new wah idea

[vendettav]

so erm i was reading this thread about sound to led and got this



and  actually thought, if it's that easy, why not have a morley wah for example or any other inductorless wah and have the fixed always-on LED replaced by this circuit, this way we dont need an expression pedal and we have a auto wah/filter

will this work?

I'm thinking a pot at the pin 5 could be a range knob.. now to think of a release and attack knobs and that thing could be a killer design.. hm?

let me know if im on a right route


thanx

[deadastronaut]

yep that'll work.....quack quack...good idea on the range pot...

it could be a distortion and an envelope wah too....

would be very quacky though!...

release would just be sustaining a note...i guess..

hmmm...not sure on the attack though!.but the range might take care of that....another guess......but yeah it'll work for sure!,

[vendettav]

nice, gotta find time to put it all up together on a breadboard

how would you go about making the LEDs to light slightly slower and later than the note is being hit? mmm I think it won't be hard, I just have bunch of classes skipped i guess 

[deadastronaut]

hmmmm...no idea!...through an electrolytic maybe?...dunno on that!...

i'd be interested to find out though!...anyone?..

[Paul Marossy]

Interesting idea. I'd like to hear it in action, though.

[Joe Hart]

Very interesting.
-Joe Hart

[Earthscum]

I've tried an auto-wah with the 386... it worked pretty good, from my memory. I think I just used a cap to in, strapped an LED across to a resistor and cap to ground. To control it, I kind of remember just using a 50k pot from the other input to ground. I can't remember which input I used, though.

The neat thing I found about using the 386's is that you can control the frequencies that activate the envelope. Look at it this way, you have a shorted gain of 200, and an open gain of 20. If you make your input not light the LED at 20 gain, but it will with pins 1 and 8 shorted, then you can select which frequencies activate the envelope only within the band you select.

I just used a Colorsound inductorless wah circuit (also used the 386 playing around with the Morely wah circuits).

It just occurred to me that you could use this in the Neutron filter pretty easily!

[vendettav]

nice, when i was writing this i thought it was me only who thought it had potential but seems like i'll have to dive into further, i'll do a read about the LEDs and Caps Rob, i've been thinking about that thing for a long while (tho not for a wah or similar thing but just how it could be done you know)

[Gurner]

I think this will be trickier to pull off than you might first imagine.

The LM386 will have 1/2 VCC on its output...so 4.5V DC quiescently (assuming a 9V supply...by the way forget batteries for this implementation ...becuase as a battery fades so will the whole delicate balance of the output & forward voltage to the LEDs).

If using two LEDs in series, that 4.5V instantly rules out red & yellow leds (else they'd be on all the time)

If you do want to roll with this circuit, go with green,. blue or white leds....but even then, it'll be a tricky implementation for what you seek.

Think about it.....4.5V DC quiescently on the output of the LM386. Along comes your signal ...and 200x gain is applied & therefore the outcome is therefore pretty much binary ...the leds will be full off (4.5V) or full on (8.2v about the most you can squeeze out of an LM386 with a 9V supply)) without much going on in between.

Ok, so you could put a pot in front of the LM386 to tame the input signal (& by defaut the output signal) or try to set the optimum gain of the LM386 itself, but still...even with it tweaked for best visual 'results'....those two LEDs are gonna see 8.2V across them (when they're not really meant to have more than about 6.4V ...ie 2 x 3.2V forward voltage for say two blue diodes)...ok, so that excess voltage will be fleeting/transient, but nevertheless...not good.

So ok, you think "I'll put a limiting resistor in front of the LEDs", & yes, that might assure that the max voltage they see is 6.4V, but then the LEDs be further cutoff under 'no signal' input conditions (whereas before they were mildly cutoff)...meaning they'll then need a large incoming signal before they'll even turn on.

oh, imho for attack/release control - better to do that in front of the lm386 (alas, more parts count!) .....cos even putting a large electrolytic on the LM386 output is gonna see the charge drained rapidly by the LEDs, meaning you're gonna struggle to get much of a fade/release vibe going down.

I'm not saying it can't be done, or that the journey won't be fun....but I wouldn't expect fine/granular control of say a wah using this implementation.

