SoC Test and Verification -> Testing mixed-signal Bluetooth designs

Testing mixed-signal Bluetooth designs
By John Hansen, Product Marketing Specialist, Agilent Technologies, Palo Alto, Calif., EE Times
December 13, 2001 (11:12 a.m. EST)
URL: http://www.eetimes.com/story/OEG20011213S0030

The Holy Grail: Create a Bluetooth system-on-chip (SoC) that fully integrates processor, programmable logic, SRAM, EEPROM and the RF transceiver for less than $5.

If you think developing the mixed-signal RF-CMOS processes required to fabricate such a device weren't enough of a challenge, the SoC design is not done until the software is verified and the entire system passes the rigorous Bluetooth qualification process. As the SoC becomes more complex, the verification process can become overwhelming and fraught with errors and time-consuming test development. A strong recommendation would be for design engineers to find methods that enable their SoC tester to accept design and simulation files directly from their electronic design automation (EDA) tools, and generate test program files.

The question is whether it makes sense to build a Bluetooth implementation from the ground up, using certain components from an outside vendor, or buy a com plete solution from a major development house. When it comes to designing portable communications products, which are affected by increasingly small form factors, short product cycles and competitive time-to-market demands, the answer is a higher degree of integration--while still offering product differentiation and protecting slim margins.

When designing Bluetooth into a product, engineers must be aware that the Bluetooth system consists of both hardware and software --both of which will entail tradeoffs between power and performance.

Product development schedules and the availability of RF resources will initially dictate whether a designer opts for a modular or bottom-up design. The latter is generally a good idea when a product is incorporating emerging standards or a technology or feature that demands customization. A product's actual design might preclude the use of standardized modules, chip sets, or SoC, while an extremely high-volume design is a viable reason for building an RF solution from scratch.

The module approach is an obvious choice for companies lacking sufficient RF design capabilities. It can also expedite testing for standards and regulatory agency compliance. But, while modules can provide predictable performance and fast, low-risk implementations, they're designed for a range of applications, and may not fit a particular design. In such cases, creating a customized solution with ICs is the way to go, using highly integrated chips developed specifically for Bluetooth.

Packaging requirements are a critical part of the design process, and should be addressed early on. Where will the antenna be located? How will the product's housing, RFI shielding and proximity to the person operating it affect antenna performance? If there are other RF sources such as a cellular handset present in the product, how will they interact with the Bluetooth radio?

One example of Bluetooth design challenges is designing a product that contains multiple independent "wireless" transc eivers that are on the same printed circuit board, are in extremely close proximity on the pc board, and operate in nearby frequency bands. In this particular case, there is a Bluetooth wireless transceiver and a 3G (third generation) cellular transceiver supporting 3GPP (3rd Generation Partnership Project) and 3GPP2 (Univeral Mobile Telecommunications System [UMTS] and CDMA2000 respectively) standards. The challenge is to integrate these two complex transceiver systems into one product.

Several questions will spring to a designer's mind. How can these technologies coexist? What are the key parameters and specifications that define an optimized design? Which RF parameters should be measured, and how should they be measured?

First and foremost, the design, test and integration process should be modularized, with the components tested and integrated as they're developed. A careful examination should be made of applicable standards and specifications in terms of each design component for both the Bluetooth and 3G transceivers. Currently, the frequency spectrum to be used for 3G cellular implementations differs from country to country. In addition, the complicated intellectual-property rights that exist for the fundamental design of these technologies necessitates checking with the appropriate international standards bodies charged with specifying how to design and test these technologies.

The Bluetooth RF chip must reject the strong out-of-band signal from the cellular transmit signal originating from the same appliance. In addition, the electronics that are used to integrate the two different RF components must be immune to both environments.

Design considerations:

Antenna parameters such as polarity orientation, resonance bandwidth, placement, antenna filter.
Link budget of each technology based on the network configuration and specific application.
Optimized design tradeoffs such as sensitivity vs. intermodulation response attenuation or power amplifier co st vs. efficiency.

The CAD analysis process can be very helpful here, leveraging current designs and then examining the impact of the addition of the second transceiver. Designers can address such issues as proprietary manufacturing processes, nonlinear effects and the ability to perform parametric adjustments to other design-specific modeling variables with CAD analysis, and it's also a helpful tool when examining the specific layout issues electronically.

However, CAD analysis loses its advantage when investigating physical devices. At this point, designers should have taken the necessary steps to ensure that the expensive silicon fabrication process risks have been minimized as much as possible.

With the technological advances that are occurring in fabrication technologies, there may come a day when more than one transceiver can be integrated onto a single SoC. Using CMOS technologies for RF circuits is still relatively new. Fabrication processes, formerly used only for digital appl ications, need to be re-evaluated for use with RF applications.

With Bluetooth technology still in the early stages of rollout, designers will have to grapple with the kinks of working with a new technology when developing products for quite some time. However, there are processes that engineers can use to ensure product development and integrating multiple transceivers into a single platform is as painless as possible.

Analyze all Bluetooth connection needs vis-à-vis data rates, range and interoperability.
Review the effects of application software on power requirements. With many aspects of connection management hinging on the application software, it is wise to take stock of the various modes because of the major effect they will have on average current.
Assess the options that exist with respect to chips, modules and outside design resources to avoid becoming overwhelmed by complex RF circuit design, packaging problems or software development. Take advantage of the services and systems-level expertise of third-party houses to provide applications assistance and development tools. Finally, design engineers should keep a vigilant eye on evolving developments within the Bluetooth and 3GPP specifications as myriad new products and applications burst on the scene.