Help me achieve value close to NFmin in LNA circuit

[spbhu]

Hi, I have built my LNA circuit, the transistor I used is 40/0.18um, the biasing current is 0.5mA. And Vgs=535mV, from the simulation result, I found nf(2)=~3.4~3.6dB, while NFmin is about 2~2.5dB. The resonant frequency is 2.5GHz. Can I adjust the circuit to make nf(2) approach NFmin? How to change the circuit, can anyone help me?
Another problem is that, since the low current consumption requirement, the biasing gate voltage is quite low, and Vod<100mv, does it matter? Since it is already in moderate region? What could be the problem for low Vod?
Many thanks.

Added after 2 hours 6 minutes:

Some additions:
The topology I used is CS inductive degeneration. The result from my simulation shown that, the resonant of NFmin at 0.8GHz, while the nf(2) of gain and s11 resonate at 2.5GHz, what is the reason? Does it mean nf(2) consider the parasitic resistance of the inductor Lg and Ls, while NFmin donot? Thanks

 

[khouly]

Re: NFmin problem!

the NFMIN can be reached if u match the input to Zopt , u can measure Zopt , and match this input impednace to it so u get NFMIN

but u can use the degeneration inductor used increase the input impedance real part to be close to Zopt , and use the Lg to eleimnate or resonate with the complex part

by this u caould achive value close to NFMIN

khouly

 

[spbhu]

Re: NFmin problem!

Thanks for you answer. Can you explain to me what does Zopt mean? My S11 is about -12dB, so that means the input matching is quite ok for practical design, do you mean I still have to improve the input matching? I have done that, but I found even when S11 is -20dB. The NFmin still resonate at 0.3GHz~0.4GHz. While S11 and S21 resonate at 2.45GHz, which is the desired operating frequency. What could be the problem? Thanks

 

[khouly]

Re: NFmin problem!

Zopt mean the optimum resistance for minimum noise this resistance sould be seen by the LNA

khouly

 

[haitaoz]

Re: NFmin problem!

In most cases,there is a trade-off between noise and maximum gain.The matching point really depends on your applications,biasing point.However,you can not do too much for the minimum noise for the device if it is already chosen.