Wiggling around the biaspoint

[LeadingZero]

First of all: Sorry for my english, I'm from Germany.

I have a problem in understanding basic transistor behaviour. I see no way to precalculate the voltage errors, an emitterfollower creates.

1. I set the base voltage for 1mA quiecent current and Ure = 1 Volt.
2. If I raise the input voltage by half a volt, then Ure will not rise by
half a volt, because Ube raises too (by a few millivolts), as collector-
current rises.
3. At the same time, S(t) changes, as collector current rises, because
small re changes.
4. It all seems to interact, but what was first ? Egg or chicken ?

Is there a step by step explanation for what does exactly happen, when I raise Ub ?

Greetings Alex

[ilimzn]

It's not really a question of what comes first, it all happens 'simultaneously'. In actuality there is a delay but it is normally very small.
At each point, Ue = Ub - Vbe, Vbe is at given Ie = Ue/Re
More Ie requires more Ib (assuming Beta is constant, but this can only be assumed for a range of Ie and Uce), this means that as Ie rises, the change in Ube required for given Ib falls because Ib/Ube is exponential.
Also, because Beta and Ube at constant Ib depend on temperature, and all parameters vary largely between different instances of the same transistor type, your conclusion that thee xact DC offset in an emitter follower is not exactly predictable, is true. however, the question is, how important is knowing it exactly? In most cases it's good enough to get beow a certain error. There are ways around that, which involve more complex circuits, that compensate the Vbe, using additional transistors. You might want to look at the Baxandall Pair, and the Diamond buffer, which are the simpler ways around the problem.

[WuYit]

ilimzn,
looking at things from a somewhat different angle some problems emerge... The Ibe /Ice relationship is consistent all right, the problem is that Ibe is not a governing parameter, Vbe is, and the Vbe / Ibe relationship is anything but consistent, since Beta depends on Ice, Vce and temperature, the expectations for linearity have to be kept very low...

Always nice to have you around.

[cbdb]

I have a problem in understanding basic transistor behaviour. I see no way to precalculate the voltage errors, an emitterfollower creates

You understand better than you think. Its gets complicated fast because the changing Ic changes the transistors parameters.

The equations for a common emmiter stage are vo/vi=1/1+(Rs+r pi)/(B+1)RL

where Rs is source resistance B is beta RL is load res (on the emiter) and r pi is the transistor input R. And the parameter that changes with Ic.

r pi=Beta/gm and gm=Ic/Vt (Vt is a constant)

You can figure out the distortion (error) from these equations but its not easy. Much easier to use a simulator.

[cbdb]

This is only one cause of distortion. Beta changes with Ic at the same time, and betas rate of change depends on the bias point, which changes with Ic. Thats why people sim or build the circuit.

[WuYit]

You can only simulate static conditions.

[ilimzn]

Not really, depends on the model. In some cases dynamic behavior (usually LF) is actually simulated quite relevantly (I'm not going to say accurately since real world transistors have so much variation).
Completely agreed on Vbe being the real governing parameter, it's only that we tend to simplify the relationship by assuming the exponential slope is so steep it's near-vertical, and Vt is constant etc etc.
This is why I mentioned compensation schemes using similar parts rather than trying to calculate things exactly for a given condition - too many real world conditions for that. If we were to construct our circuits that way, we'd never construct anything