PCB007 Magazine

PCB-June2016

Issue link: https://iconnect007.uberflip.com/i/688506

Contents of this Issue

Navigation

Page 11 of 109

12 The PCB Magazine • June 2016 by Happy Holden PCB TECHNOLOGIST, RETIRED You may not know about landless vias. This has been a well-kept secret for the last 30 years, possibly because it is not permitted on military boards, and therefore, discouraged in all IPC standards. Consequently, when our Japanese partner, OKI Electric, showed us their landless via boards, we said, "You can't do that; the vias have to have lands!" They replied, "You've been listening to IPC again! Try it out and test it!" So we created a test vehicle and tested thousands of vias with various diameters against numerous annular ring vias. Guess what? They were right! As the annular ring got smaller, the failures oc- curred earlier until we got to landless, and then they jumped to 10X fewer failures. Not understanding this result, the PhDs of HP labs went to work and came up with an ex- planation (included in this paper). This was so radical a discovery that HP made the data, ex- planations and results proprietary and a com- pany secret. Introduction I saw my first landless via multilayer while visiting NEC at Toyama, Japan back in 1985 [1] . They were an enormous automated facility mak- ing Japanese telecom and mainframe computer boards, kind of like IBM and Western Electric rolled into one. NEC was using the liquid elec- trophoretic, positive-acting photoresist process with panel-plating. I wouldn't see another land- less via multilayer until our Japanese partner (OKI) introduced it to us in 1988. OKI was us- ing the landless vias to achieve higher density without having to pay the extra costs of finer lines. They knew about the higher reliability that resulted, and they had done their own test- ing (Figure 1), but were after the higher routing density (Figure 2) and it allowed them to route five traces between 0.100 inch PTH centers. The Electrodeposited-Positive Acting Photoresist My first introduction to the positive-acting photoresist (+PR) was at Hewlett-Packard's in- tegrated circuit fabrication facility in Palo Alto, California in 1970. HP used this in the liquid form and it was spin-coated on wafers. The pho- toresist was supplied by Shipley Company of Newton, Massachusetts. I was further exposed to the +PR after moving to the printed circuit fabrication facility. It was not being used on PCBs, but on metal photo-chemically machined Against the Density Wall: Landless Vias Might be the Answer FEATURE

Articles in this issue

Archives of this issue

view archives of PCB007 Magazine - PCB-June2016