High Performance Class A amplifier

[Aniket]

Hello Aniket,

Great work!
Simulation, design, layout, component procurement, assembly, testing and sharing.
This is much more than just DIY, its NPI + sharing!
Again, good job!!

Class-A has minimal to no crossover distortion. But it has two pain areas: low efficiency and manual setting of quiescent current. The latter is the case with any discreet design, anyway.

Have you looked into Composite Amplifier Design?
It uses a very low distortion op-amp for the input stage and a chip-amp for the output stage.
The composite amplifier then gets all the good parameters of the op-amp, like the vanishingly low distortion and high input impedance for the input side and all the required parameters like current gain and low output impedance of the chip-amp, for the output side.
So effectively you end up with best of both worlds without the two pain areas mentioned above.

If good number of people here show interest, we can put together something to begin with.

Again, good job with your Class-A build.

Thanks Ravindra,

I have designed and simulated composite amplifier much time earlier, but we can of course design better using discrete devices. some really great BJT devices available now, we can achieve ever better performance and distortion figures.
But, if people show interest we can sure design a composite Class AB amp with discrete output stage. Chip amp for output stage is a no no. they can never match the power handling, speed and gain of modern power transistors.

Regards,
Aniket

Aniket,
Do you see any advantage in a class A amp powered by tracking power supply?

It could reduce consumption drastically and raise efficiency and still be in class A all the time

Hi,

Tracking power supply would be more suitable in line level stages. However, in a push pull Class A amp or a Class AB, the law of diminishing returns on you. the performance gain is marginal and not worth the added complexity and cost. A better approach could be a modulated power supply which varies the rail voltages in accordance with the output voltage of the amp. XTI series from Crown uses this technique.

Regards,
Aniket

 

[diyaudio]

I meant a tracking supply for current, not voltage.

 

[diyaudio]

Class A is biased very high in current, the idea was to bring it down at idle. When signal demands the current could be raised

 

[diyaudio]

This pave way for higher wattage class A amps without heating the room

 

[surfatwork]

Interesting thought. Maybe it is possible to have an input signal sense circuit and move the bias point when an input signal is present. Perhaps a simple transistor based switch which opens an alternate bias current path when turned on by an input signal.

 

[diyaudio]

The bias current at idle should just bring the transistor to conduction. Then based on input signal it must linearly increase.

 

[diyaudio]

I think there was a patent or even a single chip based on such a principle, trying to locate it...

 

[surfatwork]

The outout swvices have to be biased by one leg of the circuit into conduction - this means it is a Class A. I meant for a separate transistor switch which will sense an input signal, get turned on (or off) and increase the bias current in that altered state. A parallel bias current leg which is available only when an input signal is available.
It could be as simple as a transistor in series with a resistor with the base connected to the input signal sense circuit, the whole thing in parallel to the output device biasing resistor. So when the transistor turns on, it lowers the overall resistance thus increasing the bias current.
The above is theoretical - will need one of our design gurus to reflect on it further to make a workable circuit.

 

[Aniket]

I meant a tracking supply for current, not voltage.

It could be implemented. But then you have to use separate power supplies for output and preceding stages. you have to vary the current going to the output stage only according to the demand. But, still you have to set a threshold setting of quiescent current for the desired amount of power in Class A. lets suppose for 20W in Class A, you would require ~1.5A continuous from PSU. and if it is limited to say 0.3A, the rails will eventually sag and lots of ripple and noise could creep in. add to this the added complexity and finicky just for modulating the current of the output stage. which will eventually steal the sonic character and authority of a Class A amp. just my opinion.

Class A is biased very high in current, the idea was to bring it down at idle. When signal demands the current could be raised

Interesting thought. Maybe it is possible to have an input signal sense circuit and move the bias point when an input signal is present. Perhaps a simple transistor based switch which opens an alternate bias current path when turned on by an input signal.

Well, actually i have incorporated sort of a dynamic bias control circuitry in this design. At idle, bias is set for 5W in class A, so dissipation is low at idle and as the output increases it dynamically increase the bias to the max rated 50W in Class A. after that the amp transitions into Class AB. Maintaining thermal and bias stability here is critical and depends a lot on the devices and layout.

To me the best solution to minimize heat is to rather make a Class AB amp, and bias it a little higher, so that it delivers the first 3-5W in class A and the rest in Class AB

 

[Professori]

@Aniket
Yes, I'd seen that, but the schematic or the other details were not posted; I had thought you were waiting for a breather before posting a detailed look at the project.
Would appreciate the same.
Congrats again for a great (sounding) project!

-- UKP

 

[diyaudio]

It could be implemented. But then you have to use separate power supplies for output and preceding stages. you have to vary the current going to the output stage only according to the demand. But, still you have to set a threshold setting of quiescent current for the desired amount of power in Class A. lets suppose for 20W in Class A, you would require ~1.5A continuous from PSU. and if it is limited to say 0.3A, the rails will eventually sag and lots of ripple and noise could creep in. add to this the added complexity and finicky just for modulating the current of the output stage. which will eventually steal the sonic character and authority of a Class A amp. just my opinion.

I get it, its a diminishing return.

