Sherlock Ohms is a regular feature by Design News where engineers and technicians describe their adventures of troubleshooting circuit and software problems. Quite often I can relate to the scenarios because of similar instances I've had over many years of being involved in electronics. This story by A. David Boccuti is very instructive in how to find the root cause of a problem in production that suddenly appears for no apparently good reason. In Mr. Boccuti's case it was a component refusing to be soldered properly when in the past there was never an issue. He did a commendable job in resolving the problem.
Having begun my post-USAF electronics career as a technician at the Westinghouse Oceanic Division in Annapolis, Maryland, I built and tested many circuit boards and chassis that conformed to Mil-Spec standards. The ECO (Engineering Change Order) bureaucracy (sometimes nightmare) was a necessary but complicated and frustrating ordeal. If a circuit modification needed to be made or a component value changed, it was not just a simple matter of documenting the new configuration and moving on. Multiple levels of reviews and approvals were required both by Westinghouse engineers and management and by U.S. Navy personnel. Mercifully, there was a system whereby a "suitable substitute" could be used to replace a specified component whose original specific part number and manufacturer was not available, but anything that affected "form, fit, or function," was relegated to the ECO menagerie.
In particular re Mr. Boccuti's experience, I recall a time when a printed circuit board used a grounded shielding fence along the four edges. It was a copper-beryllium metal with a nickel plating to facilitate soldering. I built a couple dozen of the PCBs with no problem. Then a second batch arrived and suddenly the solder would not wet the fence's surface. Copious amounts of liquid flux would normally solve just about any solder flow problem, but not this time. I and my comrades had been thoroughly trained in NASA-created soldering techniques and our equipment was firmly regulated and calibrated to not be able to produce enough heat to easily delaminate multilayer substrates (another way of forcing solder to wet a surface is high heat with lots of flux). After many frustrating attempts to get the fences soldered, we finally did an inspection under a microscope and discovered that the plating had very fine parallel cracks in the little tabs that stuck down into the PCB's plated-through holes. That was enough to bugger the soldering job. Unfortunately, the ground fences were a custom component and there were no replacements on hand, so management descended on both our in-house incoming inspection group and on the manufacturer since we were under great pressure to get the system delivered.
BTW, if you have an interesting story to tell and would like to have it published here in the RF Cafe "Out of Order" series, please let me know. I have a hard time getting people to participate for some reason, so your contribution would be a great help in generating interest.
Do you have a good work-related anecdote to share? Please email it to me for consideration. Thanks.
- The Singing Telephone Switch
Posted May 29, 2014