Design007 Magazine


Issue link:

Contents of this Issue


Page 73 of 93

74 DESIGN007 MAGAZINE I JUNE 2023 first integrated circuit around 1957 using three months of savings from my paper route. at purchase occurred a full year ahead of Jack Kil- by's demonstration of the first integrated circuit which had been cobbled together from a piece of germanium connected to a few resistors and capacitors using gold wires. e integration of transistors for my "integrated circuit" was a very simple printed circuit with six transis- tors soldered to it. Granted it was a rather large "integrated circuit," but it definitely inte- grated those transistors and other components and made them work together, allowing me to listen to amplified music from radio sta- tions that were in work- ing proximity. I cannot say for certain if Kilby had a printed cir- cuit in mind when he built his demonstration, but I know that he worked at CentraLab in Milwaukee, which made ceramic printed circuits for military and commercial customers. us, it seems not to be a great stretch of imagination that Kilby saw and took advantage of "the adjacent possible" to create his world-changing invention. Around the same time—and independently from Kilby—Robert Noyce, then at Fairchild, invented a process for producing defined areas of conductors and insulators on a semicon- ducting base material, a process that has been continually refined using the concept of "the adjacent possible" ever since, and in the pur- suit of extending the life of Moore's law. Printed circuits also followed the path of the adjacent possible from single metal layer circuits to double-sided boards to multilayer circuits to flex and rigid-flex. So has been the path for assembly from through-hole to sur- face mount, from one-side assembly to two- side assembly. From the late 1990s to today, the effort has been focused on making circuit features an ever-smaller transition from high- density interconnections (HDI) to ultra high- density interconnections (UHDI) with feature sizes that are smaller than those used on first- generation integrated circuits. About 15 years ago, the Occam Process was proposed, but it was not considered a step toward the adjacent possible, However, it seems clear that it fits the description. e con- cept as written up can, by taking a step toward the adjacent possible, potentially addresses and solves at one time a number of problems that have long faced the industry. Most of them are related to solder and the prob- lems inherent in the soldering process, especially those related to high temperature dam- age to circuit board features, such as plated through-holes and elec- tronic components, not to mention the myriad assembly challenges faced daily by manufacturing engineers. When viewed in the light of the work of Jack Kilby, the Occam Process, which preaches minimalism in design ("It is vanity to do with more that which can be done with less"—Wil- liam of Occam), proposes to do something that can be done immediately by simply reversing the process of assembly. at is, rather than building a printed circuit and soldering com- ponents to it, build a "component board" and build the circuits required for interconnection for those components on one or both sides of it using processes that are being developed to make HDI and UHDI boards. ese components will ideally have all the same I/O pitch to make board layout easier, as I described several years ago when I sug- gested in a paper to "dis-integrate" ICs into their constituent IP blocks. Such structures are presently being called "chiplets." It is a step in the right direction but what is largely ignored I realized I had purchased my first integrated circuit around 1957 using three months of savings from my paper route.

Articles in this issue

Archives of this issue

view archives of Design007 Magazine - Design007-June2023