Connect with us

NEWBIE: Designing a Simple Audio Amp

Discussion in 'General Electronics' started by Ricky Romaya, Jan 1, 2005.

Scroll to continue with content
  1. Ricky Romaya

    Ricky Romaya Guest

    I want to design and build a very simple audio amplifier to drive my
    8ohms speakers from (onboard) pc soundcard. I figure I'm using op-amps.
    I'm a newbie and all I know about op-amps are basic, mostly from text
    book.

    Basically, I'm a little confused. Usually I see that the type of
    amplifiers to drive speakers are called 'power amplifiers'. From the
    basic building blocks of op-amps, I could only find circuits for voltage
    and/or current amplifiers. Do those 'power amps' are really 'voltage
    amps', cos I figure the main drive factor for speakers are voltage
    variations (voltage are converted into air preassure by the speakers, and
    hence creating audible sound)?

    Further more, I figure that the input stage of the amps I want to build
    will simulate that of LINE-IN/AUX/CD-IN lines found on commercial sound
    systems. Anybody knows whether there are standards to the input voltage
    and input impendance of those lines (LINE-IN/AUX/CD-IN), and if there
    are, what are the values? And how about output impendance standards for
    driving 8ohms speakers?

    Also, anybody knows how the input parameters (voltage, current, power) to
    the speaker relate to the SPL (Sound Preassure Level) generated by it?
    This indirectly means I'm asking the relationship of the amplifier gain
    to the loudness of the sound generated by the speaker.

    Last, but not least, what are the recommended op-amp series for audio
    processing (I only know the multi-purpose u741 series)?

    Well, that's a long list of questions, and there are more. Any help is
    appreciated.

    TIA
     
  2. Yes, speakers are usually designed to be driven by voltage sources. A
    voltage source is one that regulates the voltage applied to a load
    (the speaker, in this case) while supplying whatever current is
    required to apply that voltage across the load. If the speaker was a
    resistive 8 ohms, that would mean that for every 8 volts the amplifier
    applied across the load, it would have to deliver 1 ampere through
    it. So just because it is a voltage source, does not mean that you
    can neglect the current side of the equation, it just means that
    voltage is what the amplifier regulates most precisely.

    One of the big differences between a typical opamp and a power amp is
    that the opamp has a maximum current capability somewhere between 10
    and 30 milliamps while a power amplifier that operates from the same
    supply voltage may have large enough output transistors to pass more
    than an ampere.
    There are integrated power amplifiers (essentially high output current
    capability opamps with some internal bias and gain setting resistors)
    that are available for very little cash. I think you might start with
    one of these for your experiments. The data sheets show the typical
    speaker amplifier circuit diagram and additional parts needed.

    Some examples:
    http://www.national.com/ds/LM/LM380.pdf (one of the most common)
    http://www.national.com/ds/LM/LM386.pdf (low voltage version of the
    LM380)
    http://rocky.digikey.com/WebLib/ST Micro/Web Data/TDA1905.pdf
    http://rocky.digikey.com/WebLib/Panasonic/Web data/AN7513.pdf
    http://rocky.digikey.com/WebLib/Panasonic/Web data/AN7511.pdf
     
  3. Graham Knott

    Graham Knott Guest

  4. At the high frequency end of the amplifiers capability, the speaker
    tends to look inductive (rising impedance with frequency), which
    raises the open loop gain, just as the phase shift of the amplifier is
    increasing and tending to decrease the stability of the closed loop.
    This network absorbs energy at this part of the spectrum, loading down
    the gain, so that the amplifier will not oscillate at a few hundred
    kilohertz, especially if you disconnect the speaker. At ordinary
    audio frequencies, it consumes very little power.

    See Zobel network on this page:
    http://www.rane.com/par-z.html
     
  5. Bob Masta

    Bob Masta Guest

    On 01 Jan 2005 17:04:46 GMT, Ricky Romaya <>
    wrote:

    To add to what others have said, the circuits for discrete power amps
    are basically the same as ordinary op amps, with the difference that
    instead of integrated low-power devices the power amp uses big
    discretes on heat sinks.
    Assuming the speaker impedance is constant with drive level (a
    reasonable assumption over normal ranges), then the power is
    proportional to the square of voltage just as if the speaker were
    a simple resistor: P = V^2 / R
    If you double the voltage, you get 4 times the power, which is
    +6 dB. Lower-impedance speakers (4 ohm nominal instead of 8 ohm, for
    example) will give you more power.

