Can a very simple class A amp produce great sound?

[Plecto]

I've been wondering about this. I've been reading about mosfets at university and I think I have a decent understanding of them. Take this class A amp circuit as an example:

http://i50.tinypic.com/10hugp5.png

Don't mind any of the component values. When I look at many class A amp designs out there they seemed to be littered with components and I often see a minimum of 5 transistors used in all sorts of ways, but I'm wondering why? I can see some problems with the circuit I linked though like the output cap and current source, but if we assume an ideal output cap and an ideal current source, what is flawed with this design? What kind be done to further increase the quality of the sound?

[Jay]

Originally Posted by Plecto

I've been wondering about this. I've been reading about mosfets at university and I think I have a decent understanding of them. Take this class A amp circuit as an example:

http://i50.tinypic.com/10hugp5.png

Don't mind any of the component values. When I look at many class A amp designs out there they seemed to be littered with components and I often see a minimum of 5 transistors used in all sorts of ways, but I'm wondering why? I can see some problems with the circuit I linked though like the output cap and current source, but if we assume an ideal output cap and an ideal current source, what is flawed with this design? What kind be done to further increase the quality of the sound?

What is the load requirement? 100 Ohm? Is that a headphone?

There are many already. Search for Pass Labs' Bride of Zen.

With one device, you rely on device transconductance, limited by input capacitance to charge. For speaker amplifier (around 4W) search for ZCA.

When you can accept the limited power of a single device (and the associated high output impedance), what you need is to work out the bad power supply rejection. Usually the power supply needs to be complex.

[Bigun]

search for the SEWA amplifier, I heard it is a great sounding amp. Needs a pre-amp.

[Ken Newton]

Appears okay on first glance, except you would likely want to insert a mosfet source degeneration bias resistor of appropriate value.

[Plecto]

Well I was thinking of an amp for hifi speakers so 4-8Ohms then I guess. I'm not here to find a design I could make, I'm here to learn about the theory behind it and if/why a more complex design is better.

I'm aware of the bad PSRR, but won't a voltage regulator massively improve that? Heck, two voltage regulators in series would give a ridiculous PSRR, won't it? But what if we assume that the power supply is an ideal DC voltage source aswell?

[Plecto]

Appears okay on first glance, except you would likely want to insert a mosfet source degeneration bias resistor of appropriate value.

Thats to get a little feedback in order to keep the operating point steady, right? Won't having R3 connected to drain sort of do that? I don't understand why the amp would need feedback though, what would cause the operating point to change other than if the bias current were to change? But I'm assuming an ideal current source :P

[Jay]

Originally Posted by Plecto

Well I was thinking of an amp for hifi speakers so 4-8Ohms then I guess. I'm not here to find a design I could make, I'm here to learn about the theory behind it and if/why a more complex design is better.

I'm aware of the bad PSRR, but won't a voltage regulator massively improve that? Heck, two voltage regulators in series would give a ridiculous PSRR, won't it? But what if we assume that the power supply is an ideal DC voltage source aswell?

Class-A is not efficient. You may need current more than an LM317 can provide. Then you start to add bypass transistor, complicating the power supply...

If I want to create a single transistor amplifier, I will start with finding the suitable device for the job (high transconductance, low capacitance, sufficient current or power capability). Lateral fet (2SK1056 family) is known to have low capacitance and hence requires less current than VFET (IRFP240 family).

The higher the current, the better the quality of the sound (lower distortion, more linear). Then you need complex (read: big) heatsinking.

Everything else you seem already know how to do it, no?

[Jay]

Plecto, I believe that when you read the PDF documentation of BOZ along with other Zen amp documentations, you will not be back with similar questions

Thanks to Nelson Pass who made amplifier building so easy to understand.

[Plecto]

I have read the Zen amp documentation (that's the one that made me interested in this ) and the death of zen documentation. I haven't really found my answers in those papers though as they don't explain the reason for many amp designs to have 200 components in them. The guy who made death of zen seemed to believe that BJT's was more linear than mosfets, but i'm not sure how audiable this is?

The point of the thread isn't heatsinking, power dissipation or how the power supply is made, I'm just focused on the power stage that I linked in the opening post, this is purely theoretical.

So lets say I buy one of those mosfets with high transconductance, low capacitance and sufficient current or power capability and make the design I linked with an ideal current source, ideal DC voltage source and an ideal output cap. Would this then have a distortion of way under 1% and it would be good to go as a high-end amplifier with as good of a sound as you can ever get from an amp?

[AndrewT]

I may have missed it in earlier replies, but that Power Amplifier has a maximum output of 0.7W into an 8r0 load.

If you try to drive an ordinary 8ohms speaker at -20dB (average level ref maximum power) to avoid clipping then you will just about have to clamp the speaker to your ear to hear it. Am I allowed to exaggerate slightly?