Save. Share. Connect.
Monday, May 22, 2017
VOLUME -22 NUMBER 5
Publication Date: 05/1/2007
Front Page News
People in the News
Contract Mfg. Products
Special Feature: Components & Disty
Product Preview: Electronics Distribution Show (EDS)
May 2007 Issue
Add Message Board
PC Boards: Fast-Track R&D Needed
By Robert Tarzwell, Director of Technology, Sierra Proto Express, Sunnyvale, CA
This year, we may be seeing the possible end of printed circuit boards as we have known for so long — if not the end, at least a major evolutionary change. The technology of the silicon chip has become so advanced it may soon be possible to operate computers and electronics without the need for printed circuit boards at all. A few different scenarios are actually playing right now; the chips are becoming massively powerful and are being combined in one package as multiple chip sets — replacing more traditional use of many individual chip packages. Future road maps of chip packaging show up to 12 die in one package within a few years. Another reason for this potential PC board evolution is the chips themselves; many of them now have so many transistors and are getting so power hungry they need much better cooling then just a fan. Heat removal advances include micro groves machined into the base of the silicon chip or on the PC boards, with hermetically sealed water coolant flowing through the base to a heat exchanger. Other thermal water cooled ideas are multiple chips in a 3D cube which are being packaged and cooled with water flow, reducing the need for a PC board to act as a thermal cooling and support platform.
The printed circuit board itself will steadily become smaller with finer lines and holes. We are seeing a steady increase in the number of requests for micro printed circuit boards with less then 2 mil lines and spaces and holes less then 4 mils. To answer the requests Sierra proto express is building a Micro Electronics center. We have purchased a new laser capable of creating 10 micron lines and 30 micron holes. We are researching the micro technology, and are preparing today for tomorrow's needs. There is an ever-expanding need for better thermal management of the chips and components. New nano technology has produced sheets of very thin carbon which has 3 times the thermal conductivity of copper, if we can laminate these nano thermal wonders inside the printed circuit board, it could greatly increase the heat removal capacity. Unfortunately the sheets are 30 dollars a square inch, but like most new technology, the price lowers exponentially after a few competitors show up.
Chunky Chip Packaging
Chip packaging is still fairly chunky; 100 times or so of the die volume is still needed for the package, as die densities increase with more die per package, so will the heat output. More thermally conductive area is required to handle the ever-increasing heat of the chips which means a printed circuit board of some sort will still be needed to carry the heat sinks. So we may have some life left in the PC boards as part of thermal management solutions. The new micro PC boards of the future will probably have to abandon copper as the main plating metal; its plating crystal size is too big. Microscope photos of 10 microns traces clearly show that the crystals are almost the width of the line. Different metals such as nickel, gold and palladium typically plate with finer resolution then copper and may possibly become the conductors of choice for very fine lines of less then 30 microns. In the interim, printed circuit boards will continue to move towards HDI micro stacked via technology. This is a technology that the U.S. has shunned for some reason, but today's and especially tomorrow's density will be too high for normal through hole printed circuit board technology. Research being conducted at Sierra centers on a new HDI manufacturing technique which will allow for more passive components to be integrated into the printed circuit boards; the middle FR4 core which provides a strong back bone for the HDI micro layers is thick enough to have real 0201 components placed inside. Our present printed circuit board embedded carbon screened resistor is not electrically good enough to be considered, but an 0201 chip resistor, capacitor or inductor will provide a wider range of values, low thermal drift and low tolerances at much lower noise levels.
Copper No Longer Fast Enough
Even though electrical signals in copper are very fast, in today's applications, they suffer from many problems. We may no longer need the printed circuit board to carry signals, but we still we still need the board for power and ground support. We also need to support the chip and its interface with the real world. The PC board may devolve to a simple, small, 4-layer board with ground power and a few power type signal lines. The chip sets will be able to communicate using optics or RF or some other non-copper conductor. And as the chips themselves become more complex, we will need fewer passives occupying board real estate. The chips' internal technologies will continue to reduce the need for external resistors and capacitors, eventually becoming standalone products. We have come such a long way since the vacuum tube days when 32k of computer memory would fill a good-size room. Now we carry a very sophisticated communications computer in our pocket with very few chips, and the best part is I paid nothing for it.To achieve all this wonderful integration for us techno geeks, we need ever smaller traces and smaller vias. There is however a stumbling block we in the printed circuit board engineering game have reached. Our present lithographic methods really do not support high quality yields below 2 mil lines and spaces. Upcoming technologies will take us to one-third mil (10 microns) but few shops can handle this dimension and fewer will even want to manufacture such small devices. Because of this intransigence, we unfortunately can expect to see another round of board shops disappearing. Looking ahead, Sierra Proto Express has purchased an ESI 355 nm YAG laser with a beam width of 10 microns — one of only two in North America. We will be experimenting with methods to make fine lines and vias, tricks like differential metal etching, where we start with a thin layer of one metal image and plate a different metal and etch the first one with a solution which does not hurt the plated lines. It's all beginning to sound like silicon technology, isn't it? One area presently lacking in our industry is the laminate. Our old FR4 standard is way out of date, and the glass and epoxy substrates are woefully inadequate. We need a new composite plastic laminate with high heat conductivity, good electrical properties, high temperature capabilities,and ease of fabrication — possibly some form of ceramic or diamond filled plastic may be developed. The printed circuit business has failed to spend any significant amount of R&D money on PC board technology. When visiting companies as an R&D consultant, they shudder at the thought of spending even a dime on R&D, saying there is no new technology. They don't see the future, or understand why they are losing business. The silicon business has progressed so fast and far is because of R&D money and the large number of brilliant minds working together. The large well-heeled chip companies have the ability to hire the best minds;, they give them fully stocked labs, enticement programs and perks to further the relentless move forward of silicon technology year after year. Unhappily, the same does not apply to the PCB industry which has been unable to draw venture capital money to help with its R&D efforts.
Falling Sales in Photo Plotters
Even so, there has been serious change in our industry. The people who make photo plotters have noticed a drop in film sales and new machines as printed circuit shops go filmless, we at Sierra have networked all our machines and have lowered our film usage by a factor of 6. The printed circuit designers have taken a hit in their numbers in the last 10 years and will continue to see loss of jobs to the Pacific Rim and India. Design engineers there get paid as little as $1.50 per hour. Add to this the cavalier attitude toward the environment in many of Eastern nations, and you have a perfect formula for losing more business.The PCB business is not glamorous; it does not really attract smart young people. This is something we need to change if we are going to keep up with our sister industries.Printed circuit board companies will need to invent new technology to solve the heat removing problem to take the chips to the next level. We also need to improve the method of connecting signals, Silicon's pipe idea is a good intermediate step; it frees up the chip for more heat conductive copper area, the elevated separate flex circuits gets the signal lines off of the main board , but it is not the end product I think we will get, In the near future, you will see products with low chip counts mounted to a micro 4 layer HDI power ground printed circuit boards with nano-technology heat-removing core; the signals will be handled with either light or radio waves running at high speed interfacing to all chips. We definitely need to spend more money on PCB R&D; we need to look for synergy by combining with silicon companies to look into the printed circuit future; and we must be willing to experiment with good ideas. For more information, contact: Sierra Proto Express, 1108 W. Evelyn Ave., Sunnyvale, CA 94086
408-523-9926 fax: 408-735-7183 Web: