Abstract
Today, designers are demanding an overall form-factor reduction to save board space, increase functionality, and allocate more circuit board real estate toward end-user applications – all with less space allocated to power management where not just the X–Y shrink but the 3D volumetric shrink is required. For example, in today's Telecommunications Cloud Infrastructure systems, board space and power density are challenging particularly in power supply designs where several high-current point-of-load rails are present. End equipment such as enterprise servers and switches, workstations, base stations, network attached storage, FPGA testers, network testers, and other test and measurement equipment employ several high-current CPUs, ASICs, FPGAs, and DDR memory – all of which need high power, while the available board area is steadily decreasing. This paper will discuss volumetric co-design methodology and packaging construction trade-offs for 3D SiP power modules and also introduce the straddle mounted inductor assembly technique. Also provided are details around the SiP eco-system, co-design, construction, materials, and circuit topology.
I. Introduction
A. The appeal for SiP solutions
Power supplies require expertise that many system developers lack, such as how to select the type of power supply or meet compliance specifications. It's no small wonder that developers often prefer to look for a ready-made solution that takes away the headaches of power-supply design. System-in-package (SiP) power modules provide ready-made, easy-to-use solutions for power supplies. 3D SiP modules integrate a complete DC-to-DC converter power system in a single package using three-dimensionally stacked components. The result is a cost competitive solution with increased power density and simpler designs for customers, helping to accelerate time to market and realize revenue faster.
SiP packaging technologies are easy to implement and comes with a number of performance benefits, including excellent thermal characteristics and minimized electromagnetic interference (EMI); But what truly sets SiPs apart is the use of 3D packaging techniques, where components can be mounted on top of each other inside the package. For space-conscious applications, 3-D packaging has given power modules a higher power density and significant form-factor reduction, saving customers significant board space, optimizing the overall PCB (Printed Circuit Board) footprint, and supporting the industry trend toward miniaturization. Fig. 1 illustrates the miniaturization trend over a 25 year period. Additionally, 3-D packaging techniques optimizes volumetric form factors by placing the large power inductor over the rest of the power-module circuitry, optimizing space thru the use of a stilted lead interconnect that is straddle mounted over lower profile active and passive components. Straddle mounting inductors also improves electrical performance by allowing passive components to be placed closer to the active controller device, reducing the total signal length. When these advanced 3D construction techniques are utilized, a simplified module with a best in class power density can result.
Power modules have paralleled the tremendous downscaling that has taken place in the semiconductor industry.
Power modules have paralleled the tremendous downscaling that has taken place in the semiconductor industry.
Power module eco-system illustration of package styles distributed across voltage vs. current.
Power module eco-system illustration of package styles distributed across voltage vs. current.
These system application areas include large, performance-driven systems such as communications infrastructures, data management, office equipment, building automation, industrial, and transportation and defense, right down to small, often cost-sensitive systems such as sensors, appliances, consumer electronics, and even portable and wearable systems. In short, affordable SiP modules exist for almost all electronic applications, backed by in-depth development tools that help ease module selection and implementation.
B. Complex systems require straightforward solutions
The wizardry of power design seems to grow even more mysterious as systems become smaller and more complex. Product miniaturization puts extra pressure on designers to scale down the power system, and mobility requires designs that squeeze longer running times from batteries. Efficient usage demands that newer equipment accepts higher-voltage inputs near the point of use. Highly integrated systems with complex loads require increasing numbers of voltage and current levels, implemented using complicated power trees, multiple power-supply stages, or both. Circuit analysis often yields numerous possible implementations, with various trade-offs for efficiency, size and cost.
Within the SiP module Eco-System, developers are actively partitioning the various levels of integration to minimize the development time and effort needed to achieve a competitive system solution. As the industry matures there are now partitions that allow “just enough” integration to achieve an overall system solution dependent on design resources.
Schematic level partition comparing a controller vs. converter vs. power module.
Schematic level partition comparing a controller vs. converter vs. power module.
