On the 386 datasheet -

http://www.ti.com/lit/ds/symlink/lm386.pdf

- it says "To make the LM386 a more versatile amplifier, two pins (1
and are provided for gain control. With pins 1 and 8 open
the 1.35 kW resistor sets the gain at 20 (26 dB). If a capacitor
is put from pin 1 to 8, bypassing the 1.35 kW resistor, the
gain will go up to 200 (46 dB). If a resistor is placed in series
with the capacitor, the gain can be set to any value from 20
to 200. Gain control can also be done by capacitively coupling
a resistor (or FET) from pin 1 to ground." 

Does it matter what capacitor (from pin 1 to is needed for the gain to hit 200? Also, is this a clean gain or a distortion? Can an easy dirt pedal be made by putting a pot with the cap from 1 to 8? Or a pot in place of the resistor from pin 1 to ground? Forgive me if this is really obvious...

you don't even need a cap, many dirt pedals short pin 1 directly to pin 8, some with a 1k pot for gain control. it's hard not to clip the input of a 386 for clean sounds.

*edit* for examples see: http://www.jiggawoo.eclipse.co.uk/guitarhq/Circuitsnippets/snippets.html  - uglyface/pwm front end, rambler, etc.

Hey thanks! Its nice when something is easier than you thought! Two questions:

1. I don't have a 1K pot yet - would a 5 or 10K be too high?

2. After looking at some of those schematics, I see that 386 front end is on a lot Tim E's stuff. Is he still active in DIY?

you're most welcome.

1. not too high, but you wont notice much difference beyond a certain point of rotation - above 1k-ish* funnily enough  

2. unfortunately not, as far as i am aware.

* 1k35 to be precise as that is the value of the internal resistor that you are bypassing externally with your pot that will become the path of least resistance once your pot goes beyond that point so your pot wont have too much play.

I think it sounds better to leave those pins (1 & 8 ) open and overdrive the inputs to the lm386, if you want more volume / distortion.

Andrew.

The internals are a 150 ohm and a 1,350 ohm resistor. Together (1,500 ohms) is low-gain. 150 alone is high gain.

To get in-between gains, you want to short-out the 1,350 toward 150. 10K in parallel with 1.35K is practically the same as 1.35K. To get a real change of gain, you want 1K shunting the 1.35K (as Artifus says) . So 10K is a waste of 9K. Or 9/10ths of the pot rotation does hardly-anything.

The cap: to get high gain down to 20Hz you must approach 75uFd. But we don't have 20Hz in guitar, and this is a tiny amp which can't make enough 20Hz to hear (in any practical speaker). More realistically we'd aim at 100Hz, or 15uFd. OTOH, the cap is only there to avoid a minor problem (DC-amplifying input DC error). For most musical purposes, a little (even "some") DC error is no problem. As Artifus says, we can usually omit the cap.

A sorta-ideal pot would be about 2K REVerse-Audio taper. But the 10:1 range of gain probably does not warrant a special taper. 2K Linear is fine. Since the low-gain (10) is often lower than we'd ever want, a 1K linear gives gain of 21 to 200.

As Andrew says, if you have other gain in the system, you may as well leave the LM386 fixed-gain, do your gain/overdrive action elsewhere.

Great stuff guys. I found I had a 500 ohm pot and breadboarded a circuit with the pot between pins 1 and 8, and tried different input and output caps. It certainly does work, but I'll find the ideal 1k pot soon. I'll also try what Andrew suggested and overdrive the inputs. Thanks!

I believe there is several versions of the 386 too...

LM386N-1 -> 0.3W
LM386N-3 -> 0.7W
LM386N-4 -> 1W

Not saying that more Watts equals more gain

10uf across pins 1-8...to answer your earlier question.

Probably 7-8 years ago, there was a project in Nuts & Volts that sparked my thinking about interesting ways to use the 386.  It was a TV assist for folks with some hearing loss.  It used 386s, but used a small-value cap in the pin 1/8 loop to provide more gain for higher frequencies than lower ones.  The intent was to run a speaker from the TV to where they sat and listened, but goose the top end to make the speech more intelligible. 

This is not entirely dissimilar to what the Rat does in its use of two ground legs to provide different gain for content above and below 1.5khz.  So, for instance, one could use a 220R fixed resistor and a .1uf cap in the gain loop to provide some boost to top end, but not so much that you seem to lose the bottom.

Alternatively, one can use two paths in the gain loop and adjust them.  One interesting approach might be to use a 500R pot or trimpot, tie the wiper to pin 1, and the outside lugs to a 10uf and .1uf-.22uf cap, with the other end of both caps going to pin 8.  That would always provide some boost but let you vary how muchboost is applied where in the spectrum.  Rotating it to present 0 ohms to the 10uf cap would be equivalent to the full spectrum boost you get by simply sticking the recommended 10uf cap between pins 1 and 8.  Rotating the pot the other way would give various gradations of mid/top boost.

happened to have this bookmark folder open as this topic popped up again. some 386 random resources, info and speculation:

http://www.ti.com/lit/ds/symlink/lm386.pdf

http://www.wentztech.com/radio/resources/Technical-Info/_LM386/Unleash-lm386.pdf

http://www.intio.or.jp/jf10zl/LM386.htm

http://www.ece.ucsb.edu/yuegroup/Teaching/ECE2C/Lab/Lab1a.pdf