Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Common Source versus Common Gate LNA

Status
Not open for further replies.

elbadry

Full Member level 6
Full Member level 6
Joined
May 6, 2005
Messages
347
Helped
72
Reputation
144
Reaction score
10
Trophy points
1,298
Location
Egypt
Activity points
3,884
common gate lna

Dear All,

I have a question regarding LNA's. It is often siad in literature that the inductively degenerated common source LNA is the best in terms of Noise figure...

However, When I tried to design both ( in 0.13 um tech. ), I found that for almost the same current the common gate LNA has a lower NF ( 2.7 versus 3.6 for common source ).

Is rhis normal? Or have I mal-deisgned the Common source lna ( actually there weren't much degrees of freedom in its design ) !!

Thanks
 

common source lna

The CG LNA gives more better NF as far as I know...U can refer to the LNA Design in the text book given below

Behzad Razavi, RF Microelectronics, Prentice Hall PTR.

Infact most of the recent designs use CG LNA where source degeneration is a costly effort with the inductors included, which occupy large chip area. You can also refer this paper for a recent design..

Kwang-Jin Koh, Mun Yang Park, Cheon-soo Kim,Hyun-Kyu Yu, “Sub harmonically pumped CMOS frequency Conversion (up and down) circuits for 2-GHz WCDMA direct-conversion transceiver”, IEEE J. solid state circuits, vol.39,No. 6,June 2004
 

lna common gate

I think the paper you sent implements a mixer not an LNA. Could you please clarify where you got the conclusion that common gate has lower NF than common source?

Thanks for helping:)
 

cg lna

The paper I gave is though mainly about the mixer, it contains the transcondictance amplifier stage whose explanation is given.

What makes me belive that CG input stage's advantage over the CS stage is CG stage dominating low input impedance, which is not the case for CS stage. For Noise Figure input impdance lineartly over the bandwith of concern is also important. CS input stage has a very high impedance at the input of transistor and whose input impedance matching is dependent on the tuning network and gate capacitance.

With my small experience I can guess only upto that point. If you can find anything interesting please let me know.
 

gm boosted lna

What I saw is that the NF of CG is truly much more flat over frequency than that of Common source.

The reason why the common gate is not expected to have a high NF is the limitation on the input impedance. For matching it is a must that gm*Rs = 1 ( where Rs is the source impedance ).

Hence, by simple analysis you will find that F = 1 + gamma / alpha

where gamma = the noise fitting factor and alpha = gm/gd0. So, expectedly F will be high.

This is according to the following paper:
"Gm-boosted common-gate LNA and differential colpitts VCO/QVCO in 0.18um CMOS"
Xiaoyong Li; Shekhar, S.; Allstot, D.J., JSSC December 2005 ( vol 40, no 12 ,pp 2609-2619)

Thanks..
 

common gate input match

The CS-LNA with source degeneration can match the input impedance toward the source impedance as well as match the NF to NFmin. Moreover, the real part of the impedance is realized by the reactance components, namely Ls and Cgs, which do not contribute noise as 1/gm in CG configuration. But for wideband application, CG may be a better choice. I guess you need to take into accounts both S11 and NF at the same time to justify your design.
 

common gate mixer

in theory, CS lna has better performance than CG lna. There is a technique called simultaneous impedance matching and noise matching technique in CS lna design. In practice, you can tune the Re(Sopt)(Sopt is the best noise input S parameter) to 50Ω(Sopt can be see in ADS). When you match the input impedance of lna to be 50Ω, the best noise impedance Zopt is also be tuned to 50Ω. Then the noise would be lowest. Maybe the higher noise figure of CS lna is partially due to this reason. Besides, in a design of CS lna, we need Ls to set Re(Zin) to 50Ω and Lg to eliminate the imagine part of Zin. If we use a spiral inductor which has a low Q value, the parasitic reistor of spiral inductor would introduce noise, especially in the case of Lg. So implement Ls and Lg with bondwire if Ls and Lg are small enough. Then compare the noise performance of CS lna and CG lna and tell me the result. i am interesing in CG lna, for CS lna also has its shortcoming.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top