Back in the early ’80s, I led the engineering team at JBL. When I joined the company, the average time it took to develop a new product was about 2 years and most of the product line lost money. When I left, product development time was around 10 months, product quality had improved (by design), and product profitability had improve dramatically (also by design).
In the early ’80s, JBL had a reputation for displaying vaporware products at trade shows. The J-Series was designed, tooled, and a successful pilot run completed in the three month period from October to December, 1982. When dealers ordered samples at the January, 1983, Consumer Electronics Show, they had real production units from the pilot run waiting for them when they returned from the CES. That was an important step in improving dealer and customer satisfaction with the product line. Designing the products to cost meant looking creatively at all aspects of the manufacturing process. It turned out in this case that the most significant cost saving came from improvements in our woodworking processes.
There were lots of pro audio, car audio, and musical instrument products that came out of our department during my tenure. The last products I worked on at JBL were for the high-end consumer L-Series. My favorite of that series was the 18ti.Here’s an interesting problem that I personally developed the solution for. Tweeter voice coils need to be very low mass. The best material for the tradeoff between electrical resistance and mass is aluminum, so we used aluminum voice coils in our high-frequency drivers. Aluminum is not easily solderable. We crimped beryllium-copper leads to the ends of the voice coils in order to have a solderable connection. If too much force was used to make the crimped connection, the fine aluminum wire would fracture, and the connection would be mechanically unreliable.
I suspected that the cracked structure of the aluminum would act like a point-contact diode. It did. When a high-frequency current was applied to the poorly made voice coil assemblies, the current was noticeably distorted. This provided a means of testing the units following crimping. Testing stopped bad parts from being used, and it also allowed the manufacturing engineers a means of calibrating for the maximum useful crimping force for each type of voice coil. Eventually, one of the manufacturing engineers came up with a better assembly process that allowed direct soldering to the aluminum, but the test method also worked to find poor solder joints.