Simple audio band-pass filters with sharp cut-offs

I want to split an audio signal into 5 bands by frequency range (top and bottom can be low-pass and high-pass) and send these into speakers with moderate power, say 10 watts. I can use amplified speakers if a

5-way audio amp gets complicated.

This is not for quality audio, but for educational demo purposes, so I don't need an even response and distortion is okay. I'm really looking for the sharpest cut-offs I can get, in as simple a box as I can build. Can anyone suggest how to accomplish this?

I haven't made many coils before (rather embarassing for a ham op), but I'm fine with simpler circuits. I found a post about an IC that might do the trick, but I'm doubting it'll handle audio frequencies.

Reply to
mistawizard
Loading thread data ...

Don't use coils. Use active filters. I think there is an "Active Filter Cookbook" or somesuch floating around out there. You might be able to get a copy for a few bucks.

Reply to
Charles Schuler

You'll be well served by Don Lancaster's "Active Filter Cookbook", 2ed., ISBN 075062986X.

Or, google "Sallen-Key bandpass filter" and you'll probably come up with something.

Reply to
Walter Harley

No coils are required if you go to

formatting link
and get this program. Just answer a few questions and you will be presented with a fully designed circuit. The IC's are probably going to be surface mount ones, but there are easy ways to handle that.

Jim

Reply to
the.other.Jim

There are several IC's on the market that do exactly this. You could also look at active filter circuits based on op-amps. A lot will depend on the nature of the signal, and what you want to actually 'do' with the signal. A sharp cutoff, will probably involve a several stage filter, and this will be much easier with an op-amp, or an IC. This is why inductors have almost vanished for this type of application...

Best Wishes

Reply to
Roger Hamlett

As in so much of engineering, you have competing demands which means you have to decide where the tradeoff is. Sharper cutoffs means higher pole filters, which require more components with closer tolerances. ISTR Lancaster's Active Filter Cookbook discusses this.

The 'best' solution technically would be an A/D, DSP, and five D/A's, as the DSP can be programmed to do quite sharp filters and the result will be much more stable than you could possibly get with analog components.

You wouldn't want to do this with coils.

Reply to
Ben Bradley

Beautiful! I'll check out the Active Filter Cookbook (my library has it!), but if there's a good IC for it, so much the better. I just want to take a line-level or microphone input (not powered) and have it put the sound out at line level (I'll feed them into amplified "computer" speakers), each representing a different frequency range (probably

0-200Hz, 200-400, 400-800, 800-1600, 1600+) so I can turn channels on and off, and kids can experience fourier decomposition concepts and filtering in real-time. Again, quality is not a big issue (although reliability is).

Can you suggest where to look nowadays for ICs like this, or an IC that would work well? I used to use the old radio shack semiconductor reference. It's been a while. I'm guessing the Active Filter Cookbook doesn't include ICs like this, so are there any "cookbook" schematics for these handy ICs?

Thanks for all the great advice!

Reply to
mistawizard

Is that still being printed by a traditional publisher, or does the poster have to visit Don's site,

formatting link
to get it?

Michael

Reply to
Michael Black

I bought mine from Powell's a couple years ago, and it was still in print then. It's on the Newnes imprint.

Reply to
Walter Harley

In the cook book, look for a 4 pole bandpass filter. This produces a double peak result (one near the low frequency end of the band and one at the high frequency end of the band) with sharp enough roll off at the edges for this application.

Of you can combine a 2 pole high pass (to cut the low end) and a 2 pole low pass (to cut the high end). If you choose a bi quad or state variable design for these, you will be able to adjust the cut off frequencies with a pot without altering much of anything else about the response. Good for demonstrating what more bits in the Fourier does for the result.

Or you can use just 2 pole bandpass filters (single peak response) made as a bi quad or state variable and adjust both the center frequency and bandwidth as desired. These take 3 opamps each, but since you can get 4 opamps in a single package for less than a dollar, this is not much of a problem. The LM324 is a good choice for cheap quad opamp that runs on a wide range of supply voltages with low supply current, if sound quality is not a big concern (it does produce a bit of distortion at higher audio frequencies). Common as dirt.

formatting link

Reply to
John Popelish

Aha. In this application you don't neccesarily need the steep cutoff slopes. You can use a fairly low-Q (perhaps in the 0.3 to 1 range) bandpass filters, one for each band. Another idea comes to mind: a five-band audio graphic equalizer, as made by just about everyone in the '70's and '80's. Used ones are surely plentiful and cheap nowadays. Even 7-band and 10-band EQ's are common and cheap. Look on ebay. In such an EQ, all the filters are summed to one output, but you can remove the summing network and run each filter output to a separate amplifier/speaker. This is easier than making your own circuit.

Reply to
Ben Bradley

Good recommendation for audio range active filters.

I got mine a year or so ago from Marlin P. Jones who, at the time, was selling off a bunch that were autographed by Don himself for no more than the regular price of the book.

Reply to
Joel Kolstad

One other suggestion: At audio frequencies, switched capacitor filters are quite good, and since you don't care about a little distortion anyway, you can use a dirt simple anti-image filter at the output.

The beauty of switched capacitor filters is that 'everything' scales with the input clock frequency. I.e., if you design the thing so that you end up with a 5kHz Q=10 bandpass with an input clock of, say, 1MHz, if you halve the input clock to 500kHz you'll have a 2.5kHz Q=10 bandpass filter. Another nice benefit is that most of the tuning elemenets for active filter ICs are on-chip and thereby tend to be matched (ratioed) to one another much better than you'd be able to achieve by just randomly choosing components (of the same nominal value) from your parts drawer; if you decide to start building significantly narrower filters that what you've specified, this can become significant. (I had a professor who claimed that 'all good IC designs are beautiful to look at,' and while I think that's sometimes just wishful thinking and have even read papers suggesting as much, his point was that you can derive a few simple rules for IC layout that attempt to minimize component ratio tolerances and these rules end up making symmetrical patterns that tend to be aesthetically pleasing as well...)

The Linear Tech LTC1562 might be a good place to start...

Reply to
Joel Kolstad

Good information! I looked - MPJA still sells them.

formatting link

Thanks, Ed

Reply to
ehsjr

formatting link
has many copies, a few are quite cheap, but most prices are in the $20-$30 range, where the new copy prices start.

I found this amusing entry under the title "Active-Filter Cookbook by Lancaster, Donald E.":

"TRUE COLLECTABLE, AUTHOR'S SIGNATURE Tight, straight, clean, cover shows wear as well as edges of text. This book has been used but not in your normal cooking. $19.50"

Reply to
Ben Bradley

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.