# Not understanding aspects of the LM386 amp

For the following LM386 amp i have a couple of questions.

1. Is C3 and R1 a low pass filter and if so how do i calculate the cutoff frequency for it? I.e. do i have to take the impedance of C2 into consideration?

2. What does the capacitor C2 do. When diagram one states that it removes the dc component, is this like it removing the bias? Also C2 with R1 and C3 make a high pass filter doesnt it? If so how do i calculate the cutoff frequency for it?

3. For the input voltage into the amp, say if it was a square wave from 0-5V, doesnt a high pass capacitor have to be used to remove the dc bias with a voltage divider so that Vin is 0.4 to -0.4v? If this high pass filter is used to remove the dc bias is C2 needed then? Diagram 3 shows this dc bias remover.

4. If the speaker has a max power of 0.25W and a resistance of 8ohms, does this mean i need a 1.41 V output from the amp (V=sqrt(PR))

5. If i need a lowpass filter that removes say 15Khz as the cutoff frequency can i just change the values of C3 and R1 to get this cutoff?

Thanks for helping in advance

Note 1st diagram is just something i found online stating what potentially each C2 etc is.  Datasheet for LM386: http://www.ti.com/lit/ds/symlink/lm386.pdf

## 2 Answers

1. C3 and R1 is called a zobel network - go google go.
2. C2 stops the natural DC offset from the amplifier being shorted to ground by the speakers dc resistance.
3. A input DC block capacitor ought to be used. If in doubt use it. This doesn't mean C2 isn't needed. Read the data sheet.
4. A power of 0.25 watts being dissipated by an 8 ohm resistor is an RMS voltage of 1.4142 volts, yes.
5. No, C3 and R1 is a zobel network used to linearize the impedance of the speaker and make it look more like a fixed value resistor across the audio spectrum. Nobody hardly ever uses them these days but some amps will be unstable without it.
• Regarding #3, wouldn't AC coupling the input cause it to swing below the negative supply? Is it okay since the input transistors are PNP? Wouldn't there be a problem if the input amplitude was large enough to forward bias the collector-base junction of the input transistors? – DavidG25 Mar 5 at 20:18
• @DavidG25 yes it would and this should not be a problem for the expected small magnitude of the input signal. You correctly observed that use of PNP input transistors allows this. – Andy aka Mar 6 at 8:32

You didn't link to the datasheet, so answers related to the particular amp are guesses.

C3 and R1 blah blah?

That looks like something required to keep the amp stable. R1 is a resistive load the amp apparently needs at high frequencies to not do bad things, like oscillate. Surely this is mentioned in the datasheet.

What does the capacitor C2 do?

It AC-couples the opamp output to the speaker. It keeps constant DC from running thru the speaker. That would cause unnecessary dissipation in the amp, unnecessary dissipation in the speaker, and would keep the speaker cone off-center towards one end of its range. That would make it clip peaks sooner than it would if driven with AC. It probably would keep the speaker from meeting other specs too.

blah blah input voltage blah blah doesn't a high pass capacitor have to be used blah blah?

Not knowing anything about the particular amp, I would expect a DC blocking cap on the input too, especially since the amp is running from a unipolar supply. However, you do know about this particular amp. Read the datasheet and see what it says you need to do. Perhaps there is some auto-biasing inside.

4.If the speaker has a max power of 0.25W and a resistance of 8ohms, does this mean i need a 1.41 V output from the amp (V=sqrt(PR))

sqrt(250 mW x 8 Ω) = 1.4 V RMS, so yes.

For a sine, the peaks are sqrt(2) higher than the RMS value, so 2 V. That means the output needs a 4 V swing with a pure sine. Audio has occasional short term peaks much higher than that. Usually you want about 5x or more headroom above the pure flat sine case.

If i need a lowpass filter that removes say 15Khz as the cutoff frequency can i just change the values of C3 and R1 to get this cutoff?

No. That's not what they do.

• Thanks for answering, i just have a quick question. With using this amp if i needed to cutoff frequencies above say 15Khz , could i add a lowpass filter before the input? Or would i have to put, the lowpass after the C2 cap? If i am right about THE location to place it, when calculating the cutoff frequencies would i have to take in the other impedances such as R1 C2 and C3, to get the cutoff freq? – Student May 15 '17 at 11:33
• @user6186979 The best place would be before the amplifier. Putting it after would effectively be shorting out the amplifier output at high frequencies, which would waste power and might even damage the amplifier. – Simon B May 15 '17 at 11:36
• @SimonB thanks. How would i find the values for C3 R1 and C2 then? Or do i just use the standard values that the datasheet for a 20 gain shows? – Student May 15 '17 at 11:39
• @user6186979 Unless you have a good reason to do otherwise, use the recommended ones from the data sheet. – Simon B May 15 '17 at 11:44
• @SimonB Thanks How would i find the voltage across the speaker, taking into account the impedances of C1 R1 and C2? Since it wont be exactly the same as the output of the amp? – Student May 15 '17 at 12:06