Connect with us

quick emitter follower question

Discussion in 'Electronic Basics' started by tempus fugit, Sep 13, 2006.

Scroll to continue with content
  1. tempus fugit

    tempus fugit Guest

    Hi;

    I'm thinking about using a simple 1 transistor emitter follower in an audio
    design (the audio signal will pass thru it). Does the noise figure of the
    transistor need to be taken into account, or will there be no noise added
    since there is no amplification happening (sort of like the signal just
    'passes thru' the transistor untouched)?

    Thanks
     
  2. The transistor adds noise to any signal applied to it. That noise
    does not get voltage amplified by the follower, but it still gets
    added to the signal.
     
  3. tempus fugit

    tempus fugit Guest

    OK, that's kind of what I thought too, but I thought I should make sure
    before choosing a suitable transistor.

    Thanks John.
     
  4. Ban

    Ban Guest


    Not only will the transistor add some noise, it will also add a lot of
    distortion. around 1% for 600mV peak to peak. The distortion raises
    proportional to the input level and is independent from the bias point.
    The noise is completely irrelevant compared to this phenomenon. If you want
    the signal "untouched", you should use an opamp as buffer, OP27 comes to
    mind.
     
  5. Eeyore

    Eeyore Guest

    There certainly will be noise added but not a lot.

    Graham
     
  6. John Larkin

    John Larkin Guest

    What's the mechanism? With enough DC voltage across the emitter
    resistor and a light load, one can approach constant emitter current.
    All that's left is Early voltage effects, pretty small usually.

    John
     
  7. And of course, the real issue is the application. One worries about
    noise with weak signals, with larger signals the signal swamps out
    the noise.

    A lot of places where emitter followers are used, the issue of noise
    will never come up.

    On the other hand, one reason you see transformers in low level audio
    signals even today is because they are at a point where the signal
    is weak, and introducting noise at that point is not a good thing.

    So they have those matching transformers between microphones and
    the preamp, and between those moving coil phono cartridges and the phono
    preamp, since they need to step up the voltage at those points without
    adding any noise. Once the signal is stepped up by the transformer,
    the noise of the following preamp is less important.

    Michael
     
  8. tempus fugit

    tempus fugit Guest

    Thanks for all the replies.

    Michael, since you mentioned the application.....

    I currently have a really slick switching system for my guitar effects
    pedals. To make a long story short, the only problem with it is that it uses
    relays to switch the effects in and out, and they give an audible "pop" when
    switched. It's not horrible, but I was trying to find a way to make things
    cleaner (I can't leave well enough alone). I first thought of using a simple
    transistor or JFET as a switch in place of the relay, but a friend informed
    me that it wasn't possible to have the audio pass through the transistor if
    it was also being used as a switch (or is there a way?). We came up with the
    idea of using a transistor as an amp with very little gain and switching the
    amp on and off instead of a relay. Which brings us to this discussion.

    Any thoughts on this line of reasoning?

    Thanks
     
  9. Bob Eld

    Bob Eld Guest

    Never add an open, simple emitter follower to a high quality audio circuit.
    As was mentioned they will add distortion to the signal which may be
    tolerable in a phone or other circuit where high quality audio is not
    required. This happens because the internal emitter resistance, re, is a
    function of the emitter current which in turn is a function of the base to
    emitter voltage. This makes a variable, non-linear, voltage divider with the
    load impedance. The nominal gain is one, but in reality is always less than
    one and varies with the voltage level. That non-linearity causes excessive
    distortion both harmonic and intermodulation.

    Emitter followers are used inside vitually all bi-polar amplifiers, op-amps,
    etc. But, they are ALWAYS inside feed back loops attended with current
    sources and gain stages to reduce their inherent problems. If you want a
    unity gain buffer make one with an appropriate low noise unity gain op-amp.
    Bob
     
  10. John  Larkin

    John Larkin Guest

    On the other hand, audio folks love open-loop cathode followers, which
    are far more nonlinear than any reasonably biased emitter follower.

