Amplifier peaking question?

Given a source (output voltage and impedance), with a stray capacitance to ground that is to drive the input of an amplifier with a stray capacitance
to ground, what is the coupling network that gives the largest gain-bandwidth product?
I am familiar with shunt and series peaking. I also know how to analyze a given four-terminal coupling network. Somewhere in the back of my mind, I remember that there was a theoretical maximum bandwidth that could be achieved. I do not, however, remember how to obtain such a theoretical result.
In principle, series peaking gives a bigger bandwidth than shunt peaking because it separates the two stray capacitances. The question can be asked as what is a coupling configuration that can give significant improvement over series peaking? How do you approach such a problem?
Bill
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
I see a lot of references to CFAs, (current-feedback amplifiers) when I googled your question.
This tutorial talks abot the GBWP: http://www.web-ee.com/primers/files/DesignSem6.pdf
This is a TI app note on CFA analysis: http://www.ee.nmt.edu/~thomas/data_sheets/op-amp-cf-sloa080.pdf They suggest a DC coupling for current feedback topology as the GBWP is affected by the feedback resistor.
And here is something with a section on bandwidth: http://www.e-basteln.de/index_e.htm
Let me know if any of this helps.
jason

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
in article snipped-for-privacy@corp.supernews.com, jason at snipped-for-privacy@carrollsweb.com wrote on 5/23/04 12:10 AM:

I should have mentioned that I did Google the topic.
I guess my question is really a curiosity question on circuit theory. What kind of a coupling network will give best performance when stray capacitance is present at the start?
With the cost of active circuitry being cheaper now than passive circuitry, feedback is an economic alternative to clever circuit design. I am not that familiar with modern amplifiers.
My guess is that my answer is available somewhere in the Bode theory of networks (developed for feedback amplifiers). He has some general equations regarding transfer functions. These are the equivalent of Kramers-Krnig relations for optics. I do not know how to apply them.
So, I repeat my question in a slightly different form: Given stray capacitances, what can the bandwidth be from the best available passive coupling network between two stages?
Bill
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.