Author Topic: LED VU meter  (Read 7947 times)

0 Members and 1 Guest are viewing this topic.

Offline AdhithTopic starter

  • Frequent Contributor
  • **
  • Posts: 323
  • Country: in
LED VU meter
« on: September 11, 2016, 02:47:06 pm »
hello everyone..:)
        i'm plannig to built a LED vu meter for my external speakers & i have a circuit with LM3915 & 10 LEDs.But i'm trying to add more leds in the circuit by replacing a single led with a led strip(with 5 leds) so that all together there will be 10x5=50 leds in the circuit. so is this circuit feasible?? then  what additional setup is needed?? 
 

Offline rdl

  • Super Contributor
  • ***
  • Posts: 3667
  • Country: us
Re: LED VU meter
« Reply #1 on: September 11, 2016, 04:14:01 pm »
I think it's feasible, but the 3915 has limited current capability so you will probably need some kind of driver circuit for each section of five. It won't look any better than single LEDs since each grouping will all light at the same time. A better effect could be had by daisy-chaining multiple 3915s to get more resolution.
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: LED VU meter
« Reply #2 on: September 11, 2016, 04:34:54 pm »
I think it's feasible, but the 3915 has limited current capability so you will probably need some kind of driver circuit for each section of five. It won't look any better than single LEDs since each grouping will all light at the same time. A better effect could be had by daisy-chaining multiple 3915s to get more resolution.

  Not if he wants a thicker bar to be seen from far away.

  The current capability of the LM3915 is 30mA max, and it can work with 25V supply. Max power is 1.3W. With that in mind you can crunch the numbers for your LED strips. If you could give some data about them we could help further. If you are going to build them you should put them in series, 5 LED in series depending on the color and what else would be between 10 and 15V, 10mA is usually enough to drive LEDs to be quite bright, the 5V difference would be eaten by the IC, which is about 50mW per channel, 500mW total. You could use some series resistance with the LEDs to ensure the voltage across the driver is as low as possible so very little power is loss inside the IC. If you already have the strips look for the current and voltage they need, use a PS 1V higher than that and program the current to do the driving.

  If you need to add a buffer because you are going for high power LEDs remember the output of the LM391X are open collectors, so you can't use it as a voltage output as they are, you need to add something to convert that to voltage or drive some PNP darlington for example, which you could drive directly, maybe with the help of a small resistor to turn them off. It isn't a problem, you just have to be aware of it, not much complexity in going around of it.

JS
If I don't know how it works, I prefer not to turn it on.
 

Offline AdhithTopic starter

  • Frequent Contributor
  • **
  • Posts: 323
  • Country: in
Re: LED VU meter
« Reply #3 on: September 11, 2016, 05:38:46 pm »
here is the circuit i'm thinking of . the leds in the standard strips are parallel right?? by doing this  will there be peak difference between each parallel strips ??
 

Offline AdhithTopic starter

  • Frequent Contributor
  • **
  • Posts: 323
  • Country: in
Re: LED VU meter
« Reply #4 on: September 11, 2016, 05:42:54 pm »
I think it's feasible, but the 3915 has limited current capability so you will probably need some kind of driver circuit for each section of five. It won't look any better than single LEDs since each grouping will all light at the same time. A better effect could be had by daisy-chaining multiple 3915s to get more resolution.

  Not if he wants a thicker bar to be seen from far away.

  The current capability of the LM3915 is 30mA max, and it can work with 25V supply. Max power is 1.3W. With that in mind you can crunch the numbers for your LED strips. If you could give some data about them we could help further. If you are going to build them you should put them in series, 5 LED in series depending on the color and what else would be between 10 and 15V, 10mA is usually enough to drive LEDs to be quite bright, the 5V difference would be eaten by the IC, which is about 50mW per channel, 500mW total. You could use some series resistance with the LEDs to ensure the voltage across the driver is as low as possible so very little power is loss inside the IC. If you already have the strips look for the current and voltage they need, use a PS 1V higher than that and program the current to do the driving.

  If you need to add a buffer because you are going for high power LEDs remember the output of the LM391X are open collectors, so you can't use it as a voltage output as they are, you need to add something to convert that to voltage or drive some PNP darlington for example, which you could drive directly, maybe with the help of a small resistor to turn them off. It isn't a problem, you just have to be aware of it, not much complexity in going around of it.

JS
i haven't bought the led strips yet. i'll buy an appropriate one according to the specifications. am looking for those working with 5v supply. it should be avaliable right??
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: LED VU meter
« Reply #5 on: September 12, 2016, 12:33:52 am »
Few things.
  I was proposing using series LEDs so you can drive them without external beans, Driving them in parallel will make only 6mA for LED available and would be much trickier to set all not to blow.

