Issue link: https://iconnect007.uberflip.com/i/710139
38 SMT Magazine • August 2016 ing campaigns were focused on the fact that the product did not use printed circuits. Hewlett- Packard's first push into this process world was a disaster, which brought into question wheth- er a superbly engineered product could be left to the questionable quality that would result if printed circuits were the foundation of its man- ufacturing. At the same time, it was also obvi- ous that, with transistors being designed into all new products, hand soldering was not a vi- able option, even for the short term, let alone the entire new product line. The first try at printed circuit assembly— crimping component leads onto board pads— was slow and unreliable. The next step was the use of eyelets, which had some success on sin- gle-sided boards, and was proposed as the basis for two-side connection. It's almost impossible to describe the level of controversy that exist- ed in the electronics industry of the time over whether eyelets or plated-through-holes were the best alternative. Today, that seems laugh- able, but the discussion was deadly serious at the time. The general betting seemed to be on eyelets. The reasoning went as follows: It was absurd on the face of it, that a process could be developed that would plate copper on a non- conductive surface, and do it in hundreds of small holes in a board—without missing even one—while eyelets were positively put in place and missing one would be obvious. This environment of uncertainty spawned a furious search for an alternate circuit packaging technology and resulted in some very creative proposed solutions. Most were developed for specific products, or types of products, but were often put forward as general purpose packaging solutions. Most are now lost even to history. As a recent graduate in electrical engineer- ing and a new HP hire, with three years of ex- perience in the United States Navy, my first assignment was to sort this out, propose, and develop a process that would be reliable and cost-effective, and could be integrated into the manufacturing capabilities of the company. I soon realized that they had assigned the wrong person. Instead of an EE, who could commu- nicate with product developers, and board de- signers, what they really needed was a chemical engineer who could understand the chemical issues of plating and board materials. However, I had the job and needed to do a lot of learning. First, however, I quickly eliminated eyelets, which HP was using successfully on single-sid- ed, low volume boards, because 1) the differ- ence in the coefficient of thermal expansion be- tween the eyelet and the board material meant that after soldering, the two no longer could be counted on for an intimate contact, leading to potential intermittent, or open, circuits and required considerable touch-up with soldering irons; 2) eyelets had to be installed one at a time; and 3) the smaller eyelets needed for tran- sistor interconnection were more expensive and even harder to install. That left plated through- holes, but the question remained: How could they be done reliably and consistently? I went to the corporate library and found two books that claimed to be on printed cir- cuits. One was really about silk screening, and the other was a high-level discussion with no real, or useful, information. The only actions left were to evaluate commercially available chemicals, and to see how others were doing it. Giving some support to the concerns about the possibility of developing a reliable through- hole plating process was the problem that early chemical processes for reducing copper ions were very unstable. It was quite common for the solution to "go critical" and all the copper ions in solution suddenly reduce to a lump of cop- per at the bottom of the tank. In addition, there was a distinct interface between the copper that was deposited and the copper laminated to the base material. This required a separate sanding process to remove the copper deposited on the surface, leaving only the holes with new copper. " The first try at printed circuit assembly—crimping component leads onto board pads—was slow and unreliable. " THE AUTHORS OF THE PRINTED CIRCUITS HANDBOOK SPEAK