questions about LM386 based distortions

[blanik]

i just made a LM386 based distortion and WOW those thing have all the features i like about dist/OD (cleans up well, very dynamic, a sustained chord ends up sounding clean, etc...)

the Sonic Titan from DAM and the Big Daddy and Grace from ROG all share the same design (except on how the drive control is achieved):




my questions are, what produces distortion in those, the LM386 or the FET in front of it? i keep lowering the resistor on the drain to make the fet as clean as possible but it still distorts a little (is it because the LM is maxed at 200 X gain?)
what would be the best way to achieve a clean boost on this design? and how about using a mosfet booster like the SHO or AMZ in front of the chip?

[anchovie]

The Grace has the JFET configured as a buffer, so no gain from that section. Therefore, any distortion you hear in that circuit is purely from the LM386. The Big Daddy/Sonic Titan overdrive the input of the LM386 as well.

According to the datasheet, the 386 has a maximum input voltage of +/-0.4V, so I guess anything above that gets clipped on the way in. As that threshold is so low, I doubt that using a MOSFET for the boost would make a huge difference as a component choice; the sound would be influenced more by what frequency range the booster is configured to amplify.

If I wanted a clean boost, the 386 wouldn't be anywhere near the top of my list of choices. A hot pickup is going to produce an output that would get clipped by the IC input before any amplification took place.

[blanik]

thanks for the reply, i thought the 0.4V thing was the voltage used to run the IC 

so if i understand this right, even if i leave IC pins 1 and 8 not connected to have the minimum gain of 20 i'll still probably get clipping?

[Brymus]

Yeah but its a good sounding clipping IMO
Try building a Ruby then put different boosts in front of it,TS,Rangmaster
My fav 386 OD is  D Swhartz'  Burning Crunch.

[Electron Tornado]

If I was going to re-design my OD386, I would try something more like the Sonic Titan on the front end. I would have a toggle switch to select between max gain (200) and maybe a gain of just 20 or 50. That 1k drive pot doesn't do much of anything as far as sweep. I could still put a couple of caps and a "fat" switch at C2 or C3 from the Sonic Titan schematic. After the IC, the rest would be from my OD386.
(Never quite liked how notes die off, but kind of a cool pedal.)

http://img163.imageshack.us/img163/9281/od386v21schematic.jpg

[blanik]

If I was going to re-design my OD386, I would try something more like the Sonic Titan on the front end. I would have a toggle switch to select between max gain (200) and maybe a gain of just 20 or 50. That 1k drive pot doesn't do much of anything as far as sweep. I could still put a couple of caps and a "fat" switch at C2 or C3 from the Sonic Titan schematic. After the IC, the rest would be from my OD386.
(Never quite liked how notes die off, but kind of a cool pedal.)

http://img163.imageshack.us/img163/9281/od386v21schematic.jpg

interesting design, why did you put another transistor at the end (as the actual output of the LM386 is quite high)?

the B1K as a drive control isn't set the same way on your design than on the ROGs, on the latter it's a variable resistor between pin 1 and 8 on yours 1 and 8 are permanently connected and the 1k goes to a cap and ground?! that wouldn'w have the same results? (afaik)

(and also, i've often seen the 1M resistor to ground on input followed by a cap than another resistor to ground?! i've never understood why two resistor there?)

[Electron Tornado]

interesting design, why did you put another transistor at the end (as the actual output of the LM386 is quite high)?

the B1K as a drive control isn't set the same way on your design than on the ROGs, on the latter it's a variable resistor between pin 1 and 8 on yours 1 and 8 are permanently connected and the 1k goes to a cap and ground?! that wouldn'w have the same results? (afaik)

(and also, i've often seen the 1M resistor to ground on input followed by a cap than another resistor to ground?! i've never understood why two resistor there?)

The tone control will reduce the signal strength, so the transistor boost stage at the end is there to boost the signal level. There's also a big resistor right after the 386 because the 386 output is also quite noisey.

The drive control: Read the data sheet for the LM386. The drive control is set up like Aron Nelson's Smash Drive.

Early on, I wondered why there was no DC blocking cap on either the Grace Overdive or Big Daddy, but couldn't find any explanation in any of the books I have. So I asked the question on one of the effects pedal forums and was told, yes, you need a DC blocking cap. So, I put one in there. Didn't seem to hurt, and then it needed a pulldown resistor, so the 1.5M was added as well. Notice that the Sonic Titan doesn't use an input cap either.

Maybe it's worth asking the question again - does there need to be a DC blocking cap at the input of those circuits? If not, why not?

[blanik]

for what i read, the fet was added as a buffer just because the input impedance of the LM was too low but you're right usually even the average buffer has an input cap...

on your design, would leaving out the 5M resistor change anything? (or putting it in place of the 1.5M)

[Electron Tornado]

on your design, would leaving out the 5M resistor change anything? (or putting it in place of the 1.5M)

If you breadboard this, go ahead and try leaving out the 5M resistor, and post your results. If you research the buffer circuit and find a source with a good explanation, please post that as well.

