or

An old dog learned a new trick.

I was going through the text book on experiments and I came across this...

http://www.allaboutcircuits.com/textbook/experiments/chpt-6/class-b-audio-amplifier/

It solved a problem I've been mulling, mostly due to reduced parts could. I've been over complicating this circuit from Creating a Virtual Power Supply Ground as follows.


.....................Figure 11

Anyone who has been around for a while on this site knows I like to doodle with schematics. So I have redrawn the experiments schematics.





Figure 4 is iffy, the LED drop must be under 2.3V (a red LED in other words) for it to work, otherwise the transistors will likely smoke.

I have not built these circuits, just thinking about it. The application as shown is for a speaker (say around 8Ω). Gain could be provided via another op amp, this is just a driver.

It would also make a decent virtual ground for power applications, like a home brew power supply, to create something that approximates a dual tracking power supply.

What do you think?

 

Or do the scheme in this data sheet.

View attachment 76074

edit: See the High Power Audio Amplifier example.

 

or

An old dog learned a new trick.

I was going through the text book on experiments and I came across this...

http://www.allaboutcircuits.com/textbook/experiments/chpt-6/class-b-audio-amplifier/

It solved a problem I've been mulling, mostly due to reduced parts could. I've been over complicating this circuit from Creating a Virtual Power Supply Ground as follows.

instead

.....................Figure 11

Anyone who has been around for a while on this site knows I like to doodle with schematics. So I have redrawn the experiments schematics.

View attachment 94882

View attachment 94883

Figure 4 is iffy, the LED drop must be under 2.3V (a red LED in other words) for it to work, otherwise the transistors will likely smoke.

I have not built these circuits, just thinking about it. The application as shown is for a speaker (say around 8Ω). Gain could be provided via another op amp, this is just a driver.

It would also make a decent virtual ground for power applications, like a home brew power supply, to create something that approximates a dual tracking power supply.

What do you think?

Some of the NS appnotes show the transistor drives derived across resistors in series with the op-amp supply pins, the op amp output is tied to the node between the 2 transistors which is also the point that boosted output comes from.

As I don't remember the exact configuration, I kept my description a bit vague - better to leave something out than include something that's wrong.

 

I'm curious about Figure 2. I assume Q2 & Q4 allow for higher current than Figure 1, but what's the benefit of this arrangement over just doubling Q1 & Q3?

 

or

An old dog learned a new trick.

I was going through the text book on experiments and I came across this...

http://www.allaboutcircuits.com/textbook/experiments/chpt-6/class-b-audio-amplifier/

It solved a problem I've been mulling, mostly due to reduced parts could. I've been over complicating this circuit from Creating a Virtual Power Supply Ground as follows.

instead

.....................Figure 11

Anyone who has been around for a while on this site knows I like to doodle with schematics. So I have redrawn the experiments schematics.

View attachment 94882

View attachment 94883

Figure 4 is iffy, the LED drop must be under 2.3V (a red LED in other words) for it to work, otherwise the transistors will likely smoke.

I have not built these circuits, just thinking about it. The application as shown is for a speaker (say around 8Ω). Gain could be provided via another op amp, this is just a driver.

It would also make a decent virtual ground for power applications, like a home brew power supply, to create something that approximates a dual tracking power supply.

What do you think?

There are various examples online that don't include any bias diodes for the transistors - but the transistors are enclosed in the nfb loop.

I've never tried it myself, and even if that worked - trying it with Darlington pairs might be a bit optimistic.

 

I'm curious about Figure 2. I assume Q2 & Q4 allow for higher current than Figure 1, but what's the benefit of this arrangement over just doubling Q1 & Q3?

The advantage is two fold. Major current drive with no loading on the op amp, where as a single transistor arrangement is going to pass some of the loading on to the op amp.

If the beta of a transistor is 50 or so, my preferred typical value, then the 1KΩ will drop down to 1KΩ/50, or 20. The second transistor makes it a super gain transistor, a Sziklai pair. The advantage to this configuration is that it is similar to a Darlington transistor, but only has a 0.7V B-E drop. This means the 1KΩ resistor now looks like 0.4Ω seen through the transistors.

They are both aspects of the same thing.

 

Some of the NS appnotes show the transistor drives derived across resistors in series with the op-amp supply pins, the op amp output is tied to the node between the 2 transistors which is also the point that boosted output comes from.

As I don't remember the exact configuration, I kept my description a bit vague - better to leave something out than include something that's wrong.

The idea preceded the existence of NS, I think...

See the High Slew Rate Power amplifier in the attached uA741 data sheet (page 113).

 

I've used several of these circuits, and always got lower crossover distortion with the common emitters in the power supply leads than with the emitter followers on the output. OTOH, the former needs more attention paid to compensation compared to the latter.

ak

 

It ocurs to me I am making a mistake. Instead of using diode that will be a poor match at best for theBE junctions they ar biasing just use the same transistor types and theirBE junction to match thevoltage drop It isn't like the parts are being wasted.I will post redrawn schematic in a week or so to reflect this idea.t

 

 

There still is zero static current through the output pair. While the physical mounting can get you better thermal tracking, there still is a point when the opamp feedback loop open up (briefly) as ground current transitions across 0 mA in either direction. Depending on the speed of the opamp, this can show up as a signal artifact in whatever the virtual ground is grounding.

ak

 

I was thinking of using a 4565, which is a very fast dual op amp

 

OK, this going into my cook book


Figure 1 uses 2N2222s (NPN) and2N2907s (PNP)and figur2 uses TIP100s(NPN) and TIP105s (PNP).

 

I'm curious about Figure 2. I assume Q2 & Q4 allow for higher current than Figure 1, but what's the benefit of this arrangement over just doubling Q1 & Q3?

There would be two benefits, the first in that the transistors are running in a more liner region, and that the drive requirement is reduced, while doubling the transistors doubles the base drive requirement. And if transistors are going to be stuck in parallel then they should be quite close in characteristics, and that costs more.

 

I meant to say more but I have been fighting my brain fog from my stroke. The added gain from the darlingtons should really help the op amps. I spent a lot more than I meant to drawing these puppies, sad part I may redo them again.

 

One additional thing to consider is that with higher gain, especially in a directly coupled circuit, there becomes a greater tendency toward instability, both oscillations and drift. So increasing gain is not always a best choice.

 

What can I say? I was having a bad day on post #13. Here are the redrawn schematics. I would dearly like to see a O scope picture showing the crossover distortion. When I am able I will post one myself. I am hoping there isn't any. But life is full of disappointments, The Darlington arrangement ought to be great at driving a speaker, but way too much voltage is eaten up because of Darlington voltage drops, especially using a 9V battery. Interest thing about Differential Outputs If its power source is totally isolated (as in a battery) and one of the outputs is tied to a circuit ground the other output appear to be 2X larger in amplitude. The triangle ground symbol is 1/2 Vcc.

....................

 

There may be an issue with using the darlington configuration because they require twice the Vbe to start turning on. That may not be what you want to minimize crossover distortion. Also, to avoid needing the inverter transistors in the driving section, consider some of those complementary output opamps. or even a good old LM733.
I do like the amps with the complementary emitter follower outputs, that is good for a very low output impedance.

 

The double B-E junctions should turn on both driver transistor, even in the Darlington configuration. I just don' know how well it works, it looks beautiful on paper. Oops just noticed another brain fart, redrawing again.

 

Some boost work from an expert, Jim Williams at LTC on boosting opamp
output -

http://www.proaudiodesignforum.com/forum/php/viewtopic.php?t=13

Regards, Dana.