ASICs The Way Forward

ASICs are the way forward; high development costs and long development time are things of the past.

So are you missing the opportunity that an ASIC could bring to your solutions? From Richard Mount Director of Sales at SWINDON Silicon Systems. Spreading the word of the advantages of using mixed signal ASICs is a marketing achievement that not many companies have achieved.

When speaking to companies who have never considered the ASIC approach there is always scepticism and mistrust and the tales of woes of other companies that they have heard of who have attempted the high wire act of ASIC development. This is a misconception that the industry needs to address and address at the highest level of management.

The perception that most Engineering Managers predominantly have are high development cost, long development time, project risk and the commercial viability with regards to volume.
However this is no longer the case and has not been for many years. Choosing the right partner is critical in all engineering projects and the ASIC route is no different. There are ASIC
foundries with their own design capability, design houses that sub contract the production and test and then the true fabless ASIC companies who provide a full in house turn key service.
Given today’s design and development flows, fabrication processes and prototyping services along make these perceptions redundant and out of date.

The Technical Advantages

On the technical side the advantages that ASICs provide in terms of circuit optimisation, size reduction, and power efficiency improvements are very attractive. There are some discrete component approaches that simply can not provide the correct level of performance, size and cost to be viable within its designated market and the ASIC approach is fundamental to achieving project goals. There are also benefits from an engineering team level where the customer provides a specification that outlines functionality and input and output criteria and the ASIC supplier wholly designs the circuitry. The main advantage for this is that the customer’s engineering team is the free to concentrate on other core issues and production of the end product so in real terms risk and development time is actually reduced.

‘Off the shelf’ may not be right for your application

While off the shelf FPGAs and SoCs from mainstream semiconductor vendors continue to displace digital ASICs, the situation is very different within the mixed signal and analogue market as the off the shelf offering is very limited and can not compete with an ASIC designed and produced using the correct technology for the job in hand. In any case, it’ is extremely rare for any off the shelf product to match the customers exact requirement in the precise way than an ASIC can achieve the absolute correct performance for the chosen application.

The above observations are often critical to commercial success when the end application demands the lowest possible power consumption and the smallest footprint combined quite often with custom form factor requirement. Another commercial advantage of an ASIC is the protection it provides of the customer’s Intellectual Property. To reverse engineer is rarely economically viable to achieve and this is especially important when exporting to new and emerging overseas markets.

Why Mixed Signal ASICs?

Traditionally, the ability to replace numerous discrete ICs and passive components with a single slice of silicon is the dominant factor in the economic justification for developing an ASIC. Improved reliability is a key benefit, but others include inventory reduction and less need for testing, as what are now complete subsystems are fully tested prior to delivery, size manufacturing costs and time and the virtual elimination of the age old problem of obsolescence.

With some existing applications it can be advantageous to replicate or even reverse-engineer an existing design and port it to an ASIC in order to introduce new functionality within a much smaller form factor while preserving backward compatibility. This approach can preserve existing design investment while bestowing equipment with a new lease of life. It can also be of interest if an existing supplier is making a process obsolete but there is still a long life time remaining for the end product. In this case it is possible to re engineer a drop in replacement to enable the continuation of supply on a successful product.

The starting point is a clean slate

From an engineer’s viewpoint, the real excitement and opportunity comes with the clean slate approach that ASICs offer. For the most part, the standard requirement is for minimal footprint with low power consumption but with enhanced performance and functionality. This basic set of attributes suits real world products as diverse as tyre pressure monitoring systems and sensors in
factory automation.

The automotive tyre pressure monitoring system is an excellent example: for instance, the sensor project includes an ASIC that’s capable of seeing extremely low voltage levels from a MEMs sensor and provides signal conditioning and processing along with RF data transmission while being supplied by a 3V 350mA/h lithium battery where operational life has to be guaranteed for 10 years. This is in stark contrast with the performance that a typical discrete design achieves, where it’s essential to power down as much of the circuitry as possible for as long as possible, only waking it up when necessary. Even if the discrete design could match the ASIC’s size, its relatively high current demand at turn-on would almost inevitably limit operational life.