[Paul Marossy]

I don't like an LM386 distortion circuit without an input buffer. It doesn't sound very good.

[PRR]

> I wondered why there was no DC blocking cap on either the Grace Overdive or Big Daddy

At the input?

The gate of the JFET is at ground. "IF" you know the preceeding source does not leak DC, you may simply tie ~~1Meg to ground to hold it when nothing is connected.

But a DC-leaky box ahead of it will throw the JFET out of bias, no sound.

It may be generally safer to use a blocking cap.

And when you have a blocking cap and are rapidly re-plugging stuff with the speaker on, it is wise to have another resistor to bleed any charge the cap may have picked up.

> it still distorts a little (is it because the LM is maxed at 200 X gain?)

It won't make 3V out, so the max undistorted input is 3V/200 or 0.015V or 15mV. This is a very weak guitar signal. "All" sounds except long decay are distorted. That's not in the JFET, which even if badly lamed can pass all but the strongest guitar signal.

Back off on the Drive control so gain is not so high.

Some LM386 will still be a bit farty even when not overloaded. It was designed for low power consumption, which conflicts with small-signal cleanliness. And most LM386 users (modems, answering machines) just don't care. You may need to try different sources/batches to find one which stays very clean in the "clean" zone.

> datasheet, the 386 has a maximum input voltage of +/-0.4V

That must be for low gain or a "damage" rating. Obviously at gain of 200 it can't make 0.4V*200= 80V peak cleanly. Even at gain=20 it won't make 8V. {later} Yes, the +/-0.4V is under "maximum ratings", "limits beyond which damage to the device may occur.". If you exceed this, and it dies, you can not complain. Actually you'd have to force it rather hard to do serious damage. Still, this is a "damage" rating, not an "operational" rating.

[stringsthings]

so if i understand this right, even if i leave IC pins 1 and 8 not connected to have the minimum gain of 20 i'll still probably get clipping?

someone correct me if i'm wrong, but wouldn't he get an ungodly amount of gain ( infinite resistance ) by disconnecting pin 1 from pin 8?

[Electron Tornado]

so if i understand this right, even if i leave IC pins 1 and 8 not connected to have the minimum gain of 20 i'll still probably get clipping?

someone correct me if i'm wrong, but wouldn't he get an ungodly amount of gain ( infinite resistance ) by disconnecting pin 1 from pin 8?

No, you'll get a gain of about 20. Read the 386 data sheet.

http://pdf1.alldatasheet.com/datasheet-pdf/view/8887/NSC/LM386.html

[PRR]

> minimum gain of 20 i'll still probably get clipping?

hmmmm..... yup, a not-lame guitarist/pickup can probably clip even at gain of 20.

In fact that's what the Grace diagram seems to say: "from just breaking up to....".

You can insert a resistor and reduce gain. Keep this resistor near the '386; a long wire will suck-up buzz and instability.



Or you could bypass the '386 for true-clean.

[Electron Tornado]

Paul R., 

Do you know why notes fade with a "buzz" at the end of the note? It seems like as the note fades out and signal strength decreases, that there is a "buzz" or "zip" sound as the clipped tone ends leaving the tone clean to completely fade away.

[PRR]

Add the 0.1uFd 10 ohm compensation network shown on the LM386 documentation.

Some speaker-amps will go crazy without this. The mild '386 tends to "work", but as you come out of overload it may get jittery without the cap to calm it and the resistor so it doesn't get too calm.

[Caferacernoc]

Add the 0.1uFd 10 ohm compensation network shown on the LM386 documentation.

Some speaker-amps will go crazy without this. The mild '386 tends to "work", but as you come out of overload it may get jittery without the cap to calm it and the resistor so it doesn't get too calm.

Are you referencing pin 7 to ground?
Thanks.

[Electron Tornado]

Add the 0.1uFd 10 ohm compensation network shown on the LM386 documentation.

Some speaker-amps will go crazy without this. The mild '386 tends to "work", but as you come out of overload it may get jittery without the cap to calm it and the resistor so it doesn't get too calm.

Do you mean the series RC between pins 1 and 5, or the series RC from pin 5 to ground?

Are you referencing pin 7 to ground?
Thanks.

There is a 10uf cap from pin 7 to ground.

[askwho69]

This is an old thread but stil intriguing i like the sound of 386 but still difficult to tame some buzz and decaying notes!
the reason i want to UP this thread is to know the answer of the .1nf and 10ohms said by PRR i like this genius man very helpful to the diy pedal community!!


A2

[askwho69]

I doubt that using a MOSFET for the boost would make a huge difference as a component choice; the sound would be influenced more by what frequency range the booster is configured to amplify.

Can i ask about this? why not MOSFET what is in the 386 and MOsfet?

A2

Ps sorry for double posting