This Small Business Innovation Research Phase I project seeks to develop integrated voltage regulators that are 20 times smaller than existing DC-DC converters and enable better power management for processors. Modern processors use a large number of voltage domains to improve energy utilization of processing cores and IP blocks. The problem is that each voltage domain requires a separate DC-DC converter and existing converters require board components including inductors, capacitors, power switch and a controller chip. As a result, the number of board components increases dramatically for complex processors that need a large number of voltage domains. This leads to large board area overhead, high component cost and complex board-level power delivery. This project aims to develop a single-die voltage regulator that eliminates all board components and integrates everything in a single die regardless of the number of voltages the regulator needs to provide to the processor. This new regulator will reduce board footprint, which is very important in both mobile and data-center applications, and reduce component cost. This regulator also has the potential to reduce power consumption of processors by up to 30% by enabling more efficient, finer-grain power management.
The broader impact/commercial potential of this project is reduction in size, cost and power consumption of power delivery solutions for processors and logic chips in a wide range of applications from smartphones to datacenters. In a smartphone, voltage regulators are one of the biggest components other than the battery. While customers demand lighter, thinner smartphones, increasingly more space is required for new features such as 4G communication, micro-projectors and high-resolution cameras. Reducing the size of voltage regulators is critical in creating space for new features in next-generation mobile devices. Moreover, saving processor power using the new regulator can increase battery life. Reducing power and size are not only important for mobile devices, but also for server processors deployed in data-centers necessary to support increasing web traffic. Saving processor power directly translates to lower cooling costs in data-centers and reduction in CO2 emission. By reducing the size of regulators, more processors can be stored in a limited area to create more powerful servers for data-centers without increasing real estate. Impacting processors across the globe, the new regulator has the potential to disrupt the $10B voltage regulator market and enable further innovation in mobile devices and servers.
All mobile electronics devices contain a semiconductor chip called a power management IC (PMIC). The PMIC is in charge of delivering power from the battery to all sorts of different chips inside a mobile device, such as the processor, memory, chips for RF communications etc. Fitting all these chips into a limited area is a big challenge for mobile devices. The less area the device allocates for chips, the more area it has for the battery, which leads to longer battery life. The problem with existing PMICs is that they occupy a large area. It is one of the largest chips inside a mobile device, occupying up to 25-30% of the entire PCB area that contains all the chips. The reason behind this large area is because existing PMICs require many discrete components, up to 40-50 for high-end smartphones. The problem is getting worse as high-end mobile devices need more sophisticated PMICs. Also, area contraints become even tighter in wearable devices like the Samsung Galaxy Gear or Google Glass. Lion Semiconductor develops a new type of PMIC that can reduce footprint to 1/3 to 1/5 of existing PMIC solutions. We provide a single chip solution that requires zero discrete inductors on the PCB. By saving significant area on the PCB, Lion Semiconductor's solution leaves more room for the battery and can increase battery life for customers. It also leaves more room for additional chips that can be added for new features such as faster 4G communications, micro-projectors, and health monitoring sensors. Throughout the award period, we built a proof-of-concept semiconductor chip that proves that our new idea works and we can actually build a working PMIC with zero discrete components. The next step is to build a PMIC that is more than a proof-of-concept, that can actually be used in a mobile device (e.g., smartphone, wearable device) to power up all the chips and function without any problems. Once that is done, we will build a version that can be mass produced to be used in millions of mobile devices.
