yet another 386 distortion

[kingswayguitar]

would the 4.5V bias of the 386 be too much for the mosfet?  you could do away with the bias resistors and the cap.  I also find the gain pot needs to be small just like the datasheet suggests (1k35).  I use a 1k pot.  Love the 386!

[pinkjimiphoton]

why not ed? go for it!!  i still have a couple gigs i gotta finish before i really get a chance to play with this cct

[psychedelicfish]

if it's humming and noisy on the breadboard, it may be fine once it's boxed. gonna f around with it when i get a chance. 

I don't think it will help to box it, I'm getting way more hum than I normally would on a breadboard, I really think it needs a regulated supply.
In terms of using the 386's internal bias, I tried that today, and it seems to work ok, but it wasn't as good as the trimpot bias, I kept having my output fade out and fade in again and I think it was the 386's bias turning the MOSFET off occasionally. I measured a dc voltage of about 2.5v on the output of my 386 (nothing connected to it), do any of you get something different? It might just be a quirk of my chip. The dc on the output of the 386 may change with the load, so I might try sticking an 8r:1k Trafo between the 386 and the MOSFET and see if that changes the sound (yes, I know it would stop me using the biasing of the 386)

[pinkjimiphoton]

hmmmmm...

all that's really between the chip and the fet is that 22n.

are you getting dc on both sides of that cap? if so, try another, cuz it sounds like it may be shorted. i've run into that before.

that should be blocking ALL the dc from the input of the fet, only let the ac thru.

maybe try a cap with a higher voltage rating?

do ya have another 386? just thinking....barely. i'm new at this. 

[psychedelicfish]

I got 2.5v with absolutely nothing but the MM connected to the output of the 386

[pinkjimiphoton]

that's probably fine...check it after that 22n cap. if you still have voltage after the cap, the cap is probably the problem. it's a dc blocking cap, if it leaks any dc, it's no good.

[kingswayguitar]

my breadboard version biases at 3.9v which is close to 1/2 my battery. so its not turning the mosfet off at that point (so far as i can hear). biasing with a direct coupling to the 386 was just an idea anyway.

[deadastronaut]

might want to whack a 100uf across + -...

Surely most people would know to do that anyway?

fair enough...i'll butt out then... have fun.

[psychedelicfish]

New version:

Voltage regulator not necessary, but there is a lot of hum without it. Added a tone control, and improved the gain control, replace the 100r resistor with a jumper if you want fuzz at highest gain.

my breadboard version biases at 3.9v which is close to 1/2 my battery. so its not turning the mosfet off at that point (so far as i can hear). biasing with a direct coupling to the 386 was just an idea anyway.

I measured it again, getting a solid 4.45v... might have been the load, idk.
@pinkjimiphoton
I tired adding a 100r resistor pulldown resistor, and it didn't seem to have any effect...

One problem with this circuit is that while it has a very good (to my ear) lead tone, but it has quite frankly rubbish sustain on the high notes (I get more sustain on my clean guitar). not getting any fizz at the end of notes though, which is a plus. I might try turning the 386 stage into an aussiecomp. Any thoughts?

[pinkjimiphoton]

maybe try a transistor/mosfet stage at the input to the 386?

[psychedelicfish]

maybe try a transistor/mosfet stage at the input to the 386?

I'll try that, and i'll give the aussiecomp a go too.

[psychedelicfish]

OT, but being from NZ, and with the rivalry between us Kiwis and the Aussies, perhaps I should design a KiwiComp... a compressor way better that the AusieComp, of course

[duck_arse]

psychedelicfish, are you really calling this pedal "psychedelicfish's yet another 386 distortion"? I want to know what name I'm changing from when I build one.

"psyfch" is best I've come up with.

[kingswayguitar]

i couldn't help it (after 1:20 the pedal starts)
sorry the quality's not the best
love the pedal
never thought about putting something AFTER an LM386

[psychedelicfish]

i couldn't help it (after 1:20 the pedal starts)

Excellent! someone did a sample for me!

love the pedal

Thanks!

psychedelicfish, are you really calling this pedal "psychedelicfish's yet another 386 distortion"? I want to know what name I'm changing from when I build one.

"psyfch" is best I've come up with.

I couldn't think of a name, so I left it up to anyone who builds one to name it... you're welcome to come up with a name for it though (great! more work done for me )

I want to know what name I'm changing from when I build one.

Does that mean you're building one?

[psychedelicfish]

also, @kingswayguitar, which version did you use in the sample?

[kingswayguitar]

the one first posted by LucifersTrip
except...
i also decoupled the +ve input to ground with 10pf cap
LM386 gain at minimum
no 220n between LM386 and BS170 (so none of the indicated bias network)
i put a pot on the source capacitor (I used 0.1 uF) of the BS170 and dialed it kinda low
oh yeah, i put a J201 amp in front of the LM386 with 0.68 V cutoff (hung a 10k on the drain and 220R on the source with a 1uF source cap)

[psychedelicfish]

Sounds Good. Thanks for the clip!

[duck_arse]

yeah, well, "building one" and "building one" are two different things, but I have the all important parts and only three other things "in progress", so why not, eh?

[duck_arse]

where are these other versions? I can only see the one from lucifer. and my lm386 data sheet says "Self-centering output quiescent voltage" which, to me, means half whatever your supply rail is.

hmmmmmm.

I breadboarded this in the heat today (45C), and looking on the cro thought it had too much gain. with the source bypass cap at 47uF, the mosfet is trying to amplify the air. at full gain with no signal, the 386 is outputting something like second-hand cig smoke. I'm sure this is bad for my health.

the 386 quickly turns a signal into an 8V p-p square wave. no wonder it made that noise. and that horrible trailing hf oscillation as the signal decays. I fiddled the source bypass, but there was too much sizzle and spit at any value. I tried a 1M series resistor at the gate, and a cap to ground, but I don't think it's tame yet.

but I just had a thought for tomorrow. maybe if I run the 386 at its lowest supply voltage, 4V, it will present a more manageable signal to the fet. something to think about.