VU meter issues

[Jimmah]

I'm making a VU meter for a computer stereo unit, but I can't get the ah heck to work right. Circuit I'm using

Only thing I had to change was instead of the 1.2k resistor I had a 1k, best I could do with my limited supplies. put all the led's on some cut up RCA cable (to reduce overall wire amount) and they look good, doing their thing to the music, only for a short while. I'm using a 12v dc supply.

Here's the issue. For a short while the led's will work right, but only if I totally crank up the music (using my existing speaker's output and via computer) but the led's don't move too much (its like the music is at too high a level for them but the top end led's aren't even lighting much/at all). Also, after a while all the led's just stay on, even after I unplug the audio-in.

I'm a total newb with electronics, is this a good circuit or can it be modified to make it better? Sorry about my lack of proper terms.

 

[PottedMeat]

http://www.national.com/ds.cgi/LM/LM3915.pdf

If it starts out alright, then screws up, it seems like overheating ( lower the LED drive current ) or oscillation.

See page 9 on the datasheet and run through what it suggests.

 

[Jimmah]

Ugh, I tried re-wiring it to no avail. I guess I'll stick to simple off/on circuits for now, IC's and whatnot are a little out of my league.

 

[Rubycon]

Classic dot/bar driver.

Try isolating your inputs. Perhaps you have RF interference. Also there is a 35V peak to peak limit on the input. You could also have a doggy ic. These are a lot of fun to build and when properly calibrated are rugged and reliable.

 

[bobsmith1492]

Also decouple the power input with a pair of capacitors; connect a 0.1uF ceramic capacitor and a 10uF capacitor between pins 2 and 3. Also put a 10uF capacitor where the datasheet shows: http://www.national.com/ds.cgi/LM/LM3915.pdf

The circuit you linked to is wrong, also; you need to use a lower voltage supply for the LEDs or you're going to overheat. Page 2 on the datasheet says to keep the LED supply below 7V to keep from overheating.

What are you using for a power supply? You could hook up 4 AAs or D batteries to drive the LEDs with 6V; that would work fine.

I made a pair of these on a breadboard back in the day and ran 'em off a pair of 9V batteries; hooked 'em up with a 1/8" jack on one end to a pair of RCA jacks on the other and hooked it between my stereo and CD player or computer just for fun.

 

[Jimmah]

Originally posted by: bobsmith1492
Also decouple the power input with a pair of capacitors; connect a 0.1uF ceramic capacitor and a 10uF capacitor between pins 2 and 3. Also put a 10uF capacitor where the datasheet shows: http://www.national.com/ds.cgi/LM/LM3915.pdf

The circuit you linked to is wrong, also; you need to use a lower voltage supply for the LEDs or you're going to overheat. Page 2 on the datasheet says to keep the LED supply below 7V to keep from overheating.

What are you using for a power supply? You could hook up 4 AAs or D batteries to drive the LEDs with 6V; that would work fine.

I made a pair of these on a breadboard back in the day and ran 'em off a pair of 9V batteries; hooked 'em up with a 1/8" jack on one end to a pair of RCA jacks on the other and hooked it between my stereo and CD player or computer just for fun.

For power I was using a spare 12v AC adapter, I had intentions of using whatever my speaker amp uses once I get it apart, I'm going to assume its in the range of 9-12v. After the recommendation and the datasheet info I'll definitely try to lover it a touch to 6v or so.

Curious, are there any books I could buy that would help me understand this stuff a little better? I don't know exactly what 'decouple' means in terms of using a capacitor on the powered pins (huge newb).

Thanks for the help from everyone I've ruined one LM3915 but I made sure to buy a spare, I'll be working on that one later today.

 

[Jimmah]

K i totally redid it using a new IC and with the proper circuit, still having issues with needing to crank up everything just to have half the lights light up, even then its not really that accurate, more like a mushing of all lit lights together than single lights going off. Tried it in dot and bar mode, nothing changes when I connect pins 9 and 3.

Also, LED's are mighty dim, peaks of 1.5v on my crappy Mastercraft multimeter on the brightest LED.

http://i12.photobucket.com/alb...xhiron/VUmeterdiag.jpg

Its soldered exactly that way, except the 1.24k resistor is replaced with a 1k and 330ohm, and the 8.06k is a 10k pot (which if I adjust even a little the lights go super dim then nothing).

So, if anything today, I got to practice my soldering

 

[bobsmith1492]

Your resistors are a bit off... try this.

That should brighten up the LEDs and make the dot/bar mode switch work properly.

The voltage across the LED is fine (probably red at 1.5v). LEDs are funky in that they need a certain voltage in order to turn on and then that voltage stays pretty constant. The current through the LED determines its brightness, hence the need for resistors or other means of controlling that current. The LM3915 controls the current automatically; it is set by the 1.24K resistor primarily and the other one (your 10K pot) may have some effect as well.

Adding the capacitor and resistor on the input may help the signal range issues.

Decoupling means basically connecting a capacitor across the power supply lines. What happens is when the LEDs turn on, they suck a relatively high amount of power quickly and cause the supply voltage to drop for an instant. This can mess up the rest of the circuit especially inside the LM3915. A capacitor provides the energy needed to turn on while keeping the voltage at the proper level.

With that 47 ohm resistor, I think it should run from 12V. If you can run it from 4-6 volts, though, it would be better; just remove the 47 ohm resistor and replace it with a short.

Good luck man, don't give up on it! I'm not sure about books; I learned a lot from some books from Radio Shack, a series by Forrest Mims. I don't know if they still have them though, maybe Ebay? Otherwise, there's lots of stuff online so hit up Google.