    SPL is an absolute measurement of sound power, so the efficiency
    of the speakers enter into it, as well as the geometry and the room.
    The sound measured at a particular point in space will be louder if
    the speaker is beaming most of its energy there, versus sending it
    off into other unmeasured directions. Speaker efficiencies are
    usually rated as the SPL output you get for a 1 watt input, measured
    at 1 meter from the cone. Or instead of 1 watt, they may use a
    particular voltage that gives a similar output. Typical values are
    around 90 dB SPL.

    Power amp gains are typically set so a "line level" of about 1 VRMS
    will drive them to their maximum output, which depends mostly upon
    the power supply rails... you size your devices to handle the rails.

    Hope this helps!







    Bob Masta
    dqatechATdaqartaDOTcom

    D A Q A R T A
    Data AcQuisition And Real-Time Analysis
    www.daqarta.com
     
  6. Thanks for the link, I've never heard that term before. I kinda figured
    it was being used to "dump" higher frequencies to ground, but didn't
    know if it was an oscillation preventer or just some kind of tilt
    compensation (though I would have expected that the amp would have lost
    gain as frequency increased anyway, so I didn't understand bypassing the
    higher frequencies to ground). I wasn't thinking about the changing
    (rising) impedance of the load causing an increase in gain and leading
    to potential oscillations.
     
  7. Ricky Romaya

    Ricky Romaya Guest

    Thanks for the pointer. I have more questions though.

    What is the use/significance, and how to use these pins:
    1. BYPASS pin on LM380/386
    2. FREQ COMPENSATION, RIPPLE REJECTION, and BOOTSTRAP pins on TBA20M
    3. BOOTSTRAP, TRESHOLD, MUTING, and SVR pins on TDA1905

    On TBA20M, there's 2 application circuits which shows 2 alternatives on
    connecting the load (i.e. speakers): to the supply or to the ground. Why
    two alternatives? What are the strengths and trade-offs of each
    alternatives? And does anybody care to point me to some web resources on
    how to drive a speaker (apart from the amplifier stage), what are the
    required interconnections and percautions, and how it works.

    This may be basic, but how to calculate the output power of an op-amp
    network? I mean, with those extra precautions in place (like the
    Boucherot's filter and DC gain compensation capacitors) things get a
    little hairy for me.

    Last question (for the moment), I really need to know the relationsip
    between input power (or voltage/current) to SPL to estimate how much
    power need to be fed from the amps to the speaker. I don't really know
    whether 1W, 2W, or 6W is enough for me. My only guide is by vague memory
    that the old SB16 cards uses 4W amps, and I'm not really sure.

    TIA
     
  8. THe internal bias network tries to set the output voltage to 1/2 of
    the supply, so that it can swing equally far in either direction. But
    if the supply has ripple voltage, this would also couple half of that
    to the speaker, producing hum. Once you decide which rail you are
    going to use as speaker return, you bypass the middle of the bias
    divider to that rail, to low pass filter the ripple to that reference
    point, so that the output is biased not to half the supply, but to
    half of the average of the supply over a time longer than the ripple
    period.
    Frequency compensation pin is a place to tie a bit of high frequency
    feedback from output to near the input to stabilize the amplifier as
    its phase shift increases as frequency rises. It also rolls off the
    high frequency response of the amplifier.

    Ripple rejection is similar to the bypass pin on the LM380.

    Bootstrap is positive feedback from output back to the output driver,
    to increase the output swing possible by adding to the positive supply
    during positive output swings. This allows a bigger swing from a
    given supply with lower distortion.
    Bootstrap same as for TBA820M

    Treshold (threshold) sets the logic switching threshold voltage for
    the muting function to match it to whatever logic is driving the
    muting function.

    Muting is an input that silences the amplifier (useful to get rid of
    the blast of noise an FM detector produces if no valid signal is
    detected.

