smallest 5 led mono vu meter circuit diy?

[deadastronaut]

hi guys i'm messing around with leds again..

heres what i want to do..5 leds, mono/guitar input..

o   o <  o  > o   o

centre on , 2 left/2 right...that react to louder strum..

i'm looking for a tiny circuit (for use with guitar input)  preferably transistor..no ic, (it needs to be as small as possible..

any ideas?..

[artifus]

http://www.techlib.com/electronics/bargraph.htm you could also use a 4049 but be aware of the different pin out.

[frequencycentral]

Rob you'd love LM3914.

[Jdansti]

Rob you'd love LM3914.

Rick-Would R4 in this example be used to calibrate the circuit?  For example, in a given application, I might want LED #8 to light up when the audio signal reaches a point that it would begin to cause undesirable distortion.  So would I calibrate it by playing the audio signal at the level where it starts to sound "bad" and adjust R4 to where the 8th LED just starts to turn on?  Sorry if the question doesn't make sense.   

[Fender3D]

Hey Rob,
I didn't understand if you need LEDs couples (2-4, 1-5) lighting simultaneously or if they are 5 independent LEDs, anyway, you may split/adapt these comparators



I think doing it with transistors would take more parts or PCB room, even subbing trimmers with fixed resistors...



BTW

hi guys i'm messing around with leds again..

LEDs are amazing....
ever think put one under those lovely footswitch white washers?

[earthtonesaudio]

The LM3914 simplifies the LED driving considerably, but you'll need some sort of rectification or envelope extraction for any of these.  That could be as simple as a single diode and resistor though.

[Gurner]

My vote would be for a AN6884...

http://www.e-ele.net/DataSheet/AN6884.pdf

(it's about as 'low parts count' that you can get for VU meter)

It only needs 2 caps, one resistor & a trimpot (for sensitivity) .....connect up 5 your leds, feed it a signal & you're off to the races (though you'll need a buffer if feeding a raw guitar signal as its inpunt impedance is quite low)

Not the easiest of parts to source nowadays, but available here...

http://www.cricklewoodelectronics.com/Cricklewood/product.php?productid=10017&cat=0&page=1

It's what I used in my LED sustainer driver...

http://www.youtube.com/watch?v=xXB92cSQi8E

About a year ago, I coded up a VU meter in software (a PIC)....for a mock LED pickup (6 LEDs) that has about 8 patterns (sound to light, chaser patterns such as knight Rider etc!) though going this way is enough to make the sanest of souls totally barking mad.....never again.

Edit: Doh.....just re-read that you want it to move from the centre outwards (that ain't much resolution ....2 LEDs either side of centre!) ...but rather than delete my input, I'll leave it here for others who might want a more traditional VU meter & find this thread via a search!

[.Mike]

If you end up going with the LM391x ICs, it is worth noting that the 3914 is linear, the 3915 is log (3db/oct), and the 3916 is VU.

There are a couple of dedicated calculators out there for these ICs that I can't seem to find, and don't have permission to distribute. I have a couple for the 3914, but wasn't able to find one for the 3915 or 3916. If you really need 'em, I can zip and email.

Mike

[earthtonesaudio]

[Gurner]

i can't see how that particular circuit would do what rob wants.

i think rob wants the centre led of the 5 to come on fIrst, then the leds either side of centre (as the signal gets larger) & finally the other LEDs (with the signal at its largest).

With VU meters, normally comparators are involved (one needed for each level 'step'), which involves a fair amount of supporting components.

[deadastronaut]

^ yep thats it.. centre on, 2 left/ 2 right...pulsing outwards...(looks like a cool chip, cheers man )

wow a lot of choice, i do have a 3915...i just thought i may be able to get away with using something smaller/simpler.., but like fender3d said probably end up with more components anyway..

(:fender3d : no way   )

i was messing around with the good ol 386 sound to light last night, and nearly got a faux vu...by increasing the left/right resistors..

but i'll defo take a look into those other ways...cheers guys..

edit: i'll dig my 3915 out for now and see how that looks...

