Acapulco Gold Oscillation Issue

[schrectacular]

Hello tinkerers,

Comin' at ya with another noobish question. I've put together the dual-LM386 "Acapulco Gold" pedal from tagboardeffects and as many mentioned there and elsewhere, I get some squeal/oscillation issues happening at near max volume. I've looked around and can't find any info as to WHY this is happening or how to resolve. I would love to yet again call on the experience of this group to help point me in the right direction. I know there have been several threads about this circuit, but I can't find anyone asking about this specific issue. I'm using LM386N-4's.

I've tinkered around and noticed the following:

• oscillations kick in in the last third of the A100 pot
• oscillations only start to happen when input volume is sufficiently low, as long as enough signal is coming through, there is no issue
• if there is a buffer before the pedal or I use my active-pickup guitar, there is no issue
• replacing the first IC with an LM386N-1 resolves the issue, but the pedal doesn't sound as good, more raspy and less wooly

I believe this is the layout used:


I'm slowly working through Horowitz and Hill's "The Art of Electronics" right now, trying to move from follow-directions-building to more deeply understanding what's happening. The Electrosmash articles are super helpful, more directed at the actual electronics applications I'm seeking to learn. Any suggestions on that front are welcome as well. Thanks all!

[PRR]

.....oscillations kick in in the last third of the A100 pot....

The wire from that pot wiper is going too close to the input circuits. Move it or shield it.

[Vivek]

What PRR says

Plus try placing small (maybe 100pf) caps across input and output

And make sure power supply line has bypass caps across it.

[anotherjim]

Agree with all the foregoing.
Add - Be sure to use 3PDT footswitch wiring with LED switch in the middle to separate input and output wiring.

I can't help thinking the design might be more stable if both stages used the - rather than the + inputs. There would still be a danger of overall positive feedback between input and output wires but each stage might be individually more stable.

[schrectacular]

The wire from that pot wiper is going too close to the input circuits. Move it or shield it.

This indeed solved the problem, I'm sure you are not surprised to hear! Thank you! Unfortunately for me it appears the issue lies within my test box. I tried to rewire but things are still too close. But moving the output outside the box itself resolves the issue.

Be sure to use 3PDT footswitch wiring with LED switch in the middle to separate input and output wiring.

Sure enough, inside my test box I've wired the input and output next to each other I should rewire the switch but I think that's a task for another day.

try placing small (maybe 100pf) caps across input and output

And make sure power supply line has bypass caps across it.

Help me, if you would, understand where these would go - 100pf caps should wired from main in to main out? Or from main in to ground and main out to ground?

And isn't C6 in the diagram the power supply line bypass cap?

Thanks all!

[antonis]

100pF between IN & GND and OUT & GND respective jack lugs..
2X100nF (ceramic) between pins 6 and GND (one for each amp) , as close as physically possible to pin 6..

[schrectacular]

Oh and one followup question - I've never had this happen before, is this because of the amplification power of this circuit? I've never had and issue like this in any of the other pedals that I've built.

[antonis]

Crosstalk..

https://en.wikipedia.org/wiki/Crosstalk

[PRR]

....I've wired the input and output next to each other ...

Set up a PA system. Put the microphone right next to the speaker. Turned down, no howl. Turned up, YOW.

As acoustic waves in air, electric waves in space.

Or for a sh!tty metaphor: getting drinking water right next to where you put your poop. Maybe OK for 1-2 people. In a city this makes a lot of sickness.

[schrectacular]

electric waves in space

I feel like this needs to be a motto of something, somewhere.

Thanks all SO MUCH for the help. Rewired my test box and have eliminated the crosstalk issues with this circuit, and hopefully all other future circuits!

It can be really frustrating spending so much time on things only to have them fail; I can't tell you how lucky I feel to have this forum to guide me as I stumble toward... I don't know. Mostly noisy things. I daren't say "knowledge"... how about "stumble away from mistakes".

CHEERS!

[schrectacular]

I had another question - what is the purpose of C4 here? And what would happen if a slightly higher or lower value was chosen, say a 3n9 or 10nf? I thought it might be some kind of lag compensation, but isn't that usually for a feedback setup? And also usually much lower values are used for that.

[PRR]

> what is the purpose of C4

Probably another part involved? Maybe a resistor? Have you sketched the frequency response?

(In this case you could even tap that signal and hear how it sounds half-processed.)

[Vivek]

As PRR hints

It is a filter

Along with that 68k


Search Google for "RC filter calculator" and put in those values to find out the frequency of that filter

Then look at topology and guess is that a HPF or LPF


(The low 50k impedance of the LM386 plays a part in the filter / voltage divider too)

[Vivek]

Oh and one followup question - I've never had this happen before, is this because of the amplification power of this circuit? I've never had and issue like this in any of the other pedals that I've built.

I think the gain of the first stage is 200 and gain of next stage is 20

That leads to a total gain of 4000

Which is rather high

And that makes it more susceptible to feedback creating oscillation.



There are many things that could be "improved" in this schematic.

Like better input impedance

Better frequency response at input, interstage and output.

Variable gain

etc

[schrectacular]

Ok, so it’s a low-pass filter cutting off frequencies above about 500hz. Decreasing to a 3n9 would raise the cutoff frequency to around 600hz resulting in a slightly more trebly pedal while increasing to a 10nf would lower the cutoff to around 235hz resulting in a more bassy sound.

[Vivek]

This pedal would produce a tighter sound if a HPF of 500 to 900hz was introduced right at the beginning

[PRR]

I think the gain of the first stage is 200 and gain of next stage is 20

Note that Zin of the '386 is like 50k, so the 68k resistor makes a broadband divider, even before the caps start cutting.