Don't understand LM386 - Electronic Circuits Projects Diagrams Free

Oznog · 15th August 2005, 08:30 PM

Looking over the LM386, I don't see the best way to use this in a circuit and could use some help.

I have a system which has a single 12v supply and an audio signal limited to 0 to 5v since it goes through some digital pots for mixing various sources. That signal is biased to 2.5v, I don't have the exact amplitude, this will be determined by pot settings later. Obviously less than 5v pp.

Anyways, the LM386 is an amp but not an op amp. With the basic 20 gain circuit, the 2.5v bias means an output clipping far past the +12v rail. My first thought from an op-amp perspective is of course to put the inverting input on the 2.5v bias point, but the spec sheet says +/-0.4v is an absolute max input value. So I assume this means I will need to put the inverting to ground and capacitively couple the noninverting input? That will mean the input signal will be below ground for 1/2 the cycle- is this the intended operating mode?

Now what about the output biasing? It needs to bias to around 6v ideally, before the speaker coupling cap. Does the 386 do this automatically somehow or what? Or does the DC bias of the input just get multiplied by the gain and set the output bias from there? I don't see where this seems to be addressed in schematics.

**Nigel Goodwin** · 15th August 2005, 09:53 PM

You seem to be somewhat confused?. The LM386 is an audio amp IC, as such the input should be (MUST! be?) AC coupled, so your worries don't apply.

The signal level from your digital volume controls are not related to their supply voltage (only limited by it), so that's not a problem either.

The signal applied to the LM386 should be such that it doesn't clip the output - you should ensure that is so!.

Oznog · 15th August 2005, 10:31 PM

Quote:

Originally Posted by Nigel Goodwin

You seem to be somewhat confused?. The LM386 is an audio amp IC, as such the input should be (MUST! be?) AC coupled, so your worries don't apply.

The signal applied to the LM386 should be such that it doesn't clip the output - you should ensure that is so!.

Well, that's what I'm trying to understand. Say I have the typical "minimum component" circuit with a gain of 20. With a supply voltage of 12v the ideal output bias is generally 6v though certainly if I know the headphones will never use more than 2v peak-to-peak and the chip can't drive closer than 0.3v from the negative rail then a 1.3v output bias is acceptable.

Anyways, say I need a 6v output bias- does this chip just work off a straight DC gain, in which case I need to give the noninverting input a DC bias of 6v/20, or does it set the bias itself? I didn't see where the example circuits made any effort to set bias externally like that.

audioguru · 15th August 2005, 11:05 PM

The LM386 biases itself to keep its output near half-supply if the input has 0V DC or has an input coupling capacitor as shown on every schematic in its datasheet.
Its inputs are actually at 0V.

Make a search in Google for Headbanger and you will find some headphones amp projects using LM386's. :lol:

Roff · 15th August 2005, 11:24 PM

Here's how the biasing works.

**Nigel Goodwin** · 16th August 2005, 08:16 AM

Quote:

Originally Posted by Oznog

Well, that's what I'm trying to understand. Say I have the typical "minimum component" circuit with a gain of 20. With a supply voltage of 12v the ideal output bias is generally 6v though certainly if I know the headphones will never use more than 2v peak-to-peak and the chip can't drive closer than 0.3v from the negative rail then a 1.3v output bias is acceptable.

Anyways, say I need a 6v output bias- does this chip just work off a straight DC gain, in which case I need to give the noninverting input a DC bias of 6v/20, or does it set the bias itself? I didn't see where the example circuits made any effort to set bias externally like that.

I think you are confusing yourself by making it far more complicated than it is! - it's an audio amplifier, not an opamp - just follow the suggested circuit in the datasheet and everything works fine, there's no need to consider (or even know about) biasing.

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

15th August 2005, 08:30 PM	(permalink)
Oznog	Don't understand LM386 Looking over the LM386, I don't see the best way to use this in a circuit and could use some help. I have a system which has a single 12v supply and an audio signal limited to 0 to 5v since it goes through some digital pots for mixing various sources. That signal is biased to 2.5v, I don't have the exact amplitude, this will be determined by pot settings later. Obviously less than 5v pp. Anyways, the LM386 is an amp but not an op amp. With the basic 20 gain circuit, the 2.5v bias means an output clipping far past the +12v rail. My first thought from an op-amp perspective is of course to put the inverting input on the 2.5v bias point, but the spec sheet says +/-0.4v is an absolute max input value. So I assume this means I will need to put the inverting to ground and capacitively couple the noninverting input? That will mean the input signal will be below ground for 1/2 the cycle- is this the intended operating mode? Now what about the output biasing? It needs to bias to around 6v ideally, before the speaker coupling cap. Does the 386 do this automatically somehow or what? Or does the DC bias of the input just get multiplied by the gain and set the output bias from there? I don't see where this seems to be addressed in schematics. __________________ I thought what I'd do was I'd pretend I was one of those deaf-mutes.

15th August 2005, 09:53 PM	(permalink)
Nigel Goodwin	You seem to be somewhat confused?. The LM386 is an audio amp IC, as such the input should be (MUST! be?) AC coupled, so your worries don't apply. The signal level from your digital volume controls are not related to their supply voltage (only limited by it), so that's not a problem either. The signal applied to the LM386 should be such that it doesn't clip the output - you should ensure that is so!. __________________ PIC programmer software, and PIC Tutorials at: http://www.winpicprog.co.uk

15th August 2005, 11:05 PM	(permalink)
audioguru	The LM386 biases itself to keep its output near half-supply if the input has 0V DC or has an input coupling capacitor as shown on every schematic in its datasheet. Its inputs are actually at 0V. Make a search in Google for Headbanger and you will find some headphones amp projects using LM386's. :lol: __________________ Uncle $crooge