Personal Diagnostic Achievements

The Nail
My NT Server began crashing daily around 6pm. I ran multiple diagnostics and checked for any event that ran with any daily schedule; no problems found.
Eventually, I noticed on an apparently different service call that the blue toner on our color printer was running low. On that particular model, users must acknowledge the toner low signal before the printer will print. Many users were unaware of this quirk, and simply informed the office manager that the printer was low on toner. The office manager set someone else to investigate before ordering more toner, causing a two day delay before ordering. Then, once the order was placed, it took several days to get fulfilled. So it was more than a work week before the printer would be back in production. I soon realized that it was this toner problem that was the reason the NT server was crashing.
I had already taught the users not to send large, or multi-page color documents from applications that were not configured to take full advantage of the printer's capabilities. Whenever they did, the main partition on the NT server would fill up and it would immediately crash. This problem, however, caused the main partition to slowly fill with documents waiting to print until the users acknowledged the toner low situation. At 6pm, the server's activity would drop sharply and it would try to refresh it's scratch space. Since it was too low on disk space in the main partition to do so, it would crash.
So, I had someone sit there acknowledging toner low warnings for a couple of hours, and my NT server stopped crashing. When the new toner finally arrived, the problem ceased permanently.

Memory loss
I was working bench repair and had a customer come in several times with a Mac Powerbook 5300 that wouldn't start up. Each time, I would take it completely apart, check each component, and reassemble it. Each time I did, it would work. After a couple of times, I decided to check it very carefully before dismantling it. I discovered that the memory card was slightly loose. I reseated it and the machine worked.
Apparently Apple had designed the memory card area to fit a double-sided card that never went into production. The addition of a small piece of foam kept enough pressure on the card that normal movement didn't jostle it out. I reported this to Apple and they added it to the standard repair procedures for that model.

Code Efficiency
In my Advanced C course, I developed a new algorithm that eliminated 3 steps out of 5 for the standard search and compare routine. The code required 17000 fewer cycles to complete the designated task than the best routines developed to that date. It ran through the entire exercise at an average of half the time on several CPUs.

Starting problems
My truck was having intermittent starting problems. The situation confused several mechanics. Finally, I noticed that the truck had a particularly sensitive Park notch. This, combined with the positioning of my trash bag was what lead to the problem. As the trash filled, it moved into position to interfere with the accurate placement of the shifter. When the shifter didn't quite hit Park, it wouldn't start.