One-chip Bluetooth ASIC challenges

Summary form only given. This paper describes the constraints, trade-offs and challenges for designing a one-chip (RF/radio plus baseband integrated on the same Si-die) compared to designing a separate RF/radio ASIC. The presentation will start with an overview of the background of Bluetooth technology: the original purpose was to replace wires. Consequently, the RF-proposal has to be cheap (target below U5) and the current consumption should be low. The desire to have a cheap radio, cannot be implemented at all cost: the user will require a robust system (performance), otherwise the Bluetooth system will fail in the market. Nonetheless, a "simple, cheap radio" is still a complex analog system, and requires RF and analog design resources. An overview of "easy" and "difficult" Bluetooth radio specifications will be given to show the problem areas in radio design. A single-chip radio module solution based on a BiCMOS radio-ASIC will be shown to demonstrate some particulars in integrating Bluetooth radio functionality. From this single-chip radio module, additional functionality will be described for realising a one-chip Bluetooth solution: requirements on digital functionality, requirements coming from SW implemented in ROM or FLASH and additional radio requirements. Next, different Si-processes (CMOS 0.36u, 0.25u, 0.18u and 0.13u, BiCMOS and SiGe) will be discussed w.r.t. our target. Due to the large amount of digital, a CMOS process is the only possibility. CMOS scaling for the digital functionality will be discussed. The scaling of radio functionality for the most critical blocks will also be presented. This insight is needed to judge whether a future CMOS processes still allows the radio to be integrated and if we gain something from better CMOS processes.