  LEDs can't (shouldn't, but can't) be connected in parallel, the voltage each one will drop would be slightly different for the same current, or when connecting in parallel, the current in each one would be very different, resulting in very different brightness. When you need to drive several LEDs in parallel each one should have a drop resistor to correct for this difference. Depending on how tight the specs of the LEDs being driven the smaller the resistor needs to be.

  I guess you are better with 12V strips, 5V will have bigger current per section which is the limiting factor here, if you find one which could be driven fine with 20mA (one section) you may connect up to 4 sections in series with this IC.

  The strips being driven for a voltage source have parallel sections of series LEDs, first datasheet popped out shows 36W for 150 LEDs at 12V, About 240mW for LED (AVG, some goes to the resistor). For the drawings seems to be you can cut them in sections of 3 LEDs, so 720mW per section, 60mA each section. If you want 2 sections you need 120mA, to have 6 LEDs, the closest to your desired 5. This happen to be bigish LEDs for a strip, 5050 size. 3528 LEDs would make 20mA per section, at 12V 20mA you could drive 3 of those  with an LM391X without any trouble. If you want to drive 2 of those you are biting the edge, you could drive it to a lower current but I don't know how bright they will be, or in series but you don't have much headroom for the voltage.

  At 30mA and 25V max rating you shouldn't go too close to those. Test how well the strips brights with half the rated current or 10V instead of 12V. Maybe it's enough, those things are really bright, depends on the application it might be enough. Given the added complexity you may want to go for less LEDs, dimmer LEDs, external drive, custom LED panel. I don't know if any strips goes lower than 20mA per section but in any case you will be in the same ballpark of brightness than the bigger driven softer. You could drive five 20mA LEDs in series without a resistor from the IC, the strips have internal resistors which makes it not so optimal for your case. You need to know the LED voltage and minimize the voltage in the IC for the current you want, this is probably better done empirically.

5050 datasheet: http://www.nteinc.com/LED_tubing/pdf/LED_Strips_02.pdf
3528 datasheet: http://www.nteinc.com/LED_tubing/pdf/LED_Strips_01.pdf
Comprehensive datasheet: http://www.ledlightsworld.com/datasheet/Specification-of-Flexible-LED-Strip-www.ledlightsworld.com.pdf

JS
If I don't know how it works, I prefer not to turn it on.
 
The following users thanked this post: Adhith

Offline Audioguru

  • Super Contributor
  • ***
  • Posts: 1507
  • Country: ca
Re: LED VU meter
« Reply #6 on: September 12, 2016, 01:53:31 am »
I have never seen 5V LED strips and I agree that LED strips do not have the LEDs in parallel. 12V LED strips have 3 LEDs in series and in series with a current-limiting resistor.
 

Offline AdhithTopic starter

  • Frequent Contributor
  • **
  • Posts: 323
  • Country: in
Re: LED VU meter
« Reply #7 on: September 12, 2016, 07:12:32 am »
Few things.
  I was proposing using series LEDs so you can drive them without external beans, Driving them in parallel will make only 6mA for LED available and would be much trickier to set all not to blow.

  LEDs can't (shouldn't, but can't) be connected in parallel, the voltage each one will drop would be slightly different for the same current, or when connecting in parallel, the current in each one would be very different, resulting in very different brightness. When you need to drive several LEDs in parallel each one should have a drop resistor to correct for this difference. Depending on how tight the specs of the LEDs being driven the smaller the resistor needs to be.

  I guess you are better with 12V strips, 5V will have bigger current per section which is the limiting factor here, if you find one which could be driven fine with 20mA (one section) you may connect up to 4 sections in series with this IC.

  The strips being driven for a voltage source have parallel sections of series LEDs, first datasheet popped out shows 36W for 150 LEDs at 12V, About 240mW for LED (AVG, some goes to the resistor). For the drawings seems to be you can cut them in sections of 3 LEDs, so 720mW per section, 60mA each section. If you want 2 sections you need 120mA, to have 6 LEDs, the closest to your desired 5. This happen to be bigish LEDs for a strip, 5050 size. 3528 LEDs would make 20mA per section, at 12V 20mA you could drive 3 of those  with an LM391X without any trouble. If you want to drive 2 of those you are biting the edge, you could drive it to a lower current but I don't know how bright they will be, or in series but you don't have much headroom for the voltage.

