Clipping Class AB Amplifier output

BobTPH

Joined Jun 5, 2013
9,284
The R33 emitter resistors drop over 4V themselves. Also I think the 33R from base to ground on the Darlington pairs is too low and limiting the current to the output stage.
 

Jony130

Joined Feb 17, 2009
5,539
You demanding too much from the output transistor 2SC5200/ 2SA1943. Plus a voltage drop across R9/R10. You need to have three or more pairs of 2SC5200/ 2SA1943 connected in parallel.

Also, remove this connection
Screenshot 2023-02-23 at 16-00-55 Clipping Class AB Amplifier output.png
 

Audioguru again

Joined Oct 21, 2019
6,784
The amplifier is entirely DC coupled then the slant of the clipping should be perfectly flat. Maybe the power supply is poor.
There is no clipping when the input is turned down a little causing the output to be 80Vp-p which is 200W into 4 ohms.

The simulation shows 1khz but the oscilloscope shows 200Hz. Then is the load a speaker a little above resonance?
 

Thread Starter

henry4700

Joined Feb 23, 2023
7
1677168017192.png
It is the requirement of the project (A 50W up to 300W class ab amp)
Maybe it is not possible to change speaker.
 

Audioguru again

Joined Oct 21, 2019
6,784
Clipping is normal in an amplifier when the plus and minus voltage output swings to as high as it can go, but the input is too high which is trying to make it go impossibly higher.

Your revised circuit produces 100Vp-p into 4 ohms which is 312.6W!
 

BobTPH

Joined Jun 5, 2013
9,284
Can you explain R5 and R6? I know Darlington's often have a resistor from the base to emitter of the output transistor, but I have never seen them connected like this, without connecting to the emitters. TS had them both connected to ground @Jony130 suggesting removing the ground connection, and AG did that, as well as increasing the by 10X as I suggested.
 

Thread Starter

henry4700

Joined Feb 23, 2023
7
You mean r5 and r6 will neither connect as the first circuit I showed nor the latest one.
I usually connect the middle of r5 r6 to output as the first circuit.
(r5 and r6 should be combined to 1 resistor if I use the latest version)
 

Veracohr

Joined Jan 3, 2011
772
The calculated parameter "amp" is what your theoretical peak output voltage should be, and it works out to 63V with the values given. You have only 55V rails. Also, you stated the design should be "up to 300W", but the given power output used in the calculation is 500W.

Unrelated to the clipping but also a concern: you may want to adjust the Vbe multiplier Q5. They way it's set you'll have a lot of DC current through the output transistors and waste a lot of power when there's no signal.
 

MisterBill2

Joined Jan 23, 2018
19,466
A DC amplifier is easy to overdrive into clipping, which, as has been explained is exactly what is happening. But since that clipping looks to be about 10% total harmonic distortion, that is OK, because many "junk" class amplifiers are rated at 10% distortion.
So, exactly as others have suggested reduce the input and the clipping will cease.
OR, as others have explained also, increase the supply voltage so that it is greater than the required output voltage.

Aside from that, I simply do not like DC coupled music amplifiers because one failed part often destroys the whole string of transistors, or at leastt the expensive ones.
 

Ian0

Joined Aug 7, 2020
10,277
Can you explain R5 and R6? I know Darlington's often have a resistor from the base to emitter of the output transistor, but I have never seen them connected like this, without connecting to the emitters. TS had them both connected to ground @Jony130 suggesting removing the ground connection, and AG did that, as well as increasing the by 10X as I suggested.
It's mentioned in Douglas Self's book. The single resistor situation means that when the output transistors are required to turn off the base charge is removed more quickly if connected to a lower voltage (for the upper transistor).
Alternatively, if using a very low base-emitter resistor on the output transistors, it can be made so that the output transistors work in class C and the driver transistors supply the load during the time both output transistors are off. That way, setting the bias is less critical.
 

Ian0

Joined Aug 7, 2020
10,277
You can gain a little more power by splitting the power supplies into two. A lower current 60V supply to supply everything from R17 to the left and a high-current 55V supply to supply the output stage. A further advantage is that the ripple does not get to the input stage.

Another point: Your zobel capacitor is far too big. 10uF will burn the associated resistor. Use 47nF.
 

Ian0

Joined Aug 7, 2020
10,277
R5 and R6 have resistances too low. They are shown on darlington transistors datasheets.
I disagree - I'm not saying that they must be that low, but the lower the better if you want to get the power transistors turned off as quickly as possible, provided that the driver stage can deliver the current. I used 47Ω for a single MJ15003/4 output stage and 22Ω for a double.

Two more points:
1) A Baker clamp on Q4 will get rid of that nasty latch-up as it comes out of clipping from the top rail.
2) Q6 needs a current limiting resistor in its collector. Otherwise it will blow up when driven hard into clipping on the negative rail (and Douglas Self forgot that one as well).
 

MisterBill2

Joined Jan 23, 2018
19,466
From post #7 we see that this is a school project. So perhaps the goal is to learn about what does not work in amplifier circuits.

Certainly there are some interesting insights provided, but the whole circuit seems a bit excessively complex, although a DC coupled amplifier certainly needs a lot of extra protection to avoid destruction when driven a bit hard.
Which Is why consider a DC coupled audio amp a poor choice.
 

Thread Starter

henry4700

Joined Feb 23, 2023
7
After read all the comment, I change
(1) Vcc from 55V to 110V (It is 70V for 300W)
The peak voltage around 65.9V, 55V is not enough.
(It maybe early clipping, is 110V too high?)
(2) R5 and R6 bigger
(3) Keep output transistor 2SC5200/ 2SA1943 1 pair first
If it is enough for 1 pair, solve the other problem first

I need to focus on distortion (only appear the top of +ve half cycle)
螢幕截圖 2023-02-25 上午10.04.29.png
螢幕截圖 2023-02-25 上午9.52.12.png
-----------------------------------------
Remark:
Loading:4ohm
Signal type:Sine wave
Full Power:THD+N<1% (>300W)
-------------------------------------------
I test with the Vin amplitude = sqrt(Power*Rload)*sqrt(2)/gain = 3.16V
Power = 500W (up to 500W should fulfill all the power requirement of the project)
Rload = 4
Gain = 20k/1k

Thank you everyone
 
Top