dears ! i am trying to design an audio amplifier using 2n3904. according to my calculations , by pass capacitor (Emitter capaciotor)=33uF
but when i connect this capacitor, it generate noise at the output. if i remove this then gain reduces too much. what should i do>>??? please help me. thankx advance

 

My crystal ball is in for cleaning.
Please post your circuit diagram.

 

My best wild guess in the absence of any evidence to the contrary (or any evidence at all): Amplifiers amplify everything - noise and signal. The higher the gain the higher the noise.

 

Example
One transistor with gain of 100. =
Two transistor stages each with a gain of 10

 

thanks to all. but my point is this, how emitter bypass capacitor causes noise.? and one thing more, which type of capacitor is best for bypassing? i mean polyester capacitor, ceramic capactor etc.....

 

It causes noise by turning the gain up. Place a resistor in series with the capacitor to tame it. Even a variable capacitor so you can observe how the noise changes by turning the knob. About 1000 ohms should work.
Which type doesn't matter in this case. Your circuit is too simple to discern the tiny variations in capacitors.

 

It causes noise by turning the gain up. Place a resistor in series with the capacitor to tame it. Even a variable capacitor so you can observe how the noise changes by turning the knob. About 1000 ohms should work.
....................

Adding a resistor in that fashion will roll off the lower frequencies giving a high-pass filter (sort of like the treble control turned up) so you will still have high frequency noise amplified with a higher gain.

If you want to reduce the gain and still have a relatively flat response then break the emitter resistor into two resistors and only bypass one of them with large capacitor.
At higher frequencies the gain will then roughly be the collector resistance divided by the unbypassed emitter resistance.

 

If you want to reduce the gain and still have a relatively flat response then break the emitter resistor into two resistors and only bypass one of them with large capacitor.

Or to make it variable, use the track of a pot as the emitter resistor with the capacitor from the wiper to ground.

 

Adding a resistor in that fashion will roll off the lower frequencies giving a high-pass filter (sort of like the treble control turned up) so you will still have high frequency noise amplified with a higher gain.

Please forgive the very beginner way I answered that. I thought it fit the audience, and it seems to have worked.
However, if you use a 100 uf cap in parallel with the emitter resistor and add a resistance in series with that capacitor, the 1000 ohm Xc is at 1.59 Hz and the 100 ohm Xc is at 15.9 Hz.
I claim that these frequencies are not considered to be in the treble range.

 

Please forgive the very beginner way I answered that. I thought it fit the audience, and it seems to have worked.
However, if you use a 100 uf cap in parallel with the emitter resistor and add a resistance in series with that capacitor, the 1000 ohm Xc is at 1.59 Hz and the 100 ohm Xc is at 15.9 Hz.
I claim that these frequencies are not considered to be in the treble range.

Thinking about it further, I believe the results are basically the same for both configurations.
The difference is that your scheme puts the two emitter resistors in parallel above the corner frequency and mine puts them in series below the corner frequency, but the net result, in both cases, is that the equivalent emitter impedance goes down (and thus the gain up) at the respective corner frequencies.

Your scheme has the apparent advantage of requiring a lower capacitance for a given emitter resistance and corner frequency.

 

Your scheme has the apparent advantage of requiring a lower capacitance for a given emitter resistance and corner frequency.

I came up with that for a j-fet input stage. I would bet you've seen my small signal amplifier, even if you don't remember. It was labeled as something about a Rhodes piano pre-amp at one time. I needed to get 5-10 mv (p-p) up to around 5 volts (p-p) and I know percussion musical instruments have a lot of dynamic range, so I needed a volume control. The source resistor was already divided to get the right idle current and input impedance in the j-fet stage so I installed the series RC circuit in parallel with the source resistors to avoid having too many things going on at once, like redesigning the source resistors to accommodate the gain adjustment.

Yeah, it's doable, but the math gets real simple if you just do it my way; a capacitor that is nearly transparent to audio frequencies and a pot in series to diminish its effect. Worked like a charm. As quiet as a TL081 and I could run it with a 9V battery with a low voltage fail point of 7 volts.

 

what should i do>>???

1. Learn to design a circuit;
2. Learn to ask for help.