  At 30mA and 25V max rating you shouldn't go too close to those. Test how well the strips brights with half the rated current or 10V instead of 12V. Maybe it's enough, those things are really bright, depends on the application it might be enough. Given the added complexity you may want to go for less LEDs, dimmer LEDs, external drive, custom LED panel. I don't know if any strips goes lower than 20mA per section but in any case you will be in the same ballpark of brightness than the bigger driven softer. You could drive five 20mA LEDs in series without a resistor from the IC, the strips have internal resistors which makes it not so optimal for your case. You need to know the LED voltage and minimize the voltage in the IC for the current you want, this is probably better done empirically.

5050 datasheet: http://www.nteinc.com/LED_tubing/pdf/LED_Strips_02.pdf
3528 datasheet: http://www.nteinc.com/LED_tubing/pdf/LED_Strips_01.pdf
Comprehensive datasheet: http://www.ledlightsworld.com/datasheet/Specification-of-Flexible-LED-Strip-www.ledlightsworld.com.pdf

JS
does the LED strips need more voltage & current than connecting normal LEDs in series ?? is this the circuit that you are talking about?? can i supply additional power required by the LEDs by some how providing the strips with another power supply circuit??
« Last Edit: September 12, 2016, 07:14:42 am by Adhith »
 

Online macboy

  • Super Contributor
  • ***
  • Posts: 2254
  • Country: ca
Re: LED VU meter
« Reply #8 on: September 12, 2016, 06:37:46 pm »
Put the LEDs in series, not parallel. Make sure that the power supply to the LM3915 is at least a few volts higher than the combined Vf of the five LEDs. For typical red LED maybe Vf = 1.7 V so five of them Vf = 8.5V.  Then make sure that the power supply is say 14 V or higher and you will have no issues. The LM3915 will actually dissipate less power (less heat) than with single LEDs!
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: LED VU meter
« Reply #9 on: September 12, 2016, 07:20:30 pm »
does the LED strips need more voltage & current than connecting normal LEDs in series ?? is this the circuit that you are talking about?? can i supply additional power required by the LEDs by some how providing the strips with another power supply circuit??

More voltage, yes. The strips are 3 LEDs in series + a resistor, so it needs the voltage for those 3 LEDs plus something to the resistor.
More current, yes. If you use more than a single section of 3 LEDs from one strip you'll be having parallel circuits, each one taking the rated current, multiplying the current for the number of sections.

Put the LEDs in series, not parallel. Make sure that the power supply to the LM3915 is at least a few volts higher than the combined Vf of the five LEDs. For typical red LED maybe Vf = 1.7 V so five of them Vf = 8.5V.  Then make sure that the power supply is say 14 V or higher and you will have no issues. The LM3915 will actually dissipate less power (less heat) than with single LEDs!

  I don't see the need of using 14V for a <9V Vled, the common collector would be happy with 1.5V at any rated current, any higher than that will be having excess power on the IC. I could live with 3Vm but more than 5 seems excesive, as it will end hitting the max power rating than the max current rating. I' guess I'd try to go for saturation and a series resistor with the LEDs if it fits in the 25V limit. This way the power in the IC will be minimum. If using the current setting I'd still try to go for as low voltage as needed.

  If you want multiple colors, each color will have a different voltage, do the math for the higher voltage required and then you could add a resistor to the lower voltages. (red are lower, then you go up to green, blue, white, you need to double check for specific LED)

JS
If I don't know how it works, I prefer not to turn it on.
 

Offline dannyf

  • Super Contributor
  • ***
  • Posts: 8221
  • Country: 00
Re: LED VU meter
« Reply #10 on: September 13, 2016, 12:49:10 am »
Quote
so is this circuit feasible?

you are unlikely to find a signal with that kind of dynamic range. So lots of those leds will be lit.

A more flexible way to do that you are trying to do is through a mcu: it adc's the input signal and drives an led array. I recently shows a generic way of implementing this: https://dannyelectronics.wordpress.com/2016/07/16/mcu-based-led-vu-meter/

In this particular example, I drove 12 leds via 8 pins, in a 6x2 matrix. For that particular chip, you can easily expand it to drive 20 leds (5x4, two input channels, and one pattern select pin).

It is not difficult to expand it to cover more leds, but the basic principle is the same.

================================
https://dannyelectronics.wordpress.com/
 

Offline Audioguru

  • Super Contributor
  • ***
  • Posts: 1507
  • Country: ca
Re: LED VU meter
« Reply #11 on: September 13, 2016, 01:30:20 am »
You said you are using LED strips that probably have 3 LEDs in series and in series with a current-limiting resistor to operate from 12V. Then why are people replying about using individual LEDs??
The LM3915 limits the LED current so the circuit does not need LED strips that have a current-limiting resistor.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf