The market constantly demands shorter development times, especially on innovative products with short life cycles. Inicore, using the example of the development of a hearing aid, shows how the time to market of complex systems can be reduced and how different disciplines can be optimally managed and brought together.
We are going new ways in electronic system-design. This means the rapid implementation of ideas in manufacturable products, recognizing the potential of new technologies and benefiting from these points. Our strengths are basically "methodology", the capability to deal with partners in order to generate synergies, a highly developed entrepreneurial culture and a well-harmonized team. Having this, we are convinced of achieving the fastest time-to-market in the industry. The example of our hearing aid development shows how pragmatism, industrial behavior, high level design methodology, technology and teamwork can be brought together.
The idea of bringing DSP power to the ear was already around in the 1980s. However, only todays technologies allow the realization of fully programmable DSP solutions so miniaturized that they fit within the physical constraints of the human ear. In addition, the power consumption of such a device must be extremely low. The vision was to use high level design methodology to penetrate a market which was, until today, only the domain of "technology gurus". The high level description retains the advantage of an excellent time to market over several years because the latest technology can be carried over to the design base as soon as it becomes available. Full custom solutions, however, are technology related and need a much longer development time and are very problematic to transfer to new technology.
In addition, one must master DSP technology in order to be able to adapt the architecture to the needs of the application. This is to ensure that the potential of the DSP can also be used in the future. The know-how of the Canadian DSP factory was
combined with the design experience of Inicore so that algorithm could now be handled optimally on the iniDSP. A FFT-coprocessor as well as an analog chip were developed with the same methodology in a university environment (Uni Neuchatel, Switzerland) and they are a
good example of advanced technology transfer.
Description of the system:
The hearing aid system consists of an analog chip with converters, a digital chip with DSP, coprocessors and different interfaces, as well as a standard EEPROM, which is used as memory for application software. The analog and digital parts have been separated –
contrary to the often discussed mixed-signal-approach - for several reasons:
- There are no technology related compromises. The advantage of the leading edge digital technology 0.18u is connected to lowest-power-analog-silicon.
- The advantages of densities and price, as well as the special capabilities of EEPROM technologies, are well utilized through the use of a "standard product".
- New flexible substrates combined with flipchip assembly guarantee a high degree of miniaturization.
- The development time can be shortened tremendously because we have the relevant technology handling in-house and we can therefore proceed with concurrent engineering.
Another important point during the realization of challenging projects like these is the building of trust with the customer. At the beginning of the project, this was based on our "history of success". As the work became more concrete, the basis became the computer simulation and finally it was the construction of the hardware demonstrator, where the customer has a prototype of the system in his hands. These hardware-demonstrators use "big volume" FPGAs for DSP and coprocessors while analog functions have already been realized on MPW-silicon (Multi Project Wafer). We guaranteed the customer that the database of a future ASIC would be an exact logical copy of the hardware demonstrator. With this methodology, software can be written and audiological tests run months before the ASIC is manufactured. Throughout this ongoing concretization of the product, mistakes can be detected and corrected with minimal time and financial costs. This leads to confidence from the customer and a real team culture builds up, even over long distances, as in this case between Canada and Switzerland. High level descriptions claim that they are not technologically related. However, functionally correct VHDL description is the beginning of a complex synthesis and mapping process. We believe that the timely correct completion of an ASIC implementation can only be guaranteed, if you master all steps which lead to an ASIC. Concurrent engineering is the name of the game. We can then avoid procedural surprises during implementation. Examples of such surprises can be test structures, availability, price and the final size of the future chip. Elegant technical solutions very often fail due to the lack of interdisciplinary communications. This is a matter of people motivation. The entrepreneurial environment is a crucial factor for achieving this kind of success.