LNA design help

Thread Starter

RFBeginner

Joined Mar 16, 2007
9
Hi,
I am interested in learning about LNA's and I am trying to design an LNA according to the following specs:

NF < 4dB
IIP3 >5dBm
Gain >20dB
Input impedance 40-60 Ohm
Output load 500fF series with 500 Ohm
-3dB Bandwidth >30Mhz

I am trying to do some hand calculations with equations in Tom Lee's book(the design of CMOS radio frequency integrated circuits).

Any help on where to start and how to do this?

Thanks,
RFBeginner.
 

Papabravo

Joined Feb 24, 2006
21,157
You did not mention where in the RF spectrum you would like the 30 MHz, 3 db bandwidth. I would start with devices that have appropriate s-parameter characterization in the frequency range of interest. Then I would select a configuration. Then I would suggest that you understand the concept of conjugate matching. That should get you started.
I've found the book by Gonzalez helpful in these matters.
http://www.powells.com/cgi-bin/biblio?inkey=65-0132543354-1

Be prepared for some advanced mathematics.
 

Thread Starter

RFBeginner

Joined Mar 16, 2007
9
Sorry, I forgot to mention, it should work at 2.45Ghz.

About the s-parameters, will they be available in a data sheet or will I have to run simulations to get that.

Also, I have decidede to use the cascoded source degeneration common source amplifier (for better gain)

From what I understand, I will need to have a matching network at both the input and output (to match 50 Ohms and 500fF+500 Ohms respectively.)

I understand my first step should be to find a biasing point, how do I pick the best point for minimum power?

Thanks.

(I have found the book in my school library and will look through it)
 

Papabravo

Joined Feb 24, 2006
21,157
Datasheets vary in the amount of information they require. Simulation can get you part of the way there, but there is no substitute for characterizing the device in your particular layout.

With respect to power consumption. Are you are asking how to minimize the power consumed in the absence of a signal? That is the DC power consumption.
 

Thread Starter

RFBeginner

Joined Mar 16, 2007
9
I do not have access to a simulation tool right now. I'm confused as to what u mean by "characterize the device in my particular layout". I do not have any values for components yet because I cam not able to use the equations in Tom Lee's book until I have selected a bias point. (this is what is giving me the most trouble)
 

beenthere

Joined Apr 20, 2004
15,819
Methinks that biasing is pretty trivial compared to things like trace layout. Will your CAD program handle microstriplines? I'd hate to do all that by hand calculations. Seriously, just producing a fixture that can work at those frequencies makes biasing for max efficiency seem like child's play.
 

DrNick

Joined Dec 13, 2006
110
Do you have access to a network analyzer or noise figure measurment device? Without these you wont be able to test to see if your LNA works.

The following topics will get you on the road to understanding how to make an LNA:

Matching networks
Noise Figure Circles
Gain Circles
Stability Circles

Once you understand how to create (or simulate using cad software like ADS) you can start designing your LNA.
 

Papabravo

Joined Feb 24, 2006
21,157
I do not have access to a simulation tool right now. I'm confused as to what u mean by "characterize the device in my particular layout". I do not have any values for components yet because I cam not able to use the equations in Tom Lee's book until I have selected a bias point. (this is what is giving me the most trouble)
"characterize the device in my particular layout". means to confirm the information on the data sheet by building a fixture to measure the s-paramaters over the frequency range of interest. The datasheet is only a starting point and the final results will be highly layout dependent at these frequencies. Without appropriate CAD tools and test equipment your chances of success rapidly approach a very low order of probability. Sorry, but this stuff is not for the faint of heart or those with an insufficiently fat wallet.
 
Top