Audio Spectrum analyzer help

What about one LM3915 log bargraph driver with a bandpass filter per channel, plus some muxing for the outputs. Perfectly do-able.

Dave :)

I want to build a cheap audo spectrum analyzer, but im not actually going to be doing true analysis with it or anything like that.

More or less, its going to be used for looks. (like the ones in stereo systems).

I am building a little Stereo amplifier for a school project, and I want to build this spectrum.

I have a 12-bargraph VFD display, that is multiplexed array.

All anodes of the bargraph are common to one another, but the grids are seperate for each bargraph.

I was thinking about using a 4017 to scan the display, wiht a shitload of active bandpass filters and 4016 ICs to select each bandpass filter in sync with the scan.

This would take alot of parts, but if there is a simpler way to do this, let me know.

I seen something on switched capacitors, but I know nothing about them, or how to multiplex them into the display.

Any thoughts? or ideas?

Hello Mike,

A simpler way these days would be a DSP or a powerful micro controller. If going the uC route you'd need one with a hardware multiplier to make it anything close to real-time. And a good ADC.

If you aren't inclined to jump into weeks of programming Tim's suggestion with switched capacitor filters is worth checking out. Look at the National Semiconductor LMF100, but it'll cost a few Dollars. Their MF10 is cheaper though I am not sure about availability.

There has got to be some chip that does this. We used to have a stereo that had such a 'frequency bin display' and it wasn't too high end. Not

12 bars but maybe eight or so. Unfortunately it ain't here anymore so I can't open it up and look. IIRC it was from Kenwood and there wasn't anything really massive in parts where the cable of that display went.

Regards, Joerg

I once made a bank of 8 active filters that I lifted directly out of Don Lancaster's Active Filter Cookbook - the only difference was I had a handful of TL084s on hand.

Then, I detected the eight band levels, and, possibly to the chagrin of the solder programmers, just read them directly with the 8-channel ADC in a 68HC11, and massaged them a little, talked to a custom parallel port, and displayed them graphically on my old DOS PC.

You still need to translate the analog into some kind of bar-graph signal, and the 3914 or so comes to mind, but how you'd interface them to a VFD is beyond me. Also, you'd need either 12 (or however many) 3914s, or some kind of analog mux that makes my brain hurt.

Use a big micro with 12 channels of ADC, and slap together some active filters - you could do those in an evening or so, with the right equipment and parts and stuff.

Good Luck! Rich

Have a look at this guy's stuff. It is a really impressive site.

Chris

Im not going the upc route. it would be nice, but too much work, and I am horrible with processors.

I already got the display, its that VFD with 12 bars, wired as an array. All grids are seperate for each bargraph, and the 7 anodes are common to one another.

I got a BA IC driving the anodes, and it moves to the music like its supposed to, but all the grids are tied to positive, the entire display moves.

So, I would have to do sequencial scanning, and switched bandpass filters.

any ideas? more importantly, anyone have a simple active banpass filter? I dont want to have parts the size of japan.

going

of

or

no, no, no, its a multiplexed VFD.

Its a 7-point bargraph.

But, all 7 points are tied together to all 12 bargraphs. but each bargraph has a seperate grid. (think of a LED matrix, or keypad matrix, thats how the VFD is designed).

So, i would only need 1 driver, but need a 12 position decimal counter running at a high speed to scann the display.

going to

of

sync

this, let

umm, NO.

"I am building a stereo for school"

Cant fit a laptop and soundcard in a little stereo amplifier.

to

to

sync

let

or

"Mike Bates" a écrit dans le message de news: Zn%nf.190604$ snipped-for-privacy@tornado.ohiordc.rr.com...

A link, even if not what you're looking for, to my own audio spectrum analyzer project... using a PIC and a VGA display :

Just for the reference, as this pic is now a little obsolete ;+) Friendly,

--
Robert Lacoste
ALCIOM - The mixed signal experts
www.alciom.com

Cheap laptop with a sound card capability and freely available software?

On Thu, 15 Dec 2005 03:19:34 +0000, Mike Bates top-posted:

OK, then, you'll only need the one 3914 bargraph chip, ... Hmmm - the analog can come right from the filters, through an analog mux + S&H...

But you still have to do the scanning logic.

And I'm not a DSP guy, so I don't know about doing all 12 channels of filtering entirely digitally.

Good Luck! Rich

In article , Joerg wrote: [...]

You can get away without the multiplier if you don't mind a little channel to channel cross talk. By making the code multiply by a function that looks like this:

+1 ----- ---- 0 -- -- -- -1 -----

You can make the system directly reject the 3rd harmonic. This leaves the

5th and up. Before displaying, you need to subtract a fraction of the harmonic frequency from the lower channels. Depending on how you've set the bands, there may not be that many cases where the harmonics are in the band of interest.

--
--
kensmith@rahul.net   forging knowledge

Hello Rich,

Just run 12 DFTs. A DSP or a maybe even some modern uC with HW multiplier shouldn't have a problem with that, plus it can do the output muxing.

Regards, Joerg

the 3914 wont drive a VFD, becuase its a current sync open collector. So, in the past, the only way I was able to drive a VFD with that, was to use pull-ups and inverters.

Instead, I got a single BA6146 driving it, and works quite well, still need filters and scanning logic.

bargraph

the