"Signal" LED

[Gurner]

A you say, a simple 'A' on the end makes all the difference (& I'm so used to working with the 'A' variant that I'd overlooked yours wasn't!)...well done mate, I doubt anyone can trump a one component circuit in a lowest parts count competition! (that said most want to light LEDs...so you gotta add in at least one resistor!  )

[Seven64]

cool, i was just using single op-amps.  i got some 386's on the way, hopefully this is as simple as it sounds!

[deadastronaut]

it is..

[Seven64]

it seemed kinda buggy..... sometimes the LED locks on, and i cant get it to stay off.  I was also trying to figure out how to do it with less than 3 LED's, but it seems like it gets more unstable.  Everything i tried with 1 led just leaves it always on.....

[mistahead]

Increase the resisance to the one LED?

Sound like its not allowing the voltage to dip low enough to switch the LED off.

[Gurner]

it seemed kinda buggy..... sometimes the LED locks on, and i cant get it to stay off.  I was also trying to figure out how to do it with less than 3 LED's, but it seems like it gets more unstable.  Everything i tried with 1 led just leaves it always on.....

It's not buggy as such, but very sensitive to battery level, biasing & LED colour used.

the output of the LM386 will be about half your supply voltage...so if using a 9V battery, the LM386 output starts out at about 4.6V  (a fresh battery being about 9.2V), but pretty soon the battery will fade, therefore so will the LM386 output voltage (which will start chaning the LED's response)  - therefore if you are wanting to use such a simple circuit, make sure you use it with a regulated power source!

next problem... each batch of LEDs wil likely have variances in characteristics.

Let's assume you want to use RED LEDs (which typically spring to life between 1.8V & 2V)....ok let's assume you are using a regulated 9V supply...there'll be 4.5V on the LM386 output...that's too much for one RED LED (it will be always on), it's even too much for two RED LEDs in series. Therefore if you are using RED LEDS, you need to use three in series for them to be off with no AC signal present. but they won't turn on until the AC signal gets above 5.8V-6V, so there's a 'dead band' in play here (i.e. between the quiescent 4.5V & the 'turn on poin of 5.8V). If you wish to use Blue LEDs, they normally have a fwd voltage of about 3.3V, so two in series would be 6.6V...an even greater dead band.

So whilst the circuit is  simple, you need to cater for the LED type & for your own particular supply voltage level.

Ideally what you want to do is glean the voltage level where the LEDs you are using just turn visibly off off, & then arrange your supply level at twice that voltage....then when you attach the LEDs to the output of the LM386, they should be just off with no AC signal present...and no deadband until they turn on.

IMHO puretube's offering is both simpler & more flexible (for a start there's no DC voltage across the LED in the absence of a signal, which means the LEDS used will definitely be off when no signal is present!)...just remember to put some series resistance in line with the LEDs....and bear in mind the TDA7052A swings to about 15V off a 9V battery!

[samhay]

If you have that much voltage swing to play with, why not try ac coupling the LED? - put a cap between the power amp's output and the LED (you still need a series resistor).
Edit - what is the reverse voltage of and LED?  You might need 2 anti-parallel with this approach.

If that doesn't help, take a look at how the Tremulus lune's LEDs work - you can play with the DC levels with 2 resistors - one in series with the LED and the other parallel with the LED.

[Gurner]

If you have that much voltage swing to play with, why not try ac coupling the LED? - put a cap between the power amp's output and the LED (you still need a series resistor).

That's one way, but by doing so, you're essentially increasing the 'dead band', because with such a an arrangement, the LED will be hard off with no AC signal...& the AC signal will need to swing up to the fwd voltage of the LED before it'll even turn on. So with an LM386 at 9V, the AC sits on 4.5V, for a single blue LED it'll need to go up to 7.8V for the LED to start illuminating.....an LM386 isn't a rail to rail device, so that's quite tight @9V supply.

[deadastronaut]

if you want to get fancy try the AN6884 5 led vu meter instead...

very cool..

as recommended by gurner

http://www.aronnelson.com/gallery/main.php/v/chickpea/5ledvumeterpcbDA2013.jpg.html

i had it  on breadboard, and swapped a trim for 2 resistors..works great.

and had 10 leds running <<<<<>>>>> like so..

[samhay]

If you have that much voltage swing to play with, why not try ac coupling the LED? - put a cap between the power amp's output and the LED (you still need a series resistor).

That's one way, but by doing so, you're essentially increasing the 'dead band', because with such a an arrangement, the LED will be hard off with no AC signal...& the AC signal will need to swing up to the fwd voltage of the LED before it'll even turn on. So with an LM386 at 9V, the AC sits on 4.5V, for a single blue LED it'll need to go up to 7.8V for the LED to start illuminating.....an LM386 isn't a rail to rail device, so that's quite tight @9V supply.

Right you are. Might be an interesting way to do a clipping monitor.

The 'Tremulus Lune' suggestion should work reasonably well though. If you use 1k resistors, then you will have about 0.1 mA through the LED without signal. If it is not an ultra-bright, then this might look off.