The design of a new radio telescope known as the Square Kilometer Array (SKA) is well under way in the international astronomical community. lt is expected that the telescope will require a large number of antenna elements, significantly impacting the overall tele­scope cost. This thesis presents a feasibility study to employ relatively inexpensive and rapidly developing CMOS technologies, rather than exotic GaAs and lnP semiconductor technologies, to implement the low noise amplifiers (LNAs) attached to each antenna ele­ment.

To demonstrate the feasibility of using a wide-band CMOS LNA in the SKA telescope, a number of steps were taken in this work:​ LNA optimization procedures based on different design constraints were created, a wide-band LNA topology was introduced, and issues associated with the integration of SKA antenna elements with the LNAs were investigated.

The new LNA optimization technique improved the widely accepted narrow-band cas­code optimization by incorporating the noises due to losses in the gate inductor and the gate finger resistance. This optimization was further improved by using a two-port approach to model the noise parameters and electrical parameters of the LNA.

The new wide-band topology was described and experimentally verified. Two designs of the LNAs based on the wide-band topology are presented. A 90-nm CMOS 50 n design nearly meets all the specifications required by the SKA and achieves an excellent wide­band power match and the lowest noise figure published for CMOS LNAs prior to this work. A differential version of this LNA was created to assess the feasibility of using a differential LNA in the SKA telescope.

The feasibility of allowing for a non-standard impedance of the custom made antenna elements is also discussed, culminating in the design of a 90-nm CMOS LNA achieving sub-0.2 dB noise figure in the 800 MHz to 1400 MHz frequency range and 511 <​ -11 dB while consuming 43 mW of power. This LNA meets all requirements of the SKA telescope and confirms that newer CMOS technologies are achieving noise figures low enough so that even sensitive radio telescopes can benefit from the low cost, high degree of integration offered by these technologies.