Check my Lm3915 vumeter circuit

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Hi!

The signal level was not hight enough to turn them on, maybe i was a bit conservative. but i will try again today.

Jay x
 
Hi Ian!

Yes!. But i placed a 130R resistor instead, as I had some. What happens is that when the sixth led of the second string lights up, the first five leds remain lit up but with a sligth flicker. But when more leds of the upper string are lit up i don't notice this flicker. So I suspect that changing the resistor to 100R will solve this as it will provide more current.: 0.75/100*1000= 7,5ma. Right now I have 5,77ma. 6ma is a usual figure i see in some schematics.

Jay x
 
Heeelp!!😗

I thought it worked....but no. The last two top leds do not light up!!. Despite reducing the 220R resistor to 100R to give more current. I could light up to led number 9 for an instant, but when attempting to light up the last one, all leds went off....🤔

My led scale goes from -12dbu to +15dbu in 3dbu steps. Chip is an LM3915. I can light up to +9dbu, but +12dbu and +15dbu does not light up. My audio interface can put out +5v rms, with distortion, but this is more than the 3,083v for +12dbu or 4,35v for +15dbu.

My voltage divider is 7k5/1k2. Ref voltage: 10,225v

Any help much appreciated...!!
Jay x
 
Last edited:
Well, the circuit works fin for the first two LEDs of the second bank and the first bank are all on so the transistor current source seems to be OK. In principle there is no reason why the last 2 LEDs should not work. What is the signal level on pin 5?

Cheers

Ian
 
Hi Ian!

Tomorrow I will measure again, and tell you. Let's see what happens..

Jay x
 
Hi!

I changed the divider resistor: 1k2/4k7. But still does not work.

I measured pin5 voltages DC. First i st up the 0dbu point WRT 0,775v RMS. ok so far.

Then i started mesuring DC volts, wrt to dbu scale i have.

0dbu/ 1.288vdc
+9dbu/ 4.47vdc
at 4.50vdc still all 8 leds up to +9dbu are ON.
at 4,51vdc all leds, turn OFF.

Jay x
 
Did you check the orientation of those leds? If the ninth was reversed, i would expect this behavior.

It could also be that the ninth led failed open?

Thomas
 
Hi!

Yes!, I checked orientation, is correct. And also checked the leds with the multimeter in diode mode. They lit up. Maybe the chip itself?. I will check tomorrow..

Jay x
 
Hi! Some progress...

1. I changed the voltage divider to 1k2/3k3 to better match the voltage range i want to represent, between 0.194/-12dbu to 4.355vRms/+15dbu.
This way Icould lit up the +12dbu led.

2. The top led still wants not to light up!. When i increase the signal level above a certain voltage even above the range, all leds turn off!.

3. I measured the voltage at the led strings, with all except tenth led light up.
The first 5 led string measures 10,13v.
The second led string with 4 leds on, measures 8,18v.
And when i increase audio level to try to light up the last led, the second string voltage measures 16v.

I don't know how to interpret this measurements.

I have tested 7 lm3915 chips i have. Only one, because of its different marking style, gave problems (only allowed to light up 8 leds/10).
All the rest reached to light up the nineth led. So i suppose they are ok.

4. With power switched off, I also checked led orientation and diode mode. All light up except the first and the last one. The first one is from the bottom string, when powered, lights up normally.

From there I have no clues...

Any advise much appreciated!!

Jay x
 
Hi!,

Still experimenting... I built a typical vumeter to discard the chips are bad, and it worked perfectly with all 7 lm3915. All leds light up nicely.

There should be something wrong with the current source...just guessing...🤔

Jay x
 
Hi!

More experimenting...

I rebuilt the Marton circuit attached in the message , with a rectifier buffer, on a test board, so i can pull out wires, change them etc..
I had no 2n3904/2n3906. I used BC547/557, and also try MPSA42/92. no difference in behaviour.

