When my wife and I got married, we bought a house in the Phoenix area that came with a solar water heater. Almost all new houses came with solar water heaters during those days due to various government credits. After a few years, when it was just a month or two out of warranty, it stopped working. A little debugging with a volt/ohm meter and some jumper wires revealed that all the components in the system were working except for the controller. The pump would pump when connected to power. The thermal sensors were producing values that seemed reasonable.
So, I did what seemed the only sensible thing -- I took the controller out with the intention of taking it to work for further debug. My wife, however, was skeptical that I would be able to fix it. I was sure any repairman would junk the whole controller rather than debug it to the component level, which is what I intended to do.
The guts of the controller were made up of an inexpensive single-sided PCB that used op-amps, resistors, and diodes to perform basic logic functions (for example, if the water in the collector is warmer by a certain amount than the water in the storage tank, turn on the pump). These days all that would be done with a microcontroller.
I traced a schematic of the board and broke the resulting circuits down into analog and logical functions to learn how it worked and to help find the fault. During a visual inspection, I could see a couple of diodes that were obviously charred. I analyzed the circuit performance and discovered that they were very underrated for the current they would have to carry when forward biased. Further bench testing confirmed they were toast. I removed them and replaced them with higher-rated diodes from lab stock and buttoned everything back up.
That evening, when I put the controller back into the system, it began working again. It continued to work until we sold the house a few years later. As a relatively new engineer, it was a boost to my confidence to be able to fix the controller, especially since I avoided what would have probably been a $200 repair bill by replacing components that only cost a few cents. It also reinforced the lesson that components have specifications for a purpose, and violating those specifications has consequences.
This entry was submitted by John Roe and edited by Rob Spiegel.
John Roe has a master of science in electrical engineering from Georgia Tech. At the time of this story, he was working for a medical electronics company. Currently, he works for a semiconductor company designing high-performance networking and computing ASICs.
Tell us your experience in solving a knotty engineering problem. Send stories to Rob Spiegel for Sherlock Ohms.