Connect with us

Class AB Power Amplifier Design

Discussion in 'Audio' started by omartarek206, Jun 25, 2019.

  1. omartarek206

    omartarek206

    1
    0
    Jun 25, 2019
    This is my first time doing electronics projects..
    The project is a three stage power amplifier:
    1. CE pre-amplifier stage
    2. CC Buffer stage
    3. Class AB power amplifier
    I did DC analysis to obtain almost every resistance in the circuit.. and the values of the capacitors were already given.
    My question is are there any design rules for the power amplifier,
    like the (1/3 + 1/3 + 1/3) rule or should this rule be applied to the amplifier in the first place.
    In short, are there any rules to calculate the values of RB3a & RB3b or should I assume them to be equal as I have seen in many circuit designs.
    **Excuse me for any mistakes.. this is my first time.

    upload_2019-6-25_19-41-50.png
     
  2. Audioguru

    Audioguru

    2,550
    576
    Sep 24, 2016
    1) Your transistors Q3 and Q4 are tiny and will be overloaded when they drive a 4 ohms load with a 9V supply. Their current will try to be 3.5V/4.5 ohms= 778mA but their datasheet shows that they work poorly above 200mA. They will also overheat.
    2) Transistor Q2 usually replaces your RB3a. Then RB3b needs a value that will produce enough base current to Q3.
    3) RB3 is usually replaced with a resistor from between the output emitter resistors to provide AC and DC negative feedback.

    I found a better amplifier circuit like my text above but it is old and uses older transistors. Newer medium power transistors and the first transistor will have higher current gain so that higher resistor values can be used.
    Your first transistor stage is fine as a microphone preamp.
     

    Attached Files:

    omartarek206 likes this.
  3. duke37

    duke37

    5,211
    718
    Jan 9, 2011
    Q3 and Q4 are emitter followers and the emitters should be close to half the supply voltage so that clipping will be symetrical for maximum power output. RB3b and RB3a will therfore need to be the same.
    The resistors should be low enough to give sufficient current to drive Q3 and Q4.

    RB2 controls the current in Q2 as well as acting as a tone control. Q2 should have a stable current.
     
    omartarek206 likes this.
  4. Sunnysky

    Sunnysky

    463
    114
    Jul 15, 2016
    upload_2019-6-26_22-31-18.png
     
  5. aurelZ

    aurelZ

    56
    7
    Jun 10, 2019
    16 W RMS on 12 V with 112 uA
    that is not possible..maybe 1.2 A
    ok i see it is simetric power supply then is 24V
    but then should get cca 25W ?
     
  6. Audioguru

    Audioguru

    2,550
    576
    Sep 24, 2016
    16W RMS into an 8 ohm speaker requires a signal that is 11.3V RMS and a current that is 1.42A RMS. A class-AB amplifier is about 55% efficient so when the output power is 16W then the wasted heating power is 13W so the total power from the power supply is 16W + 13W= 29W.

    11.3V RMS is (11.3 x [root of 2] x 2=) 32V peak-to-peak. An audio amplifier does not have a rail-to-rail output voltage swing so the power supply voltage must be about 36VDC. The maximum continuous power supply current must be 1.9A. That is for only one channel.

    If the power supply voltage is only 24V then the amplifier output voltage swing will be about 20V peak-to-peak which is 7.1V RMS which produces a power in an 8 ohm speaker of (7.1V squared)/8 ohms= 6.3W.
    If you want the power to be 25W into an 8 ohms speaker then the power supply voltage must be about (14.2V RMS x [root of 2] + 6=) about 46VDC.
     
  7. aurelZ

    aurelZ

    56
    7
    Jun 10, 2019
    Hello Audioguru and thanks for calculation!
    it is very informative and good to know.
    I suspect that might be like that .
    Anyway
    Is there a difference in class AB with bipolars vs MOSFET?
    I heare some MOSFET amps which are promoted as better than bipolar power transistors versions
    but they sound little bit strange to me
     
  8. BobK

    BobK

    7,599
    1,641
    Jan 5, 2010
    Mosfets do not have a very well controlled Vgsthreshold, so it you cannot bias the output stage the using two diodes like you can with BJTs. A pot to adjust the voltage between the two gates can overcome this problem, but it must be set manually for each one you build.

