From: email@example.com (carltons)
Subject: Re: transistor stability and reflection co-efficients.
User-Agent: NewsWatcher-X 2.2.3b2
NNTP-Posting-Date: Wed, 13 Nov 2002 07:53:12 GMT
Date: Wed, 13 Nov 2002 07:53:12 GMT
In article , "Richard Webb"
> Hi all,
> I'm trying to design a simple RF amplifier, but I'm not all that
> experienced with RF stuff! If the stability factor of a transistor at a
> certain frequency is below zero then the device is only conditionally stable
> (i.e. it could oscillate). Now does this mean that the reflection
> co-efficients of the input and output of the transistor will be zero? I
> didn't think that it would but my CAD tool is telling me that this is the
> case! If I stabilise the transistor with a series resistance and get the
> stability factor up above 1 then my CAD tool reports sensible co-efficients.
> Can anyone tell me why the reflection co-efficients of a conditionally
> stable transistor are zero?
> Many thanks,
> Richard Webb.
I have never tried to answer this question in writing, I usually
demostrate it on the board to someone, but here goes. If you have a good
match to your input or output impedance, your reflection coefficient will
be 0, meaning the reflected power is way down from the ingoing or incident
power. If your input match is lousy and looks like a short or an open,
your coefficient will be -1 or +1, meaning all the incident is coming back
to you. Where stability comes in is if your reflected power is more than
your incident power. Since the laws of thermodynamics still apply at all
times, you are getting extra power from outside. This extra power is
being generated by your active circuit (rf amp). You are trying to get
the rf amp to amplify without having the input or output impedance
generating another batch of, usually, extraneous power. This is a very
simplistic answer, but I hope you get the idea behind it. Remember also
that you are talking about the whole circuit. The transistor, tube, what
have you, is only one component. Design your circuit and put in the
values you think are optimum, but try to put in real components and
include the parasitics. Then sweep your circuit on the simulator from DC
to blue light. If you get S22's or S11's greater than 0 db, find a
solution to change that.
There's a whole lot more, but let's not get carried away at this point. :-)