Chapter 23
Hangzhou, September 2019
The battery mount was her subject, and she put it on my desk in June: a resonance risk between the pack cradle and the frame rail, with the analysis showing a fatigue window if the mount bushing ran warm. She had nine years in the trade and the least time on my floor, to hear me tell it that morning.
I screamed at her. I want the memo to say only what happened: I told her the mount was not her department, that the supplier had been validated, and that she was the least experienced person in that room. My voice did the rest. She stood there with the plots under her arm and did not defend herself, because there was no room in my face for defense, and she left the building on a Friday in September and went to Shenzhen.
There were four others in the room. I know because I have counted them in the years since — counted them the way I count the features on a failed part, looking for where the load went. Nobody spoke after she closed the door. Somebody coughed. I said, let’s get back to the agenda, and we got back to the agenda, and the agenda is what I am offering as the explanation I do not have.
Test seven ran in October without her. The mount failed at the predicted hour — a crack, textbook, exactly where her paper said it would be. I have the test record. I have had the test record for six years and I have never once been able to decide what to do with the fact that she was right at seven.
We ran the line 9 to 9, six days, that year — ninety-nine-six, the way the whole industry counted its hours by then. Nobody asked. Everybody came.
The mount was redesigned in November, by a consultant, to a geometry that differed from hers in two dimensions and from mine in none — which is the sentence I have never been able to finish out loud, because its last clause is the confession and its first clause is the credit, and the two of them do not fit in one breath. The redesigned part went into test seven and lived. I signed the drawing. The signature block has room for two names and the second box is empty on every revision we filed, and the empty box is the truest document my company produced in the year of the high-water mark.