All Write, Already!

I wish there were a Mapquest for the routes I'm tracing. Following the path of electrons on a printed circuit board is bad enough when you're the designer; when you're a reverse engineer, the work's about ten times as hard.

Lately I've been running across nanosecond timing glitches in the prototype we've built of one of our competitor's microprocessors. A billionth of a second seems like nothing up here, but down there, where a resistor towers like the Empire State Building, it's the difference between $168 million of pure profit (what the other guys have) and being $6 million in the hole for the project (where we're currently sitting). Since neither I nor anyone else on the fab team can figure out why we're glitching so badly on the new traces we've run, it's back to the competitor's working model to work it out.

Even when I compare the two under the microsope, the two look exactly the same. But somewhere in the two millimeters' worth of copper, there's a traffic jam causing the electrons to slow down. I've spent the last four days trying to break that jam, and I've been forgoing food and sleep to do it. Now I'm working at the scanning tunneling microsope, where individual atoms are the size of beach balls.

There! A cut in the trace, about four or five atoms across. The dies are bad, and we can fix this. I leave a note on my boss' desk:

Tim:

Rebuild the die on PCB-CM-104. Problem's solved.

Kevin

P.S.: I quit. I'm moving to a commune in Vermont, where I will deal with no span of time shorter than a season.