Oh, also, to make sure you don't overheat the chip: keep your finger on it when you turn it on and play with the input signal. If it gets too hot to touch, shut it off and increase the 47 ohm resistor a bit (maybe 100?).

 

[Jimmah]

I did what you suggested, it 'mostly' seems better minus the led's are still a touch dim, might just do away with the 47ohm resistor completely (I'm using 3.2v purple led's). Also, the inline .1uf ceramic cap was a bust, with it installed the circuit just lit 2 led's on full with them fading ever so slightly. I'm much more positive about it now, so I'll try pulling the 330ohm resistor and just keeping the 1k, might help my issue of brightness.

Thanks for the help from everyone, this little project and the suggestions have really helped my understanding of circuitry

 

[Rubycon]

Originally posted by: Jimmah
I did what you suggested, it 'mostly' seems better minus the led's are still a touch dim, might just do away with the 47ohm resistor completely (I'm using 3.2v purple led's). Also, the inline .1uf ceramic cap was a bust, with it installed the circuit just lit 2 led's on full with them fading ever so slightly. I'm much more positive about it now, so I'll try pulling the 330ohm resistor and just keeping the 1k, might help my issue of brightness.

Thanks for the help from everyone, this little project and the suggestions have really helped my understanding of circuitry

Note highlighted item. These drivers were designed for standard LED's. Blue and white LED's require a higher forward voltage for proper operation. Try a 20mA (typical) 100mcd medium high brightness red LED and see how it works.

 

[Jimmah]

Originally posted by: Rubycon

Originally posted by: Jimmah
I did what you suggested, it 'mostly' seems better minus the led's are still a touch dim, might just do away with the 47ohm resistor completely <(I'm using 3.2v purple led's). Also, the inline .1uf ceramic cap was a bust, with it installed the circuit just lit 2 led's on full with them fading ever so slightly. I'm much more positive about it now, so I'll try pulling the 330ohm resistor and just keeping the 1k, might help my issue of brightness.

Thanks for the help from everyone, this little project and the suggestions have really helped my understanding of circuitry

Note highlighted item. These drivers were designed for standard LED's. Blue and white LED's require a higher forward voltage for proper operation. Try a 20mA (typical) 100mcd medium high brightness red LED and see how it works.

I checked the voltage coming to the led's and it is 1.5v steady, and I agree after reviewing my LED choices and how these little things are designed is why I'm having issues with them being so dim.

On a new note though, after spending another 3 hours or so tonight monkeying with it all I got it all together, it looks like pins 1-16-17-18 are fried, no voltage going through them at all. Which, is very displeasing, since I've spent well over 12 hours putting this all together (plus the 30$ in parts I've had to order - Ottawa has little in the ways of useful shops) I'm giving up and shelving it for a later date.

Kinda sad, but at least I found out what a voltage divider is

Back to messing around with simple LED circuits for my project.

Thanks everyone for helping, it has been a very insightful experience.

 

[bobsmith1492]

Originally posted by: Rubycon

Originally posted by: Jimmah
I did what you suggested, it 'mostly' seems better minus the led's are still a touch dim, might just do away with the 47ohm resistor completely <(I'm using 3.2v purple led's). Also, the inline .1uf ceramic cap was a bust, with it installed the circuit just lit 2 led's on full with them fading ever so slightly. I'm much more positive about it now, so I'll try pulling the 330ohm resistor and just keeping the 1k, might help my issue of brightness.

Thanks for the help from everyone, this little project and the suggestions have really helped my understanding of circuitry

Note highlighted item. These drivers were designed for standard LED's. Blue and white LED's require a higher forward voltage for proper operation. Try a 20mA (typical) 100mcd medium high brightness red LED and see how it works.

The LM3915 does constant current so it's not designed for any particular LED. As long as the input voltage is high enough it will drive anything within its thermal envelope.

I checked the voltage coming to the led's and it is 1.5v steady, and I agree after reviewing my LED choices and how these little things are designed is why I'm having issues with them being so dim.

Like I said before the voltage across an LED gives little to no indication of how bright it will be. Also, if they read 1.5V when on, they are not 3.2V LEDs or they would likely NOT be on at all. Plus, I've never heard of a purple LED but there's always time for something new.

 

[Rubycon]

Originally posted by: bobsmith1492

The LM3915 does constant current so it's not designed for any particular LED. As long as the input voltage is high enough it will drive anything within its thermal envelope.

This IC was designed long before the advent of white/blue/purple LED's which are high(er) voltage devices. I don't get to "play" with semiconductors like I used to but I would be willing to bet it's not going to put out the forward voltage required.

 

[bobsmith1492]

Originally posted by: Rubycon

Originally posted by: bobsmith1492

The LM3915 does constant current so it's not designed for any particular LED. As long as the input voltage is high enough it will drive anything within its thermal envelope.

This IC was designed long before the advent of white/blue/purple LED's which are high(er) voltage devices. I don't get to "play" with semiconductors like I used to but I would be willing to bet it's not going to put out the forward voltage required.

It acts only as a current sink so it doesn't put out any forward voltage. It operates on the low side (pulling to ground). Look at the schematic; the supply voltage runs straight into the LEDs then through the chip to ground. As long as the (supply voltage - LED forward drop) > minimum output drive voltage, it will regulate the LED current.

Lots of LED array driver chips do the same; you can put any voltage up to the chip's limit (say 20V) onto the end of a string of LEDs and the chip drops the remaining voltage to control the LED current.

 

[Rubycon]

After looking over the data sheet I believe you're correct. It should not have any bearing on LED type. :light:

If the circuit did work initially however I believe the consistency between LED's would change as these near UV emitters are susceptible to output degradation. Their lifetimes are very short compared to what's expected of LED's and their rate of degradation tends to be unique per device. Setting in dot mode would definitely make more sense here.