For the circuit, as the linearity is high, the gain is usually low and vice versa. However, why will this happen? How could I explain in transistor view? Actually, I just know this trend only.
it is a simple tradeoff principle. linearity and gain are inversely proportional to each other. i will try to explain it in the context of a transistor. if u have a high gain the power levels will be high. that means that the temperature levels will also be high. so the output will not be linear. and for a perfect linear behaviour u have to lower the gain.
for the mathematical equations involved refer to a good book of semiconductors
I have designed a LNA long time ago. it's interesting that linearity, gain and NF are opposite to each other. I don't quite understand but it's about the biasing of the transistor. To make the amplifier more linear, I increase current to bias it to a linear region but at the same time the NF is poor because of increase of thermal noise.
I agree. If you want better linearity, you increase transistor current. Suppose we double the current. gm is proportial to square(Ids), ro inversely proportional to Ids. So the intrinsic gain of a single stage or gm*ro is inversely proportional to square(Ids). So doubled current means lower intrinsic gain!
The MAG of a device is a function of IDS. It looks like a bell so the Max of the MAG is somewher close to VGS=-0.2V (different foe different FABs).
As IDS is higher the wave swing can be higher ant it means more linear. But as IDS goes up the MAG goes down from the top point.