Power-supply topologies also complicate development decisions. Traditional linear voltage regulators are relatively straightforward and flexible for designers, but they require significant airflow and board space for cooling. By contrast, switched-mode power supplies (SMPSs) are becoming increasingly attractive in many applications. The high power efficiency of SMPSs limits the space needed for heat dissipation, prolongs battery life in portable systems, and helps lower operating costs for line-powered equipment. Developers must carefully control the timing in high-frequency switches, while simultaneously preventing them from interfering with low-frequency circuitry in the rest of the system or transmitting back onto the input power line. High-frequency switches also require protection from external noise and internal parasitics.
All of these factors affect the difficulty of and time required for development, design debugging and manufacturing test, while power has its own safety requirements that complicate the process further. System developers may find the complexity overwhelming, especially small development teams that do not have an expert devoted to power-supply design. SiP modules remove the difficulties of power design and smooth the development process, allowing designers to concentrate on areas where they can add maximum value to their products.
The effect on development time when using a SiP module can be decisive for the success of a product. A market report by the independent Darnell Group [2] found that a SiP module-based design requires 45 percent fewer man-hours to complete than a design based on a discrete DC-to-DC regulator [2]. Such a large saving in development time can easily make the difference in realizing increased revenue and profit from being early to market with a new product. In addition, a reduced bill of materials, along with fewer mounting steps, helps simplify manufacturing, increases pass rates during test, and improves overall reliability. Managing the cost of buying and taking inventory of extra components is also significantly reduced. For all of these reasons, SiP modules are not only cost-effective compared to discrete solutions – they may even be the key to succeeding with the product in the marketplace.
II. Types of SiP module packages
A. QFN Modules
In quad flat pack no-lead (QFN) packages, an example of a 3-D construction technique employs a stilted (raised) inductor straddle mounted over the IC (integrated circuit) package. Straddle Mounting places the inductor with sufficient clearance over the lower profile IC and additional passives using a 3D package-in-package (3D PiP) stacking technology. This is a relatively simple manufacturing process incorporating a copper lead frame substrate for mounting the active and passive devices which are then subsequently encapsulated with plastic molding compound. This combination of advanced materials enables the lowest thermal path for superior safe-operating-area performance. In addition, QFN SiPs exhibit excellent thermal capability incorporating a simple pin-out, making them easy to use. The short electrical paths combined with closed-loop magnetics provide best-in-class EMI protection with all signals accessible, and with compensation and programming already integrated. Illustrated in Fig. 4 are examples of 3D PiP construction which is considered another subset within the 3D SIP ecosystem identifiable from the finished packages integrated internally. One can further consider the significance in this partition since the infrastructure required to produce a 3D PiP which only requires an SMT line in the frontend and molding plus singulation for the backend of line greatly differs from a 3D SIP incorporating an unpackaged wirebonded device.
For comparison Fig. 5 illustrates a typical 2D QFN module where the active devices and passive devices are placed next to the power inductor. This side by side approach is effective as a plug and play solution but doesn't offer the volumetric density in comparison to the 3D PIP.
B. Embedded Die MicroSip
Another great use of 3D packaging is embedding active die inside laminate substrates and placing passives on top of the laminate. An example is the MicroSiP module, a tiny module that occupies very little board space and has industry-leading current density. MicroSiP is perfect for miniaturized systems like those found in wearable and personal electronics.
3-D MicroSiP package cross-section with IC die embedded in substrate and passive components placed over the substrate.
3-D MicroSiP package cross-section with IC die embedded in substrate and passive components placed over the substrate.
C. Leaded Power Module
There is also the very innovative, yet easy to use, transistor outline (TO) package power module that takes advantage of 3-D packaging. The TO modules have dual lead frames with active die and passives placed on either side of the lead frame for a 3-D leaded package-based SiP module solution that takes up little board space. TO's are great for industrial systems that operate under harsh conditions. TO package power module external leads make the assembly process and mounting very easy for power supplies, especially for assembly sites that do not have advanced manufacturing capabilities.
