Connect with us

Definition of Colpitts Oscillator

Discussion in 'Electronic Design' started by sjcma, Nov 14, 2004.

Scroll to continue with content
  1. sjcma

    sjcma Guest


    I'm hoping someone can enlighten me on the definition(s) of a Colpitts
    oscillator. I've seen many circuits diagrams that show a Colpitts
    oscillator, but they don't always look the same.

    Assuming an NPN device, some show the capacitive feedback from collector
    to emitter, some show the capacitor feedback from emitter to base, and
    some show the tap between the 2 tank capacitors grounded. I've seen
    circuits that are common base, common emitter, and common collector all
    claming to be Colpitts.

    What's the common link between all these circuits besides tapping the
    tank in between two capacitors and connecting it somewhere (sometimes,
    even ground!).

    If someone feels like typing a lot, perhaps a quick explanation of the
    advantages of each topology would be nice :)

    Thanks in advance.

  2. BFoelsch

    BFoelsch Guest

    Nothing. The "tapped tank capacitance," as you call it , is the definition
    of a Colpitts oscillator. If you take the same circuit, but tap the
    inductor, you have a Hartley oscillator.

    Both oscillators may be common emitter, base, or collector. Oscillators like
    that are identified by the AC circuit, where the circuit happens to be
    grounded or how the power is applied is irrelevant.
    Maybe in the morning, I'm pooped right now.
  3. Tim Wescott

    Tim Wescott Guest

    The best illustration of this I've seen is to draw the circuit without
    bias networks, like this:

    | | |
    | --- |/
    | --- .---|
    C| | | |>
    C| o----' |
    C| | |
    | --- |
    | --- |
    | | |
    | | |
    created by Andy´s ASCII-Circuit v1.24.140803 Beta

    Now ground whichever node you want, add power and bias, and Presto! you
    have a working amplifier circuit! (if you want an oscillator you'll
    have to try to make a stable amplifier).
  4. Rich Grise

    Rich Grise Guest

    As far as I know, there isn't. The way I learned it, the definition
    of a Colpitts is a parallel resonant circuit with a tapped capacitor,
    with signal ground at the junction of the caps, so the ends are 180
    degrees out of phase, so it's a natural feedback path. Or something
    like that. One end to the output and the other to the input, basically.
    And the Hartley is the same thing with the coil and caps swapped. ;-)

    The rest is window dressing. :)
    All I can say to this is, of course, "It depends." :)

    But I'd bet credits to navy beans that a google search will turn up
    reading material that can keep you occupied for awhile. :)

  5. Well, I disagree with the other definitions in this thread on the taped
    cap, especially the one by Tim with no cap directly from collector to
    emitter. Here's the real deal.

    Consider 3 impedances on a transister. Zcb, Zbe Zce.

    The load on the collecter is Zce || (Zcb + Zbe).

    The loop gain is therefore:

    Av = gm.(Zce || (Zcb + Zbe)).Zbe/(Zbe + Zcb)

    Setting this to one, one finds that the only too solutions are where
    Zbe, Zce are capacitice with Zcb inductive or Zbe, Zce are inductive
    with Zcb capacitive.

    That is, a Colpits oscillator is when, topological, there are caps
    across base emitter and collector emitter, with an inductance from base
    to collector (Colpits), or where the caps and inducters are swaped

    Kevin Aylward
    SuperSpice, a very affordable Mixed-Mode
    Windows Simulator with Schematic Capture,
    Waveform Display, FFT's and Filter Design.
  6. Tim Wescott

    Tim Wescott Guest

    @#$%! See Kevin Aylward's comments -- I got my transistor in there
    wrong. All other comments apply.

    | | |
    | --- |
    | --- |
    C| | ---
    C| o-----v \--.
    C| | |
    | --- |
    | --- |
    | | |
    created by Andy´s ASCII-Circuit v1.24.140803 Beta
  7. Tim Wescott

    Tim Wescott Guest

    Argh. Yes, I got the transistor in there wrong -- I was concentrating
    on the nifty notion that you draw the circuit free of any entanglements
    from bias networks, then you ground whatever point is most convenient
    for you.

    Oddly enough I almost never do that in practice -- its only when I have
    an audience that I screw up in such a stupid way.
  8. I thought I made a mistake once, but that was an error.

