AB-class amplifier schematic analysis & optimization

[ratatax]

Any audio-related discussion inevitably ends like that

[Zero999]

Any audio-related discussion inevitably ends like that

Yes it does get on my nerves how audio threads invariably turn to audiophoolery and trolling. I apologise for my part in it and am no longer responding to it.

I know your suggestions here may be using higher quality parts, but what about this one :

http://www.ti.com/lit/ds/symlink/rc4580.pdf

It's cheap, has some oomph to be used as a headphones amplifier, has low enough THD/noise... and the implementation seems dead easy.

I'm a musician but I like when choices are made with measurable, scientific proofs, since audio as it is perceived by our ears is way too much subjective. Some opamps with thousand of V/uS slew rate and hundred of Mhz bandwidth may sound very good but I feel it's just unneeded complications (unneeded costs too), like you can buy a race car to drive at 30mph in a city...but WHY ? 

That's a good idea. I'll add that to my list of decent op-amps.

How about the MC33078 or use two channels of the MC33079 in parallel?

The NJM2060 also looks good.

[Alex Nikitin]

Same, I am sorry for this off-topic, stops here.

Cheers

Alex

[David Hess]

driving a speaker coil is not equivalent of driving a resistive load, that is one of reasons why amplifiers with similar traditional measurements could sound different in real life

The reactive part of the load changes the phase relationship between the current and voltage which can be seen in a class-b design where the crossover distortion will move away from zero volts.  But in a well designed amplifier, it makes no practical difference assuming that the reactive load characteristics were designed for.

A more important effect is that the reactive load shifts the current to voltage across the output devices which can result in violation of their breakdown rating.  The protection circuits and output device derating need to take this into account.

True that, but after all those years of hearing about it I would like to finally see a demonstration of two amplifiers which both show no perceptible noise, 0.1dB gain flatness from 10Hz to 50kHz and some low THD figure, all measured into the actual demo speaker, and yet sound different enough that somebody (bonus: everybody) can tell them apart.

I'm not aware of any instances of such demonstration actually being done. I suppose it would be all over the Internet if it happened, given the amount of controversy.

I have seen differences in amplifiers with otherwise similar characteristics in two areas:

1. Harmonic and intermodulation distortion tests will not reveal low full power bandwidth which will create additional distortion in real content.  Production designs should never have this problem.

2. Overload recovery after clipping can be significantly different.  This is especially a problem with naive designs which use frequency compensation to control noise and bandwidth and I suspect this is where negative feedback in audio amplifiers gets a bad image.

As far as the like of rigorous scientific tests, maybe they are allergic?

[Alex Nikitin]

driving a speaker coil is not equivalent of driving a resistive load, that is one of reasons why amplifiers with similar traditional measurements could sound different in real life

The reactive part of the load changes the phase relationship between the current and voltage which can be seen in a class-b design where the crossover distortion will move away from zero volts.  But in a well designed amplifier, it makes no practical difference assuming that the reactive load characteristics were designed for.

A more important effect is that the reactive load shifts the current to voltage across the output devices which can result in violation of their breakdown rating.  The protection circuits and output device derating need to take this into account.

The main problem is not that loudspeaker is a reactive load (at least that is a reasonably measurable and easy to deal with effect) , but that it is a very non-linear load, and not only load but a source as it acts as a microphone picking up vibrations and sounds. All this rubbish gets back to the amplifier and enters the feedback loop, as well as other drivers in a multi-way speaker.

Cheers

Alex

[David Hess]

The main problem is not that loudspeaker is a reactive load (at least that is a reasonably measurable and easy to deal with effect) , but that it is a very non-linear load, and not only load but a source as it acts as a microphone picking up vibrations and sounds. All this rubbish gets back to the amplifier and enters the feedback loop, as well as other drivers in a multi-way speaker.

I have difficulty imaging that being a problem with the low dampening factor at the feedback point of a solid state amplifier combined with the cable and speaker coil resistance.

