Implementing a Bluetooth Solution – Chip Sets or ASICS ?

by Yan Goh – NewLogic Technologies Inc., USA
Michel Eftimakis – NewLogic Technologies, France

Using either commercially available Bluetooth chip-sets, or integrating Bluetooth functionality into an existing ASIC are two possibilities for adding Bluetooth functionality to an end product.Within this brief article we will look at the pros and cons of both approaches with a special focus on cost, risk and time-to-market. While using a Bluetooth chip-set would appear to be technically straightforward, a Bluetooth enabled ASIC approach would seem to be more difficult, especially for companies that have no Bluetooth or wireless system expertise. So, are the uses of Bluetooth enabled ASICs of any use for those companies who want to get ahead in Bluetooth Wireless Technology?

Implementing Bluetooth Wireless Technology in End Products
Over the past few years, Bluetooth technology has been greeted with great enthusiasm from a variety of companies who are keen to implement the technology into their end products. On the whole, Bluetooth is meant to be simple and easy to implement, have low power consumption and be relatively inexpensive. Currently in the market, we are starting to see Bluetooth products (for example, headsets for mobile phone applications, Bluetooth enabled mobile phones and PDAs) from major names such as Ericsson, Sony, Toshiba and Motorola, emerging rapidly to capture market attention. It is estimated that by the year 2005, Bluetooth wireless technology will be a built-in feature available for more than 600 million products worldwide.

It could be argued that there are three main ways to implement Bluetooth wireless technology into an end product. The first is by using a Bluetooth module. This is the easiest method to implement Bluetooth functionality into an end product. It offers the fastest time-to-market solution. However, this is also a very expensive and inflexible method.

The second method is to use a pre-qualified Bluetooth chip set. These are readily available in the market for integration into the system level of the product and major names that offer chip sets include CSR, Zeevo, Philips and Infineon Technologies. The third method is to use a Bluetooth enabled Application Specific Integrated Circuit (ASIC).

Companies can achieve Bluetooth technical capability through in-house development or the licensing of Bluetooth Silicon Intellectual Property (IP). Currently, Bluetooth solutions can be licensed as Intellectual Property (IP) from NewLogic, Ericsson, and ParthusCeva.

Bluetooth Chip Sets
As this article has already mentioned, there are many Bluetooth chip set solutions currently available in the market, CSR and Zeevo being two of the major companies that provide such solutions. They offer users a complete Bluetooth solution on one chip without compromising the size but, unfortunately, at a higher cost. Nevertheless, chip set solutions are still the favourite choice of system manufacturers such as PC and computer peripheral companies. The main advantage of a chip set is that it offers a very fast time-to-market by cutting down the development time, with a guaranteed functionality and tested performance. On the other hand, there is no incentive by the end product manufacturer to invest heavily in any in-house development. This is because it would be much more difficult to develop chip set solutions without the necessary Bluetooth know how or wireless system capability. Today, Bluetooth chip sets are still the main solution for major system manufacturers.

From an engineering point-of-view, chip set implementation is very inflexible, despite its short development time. A Bluetooth chip set will also need external components, such as crystals and interfaces, to complete its functionality and so consequently not only the size and cost of the overall system implementation increases, but also the power consumption of the whole system. It can also be argued that the use of a chip set limits the creativity and ideas of the system designers, as the system design must evolve around the Bluetooth chip set. Furthermore, there are also limitations that discourage system designers looking for ways to reduce the total system implementation size – primarily because the size of the chip set is fixed and the numbers of external components are also compulsory.

A final, but important point is that the Bluetooth specification is still evolving and will perhaps need to be revised. When there is a revision to the Bluetooth specification, considerable disruption may be caused to the pre-specification chip set users. Currently, the industry is complying with Bluetooth version 1.1, however, if the SIG decides to revise the Bluetooth specification to a newer and better version, the current Bluetooth chip set implementation will become obsolete.

Bluetooth Enabled ASIC
Over the last few years, the popularity and usage of Application Specific Integrated Circuits (ASIC) in the electronics industry has been catching up with the usage of available, ready-designed, chip sets. The main difference between chip sets and ASICs lies in the fact that chip sets are devices that are readily available as an off-the shelf item, while ASICs are ICs that are tailored for a specific application. With the growing popularity of ASICs among the electronics industry, ASIC design houses are appearing all over the world in order to cater to the needs of a variety of customers.

