الفهرس 
Introduction and MotivationOnly 14 pages are availabe for public view
 |  | from | 92 |  |  |
| | | from | 92 | | |
Abstract
The design of low-noise-amplifiers (LNAs) entails a large number of design challenges and tradeoffs, which include sustaining a good input matching along with the optimization of various LNA performance metrics such as gain, bandwidth, noise-figure, power and linearity. The optimization of the LNA block is a key factor for the optimization of the whole RF-chain as it limits the signal-to-noise ratio (SNR) and the sensitivity of the receiver chain.
Reported in literature are various design methodologies and approaches that are tailored for the optimization of RF LNAs. These methodologies use different design techniques to standardize the design process. One of the most popular techniques in the literature is the gm-over-id method where all the transistor parameters are plotted vs gm/id or stored in the form of a look-up-table (LUT). The key LNA performance metrics are then formulated in terms of gm/id seeking an optimum gm/id for each transistor in the circuit where the optimum sizing and biasing point can be looked up from the already made LUTs. A much similar technique is also adopted in some works in the literature where the transistor gate-source voltage, VGS and drain-source voltage, VDS are used as a bias point indicator instead of gm/id. Another evolving design technique is based on the well-established charge-based EKV model that was first introduced by
C. Enz, F. Krummenacher and E.A. Vittoz in the mid-90s where the transistor biasing point is expressed in terms of a parameter called the inversion coefficient (IC) which indicates the level of channel inversion and separates the transistor operation regions
xi
into three main regions which are the weak inversion, moderate inversion and strong inversion region. The main advantage that the EKV model offers is that it can be used to formulate accurate analytical expressions for the MOSFET I-V characteristics that are valid acrross the three regions of operation. This is opposing to the square-law that can only be used in the strong inversion region and the exponential expression that can only be used in the weak inversion region. In the past, most of the wireless standards were using sub-ten gigahertz operating frequencies where the available technologies transition frequencies (ft) were comparable ruling out the utilization of the weak inversion or even the moderate inversion in the design of high speed RF circuits. This situation no longer holds due to the huge advancement in the fabrication technology offering ft that are multiples of hundreds of giga-hertz while the wireless standards are still operating at relatively low operating frequencies facilitating the use of the transistor moderate and weak inversion regions to seek better power efficiency.
The ultra-wide band (UWB) standard is an evolving wireless standard that has recently find its path to various applications ranging from indoor positioning and localization passing by the wireless sensor networks (WSNs) and reaching to the RFIDs and car keys.
This thesis presents a design approach for UWB LNAs based on the EKV model and the IC. The proposed approach can also be used as a basis for Narrow Band (NB) LNAs design with slight modifications. One of the key challenges in the design of UWB LNAs is the wide-band input matching. A complete systematic solution is presented for the problem of UWB input matching with a high degree of analytical accuracy. The design approach is illustrated through the design of two UWB stacked Common-gate LNAs in 65 nm technology. The first LNA achieves S11 better than -8.2 dB over a 27.6 GHz frequency range, a gain of 12.4 dB over 16.5 GHz bandwidth, minimum NF of 4.5 dB, and IIP3 of -5.2 dBm while consuming only 530 µW. The second LNA achieves S11 better than -15 dB over a 8.8 GHz frequency range, a gain of 12.5 dB over 6.8 GHz bandwidth, minimum NF of 4 dB, and IIP3 of -4.3 dBm while consuming only 550 µW.