Speakers are specifically designed to be linear because otherwise excessive intermodulation distortion is produced by the speaker itself.  This takes the form of using a narrow voice coil within a large field of magnetic flux or a wide voice coil with a narrow field of magnetic flux.  Woofers present the most difficulty because they must support a large throw to produce good loudness at low frequencies and base-reflex enclosures help considerably in this respect by lowering the impedance seen by the driver.

Interaction between drivers should be prevented with active or passive crossovers.

But I would agree that a lot of speakers intended for the consumer market are trash.  But hey, the customer knows what sound they want.

[dietert1]

Yes, i agree there is a huge difference between somebody ignorant using some sound system and a technical person using a sound system designed or combined by himself. Sound causes feelings! I know lots of well educated technical people who go crazy about horn speakers and the selection of the right foil capacitor.
But human hearing is extremely flexible.
Think about the differential analog transmission still used in professional audio. The input stage can be made such as to yield -6 dB when one of the two contacts in the connection fails. Or it can be made to yield -60 dB in that case. Now you may wonder how many of those connections in the real world are jumping between 0 dB  and -6 dB without anybody taking notice.

Regards, Dieter

[Zero999]

The main problem is not that loudspeaker is a reactive load (at least that is a reasonably measurable and easy to deal with effect) , but that it is a very non-linear load, and not only load but a source as it acts as a microphone picking up vibrations and sounds. All this rubbish gets back to the amplifier and enters the feedback loop, as well as other drivers in a multi-way speaker.

I have difficulty imaging that being a problem with the low dampening factor at the feedback point of a solid state amplifier combined with the cable and speaker coil resistance.

Speakers are specifically designed to be linear because otherwise excessive intermodulation distortion is produced by the speaker itself.  This takes the form of using a narrow voice coil within a large field of magnetic flux or a wide voice coil with a narrow field of magnetic flux.  Woofers present the most difficulty because they must support a large throw to produce good loudness at low frequencies and base-reflex enclosures help considerably in this respect by lowering the impedance seen by the driver.

Interaction between drivers should be prevented with active or passive crossovers.

But I would agree that a lot of speakers intended for the consumer market are trash.  But hey, the customer knows what sound they want.

Yes, speakers and room acoustics are the weakest links in the chain as far as distortion is concerned and dominate over even mediocre amplifiers. The audio amplifier is a problem which has been solved a long time ago. The only improvements made recently involve reductions in sized and increased power efficiency.

Headphones are better than speakers distortion-wise, because they're so small and the path between the ear drum and transducer is short (especially for in ear buds), transmission line effects are far less apparent, so it can be treated as a diaphragm changing the pressure levels inside the ear canal, especially at lower frequencies.

Try this. Wearing a decent pair of headphones, put some music on with a decent bassline and set the volume to a safe, comfortable listening level. Notice the deep thumping bass. Now take the headphones off, hold them around 30cm away from your ears and crank up the volume until you can probably hear the music. Notice how tinny they sound: the bass has magically gone! To get bass, there needs to be a decent seal between the headphones and the ear drums, otherwise the changes in pressure which produce the bass escape and are lost.

Yes, i agree there is a huge difference between somebody ignorant using some sound system and a technical person using a sound system designed or combined by himself. Sound causes feelings! I know lots of well educated technical people who go crazy about horn speakers and the selection of the right foil capacitor.
But human hearing is extremely flexible.
Think about the differential analog transmission still used in professional audio. The input stage can be made such as to yield -6 dB when one of the two contacts in the connection fails. Or it can be made to yield -60 dB in that case. Now you may wonder how many of those connections in the real world are jumping between 0 dB  and -6 dB without anybody taking notice.

Regards, Dieter

Sound causes feelings, so many of the perceived differences in amplifiers are often down to the placebo effect, rather than anything real. There is nothing wrong with this, until people start to make extraordinary claims, which cannot be proved.

[magic]

It is getting completely off-topic, however I will only ask a simple question: can you provide a link to an example of a statistically valid controlled DB listening test with a positive result?

