Digital VU Meter

[gramo]

I've been looking around, but struggling to find an IC/method to interface with a micro controller to create an Audio VU meter

Any thoughts insights on the idea?

 

[futz]

Have you looked at the LM3914 (3915,3916)? It's not really made for microcontroller interface (tho you could do some things with a MCU), but it is made for doing VU meters and that kind of thing. Cheap and dead easy to use. The LEDs/bar-segments fade in/out and overlap a bit so they look awesome. The chips can be "cascaded" so you can make 10 or 20 or more segment displays. Makes it simple to make an amazingly good looking/working VU type display.

I was just tinkering with one tonight. I got one of **broken link removed** kits from Nightfire Electronics. Pretty generous parts bundles. They look to be mostly surplus parts, but there's useable stuff in there.

 

[ericgibbs]

I've been looking around, but struggling to find an IC/method to interface with a micro controller to create an Audio VU meter

Any thoughts insights on the idea?

hi gramo,
As suggested a LM3915 would do the job, I suppose you could drive the input of the LM3915 with filtered PWM from the PIC,
they accept 0v thru +5V OK.

Is it stereo or a mono AVU you are working on?

 

[Mike - K8LH]

Use 5 pins driving an N(N-1) mutiplexed ('Charlieplexed') display or an 8-pin PIC 'slave' driver something like the one below.

Good luck on your project. Mike

 

[gramo]

Thanks guys, all handy pointers to take into account

I've managed to find some data on the bandpass filters, and constructed a simulation for a 10 Band vu meter (Just the analogue aspect so far).

• The bandpass filters are tuned for 31, 63, 125, 250, 500, 1K, 2K, 4K, 16K
• The Q for each filter is 1 (I'm really not sure if this is an optimal value, but 4 is used for very steep cutoffs, 1 seems to do the job so far...
• The gain is left at 1 for the amps, as I figure that volume (sensitivity) can be controlled by a pot on the input

I thought I'd show you what I meant before I hit the hay;

Image 1
Image 2

Video of the simulation - I run through every tuned frequency (16KHz to 31Hz) and have all 10 channels displayed on oscilloscopes to view the amplitude variations at different frequencies... Its un-edited, sorry if there are mistakes!

 

[gramo]

I should add, that video was taken on a 19" monitor, and all the oscilloscopes took up a the whole screen... It might look weird on a smaller system.


Sorry about the piccys being a lil blurry too

 

[Mike - K8LH]

Sorry, I didn't realize you were lookin' for the AF 'front end'...

Good luck with your project.

Mike

 

[blueroomelectronics]

Sounds like a job for a dsPIC.

 

[mister_e]

A 18F should be more than enough... maybe even a 16F depending the speed, accuracy, filter algorithm & preamp you use.

EDIT: what happen if you just sit a RC filter at the ADC input and do some basic ADC averaging, then dump the results to your LEDs? Shouldn't be as this hard.

 

[Nigel Goodwin]

Don't know if it's any help?, but a few years back I wrote software and build the hardware for an LM3914/5 type 'replacement' using a 16F876.

It had two rows of 16 LED's, fed from two analogue input pins, an input pin to switch between dot and bar mode, plus another to switch from 3914 to 3915 mode.

In fact one of my first posts here was to ask for advice about the log side of the project, which I didn't get - but after a bit more thought I used Delphi to write a program that generated a log lookup table.

Unfortunately I don't know what happened to the board, or the source code for it?.

 

[kchriste]

Here's a guy that made one with a PIC which drives a VGA monitor directly!
**broken link removed**