Need help understanding transistor bias - when does signal get UNBIASED?

[Yonatan]

Just tried to post this, but the post got lost somehow.  So here is the shortened version

I understand that a transistor (for example Electra Distortion circuit) needs to be biased, because the signal coming out of my guitar swings +/- relative to 0v, but the transistor can only pass current in one direction.  So we add some bias to the input signal, to make sure that the lowest voltage in the signal is still above 0v, and keep the transistor working throughout the entire waveform.

But now, the signal coming out of the transistor swings +/- relative to some new value, right?  So what happens when that signal gets put into ANOTHER transistor stage, or gets connected to another effect?  The next effect will try to bias the signal, but won't it be both unnecessary and incorrect?  But I know that I can string together multiple effects, so obvioulsy, I'm missing something here.

(To clarify what I mean by UNBIASED - when does the signal get restored to the point where it swings +/- relative to 0v)?

[Seljer]

Theres usually capacitor in between transistor stages and effects that only lets the signal (which is AC) through, while letting the next stage be biased on its own (the bias voltage is DC)

[Yonatan]

I know that the decoupling capacitor blocks DC, but doesn't the remaining AC signal still swing around only positive voltages?  And so how will the next stages or effects work, if they also try to bias the signal?

[teemuk]

I know that the decoupling capacitor blocks DC, but doesn't the remaining AC signal still swing around only positive voltages?

Now, think about that sentence for a moment. 

[defaced]

By isolating DC between stages, the DC offset voltages don't mix.  So if a transistors is biased for some operating condition, by placing a cap between its output and the input of the  next stage, the waveform picks up the bias voltage from that next stage.  That could be 0v, it could be 4.5v, it could be whatever.  For instance, between tube stages, the bias voltage is often 0.  Between transistor stages, it's usually a couple of volts.  At the output of an effect, it's probably going to be 0v - but that depend on what it's plugged into/pull up/down resistors/etc.  So no, the signal does not always swing around some positive voltage, it often swings relative to ground and if you ever needed a negative bias voltage (power tubes in amps for instance), it'd swing +/- of that voltage.  

Also, remember biasing is for static no-signal conditions.  As soon as you apply a signal, the bias voltage is yanked around by the signal, which causes the active device to turn on/off.  This on/off action when coupled with a couple of resistors and a higher voltage supply rail is what creates amplification.  If you think of a transistor as a signal controlled variable resistance, and work out the voltage divider for whatever output node you're interested in, things will start to make alot more sense.  

[Yonatan]

Thanks, that definitely helps to make sense of this.