Connect with us

Noise in audio + opamps

Discussion in 'Electronic Design' started by Dirk Bruere at NeoPax, Mar 5, 2007.

Scroll to continue with content
  1. Looking at the OP37 (for example) with a noise figure of 3nV (root
    If I were using this as an audio pre-amp with a gain of 5 how does
    this translate into the kind of S/N figures usually specced for amps
    eg -NN dB ?

  2. You have to do the sums
    page 150 gives you the lowdown.
    depends on source Z, bandwidth, temperature etc. +-1audiophool

  3. Thanks
    Quite interesting.

  4. Eeyore

    Eeyore Guest

    The first thing you need to do is specify the reference level for measurement !

    National's 'Audio Handbook' deals with all of this reasonably well including a
    discussion of adding the effect of source impedance and the influence of bias
    currents (noise component of ) too.

  5. I'm just looking for a noise free way to adjust non standard line-in
    to non-standard line-in!
    Since the gain will be very unlikely to exceed a factor of 3 maybe
    just a common emitter transistor as an amp?
    I really want very little noise and to run from a 5V computer PSU.
    I expect this is where someone mentions noise on the 5V, rejection
    ratios and all that.
    Alternatively a neat 5V opamp solution would be appreciated.

  6. Joerg

    Joerg Guest

    Many devices from the TI (formerly Burr-Brown) OPA series seem to be
    suitable but during a recent similar search I found lots of them to be
    out of stock. AD has nice low noise amps that are higher in BW but be
    careful, many of them cannot tolerate more than a diode drop in
    differential input. Might not be an issue for a linear amp but other
    than a small hint in the text there was no mention of that.
  7. Eeyore

    Eeyore Guest

    Do you really want to crush your audio performance ?

    5V isn't enough to run pro-level audio without clipping. Heck, it's not even adequate
    for CD level outputs (2V rms @ 0dBFS) !

    You should use a split supply too unless you really want to make life hard for

    The good news is that supply noise isn't too much of a problem with op-amp circuitry.
    You're going to have to think this one through a bit more thoroughly.

  8. Eeyore

    Eeyore Guest

    The standard commodity NE5532 or NJM4560/80 is way more than adequate for line level

    Esoteric stuff isn't needed.

  9. Fred Bartoli

    Fred Bartoli Guest

    Eeyore a écrit :
    With 0-5V supply it might be so :)
  10. Joerg

    Joerg Guest

    Yep, the NE5532 needs at least +/-5V.
  11. Eeyore

    Eeyore Guest

    5 volt supplies are a non-starter for serious audio.

  12. Joerg

    Joerg Guest

    It'll be tough. I am currently doing one and it will have +/-12V.
  13. Eeyore

    Eeyore Guest

    That's reasonably OK as long as you're not doing any serious summing typically.
    +/- 15V is more the norm but I run my own supplies at +/-17 V to get that
    elusive last dB or so of headroom. Even so, the internal signal levels I usually
    normalise to -6dB wrt external levels.

  14. Chris Jones

    Chris Jones Guest

    You need to think about the impedance of the source before choosing the
    op-amp, in case you have not already.

    You would probably benefit from reading Douglas Self's books, in case you
    don't already have them. On page 83 of "Self on Audio" he mentions that in
    one particular circuit, the 5532 was measured to be quieter than an OP-27,
    even though the datasheets would lead one to expect the opposite. I think
    it is something to do with the bias current cancellation scheme in the
    OP-27/OP-37 which introduces common-mode noise current, though that only
    matters when the impedance presented to the two input pins is not equal.

    Here is some stuff about op-amps and audio

  15. Joerg

    Joerg Guest

    So what's your opamp du jour for this stuff? AFAIK the NE5532 doesn't
    come in TSSOP and not in quad pack. If something doesn't migrate to
    TSSOP these days I become a bit concerned about remaining production life.
  16. I'm feeding into a soundcard on a PC, so it's not rocket science.
    The best I've found is this:
    1nV/rtHz isn't bad.

    Dirk - The UK's only occult talk show
    Presented by Dirk Bruere and Marc Power on ResonanceFM 104.4
  17. john jardine

    john jardine Guest

    The cheap TS974/2/1/ (Farnells) does a nice 4nV-root hz.
    Wide bandwidth, low distortion and runs down to +/- 1.35V.
    Used 8 in series for a particular x100,000 bandpass filter (30kHz bandwidth)
    and was seeing 100mVpkpk noise at the output. (as should be).
  18. Eeyore

    Eeyore Guest

    The 5532 wasn't originally available in SMT because AIUI the die used to be too large.
    It is now however available in SOIC and has been for some time.

    The other regular candidates are the NJM 4560 and 4580 along with the venerable TL07x.

    I rarely use any of the esoteric stuff since I'm normally designing for low cost and
    the above do a fine job.

  19. Eeyore

    Eeyore Guest

  20. Because I'm already outputting from it.
    See other thread:)

    Dirk - The UK's only occult talk show
    Presented by Dirk Bruere and Marc Power on ResonanceFM 104.4
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