LM386 Audio Amplifier with Gain 20

The input signal VIN goes to the top of RV1, a 10 kΩ potentiometer whose other end is grounded. The wiper picks off part of the signal and feeds it to pin 3 (+INPUT). Pin 2 (−INPUT) is connected to ground. Both inputs are biased to ground inside the LM386 through 50 kΩ resistors. Figure 9-1 draws no input coupling capacitor; TI notes that adding one removes input offset problems entirely. Keep the signal at pin 3 within −0.4 V to +0.4 V, the absolute maximum input voltage; turning RV1 down is how this circuit does that.

Pins 1 and 8 are left open. In that state the internal 1.35 kΩ resistor sets the voltage gain to 20, which is 26 dB:

A capacitor from pin 1 to pin 8 would raise the gain to 200 (46 dB); this circuit does not use it.

The output stage biases itself to about half the supply voltage, about 4.5 V with a 9 V battery. C1 blocks that DC level so no steady current flows through the speaker, and passes the audio signal. Its positive terminal faces pin 5. Together with the 8 Ω speaker it forms a high-pass filter with a calculated corner frequency of about 80 Hz:

Frequencies below the corner reach the speaker attenuated.

C2 (0.05 µF) and R1 (10 Ω) in series run from pin 5 to ground. TI places such a network on the output in most application circuits of the datasheet without describing its purpose. Such a series RC is commonly called a Zobel or Boucherot network: at high frequencies, where the speaker coil's impedance rises, it gives the amplifier a load of about 10 Ω, which helps keep it stable.

The LM386N-1 operates from 4 V to 12 V, and draws 4 mA typical (8 mA maximum) with no signal at 6 V. TI recommends a capacitor to GND close to pin 6; the datasheet does not give its value and Figure 9-1 does not draw it.