SE ClassA with low dissipation @ idle - is this it ?

[Bernhard]

Hi,

in this single ended classA circuit the lower part is a dynamic current source with 100mA or 200mA idle.

So power consumption of the amp is low.

The voltage follower fet is cascoded by the upper fet.

As we know, voltage follower fets like to conduct current

1A sounds much better than 100mA

The only little problem is that idle consumption will be 10 times higher

So there is a 6V floating psu connected to the voltage follower fet, this fet will conduct always about 1A current.

The idle current trough the current source & cascode fet is 200mA, while through the voltage follower fet it is 1A

Also the diode of the floating psu will never switch off



Do not look at the lower part of the circuit, the dynamic current source - quick and dirty - is voltage controlled, I know should be current controlled, no problem.

Also there is a dead part with another fet on both sides, do not look at that...

This works on the sim.

Also the voltage follower fets see a relative low and constant drain source voltage and thus do not need to be paralleled even @ higher output powers.


Any opinions ???


Greetings, Bernhard

[uli]

Looks like those old thing Technics Class AA or so.
I would not waste time and build such an amplifier
sounding like any Japanese 500 E (or USD) integrated amp.

Uli

[Nelson Pass]

The marketing departments of all amplifier companies
would love to be able to claim SE Class A for their output
stages, but when the idle dissipation is less than twice
the rated output, I think we all know that some sort of
comromise has been made, and the lower the dissipation,
the more egregious the claim.

[Bernhard]

Nelson, and what about my circuit ?

[Pjotr]

Hi Bernhard,

If you take it to definition that a class-A amp is something of “No current carrying device is ever cut off”, then when it comes to low idle current, yes a class-A amp is possible. But such an amp will be highly non-linear somewhere and needs lots of feedback to get rid of that non-linearity. Also have a look at this device:

http://www.linear.com/prod/datasheet.html?datasheet=113

Cheers

[Bernhard]

Hi Piotr,

I know that low idle current is not good.

But in this circuit the idle current is only low for the current source fet and the cascode fet. Both of them do not affect sound.

The idle current for the important voltage follower fet is high instead.

And it will never cut off and it will never go much below 1A or more if one likes...

I will make a drawing tomorrow.

Greetings, Bernhard

[Nelson Pass]

Quite some time ago the audio community seems to have
divided Class A into two varieties: "Pure" and "Other".

For many, "Pure" means that the idle dissipation is at least
twice the output rating. Another definition of "Pure" rests on
whether a sine wave output to the load results in a sine wave
current appearing through the gain device(s).

I would say that if you meet both those criteria, you definitely
have "Pure". Arguably, everything else is "Other".

[Pjotr]

Originally posted by Nelson Pass

... Another definition of "Pure" rests on
whether a sine wave output to the load results in a sine wave
current appearing through the gain device(s).

Hello Nelson,

A remark here. Many amps rely on the fact that the difference of two quadratic transfer functions is linear: PP tube amps as well as some FET output stages in class-A. So I would state that a pure class-A output stage is something where a sinus input generates a sinus output in a resistive load WITHOUT overall feedback and where no device is ever cut off at max. power output. For this the current through individual devices need not to be necessarily sinusoidal IMHO.

Cheers

[uli]

Hi bernhard!

The error is that the current of about 1A runs thru the fet back
to the voltage source! Current running out of the +pole must flow
back to the -pole of the 6V voltage source. There is no current left
for the signal to output. This current comes from the upper fet
and is definitely no class A.

Running a gainstage in class A means by definition that the device
never shuts off! So called class AB amps generate 2 heavily
distorted waveforms which are added at output. So if you want
no distortion in your class A device you cannot squeeze more
than double the idle current out of it!

Uli

[Nelson Pass]

Originally posted by Pjotr
A remark here. Many amps rely on the fact that the difference of two quadratic transfer functions is linear: PP tube amps as well as some FET output stages in class-A. So I would state that a pure class-A output stage is something where a sinus input generates a sinus output in a resistive load WITHOUT overall feedback and where no device is ever cut off at max. power output. For this the current through individual devices need not to be necessarily sinusoidal IMHO.Cheers

I would beg to differ. I would not consider a dynamically biased
Class A amp (like the Threshold 800A) a pure Class A amplifier,
although it meets this definition.

The requirement that inidividual gain device currents look like a
sine wave (for a since wave output) is loose enough to cover
the relatively small linearity cancellations occuring in
complementary circuits IMHO.