3-D TO Power Module package cross-section with dual leadframe and active IC die and passives placed on either side of the leadframe
3-D TO Power Module package cross-section with dual leadframe and active IC die and passives placed on either side of the leadframe
D. Straddle Mounted Inductor
In order to realize a volumetric shrink in solution size, the technique of stacked or also referred to as straddle mounted inductors are typically used. This technique requires the inductor to have sufficient standoff from the modules substrate which in this case an organic laminate based module approach is used (see Fig. 8). In order to avoid contact with the IC and any additional passives below a stilted inductor is used. Inductor stacking is synonymous with PoP (Package on Package) stacking of memory BGA's onto Processor BGA's which maximizes efficiency, reduces signal lengths in routing, and optimizes volumetric solution size.
The assembly flow for a single pass reflow process is illustrated in Fig. 9 incorporating a power inductor straddle mounted over a QFN power IC device [1]. The primary consideration is to insure the standoff of the inductor is adequate to avoid contacting the QFN package below. Regarding the best reflow profile to use, the solder paste suppliers recommendations should be the primary source since the flux performance is critical to achieve robust solder joint geometries [3]. In the event a single pass reflow can't be used due to QFN device inspection requirements prior to stacking such as AOI (Automoted Optical Inspection) or inline X-ray inspection the process flow can be decoupled in a 2-Pass reflow process requiring the inductor to be placed in the 2nd pass thru the use of dispensed solder paste. Regarding inline X-Ray inspection the Power inductors typically incorporate a ferrite alloy which requires a high X-ray intensity to penetrate making inspection difficult with most conventional X-ray systems.
3-D TO Power Module package cross-section with dual leadframe and active IC die and passives placed on either side of the leadframe
3-D TO Power Module package cross-section with dual leadframe and active IC die and passives placed on either side of the leadframe
III. Conclusion
Many different 3D techniques are employed in power module architecture. Ingenuity and persistence are the keys to designing a high density, easy to use, power module. The pressures of increased complexity – including miniaturization, multiple power rails and the need for more power-efficient topologies – make power-supply system design seem even more difficult. Fortunately, SiP power modules offer an easy-to-use complete DC-to-DC converters with many options for different applications and manufacturing requirements. SiP modules increase power density and help speed time to market, making them a cost-effective option for new systems in a wide variety of application areas. From giant multichannel equipment to the tiniest wearable electronic accessories, SiP modules are powering innovative systems for the future.
Acknowledgments & Bio
Steven Kummerl received the B.S. degree in Mechanical Engineering from University of Texas El Paso in 1995. He is a member of technical staff at Texas Instruments supporting semiconductor packaging R&D. He has worked in the field of high volume/high mix surface mount assemblies for more than 13 years and over 15 years at TI supporting packaging R&D. In his most current role Steven has designed numerous complex 3D packages, researched materials for robust second level reliability performance, and created technology roadmaps for TI. He holds over twenty patents in the field of package design and has authored multiple publications focused in SMT package assembly & reliability.
Charles DeVries received the B.S. degree in Electrical Engineering from University of Illinois Urbana-Champaign in 1994. He is the Power Module Technology Manager at Texas Instruments, and is a member group technical staff. He has over 20 years' experience in systems development of battery charging, LED lighting, wireless charging, and over 15 years in POL and DC/DC power modules. He has authored multiple publications on power module design and usage, has two patents, and won 2006 EDN Power Sources Innovation of the Year.
Usman Chaudhry received the B.S. degree in Mechanical Engineering from Georgia Institute of Technology in 2002, and the M.S. degree in Packaging of Electronic devices from Southern Methodist University in 2006. He manages the High Voltage, Isolation and System in Package Modules group in the Semiconductor Packaging organization at Texas Instruments. He has over 10 years' experience working in the Semiconductor Packaging field covering various Wirebond, Flip Chip and System in Package technologies.
Chong Han Lim received the B.Eng. degree in Electrical and Electronic Engineering from University of Leicester in 1996. He has over 20 years of experience in Semiconductor Assy/Test Manufacturing, and is currently managing Texas Instruments' Asia Packaging organization.