 

[diyaudio]

Well, actually i have incorporated sort of a dynamic bias control circuitry in this design. At idle, bias is set for 5W in class A, so dissipation is low at idle and as the output increases it dynamically increase the bias to the max rated 50W in Class A. after that the amp transitions into Class AB. Maintaining thermal and bias stability here is critical and depends a lot on the devices and layout.

To me the best solution to minimize heat is to rather make a Class AB amp, and bias it a little higher, so that it delivers the first 3-5W in class A and the rest in Class AB

Any particular reason for transition power points to be be 5W/50W? Do you see any sonic drawbacks of reducing them to 2W/10W or something like that? Especially with higher sensitivity speakers/drivers, it would be indistinguishable sonically.

 

[Aniket]

@Aniket
Yes, I'd seen that, but the schematic or the other details were not posted; I had thought you were waiting for a breather before posting a detailed look at the project.
Would appreciate the same.
Congrats again for a great (sounding) project!

-- UKP

Thanks UKP,

some time back, one of my design was copied by eight audio guys and they are selling it commercially. So, I refrain from posting schematic and pcb layout here. any particular details you need I would share.

Regards,
Aniket

 

[Aniket]

Any particular reason for transition power points to be be 5W/50W? Do you see any sonic drawbacks of reducing them to 2W/10W or something like that? Especially with higher sensitivity speakers/drivers, it would be indistinguishable sonically.

Not any, only the end user required 50W max in Class A. I set the lower threshold at 5W, as even with moderate sensitivity speakers 5W is sufficient enough for casual home listening. Reducing to 2W/10W wont affect the sonic character much. Only the crossover distortion would increase marginally. Higher the efficiency of speakers, the better.

Regards,
Aniket

 

[Professori]

Thanks UKP,

some time back, one of my designs was copied by eight audio guys and they are selling it commercially. So, I refrain from posting schematic and pcb layout here. any particular details you need I would share.

Regards,
Aniket

I see...
Eight audio is seen purveying their kits online and also in sites like diyaudiocart. I never thought they would steal things. The decent thing is first to acknowledge the designer, and also to get his/her permission for marketing, and if the sales really take off, to share a percentage with the designer. There are copycats everywhere... :-(

I would like to see the circuit and also follow the design highlights so that I could, later, decide whether to try it or not. Under the circumstances, I cannot ask you for the same. Too bad... Let us wait for a day when you could share it at least in a closed circle. I am not too worried as I have my hands full at the moment with various projects in different stages of completion!

I had wanted to try some lateral mosfet amplifiers, and I happened across the PeeCeeBee in Diyaudio (hope you are familiar with Mr Shaan's effort) and, also have asked Mr Sadik to send me his latest Hiraga pcb, which you have designed and you two, I understand, will be organizing a GB. I have a couple of other projects too...enough to keep an old man busy and dreaming!

There are very enterprising people around who make good money selling for example, Elektor magazine kits etc in my place too. And most of the present breed of home theatre guys over here are the ones who consume any latest thing that they could sell. Some of these guys also lurk around in the forums as genuine enthusiasts, but their intentions are totally different. What to do.. such are the times.

What is the progress you have made regarding the Hiraga researches, esp. in relation to the use of CCS, Kubota or other regulators for the input side etc?
Anyway, let us be in touch and share ideas, which in itself is a happy and fulfilling exercise.

Warm regards,

--UKP

 

[Aniket]

I would like to see the circuit and also follow the design highlights so that I could, later, decide whether to try it or not. Under the circumstances, I cannot ask you for the same. Too bad... Let us wait for a day when you could share it at least in a closed circle. I am not too worried as I have my hands full at the moment with various projects in different stages of completion!

I had wanted to try some lateral mosfet amplifiers, and I happened across the PeeCeeBee in Diyaudio (hope you are familiar with Mr Shaan's effort) and, also have asked Mr Sadik to send me his latest Hiraga pcb, which you have designed and you two, I understand, will be organizing a GB. I have a couple of other projects too...enough to keep an old man busy and dreaming!

There are very enterprising people around who make good money selling for example, Elektor magazine kits etc in my place too. And most of the present breed of home theatre guys over here are the ones who consume any latest thing that they could sell. Some of these guys also lurk around in the forums as genuine enthusiasts, but their intentions are totally different. What to do.. such are the times.

What is the progress you have made regarding the Hiraga researches, esp. in relation to the use of CCS, Kubota or other regulators for the input side etc?
Anyway, let us be in touch and share ideas, which in itself is a happy and fulfilling exercise.

Warm regards,

--UKP

Hi UKP,

I would share you the details on mail regarding this 50W class A amp.
you can of course try PeeCeeBee, excellent work by Shaan.
Regarding Jean Hiraga mods, I just added simple active CCS for the input stage. that will take care of the noise. rest there isn't much to improve. although we can make a much better amp and complex amp out of it. But then it would no longer be Jean Hiraga.

To be honest, I am not a big fan of Class A amps, I designed it as it was required. I would prefer a higher biased Class AB amp instead.
BTW, i am working on some new Class G and Class H projects. 800W and 1500W +


Regards,
Aniket