Today, ASIC solutions have made great advances, primarily due to new, efficient, IC manufacturing technologies, as well as much-improved Electronics Design Automation (EDA) tools. As a result, the cost of ASIC development and implementation has fallen to a level that is also affordable to many small and midsized companies. Furthermore, there is also a growing trend in the popularity and usage of Silicon Intellectual Property (IP) in IC design as well as System on a Chip (SoC). SoC is a new design methodology where several systems can be integrated onto a piece of silicon. For example, in the case of a Bluetooth system, the microprocessor, Bluetooth processor, Bluetooth radio, memories and interfaces are all placed together onto a single piece of silicon. There is a growing dependency upon IP in the ASIC industry, because most ASIC or SoC design houses prefer to concentrate upon IP assembly and offer an instant and fast time-to-market solution. For this reason, there is a growing trend towards the development of IP solutions that are concentrated upon, and actively engaged by, pure silicon IP providers.

Due to the advantages provided by ASIC solutions, the demand for Bluetooth-enabled ASICs is growing as a consequence. There are two options that a company should consider when deciding whether or not to develop a Bluetooth-enabled ASIC. Firstly, Bluetooth wireless technology could be developed in-house – although this could also result in high development cost, long lead-time and the need for additional experienced resources. Secondly, Bluetooth technology could be licensed from major third party IP vendors such as NewLogic. These Bluetooth IP providers offer up-to-date Bluetooth technology (which complies with the latest Bluetooth specification), a silicon proven solution and a cost-effective way to implement Bluetooth wireless technology.

Bluetooth enabled ASICs versus Bluetooth Chip sets
The remainder of this article will spell out the main advantages of ASICs with Bluetooth functionality over standard Bluetooth chip sets. It should also give the reader a much clearer view on why the electronics industry is adopting and welcoming ASIC solutions, not only to implement Bluetooth technology but for other systems as well.

Reduction of physical dimension
One of the main visible advantages of using a Bluetooth ASIC is the reduction in physical dimension. This miniaturization can be pushed to the limit, helped by the fact that currently it is possible to squeeze several standard ICs and systems into just one single IC. For example, in a Personal Digital Assistant (PDA), one single ASIC can contain the microprocessor, Bluetooth baseband, the interface components, memories and any other external components needed for a totally functional PDA system. The reduced signal path between components also reduces the physical system dimension. This advantage is far more difficult to achieve using a standard Bluetooth chip-set.

Reduction in power consumption
Power consumption is an important element of any Bluetooth wireless system. By replacing most major components with an ASIC including advanced CMOS process technology, power consumption is reduced significantly . Most Bluetooth IP's are designed using CMOS technology, primarily due to its low power consumption and low cost.

Cost of Bluetooth implementation on ASIC
Bluetooth IP is a zero cost implementation for Bluetooth enabled ASIC. From a technical point of view, once a customer licenses a Bluetooth IP or obtains it through in-house development, the technology and solution is essentially there and free for them to implement on any ASIC. For example, if a particular system needs to be Bluetooth enabled, all that is required is to place the available Bluetooth IP into the available ASIC.

Since Bluetooth IP is a proven working solution, designed for easy integration in ASICs and available at nearly zero cost, why use Bluetooth chip sets?

By using ASICs, several components or systems can be integrated on a single chip. There is no need to purchase off the shelf components (needed when using chip sets), and this once again reduces implementation cost. Furthermore, the total cost of Bluetooth enabled ASICs will continue to drop with a high volume production.

Performance
By using a Bluetooth ASIC to implement Bluetooth wireless systems, no unreliable and unstable external components are needed. This has the effect of increasing the reliability of the system as a whole. Furthermore, an ASIC also uses less external connections (wires or copper strips), which may induce unwanted electro-magnetic fields affecting system performance and functionality. These advantages cannot be achieved when using Bluetooth chip sets. An external antenna is the only external component required when using a Bluetooth-enabled ASIC.

Reuse Methodology
Bluetooth solutions developed in-house or Bluetooth IP used in ASICs allow reuse methodology in System on a Chip (SoC), which is impossible to implement when using Bluetooth chip sets. By taking advantage of reuse methodology, this will significantly cut down development time and costs, and result in greater efficiency, primarily due to the fact that the Bluetooth solution has already been used and the development team will be familiar with the methodology.

Furthermore, reuse methodology in multiple projects allows the Bluetooth IP license to optimize the initial investment or licensing fees, moving towards a zero cost Bluetooth solution. Bluetooth enabled ASIC trade off

Like any technology, there will be some drawbacks when using a Bluetooth enabled ASIC. The main disadvantage is the high initial development cost, which consists of licensing the Bluetooth IP or in-house development. There is also the non-recurrent engineering cost (NRE) for chip fabrication, packaging and testing. Usually, an ASIC solution has a slightly higher risk when compared to using chip sets and the development time is also longer when compared to buying off-the-shelf Bluetooth chip sets. However, the advantages of Bluetooth enabled ASIC should certainly not be overlooked. In the long term, the high initial cost and development time will pay off with a substantial reward when the volume production of Bluetooth-enabled ASICs rapidly takes off.

