yet another 386 distortion

[kingswayguitar]

i find i need to run the LM386 at close to its lowest gain too.  and i keep my pot on the mosfet;s source bypass capacitor pretty low as well.  and i keep the output volume way down.  this thing's got way too much gain 
but i love it
i wonder if some global feedback would help

[duck_arse]

there is another gain-control method mentioned in the 386 data sheet, with a cap and pot to ground at pin 1. I thought that would replace the pin 1 to pin 8 control, but I see other circuits (the /dev/null 2) have both.

so I tried it. more gain? is this possible? and now it does something I've never heard before.

running the 386 at 5V, with 10uF and pot between 1 + 8, and 47uF and pot between 1 + gnd, a 22k series gate resistor and 56k//3n3 gate-bias, 330nF/10uF/220uF source bypass cap, a humbucker input, and with gain controls both at 10 (everyone's favourite) the output sustains down to near 0 input, when the oscillations and filth comes up and over eveything. it just keeps fuzzing and fuzzing and fuzzing.

maybe we got something here after all.

about the output offset: I measured 4V1 at the output pin with 8V99 supply and 2V2 out with 5V0 supply, so half rail is not quite half rail, at least for my "gl386".

[azrael]

Surely most people would know to do that anyway? I'm starting to think it needs a voltage regulator, I have about 3 100uf caps across my voltage rails, and i'm still getting a lot of hum\

The 386 just seems to be quite noisy. I'm working on a circuit using one, and it hums too.

[kingswayguitar]

Surely most people would know to do that anyway? I'm starting to think it needs a voltage regulator, I have about 3 100uf caps across my voltage rails, and i'm still getting a lot of hum\

The 386 just seems to be quite noisy. I'm working on a circuit using one, and it hums too.

I got no more noise than any other circuit (pretty quiet really).  Sorry I don't know all the theory but I do know from practice that their are designs (like this one) that need the unused input to be decoupled from ground (i.e. capacitor from input to ground).

hope this helps

[psychedelicfish]

I got no more noise than any other circuit (pretty quiet really).  Sorry I don't know all the theory but I do know from practice that their are designs (like this one) that need the unused input to be decoupled from ground (i.e. capacitor from input to ground).

Did you do that when you built it?

[kingswayguitar]

I have a few LM386 designs that use +ve input and each one I checked left the unused -ve input unconnected.  I have one freaky distortion that uses -ve input and the unused +ve input is connected to gnd via a ceramic cap.  When I was getting a bit of noise with this one I stuck a cap on the unused +ve input to gnd.  Noise gone.  This (psychedelicfish's) cicuit is still on the breadboard BTW.  I think it will be my next PCB (just finished a fuzz and I need a break from mindracking delays and phasers).

[azrael]

This is the one I'm working on, it's a design by someone on here, I forget.
http://i131.photobucket.com/albums/p315/naazrael/tophatschematic.png
unused input grounded.
Still noisier than I would like.


don't mean to threadjack, btw.

[duck_arse]

I was getting no noise problems except at MAXX gain, and using the humbucker cut most of that. unused input grounded, like the specs say.