    John
     
  11. Eeyore

    Eeyore Guest

    It's daft ! For starters you'l get a 'DC thump' every time the follower stage is
    switched on or off.

    JFETs make very very acceptable audio switches when used correctly.

    Graham
     
  12. Bob Eld

    Bob Eld Guest

    Don't you mean Ear-Queers? They like non-linear circuits. Some of them even
    prefer single ended triode power stages sans feedback...go figure. Some
    distorions can be perceived as quite musical. After all if you are just
    adding to or rearranging a few harmonics in an already harmonically rich
    signal, who's to say it doesn't sound better?
     
  13. tempus fugit

    tempus fugit Guest

    Thanks again for all the replies.

    OK this answers one of my other possible problems - I wasn't sure if the
    switching on/off was going to cause a bigger noise than the relays popping.
    I had planned on ramping up the voltage to the follower, but I don't know if
    that would help or not.


    I'm intrigued. Any suggestions on how to use one for this application?

    Thanks
     
  14. Eeyore

    Eeyore Guest

    Dead easy.

    Use a nice low Ron jfet like a J108 and pull it off with a negative voltage on
    the gate.
    http://www.fairchildsemi.com/pf/J1/J108.html

    Use an open-collecter arrangement to drive the gate with say 10M from gate to
    source. The signal path is from drain to source ( bidirectional ). Use a 10k
    'pull down' R on both in and out to avoid any DC levels getting into the switch.

    If finding a negative voltage is a problem use a p-channel device like a J174
    and pull the gate positive.
    http://www.fairchildsemi.com/pf/J1/J174.html

    The n-channel parts have lower Ron as do the lower numbered devices in each
    family. The tradeoff is a higher Vgs to turn them off.

    Graham
     
  15. Eeyore

    Eeyore Guest

    Audiophools(tm)

    Graham
     
  16. Eeyore

    Eeyore Guest

    You could also use a constant current load too.

    Graham
     
  17. John Larkin

    John Larkin Guest

    Even better. At 10 mA emitter current, output impedance will be under
    3 ohms, so a reasonable load will hardly pump the b-e junction
    nonlinear.

    John
     
  18. Eeyore

    Eeyore Guest

    With Re as low as that you'll most likely be seeing the reflected ( terminology ? )
    impedance of the previous stage in fact.

    Graham
     
  19. Ban

    Ban Guest

    To drive the output load the stage needs a varying current, which in return
    requires a slightly different Vbe. The smaller the load, the more the
    variation in Vbe, which has an exponential characteristic. The inherent
    feedback helps linearizing, but for lower R_load and higher capacitive
    loading distortion shows up immediately.


    That would be a cure, but then any load impedance will annulate the effort.
    The output impedance is (1/gm + Rg/Hfe)||Re. With 50k input impedance we
    will hardly come below 100R.
    So even if the simple follower has a very high input and low output
    impedance, its properties cannot be used simultaneously. Those disadvantages
    can be overcome with some more sophisticated circuitry like the
    White-follower or this differential input stage:
    +---+------o
    | | +Vb
    / \ |
    ( 1m) |
    \_/ |
    | |
    | |/
    +-|
    | |>
    | |
    .---)---+ out
    | | +------o
    |/ \| |
    o---| |-+
    |> <| |
    | | |
    '-+-' |
    | |
    .-. |
    | | / \
    | | (10m)
    '-' \_/
    | | -Vb
    +-----+------o
    (created by AACircuit v1.28 beta 10/06/04 www.tech-chat.de)
    But for driving long cables I would recommend rather the opamp buffer
    solution with a complementary output stage.
     
  20. tempus fugit

    tempus fugit Guest

    Hi Ee

    I actually discovered the J105 and was thinking of trying to use it somehow
    already.

    I haven't really worked with JFETs much before, and am only now becoming
    familiar with the open collector thing (which would be open drain in this
    case?). Could we move this discussion to ABSE and I can post a schematic of
    what I think it should be set up like? I'm also unsure of the biasing
    required.

    Thanks
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-