    Bob
     
  9. Audioguru

    Audioguru

    2,550
    576
    Sep 24, 2016
    Mosfets need a much higher gate-source voltage (about 6V) than a bipolar transistor base-emitter (about 1V) so the power supply voltage must be about 10V higher for the Mosfets amplifier compared to the bipolar transistors amplifier for the same output power.

    A class-AB audio amplifier made with Mosfets or bipolar transistors is not supposed to sound strange. They should both have the same flat frequency response and same very low distortion. Maybe you heard a Mosfets amplifier that was poorly adjusted so it produced crossover-distortion.
     
  10. hevans1944

    hevans1944 Hop - AC8NS

    4,201
    1,988
    Jun 21, 2012
    If beginning a design from scratch, consider using BJTs for class AB linear operation to obtain low distortion but at relatively modest power output and poor efficiency. Eliminating cross-over distortion can still be a tricky design problem so a distortion analyzer and perhaps an audio spectrum analyzer could be of some help during the testing and bias adjustment phase.

    Analog audio basically became obsolete near the end of the previous century when Detroit, and other world-class automakers, abandoned it for digital techniques to be used in automobile sound systems. There is not a larger mass market for audio products than automobiles, so follow the money.

    I only pursue analog as a hobby now, sometimes reverting back to vacuum tubes just because I like the soft glow of the filaments and the distinctive odor of this type of electronics as it "warms up." That's probably dust cooking off, but I think it is a pleasant smell, maybe because it reminds me of early childhood winter experiences, sitting between a red-hot, coal-fired stove and grandmother's short-wave/broadcast band radio, occupying a place of honor in her living room just off Bull Creek in the mountains of McDowell county West Virginia, listening then to what are now politically incorrect serial westerns and comedies. Um, you betchum, Red Ryder.

    For really significant power (hundreds to thousands of watts) with high efficiency, consider using power MOSFETs in class D operation with "carrier" frequencies above 60 kHz. Good magnetics are required to filter out the pulse-width modulated carrier and recover the audio, but it is well worth the effort if power output is your main objective, as it would be (for example) in stage-amplifier design for performances before large audiences of thousands of people. There should, of course, be zero cross-over distortion with class D amplifiers, but there are other problems to solve, just to make things interesting.

    @omartarek206, you appear to mostly know what you are doing. Coming here to smooth off some of the jagged edges of your design with knowledgeable advice from @Audioguru, @Sunnysky and @BobK is a good idea. So take what you need and leave the rest, but most of all: Have fun!
     
  11. dave9

    dave9

    738
    166
    Mar 5, 2017
    ^ There are lots of analog audio products out there, produced new today, just not so much using discrete amp stages. Just the other day I bought a Ryobi P746 boombox using a pair of YD2030 (TDA2030 equivalent).

    I wouldn't call analog obsolete. Yes it's less efficient, but power here is inexpensive and I don't play audio loud enough for the inefficiency to amount to more than a few watts. I would have preferred the Ryobi boombox be class D because it's running off a battery, but the runtime is still acceptable.

    Things powered by a vehicle alternator or mains AC, don't concern me much. If anything I expect an analog amp to function without failure for decades longer... just about anything with digital switching power, built within a consumer grade product BOM limitation, tends to have problems within the first couple decades.

    Analog still has a few things going for it. Simple and inexpensive (under some wattage threshold, say 100W if you have a suitable transformer already) to make, using little more than a single IC and a few resistors and capacitors that many of us have lying around. Minimal PCB layout concerns beyond decoupling and avoiding ground loops, minimal concerns about EMI. These are not so much issues for a mass produced product with subsequent testing and redesign, but for a homebrew hobbyist, they just want to bang something together and have it work on the first try.

    Some feel they sound better, but that's a whole other can of worms and impossible to quantify.

    There is obviously still a significant retail presence for analog amps, or else the major players wouldn't still be making the chips. ST or TI wouldn't keep a factory line running to produce few pieces/year for hobbyist one-off projects. On the other hand it wouldn't surprise me if they do shut down these lines soon, with products using off-brand clone chips like a YD2030 instead of ST2030.
     
  12. aurelZ

    aurelZ

    56
    7
    Jun 10, 2019
    Some feel they sound better
    Yes I agree, and for 100W good analog amplifier you can find anything.
    And if you have enough enthusiasm you can make magnetostatic speakers to discover new sound!
     
  13. Audioguru

    Audioguru

    2,550
    576
    Sep 24, 2016
    Guess what? The old TDA2030 and the improved TDA2030A are already obsolete and Digikey has none.
    TI still make the LM1875 and LM3886 linear class-AB amplifiers.

