Inductive Peaking:Increases BW but whats the cost?

I have a question regarding inductive peaking circuits.
Many variations of inductive peaking such has series, shunt and combination of the two have been used widely to increase the bandwith of amplifiers, but what is the price we pay in return....ofcourse chip area is one but what about delay of the signal from input to output ?

Lets say that a combination of series and shunt peaking is used in a differential amplifier, and 4 stages of such amplifiers are cascaded to form a ring oscillator...Will the oscillation frequency of the ring oscillator increase or decrease?

In my opinion the Oscillation frequency should decrease as the delay is being increased by inductors, however the BW is increased? Can anybody confirm/elaborate on this?

Regards

i have a question how the BW is increased and the delay is increased too

You can read Chap 8 of T.H. Lee's RFIC book for bandwidth extension techniques.

However, i am concerned about the effect of such techniques if they are used in designing Ring Oscillators....where delay is an important factor...the book does not mention about this aspect?

Any explanations to my first post ?

Regards

hello again,
what i meant is that as far as i know increasing the BW and increasing the delay are two contradecting issues (i.e. they cannot be existing simultaniously)

if u want another price beside area is the headroom consumption as i think it cannot be done on small supplies

Added after 1 minutes:

sorry but could u provide a link or something for the book that u refered to.
thanks alot

The book i was referring to is a 'must have' for RF designers...the link is below:




Regards

You can get small amounts of bandwidth increase but the cost is extra hardware. If you try for more bandwidth increase you get time domain overshoot.

It is important to select propper degre of compensation. There exists a degree of compensation where group delay and total delay is allmost constant up to the end of passband and is lower. It looks like realization of Gaussian filter. This degree of compensation does not give max increase of bandwidth but a little bit les and does not increase delay.

the cost is it will bring more overshot of frequency response and worsen linearity of group delay

Another cost is the performance of the transient. When you do inductive peaking, you have peaking in freqency response, that means your transient will has overshoot, ringing.

Basically you may use two criterias to design peaking. One is maximal bandwith and other is flat group delay. First method is resulting in non-constant group delay and thus causes pulse ringing. Second method gives less bandwidth but pulse shape is clean. More realistic is to compromise between those two methods.