Sorry, I previously posted the AC and DC output of the transistors. Here is the AC output into the speaker:

 

Then the driver transistors do not need to drive a high current in your 100 ohm resistors that are not needed.

Yes, my simulation shows those resistors dissipating about 2.5 and 3Wrms.

 

Instead of using driver transistors to turn off the output transistors, I use the driver transistors to turn on the output transistors.
Then the driver transistors do not need to drive a high current in your 100 ohm resistors that are not needed. I did not bother adding a 1k base-emitter resistor to each output transistor.

Instead of reducing the gain with an emitter resistor on the 3N3904 that also reduces the output voltage swing, I use the 100k resistor to provide AC and DC negative feedback.

Of course I biased the darlingtons with 4 diodes.
Bootsrapping will add a little more output voltage swing.
Your 100uF input capacitor had such a high value that it would take "an hour" to charge and it would pass earthquake frequencies.

This circuit does not work!

 

Yes, my simulation shows those resistors dissipating about 2.5 and 3Wrms.

You changed the value of one resistor. The circuit does not work.

 

You changed the value of one resistor. The circuit does not work.

Works for me, as you can see.
Post your circuit that "does not work."

 

Works for me, as you can see.
Post your circuit that "does not work."

Here it is with RB2 changed to 10.1K.
At least that was what I could see on the schematic as it was so small.

 

Here it is with RB2 changed to 10.1K.

Works fine for me (below).
The only change I made was to change to the non-C versions of the transistors since I didn't have the C model, but all that should do is change the transistor maximum voltage rating.

 

Works fine for me (below).
The only change I made was to change to the non-C versions of the transistors since I didn't have the C model, but all that should do is change the transistor maximum voltage rating.

View attachment 166269

Please upload a copy of your .asc file as that is the only way I see to figure out why our results differ.

The schematics are identical except for the one resistor.

Perhaps our transistors have different characteristics.

Thanks,

David

 

I don't have models of TIP transistors so I used similar Japanese one in the LTspice library.
I used Sziklai pairs instead of darlingtons, added bootstrapping and AC and DC negative feedback:

 

Attached is the .asc file from post #27.

 

hi Carl,
Tried with 'C' version [ bordodynov lib], it is clipping at approx 10V.
Guess its due to model parameter differences.?
E

 

Guess its due to model parameter differences.?

Must be.
There shouldn't be that much since the only difference I could see in the data sheet was the max Vce voltage rating.

 

I don't have models of TIP transistors so I used similar Japanese one in the LTspice library.
I used Sziklai pairs instead of darlingtons, added bootstrapping and AC and DC negative feedback:
https://forum.allaboutcircuits.com/attachments/class-ab-amplifier-with-sziklai-pairs-png.166271/

But your schematic contains an error. Q4 and Q5 are OFF all the time.

 

If you use Pmaxout(peak) = Vcc^2 / RL = 30V^2 / 8Ω = 112.5 watts peak. In reality, the sim supplies 4.85 volt peak output

What have you expected if you without any critical thinking substitute values into the formula?
First of all, your amplifier circuit is supplied from a single supply (30V). It should be obvious to you that it is impossible to get 30V peak at the output form 30V single supply. The theoretical maximum you can get is 15Vpeak for ideal push-pull amplifier. To gest more power you need a bridge amplifier
https://www.allaboutcircuits.com/te...ge-amplifiers-for-single-supply-applications/

But in your circuit, the voltage swing is limited by the first CE amplifier stage. Have you done any calculations?

Let us try to do a rough calculation and find the Q1 collector voltage when Q1 is in saturation.

Ic1_sat ≈ (30V - 0.2V)/(1.7kΩ+ 220Ω) ≈ 15.5mA and the voltage at colector will be equal to Vc1_sat ≈ 30V - 15.5mA*1.7kΩ ≈ 3.65V

Therefore the maximum negative voltage swing at the output is Vmax_neg_swing = VcQ1 - Vc1_sat = 9.36V - 3.65V = 5.7V
For Vin_max = 5.7V/(1.7k/0.22kΩ) = 0.74V

As you can see even this super simplified calculation clearly shows that it is impossible to get 15V peak at the output.
The best we can get is 5.7V peak so, the maximum output power is 5.7V^2/16Ω = 2W

Why would the estimate of 112.5? be so different than 2.94 w?

Try to read this
https://forum.allaboutcircuits.com/...ign-basic-questions.104719/page-6#post-804580
https://forum.allaboutcircuits.com/...ign-basic-questions.104719/page-7#post-805029

 

hi Carl,
Tried with 'C' version [ bordodynov lib], it is clipping at approx 10V.
Guess its due to model parameter differences.?
E

I ran the crutschow2 sim with my 31C and 32C models and got an even worse
result. I guess my models are incorrect. Where can a person get reliable
models?

 

Post an image of the 'even worse' sim plots, I will compare with what I see.

 

Post an image of the 'even worse' sim plots, I will compare with what I see.

 

Tried all sorts of fixes, I cannot make my sim plot as bad as yours.?
Will keep trying.


EDIT:
Lets try a different power transistor, the B139 and BD140 are in the LTS models.
Run this and see what we get.

 

Tried all sorts of fixes, I cannot make my sim plot as bad as yours.?
Will keep trying.

Thanks Eric. Your help is really appreciated.

 

refer post #38 edited.