common gate lna circuit
Common gate (or its equivalent common base ) amplifier are really not low noise. It is barely possible to get better then 5dB NF. The reason is that that for Fmin , Gamma Opt and S11* are too far apart. A simulation with the correct Data will show that Fmin << NF becausee of the matching needs. A better choice in my opinion is a cascode with the first Fet with source and gate degeneration or a derivative of this. A NF of about 1 dB is possible using a Chebychev polynomial derived wideband matching filter, if possible external to the chip. Recent IEEE papers of the MTT Journal give more insight.
While the above mentioned grounded gate application is a very nice treatment, it does not provide the lowest possible NF.