diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   no mercy distortion killer circuit (http://www.diyaudio.com/forums/solid-state/24192-no-mercy-distortion-killer-circuit.html)

Bernhard 13th December 2003 07:07 PM

no mercy distortion killer circuit
 
Hi ,

I got a circuit on the sim that:


- cancels* distortion of the output stage

- cancels* offset voltage of the output stage

- cancels* occuring clipping (kind of severe distortion) of the output stage

- probably even cancels* overshoot on square signals ( not tested yet )

- makes it possible to run the output stage with very little or no feedback if desired.


I think I will post schematic soon :)

Anybody interested to discuss this seriously ? ;)

Or does anybody have an idea how I do that ? :devilr:






*aproximately :xeye: which means that on the sim, even under worst conditions as without feedback or while clipping, the output curve is absolutely matching the input curve :D

Upupa Epops 13th December 2003 07:36 PM

Go on !:clown:

johnferrier 13th December 2003 07:39 PM

With a simple model/simulation, the concept of negative feedback cancels distortion. The hard part, of course, is cancelling distortion in reality. So how well does your simulation model reality?

And, of course, people will want to see more details: your schematic.


JF

Pedja 13th December 2003 07:41 PM

The output curve while clipping matches the input curve?

Bernhard 13th December 2003 08:08 PM

Quote:

Originally posted by Pedja
The output curve while clipping matches the input curve?

:D :D :D Oh yes !!!

Jan Dupont 13th December 2003 08:30 PM

Lets hear more :nod:

tschrama 13th December 2003 08:37 PM

Come on... show us...: can you post a schematic? This must be a very special topology...

Thijs

Bernhard 13th December 2003 08:38 PM

please be patient
 
Okay,

the red curve is the input , you do not see very much of it because it matches the output curve across the load which is green . The curve of the output stage is blue .

To see if the match is perfect, I made the amp a voltage follower, no voltage gain, only current.

The output stage is single ended classA and it is clipping and has got no feedback and has got terrible offset.

Still the output matches the input. :clown:

I did that to demonstrate what the circuit can do.

Of course the output stage can be with feedback and operated below clipping and with very little offset.

The rail voltages are 12V only because I just copied parts of the circuit from another of my amps.

Any circuit like this known to the public ? :confused:

I should call it "The mirror of the soul" :)

http://www.beautyphoto.de/nomercycurve.jpg

Bernhard 13th December 2003 09:00 PM

This is the circuit, it is a functional model.
Explanation follows.
http://www.beautyphoto.de/mirror.jpg

Bernhard 13th December 2003 09:18 PM

The upper left side is the SE classA output stage. The op amp x3 could provide voltage gain if desired.

X5 is an inverting buffer.

R13 & R14 make the heart of the circuit.

The inverted input signal and the signal of the output stage cancel each other, left over is the distortion of the output stage, divided by factor 2.

The distortion is reamplified by the upper left side which is again a SE classA output stage. X4 provides voltage gain 2x.

The right stage compensates the errors of the left side.

Including distortion, offset, clipping, and hopefully even overshoot.



:att'n: I know that the right side itself produces distortion.



But if the left side got some feedback so that it does good work, then the output of the right side will be only of a very very small amplitude, which means the new introduced distortion will be very very small too.

But nearly all of the left sides distortion will be canceled. :headshot:

The circuit can be further improved, a differential amp and discrete parts instead of op amps could be used.

Because the output of the right side will be small except @ clipping, the right side rails can be as low as 7V or even less.

I hope this could be as good as the susy effect :rolleyes:

Comments, critics and noble price welcome :smash:


All times are GMT. The time now is 05:17 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2