I can't because I don't care, but of course such tests must have been done if there are people talking about "scientifically established" thresholds of audibility of various defects.

But that's not the point, I don't really insist that they were valid or anything. I only remarked about the suspicious tendency of audiophiles to show up everywhere and say a lot about what they hear while making zero effort to convince that what they hear is determined by the actual sound rather than by their beliefs about what produced the sound.

Say that I swapped a few paralleled NE5532 into one of those LM7171 amplifiers of yours, would you tell it apart?
After all those years, somebody could perhaps show something like that. Maybe a vendor looking for publicity, maybe hobbyists looking to stick the finger to nwavguy and his religion, anyone

And the problem of people hearing their beliefs rather than sound is obviously real. Gold plated USB plugs, compact disk demagnetizers, NE5532 relabeled as LM4562, all that junk sells and fetches positive feedback from satisfied customers.

[Zero999]

Let's not continue the off-topic stuff.

I'm surprise no one has done a simulation.

The quiescent current was around 25mA, which seems a little high, although may not reflect reality.

Overcurrent protection could easily be added by connecting some back-to-back diodes between the output (the emitters of Q1 and Q2) and the op-amp's output. This will limit the voltage drop across R7 and D8 to the diode's foraged voltage. Ordinary diodes will clamp at 0.6V, giving a current limit of 60mA, although there may be some distortion before then. The output current will be a little higher, than that because the op-amp's output current will be added to the output, which will make around 100mA, for the NE5532. A couple of red LEDs connected back-to-back or use a bipolar LED, will limit the current to a couple of hundred mA.

[David Hess]

Say that I swapped a few paralleled NE5532 into one of those LM7171 amplifiers of yours, would you tell it apart?

I could; the LM7171 will produce noticeably more hiss because of its greater input noise.  This might not be noticed in a noisy environment however.

[magic]

I will add an LM7171 amplifying its own noise and mix it with the input signal, just for you

[David Hess]

I will add an LM7171 amplifying its own noise and mix it with the input signal, just for you

The noise (hiss) only bothers me (becomes noticeable) with no input signal.

Another configuration is to use the low noise audio operational amplifier, NE5532 or whatever, to drive the LM7171 which acts as the high current output stage.  Now the LM7171 can operate at lower gain and its noise is suppressed by the excess gain of the NE5532's feedback loop.

The noise of an LM317 regulator can be suppressed using the same type of circuit.

[Yansi]

317 is quite low noise by itself. Doubt it will have any major contribution to the amp output noise.

[Zero999]

Say that I swapped a few paralleled NE5532 into one of those LM7171 amplifiers of yours, would you tell it apart?

I could; the LM7171 will produce noticeably more hiss because of its greater input noise.  This might not be noticed in a noisy environment however.

With a gain of 23? Possibly if you have some very sensitive earbuds.

In all fairness the proponent for the LM7171, as a headphone amplifier, didn't give any details regarding their design i.e. the gain.

The LME49720 seems like a good candidate for a headphone amplifier, but has a relatively high input current noise, so will only be any good, if the source impedance is low. It also has a large enough bandwidth to cause problems, if insufficient care paid to the layout.
http://www.ti.com/lit/ds/symlink/lme49720.pdf

[Zero999]

I added the current limiting diodes and reran the simulation. The actual limit is a bit higher than I predicted at 130mA. Note that there's a point when the limiting kicks in where the op-amp is contributing a significant amount to the output current: see the Ix(U1:OUT) trace. It's a good idea to keep the current below this level, even if it isn't clipping, as there will be some distortion.

[David Hess]

317 is quite low noise by itself. Doubt it will have any major contribution to the amp output noise.

It would not matter for audio applications unless the amplifier had an unusually low power supply rejection.  I just mentioned it as an example of the same idea and some applications do require lower noise than an LM317 can provide alone.

[ratatax]

Thanks a lot for the simulated schematic, i'll try LTSpice to see what I can get from it