>(Tom)
> > > Please give me an an example of a class A or AB linear PA that
> > > will go into a catastrophic failure induced by a oscillation when
> > > tickled with exactly the right transient, but will not oscillate under
> > > proper testing for stability.
> > >
> > ...(Carl)
> > I'm sorry, but I didn't say that I could do that. I do not know enough
> > about the dynamic behavior of tubes and tube amplifiers. When I learn
more
> > maybe I will have something interesting to say about that, and maybe
not.
>(Tom)
> Then you can't possibly make the following statement:
Said statement did not pertain specifically to a tube amplifier. It was a
statement of general physics, a subject that I _am_ familiar with. I've
repeated it twice already and you haven't addressed it, so I won't bother
saying it again.
Up to this point, we're not actually disagreeing so much as talking about
different things.
>
> A linear circuit either has enough feedback of the correct phase to
> be an oscillator, or it doesn't. It simply isn't going to sit there for
> years and then suddenly, out of the clear blue sky, explode from a
> parasitic just because the right "tick" came along.
>
What you say here is true but irrelevant. I completely agree that a linear
circuit is either stable or unstable regardless of the excitation level
(assuming its component values and terminations are constant, which is of
course not the case in real life). But to argue this is to invoke a straw
man because a class AB amplifier is not a linear circuit. It has an
approximately linear relation between the input and output for signals that
are within the bandwidth of the tank circuit. Anything else can be and
probably is nonlinear. The relation of anode current to anode and grid
voltages in particular is strongly nonlinear over the duration of one RF
cycle. Any desired or undesired signal in the anode circuit can see a
nonlinear environment, as can any signal anywhere whose frequency is far
from the tank circuit resonance.
I don't say here that I believe Rich's model of spurious oscillations is
correct. I don't in fact know whether it is or not. But I am quite certain
that the possibility can not be ruled out by appealing to linear circuit
theory as you do in the preceding quote.
I'll say it once more, concisely, and then I won't bother repeating myself
later: A class AB "linear" amplifier is not a linear circuit.
Respectfully submitted,
Carl WS7L
--
FAQ on WWW: http://www.contesting.com/ampsfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|