70 db gain @ 1 Ghz opamp

[sonaiko]

Guys,

I need to design an opamp that takes a 1 Ghz AC input and amplify it with gain larger than 70 dB. The BW required is larger than 15 MHz.

Please tell me if these specs are feasible or not in 0.18 um Cmos as i have seen many papers having 70 db as dc gain but only unity gain at around 1 Ghz.

I would appreciare if you refer me to helping papers/books/threads..

Thank you.

 

[FvM]

It's neither a job for an opamp nor for a single feedback amplifier.

I fear, you'll face difficulties to implement 70 dB gain @ 1 GHz on a chip. But the RF IC guys should tell.

 

[rongo024]

CMOS: Circuit Design, Layout, and Simulation, Chp 24 , 26

 

[pavel_adameyko]

Hai
For what application did you circuit intended for? I think several series LNA's can achieve 70 dB gain at 1 GHz.

 

[sonaiko]

So it seems this is a job for LNA?

What is the difference between LNA and opamp? Arent they both supposed to amplify?

Would you help me in referring to similar researches and papers/books on this?

 

[pavel_adameyko]

So it seems this is a job for LNA?

What is the difference between LNA and opamp? Arent they both supposed to amplify?

Would you help me in referring to similar researches and papers/books on this?

yes - they both amplify the signal.
Take any book on RF microelectronics - you' ll find answers for all your questions.

 

[leo_o2]

For LNA, single stage is usually used. It also employs inductor as load.
General-purpose opamp often use differential input stage + some gain stages.

 

[sonaiko]

Why LNA is usually designed with a single input?

Also, I could not find a paper or a book for LNA with +70dB gain at 1 GHz..!

 

[RCinFLA]

70 db of gain at 1 GHz on a single CMOS chip?

How you plan to get on and off the chip without feedback coupling causing it to oscillate?

Good luck.

 

[FvM]

I need to design an opamp that takes a 1 Ghz AC input and amplify it with gain larger than 70 dB.

I wonder if this is a project assignment? If so, how much ignorance is required to release it to a student? I suggest to rephrase it: "Show, why it's not feasible to design..."

 

[sonaiko]

What do you mean it will oscillate? Ofcourse this is on a single chip..

What is the maximum gain currently achieved on 0.18um LNA?
Im expecting my wireless input to be at most few mV of amplitude at 1 GHz.

 

[FvM]

What do you mean it will oscillate? Ofcourse this is on a single chip..

I don't think, that the term needs further explanation. You can calculate, that with 70 dB gain a few pH common ground inductance for in- and output will be sufficient to fulfill the oscillation condition.

All contributors up to now assume that 70 db/1Ghz isn't feasible on a single chip. But you have been reviewing recent literature. What's the maximum gain value you could find?

P.S.: Availailable products seem to achieve 30 to about 40 db gain.

 

[sonaiko]

Actually it was 20 dB when using active inductors instead of on-chip coils.
But how am I supposed to understand my received signal if its amplitude is a few hundreds of micro volts only with an LNA of 20 db gain?

 

[FvM]

But how am I supposed to understand my received signal if its amplitude is a few hundreds of micro volts only with an LNA of 20 db gain?

I understand, that you are asking how to achieve the required total gain for a sensitive receiver?
The problem exist only for the GHz carrier frequency. A sensitive receiver will mix the signal down to a reasonable IF frequency after applying some gain.

 

[davenn]

100's of uV is a very strong receive signal ! 200uV would overpower most FM IF detector chips... say a MC3361. It would have to be attenuated before going into the chip. a ballpark figure (without refreshing memory from a datasheet) would be ~ 50uV

most decent receivers will have sensitivities down to 0.5 - 0.2uV @ 12dB SINAD
What everyone above is trying to tell you that it wil be virtually impossible to achieve 70dB of gain on a singel chip without it becoming very unstable and breaking into oscillation. ( to use an old phrase ... it would be more sensitive than a tooth ache )
Normally hi gain is spread across several stages each stage with a max of may be ~ 20dB.

cheers
Dave

---------- Post added at 13:10 ---------- Previous post was at 12:49 ----------

Altho there are one or two very new 1GHz op-amps out there like the ADA4817-1
they are run at very low gain, from unity to ~ G=5

Dave

 

[sonaiko]

Oh i understand now, FmV!
So on the same Gain-bandwidth product of the amp, higher gain is ahcieved when the carrier is down converted before the gain stage, say to 100 MHz. This way the gain is increased 10 times for the same amp design (theoritically). Plz correct me if i am wrong.

Davenn,
The threshold for thr digital circuit is around 0.7 volts. And having 20 dB gain only for a 0.2uV carrier is not enough to operate the digital circuit. Would you explain how The process goes?

 

[Zanderist]

I just want to make a comment on the frequency; Is this for wifi or video?

I can tell you right now that a 1.4-2.7 Ghz ( at the moment I am not sure of the precise frequency) signal from a home security camera is enough to jam your laptop's wifi, or more specific the signal from your router.

So in your design be sure to test it around not just the circuit you want to operate but other circuits you have lying around.