Issue link: https://iconnect007.uberflip.com/i/1381013
26 DESIGN007 MAGAZINE I JUNE 2021 idea; serious product designers considered eyelets a preferred technology. It became a competition, and people took sides. It was a slow week if Electronic News, a very important industry publication at the time, did not print an article reporting a new study, or just quoted a learned opinion, on the benefits or perils of one versus the other. Hewlett-Packard was not immune to this controversy at the time. In 1959, HP was still recovering from the misuse of single-sided boards in a new product. e poor reliability of that instrument had not only done serious damage to the company's credibility on qual- ity and reliability, but had caused serious finan- cial issues on rework, repairs, and lost sales. At this point in time, major electronics com- panies such as Sylvania Television and Tektro- nix Instruments were actually advertising that their products did not use printed circuits. In response to this problem, HP had turned to eyelets to solve the immediate issues, and that did have a benefit for those single-sided boards. A bigger problem, however, was looming as all the new products were to be designed with transistors, which meant that both sides of a board would be needed. Young Engineer, Risky Project at's when, and where, I came in. I was a newly hired electrical engineer tasked with bringing the company into a printed circuit technology that would be an advantage to the product and perceived as such by the customer. To illustrate the level of misunderstanding of the real problem, it was viewed as an electrical engineering issue because it was dealing with interconnecting electronic components and circuits rather than the mostly chemical engi- neering challenge that I soon found it to be. ey had the wrong person. However, I had the job. Aer some stress tests on both eyelet boards and plated through-hole sample boards, I became convinced that the only process that could achieve company reliability and cost goals was the plated through-hole. Getting others at the company to agree with me, even given my test results, was a much bigger prob- lem than I ever imagined. Coming off the previ- ous experience with printed circuits, product engineers did not want to abandon eyelets, since they were now working, and insisted on refining that process to the new double- sided need. eir argument was, basically, that "plated through-hole" was absurd on the face of it. How could any reasonably sane person expect to plate copper in hundreds of small holes, drilled in a nonconductive material, and not have many fails? I actually had three senior technologists each provide me with informa- tion on separate eyelet processes they insisted would work. One thing they did not seem to remember was a basic physics issue: brass eyelets and epoxy- based PCBs have a different co-efficient of thermal expansion. Intimate contact between the two aer soldering was mostly lost, leaving the solder to provide the connection between them, causing all the touch-up and rework we were experiencing in production. One prod- uct manager was so alarmed by my advocacy of plating that he went to Dave Packard per- sonally and demanded that I be fired before I destroyed the company. Packard was well aware of the company's commitment to transistorized products, and the resulting reliance on printed circuits, so he asked me to give him a personal review of where we were on a workable process. I spent two hours explaining that what I was pro- posing was not "black magic" or an "art," but was based on a chemistry that could be read- ily understood, controlled, and predicted. To show that it would work, I had two sample boards from new instruments that the product designers had agreed to include. He was con- vinced and signed off on the budget for the first shop: $25,500. To be fair, we did not need a new building, as I was able to use the space that had been made available when the general plat- ing shop was moved to a bigger facility.