I am developing a magnetometer -- my last great design, I claim. It incorporates sixteen amplifier stages, four sensors, and several controls -- and is performing fabulously. But it is a prototype, so that I keep handling it -- and it keeps suffering catastrophic failure. There is a lot to go wrong when one handles such a machine. Today, one wire came off a control. Yesterday, a shooting spot of solder caused a short-circuit. A trusted battery failed. Two leads touched. And so on. Each time, my heart sinks -- and I have to track down the fault in this maze. Thankfully it has always been a material error or failure, not the design. OBSERVATION: A lot of nuisance can be avoided by progressing step by step. Advance -- check. Advance -- check. I photograph a circuit, too, so that I can go back if I have to.