The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> S21 of class B amplifier
https://designers-guide.org/forum/YaBB.pl?num=1301742036

Message started by purplewolf on Apr 2^nd, 2011, 4:00am

Title: S21 of class B amplifier
Post by purplewolf on Apr 2^nd, 2011, 4:00am

Hello,
Is the claim that the S21 of amplifier in class-B is zero since there is no quiesent current keeping in mind s parameters are small signal. In my case, i am getting S21 non zero and i have determined the threshold voltage through id-Vgs plot. The current waveform is half cycle on while off.. Does it mean that my amplifier is not class-B but infact AB?? How to make sure a perfect class B operation

Title: Re: S21 of class B amplifier
Post by grumpy_bear on Apr 7^th, 2011, 2:58am

Hello,
could you plot the in and output waveform in transient operation? Your description is really unclear to me.
What is your design goal and what information do you want to retrieve with your S-Parameter simulation exactly?

Cheers

Title: Re: S21 of class B amplifier
Post by purplewolf on Apr 7^th, 2011, 4:39am

my wavefroms are exactly like class-B..i dont have them at the moment..
I am looking into the circuit from the input. The claim that S21 of class-B amplifier is zero is made by one of my colleagues, the reason being that there is no quiesant current in the device as it is biased at threshold keeping in mind that sparamter is a small signal parameter..according to him, for class-AB and Class-A you should have S21 non-zero but for class-B it should be zero...

Title: Re: S21 of class B amplifier
Post by Vladislav D on Apr 19^th, 2011, 1:34am

Hi,
You cannot estimate gain using S21 because this is small signal parameter. Small-signal parameters are not valid with PAs because PAs work with large signals. You should estimate power delivered to the load using time-domain or Harmonic Balance simulation. Practically, simulator can show some small-signal gain because small current may flow through a transistor. In fact, it is very easy to shift amplifier to class AB or class C operation. The best way to understand what biasing is to build a load line. For a class B it should be like in a figure I attached (VDD=20)

Title: Re: S21 of class B amplifier
Post by purplewolf on Apr 20^th, 2011, 6:02am

yes..i design PA using harmonic balance not with s parameter..but i just wanted to validate or invalidate the claim that a class-B amplifier has S21 equal to zero..

Title: Re: S21 of class B amplifier
Post by grumpy_bear on May 2^nd, 2011, 1:58am

It does not make sense at all to estimate anything in a class B amplifier with small signal S-parameter simulation.

First, you have to define what you exactly mean by "zero". Do you mean 0dB (which is 1) or "nothing", or very small.
Actually, you will not have an ideal switch or transistor, where quiescent current is zero at your biasing point. There is a "threshold voltage" , which actually is a transition from weak inversion with exponential behaviour to normal operation with quadratic characteristic of a transistor. There is no point, where the current is really zero, only very small with respect to "on" condition. Second, there will always be coupling of the voltage through parasitic gate-source capacitance, so the signal also won't be "zero".

Title: Re: S21 of class B amplifier
Post by purplewolf on May 21^st, 2011, 1:38pm

In the document (http://scholar.lib.vt.edu/theses/available/etd-07152001-172453/unrestricted/chap4.PDF), page 2 second last paragraph it says,

"This makes the design of the large signal class-A amplifier possible
using the small-signal S-parameters. However, for class-AB, B, or C, the small-signal Sparameters are not suitable for design purposes".

Hence the moral of story:
1)Class A => Highly Linear => small s parameter = large s parameter
2)Class B,AB,C => Non linear => small s parameter =/= large s parameter
3)Large signal s parameter computation depends on input power as well as bias
4) Small signal s parameter almost fully dependent on bias

Title: Re: S21 of class B amplifier
Post by vp1953 on May 24^th, 2011, 10:09am

Hi Purplewolf,

Small signal S21 for Class B would be 0 (not in db) if you had ideal components and there was no parasitic that connect from input to output. Real devices would have parasitics and so you would get some non-zero S21 even with zero bias, but it should be very small