Opencircuits

Don was one of the worst engineers I’ve ever known. As a matter of fact, he is one of the dumbest people I’ve ever known. I have no idea how that guy made it through college. I have a few stories about Don. This is one of my favorites.

Don’s boss was getting exasperated with him. He couldn’t get basic assignments done and thought he was beneath doing documentation. He kept lobbying for more tasks but would screw up anything he was given. It was getting to a bad situation between Don and his boss one day when the phone rang from the production department. They had a new batch of boards going through that weren’t working. Don’s boss told Don that this could be his moment to shine. Don eagerly jumped in his car and drove over to the production building where he went to work.

The problem turned out to be the switching transistor on a flyback circuit. The switching transistor wasn’t getting turned on. Don went to work. Engineers of Don’s caliber have the indispensable research tool – the soldering iron. Don grabbed one and got to “redesigning” the switching circuit. He was at it for a good week. The production line was shut down while Don was working away with his soldering iron and he finally got it working. He increased the input drive to the point where the circuit worked “normally.” Don called his boss and told him he had the problem solved. Don’s boss was in the production building in a matter of minutes and was ready to check things out.

Don had all of the changes written up and was ready to demonstrate the before and after boards for his boss. Before he could get started, the boss picked up one of the unmodified boards. He looked at if for a good two seconds before his face started getting red. He composed himself pretty well under the circumstances and pointed out to Don that all of the switching transistors had been mounted backward.

Before they left, they had a couple of boards reworked to put the transistors in the right way. Everything worked fine. Don’s changes to drive the daylights out of the poor little transistors never got implemented.

Tweet This Post

Years ago, I worked at an aerospace company.  Testing of finished products was always an issue.  Comprehensive tests were required for finished hardware.  The test stations were generally pretty complex systems, which had to exercise all inputs and outputs of production hardware before it could get shipped and installed in final aircraft.  Unfortunately, the engineers who ended up in the test equipment group often tended to be those who weren’t good enough to work on the actual products.

The best example of that I ever saw was a guy named John.  He was working on one of the aforementioned test setups and was assigned a very small part of a test system.  For his particular circuit, a time delay was required.  There are plenty of right ways to design a time delay, but none of these seemed to occur to John.  He obviously struggled with this one when the light bulb went on.  A 7404 hex inverter has a propagation delay of 9 nsec.  John grabbed his calculator and worked out the details of his “design.”  It turned out that eighty nine packages of 7404 hex inverters at six per package gave him the delay he needed, which was 4.8 usec.

Test equipment in most companies tends to be behind schedule, since more priority and attention is given to production product that turns into revenue.  There was no design review process for test equipment at this company and things tended to just get built.  John’s circuit that included his time delay circuit got built into three test stations.  I always felt sorry for the person in the lab that had to wire wrap all of those inverters in series.

When the tester finally got tested, John’s time delay didn’t work out so well.  Since all 534 inverters had the same temperature coefficient, the actual delay would drift way out of spec from the time the circuit got turned on until whenever the test got made.

John’s boss replaced the hex inverter delay circuit with a single 74123 monostable circuit.  It was hardly a precision time delay, but it was good enough for government work in the application it was used.  There was a lot of space left over on the board after all of those inverters were removed.  The poor person who had to wire wrap all of those inverters in the first place got the privilege of unwrapping all of them and implementing the monostable.

Tweet This Post