Question on differential cmos LNA

[lsfeng]

hi ,all,
Recently, i have investigated many kinds of differential LNA,and sorted them in two
kinds,with the biasing isource to connect to the ground ,and without it,
so i want to know ,which kinds is better?
my design target is wideband 3-5G, gain :15db,mathing :10db

thank you in advance!

 

[streamlet]

You classify the differential LNAs according to whether or not their bias I sources are connected to the gnd. I have not heard of this kind of classification before. The bias is just bias, not a key component in LNA. Why you want to do like that?

 

[MK28]

If you are restricted by current consumption then you have to consider the bias condition !

there is a few things you have to remember while designing UWB LNA

1. use inductive degenrate topoly, advantage of this topology is it is possible to generate the realy parf of the LNA input impedance is 50 ohm, and this is not a physical resistor , so do not have to worry about noise !

2. One major thing is, you cant use the traditional complex conjugate matching in wide band design, cuz it is not possible! Complex conjugate matching theory only holds for narrow band system. So while doing mathcing at the input, you have to have mismatch at the input so that the GAIN is flat over the 2 GHz BW, and also S11 will be below 10dB or so !

Remember, your IP3 dependent on Vbias ( bias condition), so check out whats your current consumption in the design. Also differential pair will consume more current than single ended LNA.

Hope this will be helpful !!