1. The ref divider i used is 1k2/3k6. I played a drum loop because i was sick of the 1khz tone...:LOL:

2. In the original circuit pin 9 is tyed to pin 11. leaving pin 9 not connected does not change anything. When the signal level is within the range of the lower string of leds, these leds light up normally. When the signal goes slightly above, to the upper string of leds, then the upper leds go steady but the lower leds blink. I could make all the leds of the upper string light up, but the lower leds did not light up steadyly... not even changing the 220R to 100R...:unsure:

2.1 Tying pin 9 to pin 3 (vcc) stabilizes the supply to the lower string of leds, but at the price of "starving" the upper led string. Even at full level 3 leds of the upper string can't be light up. If i pull out this connection, all upper leds light up nicely, but lower leds can't get steady...they are switching with the signal level.

I expected to save some milliamps building it into a 500 series preamp. But no luck by now...If anyone has a clue what could be done, here I am.

Thank you very much!!

Jay x
 

Attachments

  • LM3915 2.jpg
    LM3915 2.jpg
    114.7 KB · Views: 24
Last edited:
A couple of things. This display will only work properly in DOT mode so pin 9 must be left unconnected. Also, as the circuit stands, I am unsure about running high currents through the LEDs. With the 1K2 between REF OUT and REFADJ with REF ADJ connected to 0V, the LED current will be 10mA. The problem with this is R3 at 470 ohms with 10mA through it will try to drop 4.7 volts but it cannot because of the base emitter junction of Q1 which will try to limit it to about 0.7V. So only about 1.5mA will go through R3 and the other 8.5mA will try to go into the base of Q3. This means Q3 will be heavily in the saturation region. This may or may not be an issue because Q3 collector current is limited to about 2.5mA by R4. SO it might be worth reducing R3 to something like 240 ohms.

The other thing to remember is that the current though R5 sets the current in the first five LEDs. It has no effect on the current in the second five. So what you need to do is to set R5 so it makes the same current flow in the first five LEDs as is set by the value of R2 in order for them not to flicker when going from 5 LEDs lit to more than 5 lit. With R2 = 1K2, the chip LED current is about 10mA.

Cheers

Ian
 
Hi Ian!

Yesterday i posted the picture, but today i post the same picture with the modifications i made. MY BAD!!!. Sorry!!, with 1k2/3k6. Led current is about 2.28ma. given the formula in the Lm3915 datasheet. I will try again today following your advise!!.

I forgot to say that with the divider 1k2/3k6, with pin9 connected or not to pin 11, it works in bar mode.

Cheers,

Jay x
 

Attachments

  • LM3915 3.jpg
    LM3915 3.jpg
    130.7 KB · Views: 23
Last edited:
Hi!

Better news!!. The circuit i build in a protoboard works well, with all leds Shining nicely. I corrected a faulty connection, and voilá..a bargraph led vumeter.
Why does it not work equal in my preamp pcb?...right now i don't know.
I kept the 4k7 resistor equal, 220R reduced to 100R. 470R equal, or 330R. All these values in the protoboard.

I measured the voltages at the outputs of the lm3915. I will post an image with the values. Maybe there is a clue in the values.

Jay x
 
Last edited:
Attached are two tables with voltages measured at the outputs of the lm3915. The voltage difference between led strings is quite el evident...

At left are the values at protoboard. At right in my pcb.
Jay x
 

Attachments

  • IMG_20220121_194733.jpg
    IMG_20220121_194733.jpg
    64.4 KB · Views: 18
Good news!! Now it works!!

The cause of the problem?...Flux residue...I just re-washed the bottom layer with spray, after re-checking all solderings, and it worked!

Electronics is about details, even the most tiny thing can drive anyone crazy...Well, lesson learned!!😅

Jay x
 
There is still a question that i did not pay attention to till now... I notice that the bottom string leds take a quite long time to decay to zero after input signal ends. Much longer time than a typical LM3915 circuit. ¿What can I tweak to correct this decay time?

Jay x
 

Latest posts

Back
Top