The ability to integrate RF stages alongside measurement and control circuitry extends the possibilities yet further to address most telemetry applications, with representative projects ranging from smart metering to traffic toll collection systems to hearing protection with voice transmission and reception capability.

In most cases the building blocks can be reused and available for new design projects thus reducing development times accordingly. However there are on many occasions where the customer requirements demand a full custom design. Below is an example of what individual blocks can designed onto an ASIC. Depending upon the chosen process the die size for a ASIC with this functionality can be as small as 1.5mm x 1.5mm and housed in a 4x4mm QFN package. The advantages in size and cost are easily recognisable.

The ASIC Experience

Fabless companies have access to and experience of a range of fabrication processes that range from 0.6μm CMOS through to 0.18μm BiCMOS SiGe and this range of processes will be a match for the vast majority of applications and analogue circuits. Along with long established economies of scale with the foundries and packaging houses, a solution should be available that is both technically and commercially suited to the customer’s exact needs. By utilising processes where the designers have many years of experience the design risk and the maximising of yield is achieved through expertise and experience.

This minimisation of risk is due to the design rules being well known by the designers and fully and accurately characterised by the foundries and hence accurate estimations of development time and die size can be calculated. The accurate calculation of design resource and time ensures significantly cheaper costs to implement, both in terms of NRE and final device prices as the whole process is understood and planned prior to design start. Despite the clear technical and commercial advantages that ASICs provide, volume and logistics continue to dominate the concerns of potential new entrants in to this arena.

Clearly, high volume production always helps to reduce both silicon and packaging costs, but experience has demonstrated that it can be viable to consider ASICs for volumes as small as 50,000 devices per year in a commercial environment and 5,000 devices per year in the Aerospace and Defence market dependent upon the qualification and testing required in the application.

Test and test and then test again…

One of the major bottlenecks in the logistic chain for ASICs is the testing of the devices. Dedicated test houses cost the product dependent upon test time and the volume tested. Therefore it is not in their interest to be concerned with yield improvement or test time reduction unless the end customer pays for the engineering resource that these evolutions demand. For the lower volume customer there is also a real issue of allocation when the demand is very high. A design / supplier who controls all of these aspects and especially has its own in house low and high volume test capability is very much an advantage to the customer where the test capacity is always kept higher than the demand and the number of customers is not as high as a pure test house. This should prevent allocation and ensure continual supply to the customer to meet their demand. Constant investment both in people and equipment is required to ensure a high quality end product. Included in all device prices is the commitment by the company to continually work on yield improvement and a reduction in test time to ensure that the quality of the shipped device meets exacting standards and that the on time delivery is always kept.

Summary

The market place is an ever evolving environment and companies must continually look at adding functionality and increasing performance while reducing size and cost. The Mixed Signal ASIC route is one that is well trodden and can provide all these advantages and more. There is a misconception that once an ASIC is in production that it can not be changed. This is not entirely true. Second generation of ASIC devices are common and some suppliers have even run to the fifth generation.

One of the real issues that manufacturers also see on a daily basis is obsolescence of electronic devices. The ASIC route virtually eliminates this resource sapping phenomena and
some suppliers will guarantee availability of up to 15 years and more. So what are the main advantages of ASIC technology adoption?

1. Higher integration
2. Smaller footprint
3. Lower power consumption
4. Higher functionality
5. Intellectual Property protection
6. Ease of final manufacture
7. Obsolescence guarantee
8. Proven development and logistic routes

The advantages are real and they are ones that can add sales and profitability to most companies. The only choice the customer has to make is which partner can offer them the best product with the best performance at the lowest cost with the least risk. Investigate the options, you may be pleasantly surprised.