[deadastronaut]

oops i have a 3914 , not 3915....

http://www.eleccircuit.com/wp-content/uploads/2010/11/VU-meter-10-pieces-using-ic-LM3914.jpg

i tried this circuit,with 10 leds for now,  but its behaving weird...its lighting up by itself going up n down regardless of guitar input..

does it need an input buffer or something, or shouldit be ok as is?...


edit: i'm guessing what alex said!..


anyhow: i just bought AN6884's, got 10 for £4.00 tayda...free delivery..they look ideal. cheers G.

[earthtonesaudio]

Just to clarify my above drawing, the "LO" string would contain the LEDs immediately left and right of center, and the "HI" string contains the outermost LEDs.  Note the "HI" string contains an extra diode (could be Si or Ge) so that it requires a lower collector voltage to light up.
I was assuming the center LED would only be a status indicator (always on unless the effect is bypassed), so that's why I only used two levels.



Here's another using the same principle (diode breakpoints), but for 3 levels.  This one doesn't require a bias pot.


[edit: the above doesn't work quite right in the simulation.  Still needs peak detection.]

[Nasse]

ttp://www.spelektroniikka.fi/kuvat/vumittar.pdf there is small pcb layout for that other chip if someone needs

[Gurner]

ttp://www.spelektroniikka.fi/kuvat/vumittar.pdf there is small pcb layout for that other chip if someone needs

That circuit uses the IC (AN6884), which I recommended a couple of posts back, but then re-read Rob's initial post & realised that the leds won't react how Rob want's them to. That said, the AN6884's supporting circuit could fairly easily be adapted (just use the first three LEDs - centre LED would be LED1 - & a little bit of extra circuitry to 'duplicate' LEDs 2 & 3)

[earthtonesaudio]

Buffered Peak detector:
simulation
...Vary the attenuation to see how it reacts to different signal levels.

[Gurner]

I like it   , just a couple of thoughts....

1. The voltage follower is biased to ground (i.e. no AC signal= 0V at the output of the buffer)....I think such biasing will result in a 'dead region', where after a period of no AC signal, once the AC signal then arrives it will need to charge up that cap from 0V to near the LED's  approximate forward voltage beforethe LED will even illuminate. Probably Better to bias the input to the voltage follower to somewhere below the diode's forward voltage.

2. You might struggle to find an opamp that'll deliver the current that 5 LEDS will sink.

[merlinb]

centre LED would be LED1 - & a little bit of extra circuitry to 'duplicate' LEDs 2 & 3...

2. You might struggle to find an opamp that'll deliver the current that 5 LEDS will sink.

Just put each outer pair of LEDs in series. That way you technically only need enough current for 3 LEDs, or about 3mA, which any opamp will do.

[deadastronaut]

nice one alex: i'll give that a go later...looks ideal on the sim...

@merlin: yep i'll try them in series too..

cheers guys..


edit: tried it with an  072  and using 1n4148, and 1-3 leds for now...didn' get any fluctuation in light when playing guitar though!..

correct diode?

[Mustachio]

Sorry I cant add more tech info on this. About a year or so ago I was playing around with an LM3916 and was able to get it to work with a very simple setup.

Ill have to look for the pics i took of my breadboard when I had it working. I started moving it over to vero to slap inside a sonic maximizer clone I had built.

Still on the back burner but it did work with 10 leds and worked very nice or so it seemed. I could get the leds to react with a dry passive guitar signal into it alone and it would peak the leds out if i strummed really hard or subtle led reaction on light picking.

If i recall I used one low value resistor at the input pin and no buffers just very simple low parts count. Haha now I remember I accidentally knocked the circuit into a glass of coca cola and left it off to the side to dry and never went back to finish the vero.

I did test it on bread board on the output of the sonic maximizer and it worked really well. I was less knowledgeable then as I am now and I'm not very knowledgeable so ill have to revisit this. I was trying to read my way thru the schematics on the datasheet for the lm3916 and when I would try to feed the chip with the buffers they suggested I didn't get any signal or not what i hoped for(one led staying on others not lighting up or weak signal) . I think I used a 60-70r resistor and thats about it before input.

I know the info I gave is remedial just hoping it might help some one at some time. And if any of you can give correct info on using the lm3916 for guitar that would be awesome