Author Topic: Biasing Op-Amps Into Class A  (Read 6594 times)

Joe Kramer

Biasing Op-Amps Into Class A
« on: October 23, 2006, 06:23:49 PM »
Hi Friends,

This article may have some stompbox applications:

http://tangentsoft.net/audio/opamp-bias.html

If you've ever tried this or have ideas for implementing it, please share!

Regards,
Joe

Solder first, ask questions later.

www.droolbrothers.com

R.G.

Re: Biasing Op-Amps Into Class A
« Reply #1 on: October 23, 2006, 08:05:51 PM »
There is no mystery there.

Most opamps use class AB output drivers. This lets them give you a bigger maximum output current and low quiescent current in the same opamp. To force an opamp into class A, all you have to do is pull its output to either the + or - rail with a constant current source. The voltage feedback forces the output to stay at the same voltage, so the only way this can happen is if the opposite-polarity output transistor conducts that current into the constant current load all the time. The active internal transistor is always on, so the opamp is operating class A into a constant current load.

This is major cool - if you can hear the difference. Unfortunately, this mod is nearly always followed by a description of how the mod gives an clarity, an inner piecefulness and a zen-like crystalline quality to the sound. I'm sure that there are people somewhere that can hear that in good listening conditions.

I'm not sure that there are people anywhere that could hear it when its used in a distortion pedal.

It also increases the power supply current to the opamp by the constant current load amount, and the power dissipated by the opamp by the current load times one power supply. It also thermally unbalances the opamp's output. In a modern opamp, the layout of the chip is done to equalize thermal effects as well to get greater accuracy inside. The output transistors are placed carefully so their thermal load balance across the chip and it all heats evenly. This generally relies on the output transistors heating equally. In the constant current load Class A setup, half the dissipation is on the chip, and half is off, in the external load. So some of the layout accuracy is lost.

The class A action is also only true for load currents less than the external load. What the external load actually does is move the AB crossover point away from the middle. At currents bigger than the external load, you still do crossover to the opposite side driver.

R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.

Joe Kramer

Re: Biasing Op-Amps Into Class A
« Reply #2 on: October 23, 2006, 08:58:44 PM »
Hi Mr. RG, and thanks for your thoughts!

There is no mystery there.

The mystery is in the next paragraph. . . .   :icon_wink:

This is major cool - if you can hear the difference. Unfortunately, this mod is nearly always followed by a description of how the mod gives an clarity, an inner piecefulness and a zen-like crystalline quality to the sound. I'm sure that there are people somewhere that can hear that in good listening conditions.

You mean as in, "What is the sound of one hand clapping?"  I agree that kind of thing isn't for everybody, but to each his own ears, eh? . . .

It also increases the power supply current to the opamp by the constant current load amount, and the power dissipated by the opamp by the current load times one power supply. It also thermally unbalances the opamp's output. In a modern opamp, the layout of the chip is done to equalize thermal effects as well to get greater accuracy inside. The output transistors are placed carefully so their thermal load balance across the chip and it all heats evenly. This generally relies on the output transistors heating equally. In the constant current load Class A setup, half the dissipation is on the chip, and half is off, in the external load. So some of the layout accuracy is lost.

The class A action is also only true for load currents less than the external load. What the external load actually does is move the AB crossover point away from the middle. At currents bigger than the external load, you still do crossover to the opposite side driver.

Hmm.  From all this, it seems the Class A buzz can be procured much more cheaply with a couple of plain vanilla transistors (or even a discrete op amp for that matter).  IOW, why jack around trying to take two wheels off a car, when you could just get a motorcycle. . . .

Thanks again.

Regards,
Joe 

Solder first, ask questions later.

www.droolbrothers.com

JHS

Re: Biasing Op-Amps Into Class A
« Reply #3 on: October 24, 2006, 04:44:32 AM »
It only works w sym. power supply and is not really new, a lot of Hifi gear tuners use this for opamp tuning.

I'm just working on a class-A biasing circuit adaption for 9V asym. supply voltage for the TI TL071C. Not quite easy, the crux is to bias the PNP trannie to the correct base voltage (the hfe is crucial too). I use a zener diode for that, like in the old Hot Cake (the biasing circuit in the HC is adapted from an old tube-Hifi preamp and it's a lot easier doing this w a preamp tube than w a IC).

It works, but isn't perfect yet and I have to spend some more time on the finetunig of the bias circuit.

BTW: It only works w the TI TL071, 071 from other companies will not work due to the slightly different internal layout.

JHS




Joe Kramer

Re: Biasing Op-Amps Into Class A
« Reply #4 on: October 24, 2006, 05:10:39 PM »
Hi JHS,

I read your post about the Hot Cake in another thread.  You're saying it's possible to do this Class A bias trick with a single-supply circuit?  I'd like to see a schemo sometime, if it wouldn't be infringing on the HC proprietary schemo or your own stuff. 

That's interesting about the TI O71s being slightly different from other makers.  I've always thought that the difference in sound between TLOxxs from various makers was my imagination--maybe not, it seems.   I've read in another article that only certain chips can be Class A biased, because of the chip's output config.  Any others you know of that definitely cannot be?

Regards,
Joe
Solder first, ask questions later.

www.droolbrothers.com

JHS

Re: Biasing Op-Amps Into Class A
« Reply #5 on: October 28, 2006, 10:04:28 AM »
My prototype circuit works, this Class-A thing is a bit tricky to bias (to find the right Zener-values).
I used an old Crotwher HC-circuit as a base, and an old E-cookbook w a good description of this type of biasing.

As I wrote befrore, nothing special, very commen in the Hifi scene.

The sound of the circuit in class-A mode is softer, fatter w/o any sharp highs like a small Class-A amp.
It can be done w nearly any single Opamp and with a medium hfe PNP Si-trannie, but the TI TL071 and the BC559B used in the HC sounded best to my ears.

If anyone could tell me how I can add a schem/pic here I will post the basic circuit.

""

JHS



JHS

Re: Biasing Op-Amps Into Class A
« Reply #6 on: November 02, 2006, 07:55:06 AM »
A schem w the basic layout/principle can be found here (hope this link will work here).
http://www.bilder-speicher.de/Sohale365773.vollbild.html

JHS

Joe Kramer

Re: Biasing Op-Amps Into Class A
« Reply #7 on: November 02, 2006, 04:17:54 PM »
Hey JHS,

Thanks very much for posting this!  Nice schematic!  I'm interested in comparing how this sounds, especially in a clean mode.  Now to find some time for the breadboard. . . .

Thanks Again!

Regards,
Joe
Solder first, ask questions later.

www.droolbrothers.com