Design ideas on a DC to 5GHz CMOS low noise amplifier

Hi I need to design a DC to 5GHz LNA using 0.18um CMOS. I know distributed amplifier can provide wideband response up to 10GHz. But the powre consumption, gain and noise performance are not that good. So I am thinking of other structure like feedback.. Can any one provide any suggestion on the possible topologies for this wideband LNA? Related references or books are welcome.

Are you sure you want it from DC to 5GHz? An amp from 2000meters up to 6cM wavelength? hAVE A THINK ABOUT WHAT YOU REALY NEED

Not necessarily 0Hz to 5GHz. 100MHz to 5GHz is fine. I guess

Even 100Mhz to 5GHZ will not work....... you REALY need to think about what you want to do, you will need about 6 or more individual circuits to cover this much of the spectrum......... 100MHZ is simple, but as you go up into UHF and microwave, it get more tricky...... inductors are etched onto the board for one thing, and the type of board you can use will change as you go up in frequency.............. Give us a better idea of what you want please.......

Zencir, E.; Tekin, A.; Dogan, N.S.; Arvas, E.;"A low-power DC-7-GHz SOI CMOS distributed amplifier", Circuits and Systems, 2004. ISCAS '04. Proceedings of the 2004 International Symposium on Volume 1, 23-26 May 2004 Page(s):I - 605-8 Vol.1

Ren-Chieh Liu; Huei Wang;"DC-to-15- and DC-to-30-GHz CMOS distributed transimpedance amplifiers",Radio Frequency Integrated Circuits (RFIC) Symposium, 2004. Digest of Papers. 2004 IEEE 6-8 June 2004 Page(s):535 - 538

rf_ray said:

Zencir, E.; Tekin, A.; Dogan, N.S.; Arvas, E.;"A low-power DC-7-GHz SOI CMOS distributed amplifier", Circuits and Systems, 2004. ISCAS '04. Proceedings of the 2004 International Symposium on Volume 1, 23-26 May 2004 Page(s):I - 605-8 Vol.1

Ren-Chieh Liu; Huei Wang;"DC-to-15- and DC-to-30-GHz CMOS distributed transimpedance amplifiers",Radio Frequency Integrated Circuits (RFIC) Symposium, 2004. Digest of Papers. 2004 IEEE 6-8 June 2004 Page(s):535 - 538

Click to expand...

Are you saying you want the books? If so I cant help you, I dont have them..

For feedback topology, that means large fT. I don't believe right now, there is so high CMOS tech can do that high. Maybe HBT can work. It depends ur project is for product or just for paper. Purpose is different, the requirement is different and the difficulty is different

rf_ray said:

Not necessarily 0Hz to 5GHz. 100MHz to 5GHz is fine. I guess

Click to expand...

RF_ray, do a simple search on the google for LNA or PA amplifier in the frequency range requested. Even there are large band PA (running from 500MHz to 4GHz) the circuit must be adapted to a small frequency range with a bandwith of maximum 400-500MHz (ie 2.3GhZ-2.7GHz).
There are a few LNA's which are running from DC to 2Ghz with the same requirement. So, please detail exactly what do you want (the application and not the amplifier parameters).

greetings,

I appreciate all the comments you guys provided. But seems my question was posted on the wrong board( I posted this question on analog IC board but got deleted right way, which I think doesn't make any sense).

It seems to me that people who did the hybrid microwave/RF design are not quite familar with the different design concept in analog IC design.

Yes, the CMOS technology today does have a very high ft. For 0.18um CMOS, the ft of NMOS can be 60GHz.

Yes there does exist such LNA that can work over wide frequency range. Just check recently publications on CMOS realization of UWB front end, which covers 3.1GHz to 10.6GHz. In RF IC design, LNA is different from PA !!!. In a typical receiver , LNA is used to provide 50 ohm input matching and connected with a capacitive load(usually the mixer). It conduct the voltage amplification instead of power amplifier.

Now back to my question. I ended up designing the DC -4GHz LNA with two basic structures. Distributed structure and a resistive feedback structure. The feedback structure it alone usually not able to make the bandwidth wide enough. A simple inductor peaking technology is used to further extend the bandwidth. The simulated gain is 12dB-10dB acrossing the whole DC-4GHz. Input return loss is from 17dB to 7dB. Obtained noise figure is below 3dB.

I think your design is a Transimpedence Amplifier whick can be used in tranceiver of Optical Electronics, disk reader and writer and so on.
not a typical RF LNA!!

and there is many paper in this area, 0-10GHz TA is published in 0.18CMOS.
and I think it is not difficult to find them out

wrong..I didn't convert current signal to voltage signal. The input of my LNA is directly connected to antenna. The output of LNA is connected to a sampling mixer. It IS A TYPICAL RF LNA!!!

rf_ray said:

Now back to my question. I ended up designing the DC -4GHz LNA with two basic structures. Distributed structure and a resistive feedback structure. The feedback structure it alone usually not able to make the bandwidth wide enough. A simple inductor peaking technology is used to further extend the bandwidth. The simulated gain is 12dB-10dB acrossing the whole DC-4GHz. Input return loss is from 17dB to 7dB. Obtained noise figure is below 3dB.

Click to expand...

Well, if the simulation works, then is time for a practical approach. I've guess that at least once you have measured in the real word the excelent simulation, right ?
What it means "12dB-10dB" ?

12dB-10dB means the the gain drop from 12dB at DC to 10dB at 4GHz.

Yes, I have several active circuits tape out and conducted the measurement. Using the design kit, usually the gain will drop around 2dB and noise will rise no more than 2dB comparing with simulation results. The return loss will pretty much the same as simulation, sometimes better than the simulation results.