    A Company Txxxx invented class-T digital amplifiers but the company went bankrupt and now clones are made in China.
     
  14. aurelZ

    aurelZ

    56
    7
    Jun 10, 2019
    ...and now clones are made in China
    and one reason more to build amplifier with components not with IC
     
  15. Audioguru

    Audioguru

    2,550
    576
    Sep 24, 2016
    ICs are fine if you check to see that they are still being made and when you do not buy cheap fakes from China (Ebay, Banggood, AliExpress, Amazon, etc).

    On Elliot Sound Products website there are many amplifier circuits, most made with components. Years ago he used excellent Toshiba output transistors but here is what he says about them today:
    "Note: It is no longer possible to recommend any Toshiba devices, since they are the most commonly faked transistors of all. The 2SA1302 and 2SC3281 are now obsolete, and if you do find them, they are almost certainly counterfeit, since Toshiba has not made these devices since around 1999~2000."
     
    hevans1944 likes this.
  16. hevans1944

    hevans1944 Hop - AC8NS

    4,201
    1,988
    Jun 21, 2012
    I did not say, and I hope no one thought I implied, that analog (electronics) was obsolete. The real world (except for certain quantum mechanical interactions) requires interfacing electronics with analog signals. It's just not done (much) anymore with discrete components. Certain vacuum tube power components are still being manufactured and are in use: Magnetrons for RADAR and microwave oven use, klystrons for UHF and microwave transmitters and receivers, TWTs (Traveling Wave Tubes) for microwave power amplifiers, BWOs (Backward Wave Oscillators) for microwave transmitters, water-cooled triode, tetrode, and pentode vacuum tubes when ten kilowatts or more plate power dissipation is required, x-ray tubes in sizes suitable for small dental x-ray machines to very large (and mostly classified) x-ray tubes used to verify the operational condition of our nation's stockpile of nuclear weapons. I've been involved with all of these at one time or another during my career working for defense contractors. Analog is NOT obsolete. Vacuum tubes are NOT obsolete. Analog engineers and technicians are NOT obsolete.

    Commercial applications of analog technology may still linger as niche applications, but the heyday of their use has long passed. Have any of you old farts received your digital aural implants yet? Rush Limbaugh may have been the first person to nationally inform the public of the availability of cochlear implants to restore hearing, but there have been other interesting developments in prosthetic research. Research in this area requires familiarity with both analog and digital circuit design and application.

    I began my career as an analog experimentalist and hobbyist, sometime around 1952 or 1953 when I started living temporarily with my grandparents. I began reading everything I could find pertaining to electricity and, later, electronics to try to understand all the neat and wonderful things grandfather brought with him to Morristown, Tennessee, after he retired from a career as a deep-coal mining electrician near Welch, West Virginia... motors, transformers, selenium rectifier stacks, Edison cells for miner's lamps, carbide miner's lamps, all manner of "mysterious things" to whet an eight or nine year old's appetite for adventure. And Grandfather let me play with it all, and tried to pass on his knowledge of electricity. Unfortunately, Grandpa knew diddly about AC, so I didn't hear about Tesla until I was in high school.

    @Audioguru is spot-on about the difficulty of obtaining genuine parts from Asia, especially from China, who has the capability, and the will, to reverse-engineer and counterfeit just about anything. This isn't always a disaster: some of the "cloned" parts work as well as or better than the originals. Problem is, there is no way that I know of to determine which ones are "gud enuf" and which ones are a total POS to be avoided at all cost.

    Lest your expensive satellite transponder and its electronics fail soon after (or during) launch into orbit, purchase name-branded parts from reliable, vetted, distributors. A money-back guarantee is worthless if the failed part cannot be retrieved for replacement. With China you may get high volume, low cost, or quality control. Pick two. Then read your contract carefully.
     
  17. aurelZ

    aurelZ

    56
    7
    Jun 10, 2019
    wall of text should help?
     
  18. WHONOES

    WHONOES

    653
    131
    May 20, 2017
    Seems like OMARTAREK206 fell into the trap of trying to do it all at once. He needs to break his circuit down into small lumps (as you would any design) and get each bit working before joining them al together.
    Perhaps he should start with a simple one transistor class A amplifier and understand what is happening and why. I'm sure we would all be only to pleased to provide appropriate guidance.
     
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

-