TSSA - The Simplest Symmetrical Amplifier

[Joachim Gerhard]

As a comparison a 84dB sensitive speaker with 2,7 Ohm minimum would need 2 x 270W to sound as loud as my speakers with 2 x 3,125W into 8 Ohm.

[mikelm]

Originally Posted by sonnya

You know that is not possible

Well Actually I do not know for sure that it cannot be done - perhaps some fancy design with distortion cancellation - that's beyond my range of knowledge but I'm interested to hear this opinion from an experienced professional designer.

For the simpler designs that interest me they always seem to sound better when biased at least into heavy AB. 500 - 1000mA.

Why use fancy complicated designs to try and eliminate something that simply disappears with higher bias. For me simple is best.

[MiiB]

Originally Posted by mikelm

I would be very interested if anyone could demonstrate an amp design that when biased to 100 - 150mA had lower H10 - H20 than the same design biased into class A.

what if the output-stage was non switching...??

[mikelm]

Originally Posted by MiiB

what if the output-stage was non switching...??

yes - good point

I saw a design by a guy called chater in a collected designs book where he used op amps to make sure the o/p transistor never turned off and it was supposed to sound very good - and I can imagine that it could be great.

Do we think lower high order distortion in higher bias designs is purely down to less switching noise ?

[sonnya]

Originally Posted by MiiB

single ended Class A like the old Pass Aleph series does draw constant current. But the output current is limited to the quiescent current running in the output stage...true class A in its purest form...!! What we have here is capable of swinging more current than the bias, so It's class A-B. It will under load modulate the current draw on the supply lines.
I can be biased Hard, so it rarely leaves class A...but at the expanse or high current consumption and running hot...question is what is the optimum bias...does the high current bring any benefits....Think I remember some investigation that show and optimum distortion vise arround 120-150 mA...??

When running single supply and Constant current source this is correct. But if you run dual supply one will be constant current and the opposite site will modulate the supply.

The best way is to use a push pull stage in differential mode. This way the current from drawn from the output sinking will be added from the output sourcing. This way in theory the current draw in class A will be constant.

Optimum bias depends a lot on the gain step. with BJT it is very straight forward - 25mV drop across the emitter resistor.

For Mosfet we need to take in account what the gfs of the chosen mosfet is.

- Sonny

[sonnya]

Originally Posted by MiiB

what if the output-stage was non switching...??

Even in non- switching you will have a steep roll of in the bias current. The faster the slope is (di/dt) -> more high order harmonics will be generated.

But if you can switch a mosfet really fast off and on and slew really fast from the point where one goes off and too the point where the other goes on you can reduce the harmonics a lot and be really green. (4 - 8 watt idle for an 50 - 100 watt @ 8R amp).

I think Mikelm knows what i mean.

[sonnya]

And again in class A the same circuit will be even better.

[sonnya]

The TSSA V1.5 is not perfect. I am the first one to say it!

But given the design rules that it should be simple and stable i Think it is a good compromise between performance and simplicity.

If i would make a version of it and not listen to what other thinks the AMP would twice as many parts and it would be All bjt design. But that is not the case here.

Still with twice the number of parts it would only be 2 stage design.

The present design Will maybe get changed resistor values and lower gain when i have completed the build and test.

[mikelm]

that was a great discussion - thx Sonny.

You seemed to confirm what I suspected but didn't really know for sure.

[sonnya]

I have sent 9 proforma's now and the rest will follow this evening.