Need help understanding current sourced common emitter/source amplifiers

[GregH2]

Hello all,

I'm still having trouble grasping a basic amplifier concept I'm hoping you can help with. I'm thinking out loud here so bear with the lecture.

With a common emitter/source amplifier, as most of you will know, it is common to add an emitter resistor (Re) and collector Resistor (Rc) to stabilise the gain and make gain more or less independent of transistor variations. So gain essentially becomes Rc/Re (within reasonable limits).

So we now have a circuit that can be used in multiple amplifiers and perform in much the same way despite transistor variations.

One way to improve linearity and reduce distortion is to replace Rc with a constant current source, which if well designed has essentially infinite output impedance.

This means that gain theoretically becomes infinity/Re, but obviously gets limited by the transistor transconductance and other transistor properties.

So my question is, how do you make a simple current sourced common emitter amplifier operate independently of transistor variations?

I need to make two channels of an amplifier, and each has a current sourced common source jfet input. How do I make both input stages have the same gain without either removing the current source or resorting to a LTP input?

Any advice you can provide will be appreciated!

Regards,

Greg.

[Miles Prower]

Originally Posted by swordfishy

This means that gain theoretically becomes infinity/Re, but obviously gets limited by the transistor transconductance and other transistor properties.

The voltage gain is:

Av= GmRl

Where Rl= 1/Hoe || Rc || Ri (next stage/load resistance)

1/Hoe being the inherent collector dynamic resistance. You can usually ignore it with minimal errors since Rc is usually ~an order of magnitude smaller. An active load can increase gain by making:

Av=~ (1/Hoe)Gm (no load)

I need to make two channels of an amplifier, and each has a current sourced common source jfet input. How do I make both input stages have the same gain without either removing the current source or resorting to a LTP input?

I wouldn't worry about it. You could always match transistors, but for the most part, the gains won't be that far off for any reasonable design. How much gain do you really need here anyway? How accurate do you need to be?

[Samuel Groner]

Active load techniques are not particularly useful without feedback for the reasons you mention. They are mostly seen within opamp topologies where global feedback stabilizes the operating point.

Samuel

[AndrewT]

Hi,
the active collector load will almost certainly not have constant impedance across the whole passband of the amplifier. This results in gain that varies with frequency and confirms SG's need to apply NFB.
JLH shows a variety of one, two and three transistor amplifiers with different ways of providing AC feedback and DC feedback. All these circuits appear individually in other papers, so far I have only seen the collection of them in "The Art of Linear Electronics"

Where else can we find a comprehensive collection of feedback scenarios for the one to three transistor amplifiers?

[GregH2]

OK thanks, I will get JLH's book and look into local NFB options. In the mean time I'm going to play with some simulations and build some test circuits.

Greg.

[AndrewT]

JLH's The Art of Linear Electronics is out of print.
It was published by Newnes isbn 0-7506-3746-3

I consider this my best electronics reference book.
Better than Horowitz and Hill and the ARRL handbook.
It is a good complement to Self and/or Cordell.
Those 5books + a number of briefer pamphlets & small paperbacks cover most topic I need to consult.

[Ian Finch]

Chapter 3 of Self's recent title "Small Signal Audio Design" is all about discrete amplifier circuits in the standard categories of single, 2 and 3 transistor amplifiers and opamps. It's yet another book but this one is right in the audio ballpark with up to date analyses.

As a comment, the book is light on theory whilst retaining essentials and thorough on evaluation and choices as would be necessary for any DIYer to make. BJT in this chapter only.

The publishing details and a little more on discretes are on The Douglas Self Site

[GregH2]

OK thanks all, I will track down those books. Ordered the JLH one already.

Greg.

[GregH2]

Thanks Miles,

Originally Posted by Miles Prower

The voltage gain is:

I wouldn't worry about it. You could always match transistors, but for the most part, the gains won't be that far off for any reasonable design. How much gain do you really need here anyway? How accurate do you need to be?

Well I don't know really. I just assumed they would need to be pretty close if they're for the L and R channel of the same amp. So you think I'm worrying too much then and that I should just build it as is and match Vgs as best I can?

[Miles Prower]

^
|
|

Yes.