In house development or IP licensing for Bluetooth enabled ASIC?
What type of implementation for Bluetooth enabled ASICs should a manufacturer be looking for. There are various comparisons that can be made between in house Bluetooth development and licensing Bluetooth IP from a third party IP vendor.

On the one hand, costs can be high when Bluetooth wireless technology is developed in-house. Heavy investment is needed to source qualified resources and EDA tools; potentially somewhere in the region of at least US$ 5 million. On the other hand, the cost of licensing Bluetooth technology varies with the different kind of business models per ASIC project, plus royalty scheme. IP business models vary from low start-up fees to a high one-time payment without royalty. Therefore, the author would argue that IP licensing is a cost-effective way to realize

Bluetooth wireless technology. A high volume production will also reduce the IP licensing cost.
It can take at least one to two years to develop a working Bluetooth solution. However, Bluetooth solutions from third party IP vendors are fully functional and ready-to-use instantly. Technically, it could only take three to six months to achieve a Bluetooth-enabled ASIC. Furthermore, as all Bluetooth products need to be qualified by the Bluetooth Qualification Body (BQB), Bluetooth IP is no exception. Currently, only some of the Bluetooth IPs on the market are BQB qualified. By using a qualified IP, the time to qualify the final product will be significantly reduced. And the only additional effort needed is to integrate the Bluetooth IP into the ASIC. In the electronics field, time-to-market is a very important factor and it would appear that using Bluetooth IP ensures that the critical time-to-market equation remains low.

As with all in-house development, the associated risks can be very high and will also be compounded and affected by time-to-market pressures and the intense pressure from competitors. Using a fully functional, BQB qualified Bluetooth IP, however, comes with a very low risk. Licensees will only be responsible for the IP assembly and ASIC production. Third party IP vendors will assure licensees of the functionality and maintenance of the Bluetooth IP. Furthermore, licensees will also have the option to upgrade the Bluetooth IP to a future specification; the risk of using an out dated Bluetooth technology is therefore eliminated.

In-house Bluetooth development will also require in-house expertise within the organization. Currently there is limited wireless expertise in the electronics industry and the opportunity cost is also high. When using Bluetooth IP, the only required expertise is the IP assembly. It does not require heavy commitment and investment in Bluetooth wireless development. Furthermore, the current business model trend in the electronic design industry is to engage actively in outsourcing. Bluetooth wireless technology is no exception.

While it is true that in-house development will create a very high technical capability within the company, depending on the type of IP licensing model, technical transfer could also be available. For example, source code for the Bluetooth IP is normally part of the deliverables, as is training. It ensures IP licensees have total control over the core technology and can achieve a high technical capability in using the IP. Hence, some Bluetooth IP providers not only provide Bluetooth IP solutions, but also offer technology and skill transfer with an option to upgrade to the revised Bluetooth specification.

In order to sum up some of the arguments made in this article, the table (Figure 1) summarizes the types of Bluetooth technology implementation against the arguments made earlier in the article. System on a chip (SoC) methodology is a widely accepted approach in the ASIC industry and is used to develop Bluetooth enabled ASICs. This has also greatly helped to develop the Silicon Intellectual Property (IP) field, especially in new emerging technologies such as Bluetooth wireless systems.

It is believed that the popularity of Bluetooth enabled ASICs will surpass the usage of Bluetooth chip sets in the very near future, due to the fact that using ASICs to implement Bluetooth technology could provide many more benefits with respect to a conventional solution using chip sets.Whether it is by in-house development or IP licensing, the benefits of using a Bluetooth enabled ASIC goes beyond any cost issue, due to the advantages that only an ASIC solution can offer.

Summary
By buying an IP license, the customer is essentially buying an off-the-shelf-solution, very similar to buying a Bluetooth chip set. Therefore, when choosing a third party Bluetooth IP vendor, factors such as the type of business model available, IP deliverables, training, support and maintenance should be examined and compared. With due consideration, one might find a Bluetooth IP vendor that provides all the necessary benefits that come with the IP license.

References:
1. Bluetooth Handbook 1.1, Andrew Griffin, Merrill Lynch, June 2000.
2. Reuse Methodology Manual, Michael Keating and Pierre Bricaud, Kluwer Academic Publishers, 1999.
3. Bluetooth Demystified, Nathan J. Muller, McGraw Hill Telecom,2001