chassis.  Trouble shooting these doesn't appear easy, especially in O1 time.

It makes sense that we tailor the model side to not trip the ISI when these less than a second changes occur.  Otherwise we should just wait to see if it is a problem.  Looking back a month, the first time it occurred was during the power outage but that was legit.  Otherwise, it occurred Tuesday but Richard says that might have been him exploring, experimenting or some other exing.  And then yesterday.  So other than the model change which we'll discuss that what should be a simple code adjustment, we'll leave it til it breaks for now.