    Kevin Aylward
    SuperSpice, a very affordable Mixed-Mode
    Windows Simulator with Schematic Capture,
    Waveform Display, FFT's and Filter Design.
  9. Don't worry, Tim. _I_ saw the one that got away. ;-)
  10. sjcma

    sjcma Guest

    Hi Rich,

    I've seen circuits like that described as being a Colpitts oscillator.
    But I've also seen the exact same circuit being described as a Pierce
    oscillator, where an inverting amplifier (could be a digital inverter)
    is attached to a parallel-C, series L or crystal, and parallel-C before
    feeding back to the input of the inverting amplifier.

    So is it a Colpitts or Pierce? Or is the Pierce name strictly for xtal

  11. sjcma

    sjcma Guest

    Hi Kevin,

    Thanks for a brilliantly clear explanation.

    So how should one go about choosing which of the three terminals of the
    transistor to ground? What are the pluses and minuses of each of the
    three options?

  12. Rich Grise

    Rich Grise Guest

    I'd have to look that up, but from memory, the difference would be
    the tapped capacitor. The frequency-determining circuit in the Colpitts
    is the parallel resonant ciruit formed by the inductor L in paralled
    with a capacitance C which is the total capacitance of the series

    The way I understand it, in the Pierce, the frequency-determining
    element stands on its own merit, and the cap on the other end of
    the tank, while providing feedback, doesn't participate in
    determining the frequency of oscillation. I'm sure you can look
    this stuff up - and while you're at it, look up Clapp and Armstrong.

    Here's a leg up:

  13. Active8

    Active8 Guest

    What about - I've seen this, but it's an n-fet, not an npn. The c-e
    cap isn't part of the tank, per se, but it's a bypass - ?

    | |
    | Bypass |
    |/ ---
    +----------+------| ---
    | | |> |
    | --- | |
    | --- | |
    C| | | |
    C| +--------+---- out |
    C| | |
    | --- |
    | --- |
    | | |
    +----------+------------------+-- GND, IIRC

    created by Andy´s ASCII-Circuit v1.22.310103 Beta
  14. Active8

    Active8 Guest

    My guesses:

    CE - big swing moderate output Z.

    CC - low Z out for driving a buffer. Osc isn't loaded or Miller
    f*cked by the load.

    CB - keeps the swing from exacerbating the Miller / Early effects.

    'twas a good qweshtion :)
  15. Rich Grise

    Rich Grise Guest

    Yes - the cap is a short for RF, so from the RF's POV, it's the same
  16. Active8

    Active8 Guest

    Then there are those osc tranny specs/notes where they hardly add
    any components - it runs off junction caps. I think the output cap
    is reflected to the input as inductance in significant enough
    amounts to affect the tank.

    It's one of the amps that do.
  17. Tim Wescott

    Tim Wescott Guest

    There are any number of ways to make an oscillator using the parasitic
    components in the transistor. Other than overtone crystal oscillators
    which use the parasitic C-B capacitor to form a Hartley oscillator
    (actually a tuned-base tuned-collector oscillator) between the crystal
    on the base and the tank circuit on the collector I haven't actually
    used any of that sort.
  18. On one of my first PCs, I had a 2-serial/parallel board, but the
    parallel port on the board used some Chinese lame-chip, that died
    if you looked at it wrong. This wasn't a problem for me, except
    that it also hosted the oscillator for the two UARTS. Since this
    was back in the days when I was young and had energy, I built
    a daughterboard with TTL that duplicated the circuit right out
    of the IBM Tech Ref, except plugged in in place of the Chinese
    chip. The oscillator I did separately on the same board, and it
    consisted of one transistor (2n2222 equiv.), the crystal, a cap,
    2 resistors, and a choke for the collector that I made by winding
    about 25 turns of about #36 wire on a 3/4" piece of toothpick
    with cut-off resistor leads glued into little holes I poked in
    its ends. Oh, and Vee-Vcc clamp diodes.

    I think it was a Pierce or Clapp oscillator - about the simplest
    one-tranny xtal osc you can make.

    I never blew up the LPT port after that, and the UARTS worked!

  19. Max Hauser

    Max Hauser Guest

    "sjcma" in news:...
    You don't even have to ask, you can look it up in the archives. For
    example, Question #16 in

    Google has this as the first newsgroup posting on the subject.
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