    SVR (supply {ripple} voltage rejection) similar to bypass on LM380.
    I am not sure I have figured them all out, but the speaker to positive
    rail saves one capacitor (the bootstrap) since the speaker coupling
    cap can do both duties. But in many applications, you have a speaker
    that is already tied to the negative rail by some other requirement,
    so they show that configuration, also.
    Wire (twinned pair or twisted pair) rated for the current and voltage
    is all I usually worry about. What else do you have in mind? If the
    run is long, I use oversized wire, to keep total voltage drop a small
    fraction of the total.
    Instantaneous power to the speaker is amperes times volts. The
    speaker impedance tells you the ratio of volts per ampere.
    That depends on the efficiency the speaker and enclosure have at
    converting electrical energy into sound energy. If you are sitting
    right in front of a typically efficient speaker (say, a 6" by 9"
    automotive type speaker with big magnets in a cubic foot or so box), 1
    watt will shake your teeth a bit. If you use tiny computer speakers
    or want to fill a large room with sound, 1 watt will probably be weak.

    You almost have to start with the speakers and space and work
    backwards ot the power needed, then make sure those speakers can
    handle that power without distorting or burning up.
     
  9. Rich Grise

    Rich Grise Guest

    From my intuitive analysis of the two figures and the schematic,
    The chip would rather drive the speaker that's returned to the positive
    supply, because it just feels better that way. ;-)
    ....
    Any reasonable speaker, i.e. not an audiophool model, can fill a room
    uncomfortably loud with one watt. Well, maybe you need a really good
    enclosure. :)

    I vote for the TBA820M 'cause it's the coolest. :)

    Cheers!
    Rich
     
  10. Especially if you glue a piece of aluminum on top of it. ;-)
     
  11. Ricky Romaya

    Ricky Romaya Guest

    (Bob Masta) wrote in
    Care to elaborate on this? I've noticed that the power amp op-amps have
    their gain preset. Do you mean that to drive them to max output, you need
    1 VRMS input? Take TBA820M for example, it's gain is preset to 20 (if not
    mistaken), then a 1 VRMS input will give 20 VRMS output, and the power on
    the 8 ohm speaker is then (20^2)/8 = 50W? But it says that TBA820M is a
    2W power amp (on 8 ohm load and 12V rails).

    TIA
     
  12. Ricky Romaya

    Ricky Romaya Guest

    I mean what is the guidelines of connecting the amplifier stage to the
    speakers. What are the precautions. Like in text books (ones I read at
    least) when explaining op-amps, it only explains the op-amp in both open
    and close loop configuration. They seldom explains about the capacitors
    networks for unity gain only for DC range, or baucherot's filter for
    example. They only explains op-amps in terms of cute triangles with 2
    input pins and output pins, and resistor feedback networks. This made me
    confuse when comparing real schematics with what I get from the textbook.

    Yes, in theory the only thing needed to couple the amp stage to the
    speakers are a single wire. But from what I saw in the app circuits of
    the datasheets, there are capacitor networks involved as well. This is
    what I want to know, or if there are some resources out there explaining
    all the parts (this device/network do what) step by step of a simple
    power amp (like the app circuit of TBA820M).
    What I meant is does those precaution networks, like baucherot's, will be
    involved in the equation, cause the speaker is connected after it. By
    your reply above, it's still unclear to me, what voltage and amperes to
    use. Is it the input voltage times the op-amp network gain (for voltage),
    and the rated output current from the datasheet?

    TIA
     
  13. Those details are amplifier specific.
    Each amplifier design may have different requirements.
    Only for signals that don't saturate the amplifier, For a single
    supply amplifier, like the TBA820M, the largest voltage the amplifier
    can impress across the speaker is half the supply, minus the output
    device saturation voltage. A general estimate for the saturation
    voltage is about 1.5 volts. So if you power the amplifier with 12
    volts, the largest voltage it can impress across the speaker is about
    12/2-1.5=4.5 volts. If the speaker has an 8 ohm impedance, the
    implies a peak power of 4.5^2/8=2.5 watts. If that peak is part of a
    sine wave signal, the average power is half that. If the amplifier
    has a gain of 20, this implies that the output will clip if the peak
    input voltage is greater than 4.5/20=.225 volts.
     
  14. Bob Masta

    Bob Masta Guest

    Sorry for the confusion. I was speaking of discrete power amps,
    like for home stereo. But the numbers are about what you mention:
    a gain of 20 or so. Most home stereo discrete amps are in the 50-100
    watt range.

    Best regards,


    Bob Masta
    dqatechATdaqartaDOTcom

    D A Q A R T A
    Data AcQuisition And Real-Time Analysis
    www.daqarta.com
     
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

-