A semiconductor ecosystem that applies the principles of open, collaborative software development to hardware. Why is this interesting and important for us as a society? Alexander Stanitzki, Head of Industry at Fraunhofer IMS, and his team are researching on open source semiconductor solutions for various use cases in industry, thus also participating in a potential revolution in microelectronics development. »Due to closed hardware systems, there are often hurdles in the research and development area to implement good, sustainable ideas across the board,« explains Alexander Stanitzki. Find out which solutions and projects can break down barriers in the feature below.
An open ecosystem with the help of RISC-V, competence networks, and high-quality training for young people
Linux, probably the most successful example of open source development, and the investments of tech companies such as Google in open source chip design tools, indicate the potential that is hidden here.
»Since software developers are bound to their system, good solutions often have to be extensively modified for other hardware, which often increases costs and energy expenditure. This system dependency must be dissolved through open source solutions, also in IC design or for RISC-V, in order to design an open ecosystem in which we can develop industrial solutions in a climate-friendly manner and safely for the users,« says Alexander Stanitzki. In this context, the Fraunhofer IMS is also committed to the expansion of a RISC-V ecosystem in Germany and Europe. To this end, last year the institute developed a European model for a RISC-V ecosystem with the ASPECT study[1] on behalf of the German Federal Ministry of Education and Research (BMBF), in which open source solutions play a major role.
In addition, CHIPS.NRW and the Microelectronics Academy of the Research Fab Microelectronics Germany (FMD), to which the Fraunhofer IMS also belongs, have created cooperation platforms as well as training and further education centers. Both networks want to offer high-quality training and provide an exchange platform for start-ups, science, and industry to establish chip design »Made in Germany« on the basis of open source semiconductors.
Besides building consortia and conducting studies, Fraunhofer IMS is already developing its own solutions that are open source and thus flexibly adaptable. You will get to know the most important ones below.
Open source RISC-V processors with AI extension (AIRISC)
Fraunhofer IMS has already developed a wide range of embedded AI applications with the flexibly configurable and customizable RISC-V processor system AIRISC and offers sustainable solutions for versatile applications in industry, health care, mobility, and security. Application areas can be, for example, smart patches and wearables for monitoring vital parameters, current transformers, predictive maintenance of machines, AI-supported pre-processing of LiDAR data, or flight controllers of drones and similar flying objects. AIRISC is open source and available on GitHub.
IC designs based on open technologies and add-on sensors for CMOS circuits
IC designs based on open technologies are perfectly suited for industrial and medical sensors, IoT, and wearables, offering optimal performance through domain-specific and energy-efficient designs based on open source hardware. The template-based and fast system-on-chip design ensures the best possible performance per watt, significantly reducing development time and fixed costs.
With add-on sensors for CMOS circuits, Fraunhofer IMS creates integrations of sensors into any kind of chip, even from external manufacturing. The basis for this are IP cells for our post-CMOS SPADs, microbolometers, pressure, and gas sensors in common open source PDKs. Application fields can thus be integrated pressure sensors, LiDAR detectors, uncooled IRFPAS, miniaturized as well as smart gas sensors.
With its open technologies, the Fraunhofer IMS shows how versatile and adaptable sensor technology can be. The solutions are even compact, energy-efficient, and increase safety in the application. You will get to know two examples in the following. With the GaNext and PowerCare projects, the Fraunhofer IMS is pursuing two different approaches in the field of power electronics with the aim of not only making them more pleasant and safer in the end-user application, but also contributing to sustainability.
More compact, efficient, and fail-safe modules for future technologies and regenerative energies through gallium nitride (GaN), AI and RISC-V
GaNext develops an intelligent power module based on gallium nitride semiconductors, which is combined with gate drivers and a programmable, fail-safe control unit with integrated protection circuits. In the field of regenerative energies, for example, an enormous savings potential in the overall system can be achieved in the voltage conversion and feed-in of energy from photovoltaics and wind power due to a more efficient power electronics system.
The PowerCare project implements a new monitoring concept in the form of a miniaturized motor controller with integrated real-time failure prediction, thus laying the foundation for a new evolutionary stage of intelligent power modules. For this purpose, AI-based prediction models are implemented on a domain-specific RISC-V control SoC and combined with highly efficient GaN power transistors to form a power module. These power modules can, for example, support future technologies such as drones, electric aircraft, or collaborative medical robots in a compact, reliable, and sustainable way. Fraunhofer IMS contributes its know-how from the RISC-V area with AIRISC.
The possibilities of open source semiconductor solutions seem almost unlimited. »We are still at the very beginning,« Alexander Stanitzki concludes. »We need to continuously test, develop, and put into practice our activities in the emerging ecosystem so that the open approach can be used as an advantage for everyone in local and European industry, as well as in academia. At Fraunhofer IMS, we are strongly committed in the Business Unit Industry to transfer this vision into reality.«
[1] The ASPECT study was funded by the BMBF under grant number 16ME0656K.