I too have a lot to learn, but our experiences are obviously different. My impressions of bipolar transistor amps involved thermal runaway, irritating TIM from the negative feedback, mediocre slew rate, RF oscillation, audible protection circuitry that disallowed maximum performance, and frequent self-destruction.
I have 12 of the biggest Soundcraftsmen FET amps, each with 24 big FETs and a switching-mode power supply that's fully regulated (and another audiophile FET amp, and several bipolar pro sound amps). If you short the speaker output of my FET amps with a strand of the speaker wire, the fans go on high, the house lights dim, and your electric meter spins fast until the stray wire strand lights up red and perhaps burns out. In 30 years heavy use with these I've never blown out an output transistor. They sure don't lack punch, more like full body-slam. But I'm learning that engineering is about trade-offs and compromises, making sacrifices in areas that matter less to you at the moment in order to maximize the benefits that match your project objectives. Designers might trade device linearity for speed if the fast device suffers less ill effects from more negative feedback, ending up with a more linear amp after correction, etc. Now that I don't so sound jobs, and I have power amps with 288 MOSFETS and hefty power supplies, perhaps some day I'll learn enough to be able to turn them into lower-power class A amps...but for now IMHO they sound pretty good in A/B to me.
Friends of mine with big bipolar Crown VZ amps have their own prejudices and favorites due to completely different experiences with economy FET amps.
We're all subject to our own attribution errors when we generalize, due to limited sample sizes within our experience. That's another reason I enjoy forums with people of broader (or deeper) experience than mine. It's like gaining the experience without the personal pain and expense.
I have 12 of the biggest Soundcraftsmen FET amps, each with 24 big FETs and a switching-mode power supply that's fully regulated (and another audiophile FET amp, and several bipolar pro sound amps). If you short the speaker output of my FET amps with a strand of the speaker wire, the fans go on high, the house lights dim, and your electric meter spins fast until the stray wire strand lights up red and perhaps burns out. In 30 years heavy use with these I've never blown out an output transistor. They sure don't lack punch, more like full body-slam. But I'm learning that engineering is about trade-offs and compromises, making sacrifices in areas that matter less to you at the moment in order to maximize the benefits that match your project objectives. Designers might trade device linearity for speed if the fast device suffers less ill effects from more negative feedback, ending up with a more linear amp after correction, etc. Now that I don't so sound jobs, and I have power amps with 288 MOSFETS and hefty power supplies, perhaps some day I'll learn enough to be able to turn them into lower-power class A amps...but for now IMHO they sound pretty good in A/B to me.
Friends of mine with big bipolar Crown VZ amps have their own prejudices and favorites due to completely different experiences with economy FET amps.
We're all subject to our own attribution errors when we generalize, due to limited sample sizes within our experience. That's another reason I enjoy forums with people of broader (or deeper) experience than mine. It's like gaining the experience without the personal pain and expense.
One of the things not mentioned is not directly related at all to the class of operation. That is the spectra of distortion products. Low absolute values of distortion are not always perceived by the ear in a positive way, although they well could be.
The classic example is guitar valve amplifiers.
The distortion they express on full power is loved by many rock guitarists.
Many people have tried to emulate that with fuzz/metal/distortion pedals.
The nearest I have found is a soft limiter which mimicks the valve distortion rounded off waveforms.
The distortion they express on full power is loved by many rock guitarists.
Many people have tried to emulate that with fuzz/metal/distortion pedals.
The nearest I have found is a soft limiter which mimicks the valve distortion rounded off waveforms.
Hi,
You can't bias class aB with MOSFETs as nicely
as you can bias it with BJTs, simple fact.
rgds, sreten.
You can't bias class aB with MOSFETs as nicely
as you can bias it with BJTs, simple fact.
rgds, sreten.
BJT output stage can be biased at any current.
There is a specific current that minimises Crossover Distortion.
That amplifier with a bias value for minimal crossover distortion is referred to as an Optimally Biased ClassAB stage.
Decrease or increase the bias current and the crossover distortion will increase.
A mosFET output stage does not have an equivalent to "Optimally Biased ClassAB"
The crossover distortion of a mosFET output stage gets less as the bias current increases.
That's why he said
But I don't know why he inserted "class aB".
That I don't understand.
D.Self was the first Designer that I became aware of that discussed this "special feature" of BJT output stages. That was about 20years ago.
It is still valid today.
There is a specific current that minimises Crossover Distortion.
That amplifier with a bias value for minimal crossover distortion is referred to as an Optimally Biased ClassAB stage.
Decrease or increase the bias current and the crossover distortion will increase.
A mosFET output stage does not have an equivalent to "Optimally Biased ClassAB"
The crossover distortion of a mosFET output stage gets less as the bias current increases.
That's why he said
But I don't know why he inserted "class aB".
That I don't understand.
D.Self was the first Designer that I became aware of that discussed this "special feature" of BJT output stages. That was about 20years ago.
It is still valid today.
Last edited:
- Status
- Not open for further replies.