[vendettav]

hmm so LM386 outputs only half a volt?


hmm you sure if I put 2 resistors in series to the diodes and have a electrolytic cap from between the 2 resistors to the ground won't do much of a smooth attack release on the LED?

[Gurner]

hmm so LM386 outputs only half a volt?

No, that's not what I said ....an LM386 output will always be near half VCC (VCC being your supply voltage, so if your supply is say 9V,then  your LM386 quiescent output will be half of that @ 4.5V DC)

hmm you sure if I put 2 resistors in series to the diodes and have a electrolytic cap from between the 2 resistors to the ground won't do much of a smooth attack release on the LED?

But then you're also messing with the delicate forward voltage biasing of the LEDs (ie by putting in some resistors   ...best you could hope for is very low value resistors, else too much voltage would be dropped across the resistors themselves leaving no voltage for the LEDS....but then to get any sort of impact on attack you'll need a hulking cap), so therefore I reckon it's better to approach the attack/release aspect before the LM386.

LEDs work best using PWM...this is where the voltage across them remains constant, but you simply switch this preferred voltage on & off very quickly depending on how bright you want the LEDs. I'm sure you'll have a great wheeze dabbling with this, but I'm also sure that everyone's end result will vary ....in other words repeatability with such a method is gonna have lot tearing their hair out.

[slacker]

Have a look at the Ugly Face, that uses an LM386 as a simple envelope follower driving a LED.

[vendettav]

hmm so LM386 outputs only half a volt?

No, that's not what I said ....an LM386 output will always be near half VCC (VCC being your supply voltage, so if your supply is say 9V,then  your LM386 quiescent output will be half of that @ 4.5V DC)

Oh Ok, my mistake

hmm you sure if I put 2 resistors in series to the diodes and have a electrolytic cap from between the 2 resistors to the ground won't do much of a smooth attack release on the LED?

But then you're also messing with the delicate forward voltage biasing of the LEDs (ie by putting in some resistors   ...best you could hope for is very low value resistors, else too much voltage would be dropped across the resistors themselves leaving no voltage for the LEDS....but then to get any sort of impact on attack you'll need a hulking cap), so therefore it's better to approach the attack release aspect before the LM386.

LEDs work best using PWM...this is where the voltage across them remains constant, but you simply switch this preferred voltage on & off very quickly depending on how bright you want the LEDs. I'm sure you'll have a great wheeze dabbling with this, but I'm also sure that everyone's end result will vary ....in other words repeatability with such a method is gonna have lot tearing their hair out.



[/quote]
[/quote]

hmm so what about a switch? that'd turn on/off a release/attack mode

anyways i get what you are talking here but you really think this is not worth a try?

[Gurner]

anyways i get what you are talking here but you really think this is not worth a try?

Sure, it's worth a try...go for it.

[deadastronaut]

Have a look at the Ugly Face, that uses an LM386 as a simple envelope follower driving a LED.

oh yes......of course it does!...


@gurner, dont forget the zobel...

ive  had 8 'blue' leds running sound to light out of the 386 in the past?....4 run really bright for sure!

[Earthscum]

Have a look at the Ugly Face, that uses an LM386 as a simple envelope follower driving a LED.

Aye, I think that's what started me on it that day. I set mine up output => cap => LED => Cap/Pot to ground. After building the nurse quacky, I would probably add in series resistance to shape the attack (think that's where I pulled everything... it had too much attack and not so much quack, and I didn't know a whole lot about EF's at the time). I used the LED to drive a CDS cell. Now I have some opto's, and more knowledge, I might have to try this out again.

Basically, uglyface worked for me, and setting up the output similar to the NQ worked for me. I remember that it was the first follower that actually worked for me , and it was easier. No bias, just set gain to max (pin 1-8 jumper), use a pot to ground at the input for "sensitivity". Maybe a little tone shaping before hand to avoid false triggering (Ef... false triggering... lol...). erm, set frequency selection. It's morning, I'm not quite awake yet.

[vendettav]

this is actualyl encouraging but are you sure it runs on blue LEDs? in the ugly face thread they told me to have red LEDs... hmm this tho look a bit different from uglyface... gotta try it