Do we analyse multivibrators and oscillators in the same way,
i.e. beta*A = -1 . For example, if we look at the multivibrator
at
http://www.uoguelph.ca/~antoon/circ/2qflash.htm
do we need to calculate the gain with a small signal equivalent
circuit or not. Is it a voltage-shunt feedback, what approximations
do we do to ease the analysis, and how do we get the equations
for the frequency of oscillation?
With this sort of oscillator, you have to analyze several completely
separate phases.
One phase starts the moment the output transistor begins to turn off,
and collector voltage heads toward ground by the pull down action of
the 470 ohm resistor. during this phase, D1 first forward biases and
clamps the base voltage for Q1 to no more than a diode drop and C1
charges through R2.
Eventually (this is where a bit of RC analysis comes in) the current
through D1 goes through zero and C2 changes charge rate because the
current through R1 in series with R2 sets the charge rate for C2.
During this phase, the base voltage at Q1 swings between negative one
diode drop toward positive one diode drop, where Q1 begins to conduct
as the current through R1 detours from C2 to the base emitter
junction.
Once Q1 conducts enough to start to turn Q2 on and the voltage drop
across R3 starts to increase, a positive feedback loop forms that very
quickly drives Q1 and Q2 into saturation because the positive voltage
change across R3 gets coupled back to the base of Q1 through C1 and
R2, making the current from R1 insignificant. This phase lasts as
long as C1 R2 can supply enough current to Q1 to keep both transistors
well saturated.
Once C1 charges enough that this current is no longer available, Q1
and Q2, while still conducting somewhat, fall out of saturation enough
that the voltage drop across R3 starts to sag. At that moment, the
current through C1 R2 reverses direction and Q1 shuts down and you are
at the starting point of this description.
Each of these phases has to be analyzed to predict both the on and off
time.