LM3915 VU meter, with peak hold

[danrogers]

Hi all. Been planning to make some VU meters for a number of years now lol, now is definitely the time!

There are so many circuits available on the internet - to be honest I'm lost. I some of the LM3915 videos on youtube show a very flickery leds which is not what I'm after at all.

Is there a 'preferred' circuit for this? Audio guru's peak hold circuit looks good but hes uses a microphone - I will need mine to work with line levels, so I'm not sure how the circuit would need to be adapted?

Many thanks

 

[alec_t]

I will need mine to work with line levels

Are you using music/speech to provide these levels, or something else (e.g. fixed tone generators)? The peak hold circuit components might need to be optimised accordingly.

 

[danrogers]

Hi Alec.

Sorry I should have stated, it will be line level music

 

[audioguru]

My Sound Level Indicator project has a preamp opamp with a gain of 101 and its peak detector has a gain of 1. Maybe with line level you need a gain of 10 to 33 times. I used an opamp with inputs and an output that work at 0V when tyhere is no negative supply.
The gain is the ratio of two resistors like this:

 

[danrogers]

Hi Guru and thanks. So i guess if I used something like a 1k (R9) and a 20k (R8) I should get a gain of x20, and around the middle of the gain range you suggest for line level audio?

would you mind posting the circuit with the peak detector and the lm3915 shown again?

Many thanks

 

[audioguru]

If R9 is a value as low as 1k ohms then your music source might not be able to drive it and the value of the input capacitor will need to be 33uF. Why not make R9 be 15k and R8 be 300k? Then C4 can be 680nF.
Here is my Sound Level Indicator circuit that is almost 10 years old today and still works perfectly after being turned on all the time:

 

[danrogers]

Good point about the low resistance value, thanks

How would you connect the LM3915 to your peak detector circuit? You show the output that you connect to pin 5 but what about RHi that you show on your original circuit with the mic?

Thank again

 

[audioguru]

The output of my peak detector connects to the SIG input pin 5 of the LM3915. I use a slower output of my peak detector to feed a DC voltage to the Rhi pin 6 on the LM3915 for it to change ranges so that I get an additional 20dB of loudness range.

 

[danrogers]

Oh i see great. Most schematics seem to show RHi connected to the REFout pin, I guess it would be OK to do this combined with your line level peak detector circuit, then I would just have a static range?

Thanks for the help

 

[audioguru]

The Rhi to Rlo voltage divider in an LM3915 is typically 28k ohms. I bias it with a 220k resistor R15 connected to regulated 5V so when audio levels are low causing transistor Q3 to be turned off then Rhi is 0.565V resulting in very high sensitivity. When sounds are loud then Q3 turns on making Rhi as much as 5.57V which reduces the sensitivity 10 times (20dB). I get a range of 50dB.

If I didn't use Q3 to change the range then only the top LED would light most of the time or the sensitivity could be turned down then no LEDs will light most of the time. Then the range is only 30dB with Rhi at one voltage.

 

[danrogers]

thats a great idea for auto scaling!

would you recommend trying to implement that part of the circuit into your peak detector circuit without the microphone?

 

[audioguru]

If your music is heavily compressed then it has a small dynamic range so the extra range is not needed.

 

[danrogers]

Great, thanks I will see how it looks. The music is not greatly compressed as I will be using the VU meters for monitoring levels while producing some tracks

Do you know of any alternative OP-amps to use? The MC33171 appears fairly rare / expensive?

 

[audioguru]

You could use an LM358 dual opamp since its inputs work fine down to 0V and its output goes down almost to 0V without a negative supply. But an LM358 is noisy then it might show an output (of noise) without an input.

 

[danrogers]

Ok thanks. Sounds like I will be better off sourcing the MC then

 

[audioguru]

I used the MC33072P dual opamp because it does everything I want and uses low battery current. www.farnell.com is a British electronics parts distributor with warehouses all over the world and their American office has more than 1000 of this opamp in stock and many MC33078P that is even better but uses more battery current.

 

[danrogers]

Great, thanks I will have a look about Not having any luck finding them in eagle though lol

 

[audioguru]

Are you doing a simulation instead of building the circuit? Eagle probably has the LM358 dual opamp in its files.

 

[danrogers]

I'm just drawing the circuit out at the moment and might end up etching a board