Connect with us

White noise generator

Discussion in 'Electronic Basics' started by Roger Dewhurst, Dec 26, 2005.

Scroll to continue with content
  1. I would be grateful for suggestions on constructing a simple random noise
    generator. Most of the noise should be within the audio spectrum.

  2. John  Larkin

    John Larkin Guest

    Radio Shack sells a little amplified-speaker box, $9 or something.
    It's very noisy, and should output a reasonable amount of noise via
    the headphone jack. I can't vouch for its statistics.

    The more formal way to do this would be to bias a low-power 10-volt
    zener to maybe 1 mA current, and amplify that. Figure the zener will
    make roughly 300 nV/rootHz noise density, or about 40 microvolts RMS
    in the audio KHz range.

    You can also make noise digitally, with a pseudo-random shift
    register. See AoE.

  3. Thanks. Radioshack stuff is unavailable for me. The second option is more
    in line with what I anticipated. Will one operational amplifier be
    sufficient? I am not interested in the quality of the noise but merely
    getting sufficient to drive a small speaker fairly hard.

  4. Ralph Mowery

    Ralph Mowery Guest

    Tune a cheap FM radio inbetween stations.
  5. The cheapest FM radio would still need an amplifier as the speaker has to be
    well separated from the signal generator in the application that I have in

  6. Pooh Bear

    Pooh Bear Guest

    I use a 10 Megohm resistor and amplify the f**k out of it with a low noise

  7. Pooh Bear

    Pooh Bear Guest

    Don't semiconductor junctions make excess shot noise though ?

    Nat Semi once made a chip that did that. I have a 'little box' I made using
    one. Sounds horrid. You can hear the pattern.


    p.s. I've sometimes heard shot noise referred to a Schott noise. Any idea
    which is correct ?
  8. BobG

    BobG Guest

    Any reverse biased eb junction is noisy like a zener diode. This is
    white noise... equal energy per hz... you want pink noise? (equal
    energy per octave)
  9. John  Larkin

    John Larkin Guest

    At this sort of current, zeners are pretty white and gaussian, maybe
    just a tad asymmetric. At much lower currents, shot noise is
    The sequence must have been short. A 1 MHz, 64-bit shift register
    won't repeat in your lifetime.
    I think it's "shot", like buckshot falling on the roof.

  10. John  Larkin

    John Larkin Guest

    I's suggest two opamps, each with a closed-loop gain of, say 50, to
    get you up to 100 millivolts RMS, then some sort of power amp to drive
    the speaker, one of those cheap National thingies maybe. A single
    opamp might work, ahead of the amp, depending on your numbers... x1000
    at 3KHz requires at lease a 3 MHz gain-bandwidth opamp.

  11. John  Larkin

    John Larkin Guest

    The opamp noise would probably wind up dominating the resistor noise.
    That might be OK, but opamp noise tends to be fairly non-white, with a
    lot of excess low-frequency (1/f) component and occasional pops and
    other weirdness.

  12. Pooh Bear

    Pooh Bear Guest

    I make 10 Meg = 409 nV / sqrt Hz

    Used with a TL07x type op-amp ( 12 nV / sqrt Hz ) it seems to be just fine.

    With a bipolar op-amp you would indeed likely get loads of input current related
    noise though.

  13. John  Larkin

    John Larkin Guest

    Yeah, that works. What's the input capacitance on one of those?

  14. Pooh Bear

    Pooh Bear Guest

    Seems to be unspecified. I know what you're thinking. I've used that configuration
    for test jigs - can't recall now if I did a sweep across the audio spectrum. JFET
    input btw.

  15. Brian

    Brian Guest

    Check out Also go to Google
    and put it "white noise circuits", there is a lot of good information there.

  16. Bob Masta

    Bob Masta Guest

    You might want to check out my DaqGen freeware
    at It uses the sound card on any
    Windows computer and can generate all sorts of
    test signals, including non-repeating white and pink
    noise as well as band-limited noise. You can see the
    real-time waveform or spectrum of the noise, average to see
    flatness over time. or run a histogram to get the
    amplitude distribution.

    Best regards,

    Bob Masta

    D A Q A R T A
    Data AcQuisition And Real-Time Analysis
    Home of DaqGen, the FREEWARE signal generator
  17. spudnuty

    spudnuty Guest

    I used a circuit I found that used a saturated transistor few parts and
    a 9 V. Very simple and I put it into the mike input of a portable
    radio. Worked great as a noise blanker for the babies room.
  18. Thanks. Would a 741op-amp drive a small speaker? If so would two or three
    of these in series do the job? Would you bias the + input to half the
    single supply voltage (12 volts) in each case and use 470pf capacitors
    between the zener diode and the first stage, between each stage and between
    the last stage and the speaker?

  19. Never Mind

    Never Mind Guest

    It would help if you described your application. (The PP was referring to
    using an FM radio with NO signal)

    Since you mention a speaker, do you really need white noise, or just some
  20. Noise across the audio spectrum of a small speaker. The speaker will be
    some metres away from the signal generating part of the device. The noise
    level must be controllable nearly upto the limit of the speaker.

    I have just about come around to the idea of a 12 volt Zener diode driving
    two LM741s followed by a LM368N to drive the speaker. The op-amps would
    have a 12 volt unbalanced supply with the + inputs of the LM741s biassed to
    6 volts. The parts are all available and cheap. I have the speaker. Would
    this work in your view?

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