In 1965, Fairchild Semiconductor Director of R&D and soon to be Intel co-founder Gordon Moore wrote “The Future of Integrated Electronics,” which was intended as an internal paper to define the most cost-effective number of components per integrated circuit. As he looked ahead to the next decade, Moore argued that the number of components per chip would double every year.
The paper was edited and published by Electronics in 1965 as “Cramming More Components onto Integrated Circuits”. Ten years later, Moore, then with Intel, spoke at the IEEE International Electron Devices Meeting and showed that his initial prediction was correct and estimated that the rate of increase would slow to “a doubling every two years, rather than one.”
This prediction has now become widely known as Moore’s law. It has become a tenet of the electronics community and continues to propel the industry forward at a time when the number of transistors on a chip (which was around 65,000 in 1975) now exceeds one billion. [1]
These high-powered components are common on printed circuit boards (PCB) in every day electronics from mobile devices to computers to automobiles. Recently, the Defense Advanced Research Program Agency (DARPA) announced that it will spend $200 million on the Electronics Resurgence Initiative to seek new materials and manufacturing techniques in expectation that Moore’s law will come to a natural end. [2]
Not only are the components themselves getting higher-powered, but increased demand for functionality in ever-smaller packages has meant that these components are increasingly being squeezed into tighter areas. A 2012 article on Tech Design Forums, based on information from Mentor Graphics’ Technology Leadership Awards, indicated that while PCB size had been “relatively constant,” the “average number of components has quadrupled in 15 years.” [3]
As the forum noted, “Despite attempts by IC (integrated circuit) suppliers to cut power dissipation, as IC speeds and densities increase so does the heat they dissipate. And putting these ICs into smaller and smaller form factors compounds the problem. This causes significant thermal management challenges that must be met at the IC package, PCB and system levels.”
OCM Manufacturing, a low- to mid-volume manufacturer of electronics products, offered a chart that detailed standard spacing of components on a PCB, but also added, “With that said, there are no hard and fast rules for component spacing. Tightly packed components may have very good yield and problems may arise only during rework.” [4]
Of course, all of that power will inevitably lead to increased heat across the system. Coupled with the decrease in space between components, which puts constraints on the amount of airflow across a component and leads to heat from one chip being passed on to the next, thermal management is a critical aspect of PCB design to an even greater extent than before. [5]
Heat sinks remain the most cost-effective method for cooling chips. The benefits of heat sinks, the thermal impact of different materials, and the development of new fin geometries are all discussed in depth elsewhere on this blog, but this article asks, “What is the best way to attach heat sinks, especially in a component-dense environment?”
As Dr. Kaveh Azar, founder and CEO of Advanced Thermal Solutions, Inc. (ATS), wrote in ECN Magazine, “An engineer starting the process of thermal management must first determine the cooling needed and then consider the mechanical aspects of attaching the heat sink.” [6]
He added, “The thermal consideration is foremost on our decision tree. Once we have resolved the cooling issue, including the heat sink size and the type of thermal interface material (TIM) needed, we need to ask the question of how this heat sink will be attached to the device or the PCB.”
There are several options for design engineers to consider, but each comes with its own set of challenges. Thermal tape and thermal epoxies [7] would obviously add nothing to the existing component footprint, but tape has proven better for low-powered chips and epoxies require time to cure and are essentially permanent, making potential rework more time-consuming and costly.
Push pins, threaded standoffs and z-clips are mechanical attachment technologies that are common in the electronics industry but all require expanded footprints as well as holes or anchors in the PCB, which may not be available on high-density boards. Holes and anchors also make signal routing more difficult in the design phase and there is a possibility of a standoff or solder anchor causing a short during installation that could result in damage to the board. [8]
To meet this need, ATS developed superGRIP™. The two-part attachment system features a plastic frame clip that fastens securely around the perimeter of the component and a metal spring clip that slips through the fins of a heat sink and locks to the frame clip on both ends. [9]
The system is designed to need minimal space around the component. [10] The frame clip is made of a plastic resin that allows it to be very thin but also very strong, which was demonstrated during shock and vibration testing. The interior frame profile locks securely around the bottom edge and sides of the component package. The horseshoe tabs secure the clip to ensure the proper pressure on the heat sink.
The following chart shows the superGRIP™ clearance guidelines, although custom options are available and may be needed depending on the design:
superGRIP™ was also designed and tested to ensure maximum airflow through the heat sink. In a tightly-packed system where airflow is at a premium, superGRIP™ provides the necessary attachment security with only minimal impact on the flow. In addition, the plastic used in the frame clip stays cool in high-heat environments, rather than adding fuel to a potentially combustible situation.
Unlike other attachment technologies, superGRIP™ also requires no separate tooling and can be installed or released with a common tool such as a screwdriver. [11] This makes any potential rework easier. It is important to note the direction of the airflow when placing a heat sink, so it must also be considered when placing the frame clip as well.
References
[1] http://www.computerhistory.org/siliconengine/moores-law-predicts-the-future-of-integrated-circuits/
[2] http://www.eetimes.com/document.asp?doc_id=1331974 and https://www.darpa.mil/news-events/2017-06-01
[3] http://www.techdesignforums.com/practice/technique/overcoming-increasing-pcb-complexity-with-automation/
[4] http://ocmmanufacturing.com/resources/resource/dfm-tip-spacing-components-on-a-pcb/
[5] http://www.electronicdesign.com/embedded/engineer-s-guide-high-quality-pcb-design
[6] https://www.ecnmag.com/article/2011/09/know-your-choices-mounting-heat-sinks-hot-components
[7] https://www.masterbond.com/industries/heat-sink-attachment
[8] “How the maxiGRIP™ attachment system impacts component mechanical behavior,” Qpedia Thermal eMagazine, May 2008.
[9] https://www.qats.com/cpanel/UploadedPdf/ATS_superGRIP_Launch_Release_FINAL_with_Photo_0427092.pdf
[10] https://www.qats.com/cms/wp-content/uploads/2013/12/superGRIP-Clearance-Guidelines1.pdf
[11] https://www.qats.com/cms/wp-content/uploads/2013/12/superGRIP-Installation-Guidelines.pdf
For more information about Advanced Thermal Solutions, Inc. (ATS) thermal management consulting and design services, visit www.qats.com or contact ATS at 781.769.2800 or ats-hq@qats.com.