Chips used in space, cars and medical devices face some of the harshest operating environments of any electronic products, have the highest demands for low power operation and must have bulletproof reliability.
For many manufacturers in these three sectors this more often than not leads to a decision to shy away from using cost-effective off the shelf technology and to undertake a laborious custom design process for chips. However, IP processor cores could provide the off-the-shelf design approach to low power and high reliability these industries need.
High reliability and lower power design are critical to effective microelectronic systems of all descriptions, none more so than electronic equipment designed for use in space, automotive machines, and medical electronics such an in-body implantables. Each of these industries has their own challenges.
For instance, space represents one of the harshest environments for electronic systems to operate in, and space companies must ensure that they use chips that are designed to cope with extreme temperatures and radiation. Furthermore, the need for high-reliability systems in automotive and medical electronic systems can literally be a matter of life or death, meaning every element must work flawlessly.
However, silicon chip vendors gain little value from designing chips for these low volume niche applications – whether that is developing radiation hardened versions of standard products for space applications, or ultra low power devices which pass stringent medical regulations.
Consequently, space, automotive and medical companies cannot take advantage of production economies of scale and must spend vast amounts of time and money on custom designing chips.
Custom design is a challenge
However, in these industries, even taking a custom approach to designing these chips entails significant challenges. Area constraints drive the need for system-on-chip (SoC) single chip solutions.
It is also often the case that chips must be mixed-signal ICs, which feature both digital and analogue parts, and either support large on-chip memories or interface with external solutions.
Another key requirement is that the systems must deal effectively with real-world interfaces to specific sensors and actuators, as well as being compatible with both wired and wireless communications systems.
High-reliability systems require other specialist features in order to ensure that they are 100% fit for purpose that further increases the design challenge. Any technical failure is extremely difficult to repair and can have disastrous consequences.
Failures in space hardware cannot be solved by a local repairman, for example, meaning that systems must be designed to work flawlessly for 15-20 years. Similarly, automotive and medical electronic system failures can have catastrophic effects and these chips must also have total reliability.
This high reliability is not just about the hardware performance, it also includes the software development. When dealing with high-reliability systems, software design must cater for software test pattern generators, code coverage, and validation tools.
Chips in space
As mentioned above, space microelectronics must be completely geared to deal with the high radiation environment, ever-changing extremes of temperature, and the heavy vibration and shock of the initial launch process and transition from atmospheric pressure to vacuum.
While the hostile environment drives the need for single chip solutions to be rugged as well as high-reliability systems, one additional constraint of space hardware is that it must be as light as possible, as every extra kilogram could add up to $20k to the cost of launching a satellite.
Similarly, medical and, to a lesser extent, automotive electronics are constrained by weight and size. High-performance motor vehicles are designed to be as light as possible, from the material chosen to build the chassis down to the microelectronics inside. Similarly, in-body implantables must operate effectively within extraordinary compact areas so as not to impact on other bodily functions and further complicate the surgical process.
On top of this incredibly demanding set of power, weight and size requirements, there are more detailed features the processor must support to ensure the system as a whole is reliable while consuming as little power as possible. These custom chips must ideally be able to provide the ‘big-boy’ computing features, usually only found in large, power-hungry desktop processors, in a very small silicon and energy footprint with bulletproof reliability.
To ensure energy efficiency, a processor needs a fast start-up time, allowing it to spend more time in low power sleep modes. It must offer a high code density with sensible data width and efficient stack operation, to reduce the power consumed by the memory.
Software reliability is provided by support for modern programming models, which offer protected software operating modes that isolate user code from privileged code for secure, deterministic, trustworthy and high-availability processing. It is also often necessary to design in an ability to perform software upgrades remotely, wirelessly and securely.
The cost and timescales of software development can be reduced if there is support for multiple languages, operating systems and communication stacks. Finally, in the case of medical and automotive electronics all of these systems must be completely verifiable in order to satisfy stringent regulatory standards as well.
The exact mix of these features required by a specific product probably does not exist all at once in commercial off-the-shelf (COTS) products. Existing mass-market chips will either be underpowered, resulting in a degradation of performance, or offer too many unneeded capa bilities, resulting in a design that will not fit the size, weight and power constraints.
Chip suppliers do not see a return for their investment into the design process of such unique, complex designs, as the volumes are far too low. This means that space, automotive engineering and medical device companies cannot take advantage of low-cost, high volume microcontrollers and other off-the-shelf silicon.
One potential middle-ground solution to this issue is the use of IP cores in custom chips, which can provide access to the benefits of high-volume engineering: high-reliability, maturity of design and large investment, while retaining the ability to customise the device to exactly match the requirements of the user.
One example of IP core technology that fits this niche is the XAP family of microprocessors, produced by Cambridge Consultants and licensed in over one billion chips worldwide.
Processors such as this give niche application designers the ‘off the shelf’ type of product they need to effectively utilise economies of scale to save time and cost. At the same time, such products also provide the flexibility, reliability and low power operation that these specific industries demand without compromise.
Chris Roberts is senior consultant at Cambridge Consultants