# Decreasing LM386 gain even more

#### bloguetronica

Joined Apr 27, 2007
1,453
I recently found out that even having a gain of 20 (26dB), the LM386 amplifier will have problems of distortion due to clipping. The amplifier project is on the "Projects Collection" section, here: http://forum.allaboutcircuits.com/showthread.php?t=6204

Is there a way of having a gain below 20? I heard rumors that are other way of setting the gain o this amplifier, but I don't know exacly how it is made. Just for reference, you can see the datasheet here: http://cache.national.com/ds/LM/LM386.pdf.It might be handy to see the schematic. I didn't figured out where are the positive and negative input inside this chip, since the negative input outside doesn't seem to be a true input.

#### Distort10n

Joined Dec 25, 2006
429
I for one need to brush up on my transistors, rather than looking at them as switches. I am afraid I cannot give you a better answer than, 'it appears so.'
On pg. 3 it mentions that it is only stable in gains 9 or greater so this implies there is a way to have a lower gain than 20. It also looks like the 'bass boost' circuit as explained on pg. 3 will be a good place to start with a lower gain of 20.

#### Ron H

Joined Apr 14, 2005
7,014
In the LM386 schematic, let Rf be the resistor between pins 1 and 5. Let Rs be the resistance between the emitters of the diff amp (this is 1.5k, or 150 ohms if there is a cap between pins 1 and 8). The gain, by calculation and by simulation, is
Gain = 2*Rf/Rs+1
The gains, then, are actually 21 and 201, respectively (datasheet says 20 and 200).
If you shunt Rf with another resistor (preferably in series with a DC blocking capacitor), you can lower the gain. You can use simple algebra to calculate the value, but keep in mind that the circuit may be unstable for gains less than 9.

#### bloguetronica

Joined Apr 27, 2007
1,453
I for one need to brush up on my transistors, rather than looking at them as switches. I am afraid I cannot give you a better answer than, 'it appears so.'
On pg. 3 it mentions that it is only stable in gains 9 or greater so this implies there is a way to have a lower gain than 20. It also looks like the 'bass boost' circuit as explained on pg. 3 will be a good place to start with a lower gain of 20.
Lets just say that putting a resistor between pins 1 and 8 will raise the gain. I experimented it myself. But I don't have the smallest idea of how to have a gain smaller than 20. However, like you, I think that the bass "boost" circuit is the key since does it's job by lowering the gain to 10 for high frequencies. I think, once the amplifier turns unstable if I lower the gain and use the bass boost circuit, perhaps the solution consists in dividing the voltage of the input signal.

In the LM386 schematic, let Rf be the resistor between pins 1 and 5. Let Rs be the resistance between the emitters of the diff amp (this is 1.5k, or 150 ohms if there is a cap between pins 1 and 8). The gain, by calculation and by simulation, is
Gain = 2*Rf/Rs+1
The gains, then, are actually 21 and 201, respectively (datasheet says 20 and 200).
If you shunt Rf with another resistor (preferably in series with a DC blocking capacitor), you can lower the gain. You can use simple algebra to calculate the value, but keep in mind that the circuit may be unstable for gains less than 9.
Good answer, but I forgot the detail on gain 9, and I want to use the "bass boost" circuit as well (of course, the "bass boost" circuit diminutes the gain by doing that). Is there any problem if I just limit the input voltage?

#### Ron H

Joined Apr 14, 2005
7,014
I'm not sure what you're getting at. Attenuating the input voltage will eliminate the clipping problem, but the amplifier will still be unstable if the gain (as defined in the equation I posted) is less than 9. If you are running the amplifier with pin 8 open, then the smallest resistor you can put in your bass boost circuit is 10k. The most bass boost you can get is about 7dB.