Mains noise problem and reduced sensitivity

[arnab_2014]

I had tried to build an audio amplifier using a preamp build out of a transistor and thereafter using lm386 for the audio amplifier. I had received a suitable response using an earphone. But when I use an 8 ohm 0.5 speaker the audio output gets distorted and also the sensitivity reduces. Also I am unable to eliminate the mains noise which distorts the output of the speaker. Along with it when I increase the volume control or the gain control too high i hear a sharp beep sound of 2.7 kHz. I would love to hear what improvments i could make in the circuit to make it sensitive as well as clear

 

[Audioguru]

When you make a circuit with an IC then you should read the datasheet for the IC. The datasheet for the LM386 shows how the output parts should be connected.

The LM386 has fairly low maximum output power. When its output power is too high then it has clipping distortion but with an earphone the gain is probably turned down so there is no clipping distortion.

When you use a speaker and you increase the gain then the mic can hear the speaker and the sound goes around and around (in the mic, amplified then out the speaker then in the mic again then amplified and out the speaker again, over and over) which is called acoustical feedback howling and causes the 2.7khz sound. It is prevented by recording the sounds then playing them back later with the mic turned off, or using an earphone that the mic cannot hear.

The 10nF capacitor to ground at the output of the volume control cuts all high frequencies making the sounds muffed and not clear. The 10nF capacitor and pot at the top of your schematic also cuts high frequencies and causes oscillation at a very high frequency when the put is turned down too low, remove them.

Mains noise? If it is pickup of the 50Hz or 60Hz from mains electricity then audio shielded cables are used to eliminate it. If you built the circuit on a solderless breadboard then all the strips of contacts and all the jumper wires are antennas that pickup mains hum. Build it on a compact pcb instead. I soldered all my prototype circuits on compact stripboard.

 

[arnab_2014]

Thanks . So i do not have to tweak the preamp

 

[D.A.(Tony)Stewart]

Your mic leads are not balanced so, CM noise can be a problem. If not already shielded input, you have insufficient rejection of the Common mode hum, since the input is unbalanced, especially if you are using a laptop with a floating ground ,,so take care with the grounds. If earthing the amp, reduces the CM hum, then the imbalance converts it to a differential signal.

So suppressing the CM noise by earthing, or if you prefer floating from battery power, use a large ferrite balun around mic pairs to raise the common mode impedance.

AG has addressed these and other issues.

 

[arnab_2014]

When you make a circuit with an IC then you should read the datasheet for the IC. The datasheet for the LM386 shows how the output parts should be connected.

The LM386 has fairly low maximum output power. When its output power is too high then it has clipping distortion but with an earphone the gain is probably turned down so there is no clipping distortion.

When you use a speaker and you increase the gain then the mic can hear the speaker and the sound goes around and around (in the mic, amplified then out the speaker then in the mic again then amplified and out the speaker again, over and over) which is called acoustical feedback howling and causes the 2.7khz sound. It is prevented by recording the sounds then playing them back later with the mic turned off, or using an earphone that the mic cannot hear.

The 10nF capacitor to ground at the output of the volume control cuts all high frequencies making the sounds muffed and not clear. The 10nF capacitor and pot at the top of your schematic also cuts high frequencies and causes oscillation at a very high frequency when the put is turned down too low, remove them.

Mains noise? If it is pickup of the 50Hz or 60Hz from mains electricity then audio shielded cables are used to eliminate it. If you built the circuit on a solderless breadboard then all the strips of contacts and all the jumper wires are antennas that pickup mains hum. Build it on a compact pcb instead. I soldered all my prototype circuits on compact stripboard.

So i did it andit is working fine. What changes can I make in the preamp

 

[arnab_2014]

These are some of the modifications that I did. The response is clear with a gain of 50 but still acoustic howling takes place . How do I prevent it dynamically

 

[Audioguru]

Acoustical howling can be reduced if you use a very expensive speaker and a very expensive microphone with no peaks in their frequency response and point them in different directions far apart.
A cheap speakerphone or intercom uses "voice switching" that cuts your speaker when you talk and cuts your mic when the other end talks. Frequently talking is cut off.
A very expensive speakerphone or intercom eliminates feedback by making a digital model of the acoustics (it is trained by using pink noise) then uses a digital echo canceller circuit. If something changes the acoustics like a door being opened then the circuit tries to dynamically adapt or it goes crazy making odd sounds if it cannot adapt.

A 7805 is no good with a 9V battery because its minimum input is 7V which a 9V battery quickly drops below. Use a low dropout 5V regulator that works fine until the 9V battery drops below 5.5V.

 

[D.A.(Tony)Stewart]

did you try to reverse the speaker polarity?

If feedback is in phase at >unity gain, it will cause oscillation or howling. IF phase shifts 180 deg, at different f, but at a higher feedback, then a quality factor of non-linear response as AG indicated is your problem. Generally cheap mic's have background noise cancellation by having a balanced open back vent so that only near field sound pressure is picked up. Trial and error with cheap electret mics will determine the best for you.

 

[arnab_2014]

Can I use an op amp to cancel out the noise. The op amp inputs are a scaled version of the LM386 output and the preamplifier input

 

[Audioguru]

It looks like you are using an extremely cheap speaker that will have some peaks and troughs in its frequency response so its sound level and phase will barely resemble the output from the LM386 audio amplifier.
Therefore an opamp will not reduce acoustical feedback howling caused by the mic hearing the speaker. You could make a notch filter to reduce the highest peak sound frequency from the speaker which will reduce feedback a little until feedback happens at the next frequency the cheap speaker produces a peak at. Your speaker doesn't even have an enclosure so out-of-phase sounds are coming from its rear and are fed to the mic.