Transistor buffer questions

Started by oliphaunt, May 04, 2010, 03:09:59 PM

Previous topic - Next topic

oliphaunt

I have been playing around with this transistor buffer circuit:



I like it a lot, it has a very natural sound and does it's job without a lot of coloration.  My questions are:

1. This is an unfamiliar design, and frankly I don't understand how it works at all.  Can anyone offer a brief expalanation of what is going on here?

2.  How do I calculate the input and output impedances (I am asking this a lot these days)

3. Can this be converted to a boost?

PRR

> I don't understand how it works

It's an Emitter Follower.

It has power steering and tail-fins, but it's just an emitter follower.



From this you can work-out the DC currents and voltages.

And the fact that the voltage gain is always less than One.

i.e. it can't be converted to a boost any simple way.

The output impedance is a mess of several factors.

  • The source impedance divided by Beta.
  • The transistor emitter impedance.
  • The 150 ohm resistor.
  • The 7K5, 20K, and 50K resistors.
Pretend the source is 50K and the Beta is 100-400. We have 500-125 ohms due to this.

The transistor emitter impedance can be found from its emitter current via Shockely's Law. You are gonna have to learn this to get far in transistor analysis. Say the current is 0.5mA, the impedance is 60 ohms.

And another 150 due to R8, is 660-335 ohms.

This is in parallel with 7K5, 20K, and 50K. These won't change the result much.

The input impedance is the base impedance in parallel with the bias network. Oh, and the 10Meg, but other impedances are low enough that this may be moot.

The base impedance is roughly Beta times all the load the emitter drives. It already drives 5K, so typical stompbox external loads hardly matter; the base impedance is probably over 500K.

The bias network is "bootstrapped". Here used as a cheap trick to (maybe!) raise input impedance. The Wikipedia article on bootstrapping is awful. And matchbook computations of bootstrapping are treacherous. Over most of the audio band R3 will "act like" much more than its 120K marking. Perhaps even 50 times higher! However it tumbles in the bass, and I'm not sure what C2 does to the high end.

I bet a penny it is over 250K for most of the audio band. Perhaps WAY over 250K mid-band. And that's really good-enough for real life, because most sources are 50K or less.
  • SUPPORTER

Gus

Look at the beginner project at this site.
  Look at the "Art of Electronics" There is a section on bootstrapped EFs.

PRR

  • SUPPORTER