A "war story"

 Old farts are supposed to bore you with war stories.  I told one on Facebook.   I'll elaborate on it here.

It's 1984 and I'm on a team at Motorola where marketing wants something like a Macintosh but it runs unix and lives on a network.   Motorola/Four Phase (across the street from Apple in Cupertino!) is not up to this task as a company.   But my team is going to build a system with Motorola technology that can run demos and look like an office machine.  We're gonna learn a lot and it will be hard and intense and really fun.

A few months later and I've got a workstation hardware prototype chassis running. It's got a Motorola MC68010 16-bit  microprocessor in a socket.  (Back then CPUs didn't get warm enough to need a heat sink.  You could just pry it out of the socket with a little screwdriver.)   And it's got bit mapped graphics and a mouse!

My department was about to receive engineering samples of Motorola 68020 (32-bits) and there wasn't a socket on campus to plug it into.  Due to pervasive bad management..   My department had been told we weren't getting one, and suddenly they were arriving any day now.

I made an adapter that fit into a 68010 socket and held a 68020. Designed it on graph paper, while sitting in an airline seat.   Wire-wrapped it myself when I got back to Cupertino.   It ran in my workstation chassis and in the Convergent Miniframe, and made them run faster with no software or firmware changes. I think it took half a dozen 74F or 74LS  type"jelly bean"  parts, no programmed logic device.  Met every timing spec.   Fully synchronous.   I should have stolen it for a trophy.   Nobody would have cared.  

The MC68020 was way faster than 68K.   You couldn't see that when it was working through a 16-bit socket, but it still doubled the speed of the Miniframe.   Because cache memory is awesome.   I always wondered who they would have blamed for not having a socket ready for a new microprocessor FROM THE SAME COMPANY.