I think it is the 1 k emitter resistor. Q3 can dump current into the capacitor and then to the headphone, but the 1k resistor needs to dump that charge on the negative half cycles as fast as the charge is added in by Q3 during the positive half cycles. This amplifier will load up a positive charge on the left of the output capacitor and then most of the current will stop flowing.

I'd try somewhere between 47R to 150R for the emitter resistor. A 16R load is probably a bit steep for a BC547 - I'd go for something in a TO126 package.

 

I think that is rather pessimistic. This circuit is intended to drive 8 Ohm headphones:

 

Update: I solved the problem by connecting the emmitter follower's base to the collector of previous stage with a 33K register and changed the emitter register to 220ohm. I realized that if I use 1K Re then the Base-emitter voltage is no more in the ON state and it was clipping the waveform. Now My circuit working fine. I see nice sinewave output and it can drive my 32ohm headphones. Thanks for your feedbacks.

 

In the schematic in post #3, T3 is not biased correctly. What is shown is called "dangle-biased". This requires exact knowledge of the transistor's gain and that gain must be extremely stable. The problem is that a transistor's gain wanders all over the place during operation as it warms up. Dangle-biasing is the worst of all possible methods. I would eliminate the 10 uF coupling capacitor between T2 and T3 and the 560 K resistor, and connect the T3 base directly to the T2 collector as an emitter follower.

Also, the 1K resistor in the T3 emitter is too large. The smaller it is, the better the circuit will reproduce the negative half-cycles of the audio. BUT, the smaller it is, the hotter T3 will get. Since you have a 9 V battery, a better solution is to replace everything after T1 with an LM386 audio power amplifier.

ak

Yes I realized that and see my Update which fixed it. Thanks for explaining.

 

See
Vin_amp=2mV Vout_amp=850mV

View attachment 110089

Yes, I saw that clipping of negetive cycle in the scope. I connected the T3 base with t2 collector with a 33K and changed Emitter registor from 1k to 220ohm, now for small input signals (in the FET input), the output signal is nice sine wave.

 

Here is a good push-pull stage to replace everything after your JFET stage. R9 reperesents the 32 Ohm headphones. I get 500mV+ out of a 4mV input.

Oops. You don't need both C1 and C2. I clipped out another stage from my original design.

View attachment 110054

Thank you very much for supplying this circuit. I would implement this in my next projects. One problem I see in my emmitter follower is that it continously draws ~20mA current even without presence of any input. Does this push-pull amplifier has low current consumption if I have no input ?

Thanks,
-Rihan

 

See
Vin_amp=2mV Vout_amp=850mV

View attachment 110089

This circuit looks nice..!! thanks.. Is the push-pull configuration more energy efficient than emitter follower? I am running in 9V battery. so it would be better if I use a circuit which has low consumption when there is no input signal..
Thanks.

 

Yes, a push-pull configuration is generally more efficient since the output stage draws only a small bias current when idle, compared to the large bias current a Class A (emitter-follower) stage draws.