class A to class AB switch

[pos]

Hello,

I have a simple question:
Most class AB amps are biased to operate in so called "class A" mode on low level signals. So for example lets consider an amplifier that would provide class A behavior for its first 10W. Is this limit dependent on the impedance of the speaker? It is current or tension limited?

Thank you

[Frank Berry]

The class of operation is determined by the biasing of the amplifier stages, mainly the output stage. In AB operation, the output transistors operate in class 'A' mode at low signal levels and then gradually go into class 'B' mode after the transistors reach the more linear portion of their curves.
You can operate an AB amplifier up to 10 watts in 'A' mode but it may require large heatsinks on the output (and perhaps driver) devices and possibly heftier transistors.

[pos]

So if an amplifier is able to operate in class A up to 10W in a 8 ohms loudspeaker, how many "class A watts" can it produce in a 4 ohms one?
Is the limit set in term of volts or amperes?

[Frank Berry]

It depends on the amplifier design. If the power supply can deliver the current and if the driver and output transistors can handle the 4 ohm load, the amplifier should be able to produce 20 watts into 4 ohms while operating in class 'A' mode. With a properly designed amplifier, the limit is the voltage swing.
Most AB amplifiers are not designed to operate in class 'A' mode above 1 or 2 watts.
Just enough to eliminate the notch distortion in the output stage.

[sreten]

edit :

Originally Posted by Frank Berry

It depends on the amplifier design. If the power supply can deliver the current and if the driver and output
transistors can handle the 4 ohm load, the amplifier should be able to produce 20 watts into 4 ohms while
operating in class 'A' mode. With a properly designed amplifier, the limit is the voltage swing.

No. That is not the case. The transition to class B depends on the standing output current.
To produce 10W A into 4 ohms requires 1.4 times the standing current and doubles dissipation.
To produce 20W A into 4 ohms requires 2 times the standing current and quadruples dissipation.

endedit :

Hi,

Its current limited so it would be 5W into 4 ohms. Note even though you
have lower class A power, output stage heat dissipation will be the same.

A high bias class AB amplifier, as opposed to optimum bias aB likes high
impedances. A quarter A amplifier 10W A, 40W B into 8 ohms would
deliver 20W A into 16 ohms and never enter class B.

Conversely the same amplifier into 4 ohms would deliver
5W class A and towards 80W class B if it has a stiff supply.

Basically you need to drop rail voltages and increase standing current
for lower impedance, for the same standing output heat dissipation.
(You also drop rail voltages for a 4ohm amplifier if your
pushing the thermal SOA limits of the output devices.)
Heatsinking here matters.

A quarter A amplifier 10W A, 40W B into 4 ohms would
deliver 20W A into 8 ohms and never enter class B.

Going for quarter A into 6 ohms is a fair compromise.
About 1/3 A into 8 ohms and 1/8 A into 4 ohms.

rgds, sreten.

[jcx]

Class A bias is determined by the amount of current, a typical push-pull (complementary Q) output stage can deliver up ~ 2x the bias current "in Class A"
since the current is fixed the Class A power into 1/2 the load Z gives only 1/2 the Class A power at 1x load

if you want a larger (or even the same) region of Class A operation in Watts with lower Z load you need more bias current

[pos]

Thanks a lot for your anwsers!
So this is current limited, perfect!

I asked this question because I am going to use an autoformer between my amplifier and a compression driver (active setup) to reduce the voltager the compression drivers gets and lower the hiss. The amplifier will see a higer impedance (for exemple 64ohms if I want to drop the level by 6dB on my 16ohms compression driver), and I was worried if I would still get the same amount of class A behavior...

[sreten]

Hi,

If you do that you will get more class A behaviour, a lot more.

But note that 2:1 will attenuate the tweeter by 12dB,
not 6dB, for 6dB you use a 1.41:1 ratio, x2 impedance.

rgds, sreten.

[sakis]

i presume that all the above is some kind of joke ....

i cant understand why any one will overbias a class Ab amplifier to approach class A and then put in the midle any transformer that obviously will degarde signal quality, speed and bandwidth ...

then again the word HISS i understand as noise comming from tweeters ( wrong ? ) so if a system produces hiss you try to eliminate the problem before the speaker looking at amps preamsp eqs (?) or xovers that are conected in the system

...
sakis

[sreten]

Originally Posted by sakis

i presume that all the above is some kind of joke ....
...
sakis

Hi,

No. Its a perfectly sensible way of attenuating a tweeter and
maintaining amplifier damping that you'd lose using resistors.

Whatever the bias is set to high impedance gives more class A.

rgds, sreten.