You are chasing a question that has little practical value. The question should not be about the difference between pure "current control" and pure "voltage control" for biasing since both vary and neither gives sufficiently stable results to use in practice. The question should be about the stability of properly designed bias circuit due to both Beta and Vbe variations. A good bias circuit is relative insensitive to both these factors. That is why local feedback using an emitter resistor is typically used to bias a transistor since it minimizes the effect of both.
Yes - I think, this contribution has a real practical background and is worth to be discussed further.
The necessity of an emitter resistor has been mentioned already by Eric (see post#7).
In the following I present an example, which clearly shows how dc feedback drastically reduces the uncertainties connected with the spread of beta and Vbe (see Is in Shockley`s famous formula).
For the following example the only information regarding the BJT is that the base-emitter voltage in the active region will be between 0.65 Vand 0.7 Vs.
Example:
*Transistor stage in common emitter configuration
* DC feedback with Re=1kOhm
* Supply voltage: Vcc=12 V.
* Given collector current:
Ic=1mA.
* Resistive divider for dc biasing assuming a divider current Id=0,1*Ic
(this assumption results in a resistor niveau that seems to be acceptable with regard to the input resistance of the circuit)
* Note that no information about the DC gain B=Ic/Ib is used
(a)
Assumption Vbe=0.65 V
It is an easy task to compute the values for the divider:
R1=103.5 kOhm and R2=16.6 kOhm.
Simulation result for 2N2222:
Ic=0.92mA.
(b)
Assumption Vbe=0.7 V
R1=103 kOhm and R2=17 kOhm.
Simulation result for 2N2222:
Ic=0.96mA.
_______________________________________
Summary: Based on a rough estimation for Vbe (0.65...0.7V) and without any knowledge/estimate for B=Ic/Ib it is possible to design a common emitter stage with dc feedback.
A circuit simulation for a real transistor shows that the calculated resistor values for BJT biasing lead to a collector current that deviates from the target value by only (4...8)%.
This deviation has the same magnitude as other uncertainties within the whole circuit (resistor tolerances, transistor parameters).
Of course, this deviation (error) can further reduced for increased DC feedback (Re increase).
W.