why a multimode filter can't be 24 dB?

space_nerd · Post by **space_nerd** » Sat May 24, 2008 7:29 am

I' ve got this technical question:
why a multimode filter can't be 24 dB?
the more technical the answer the better please! (as much as this forum allows)
Thanks in advance

Nikos

till · Post by **till** » Sat May 24, 2008 10:08 am

The amount of electronic parts = money.

Kevin Lightner · Post by **Kevin Lightner** » Sat May 24, 2008 10:43 am

Take two -12dB/oct filters, run them in series and you have -24dB/oct.
As Till said, this will cost more money.

space_nerd · Post by **space_nerd** » Sat May 24, 2008 4:19 pm

Thank you both for the answer,
although I was thinking that the answer was more technical related.
I remember a synth, was called "the aviator"
http://machines.hyperreal.org/incoming/ ... tor.review
that had two filters one multimode 12dB and one lowpass (Moog-clone I think) 24dB

electrical_engineer_gEEk · Post by **electrical_engineer_gEEk** » Mon May 26, 2008 6:21 pm

but wouldn't that require a dual ganged pot for the cutoff so both filters in series change at the same rate?

Kevin Lightner · Post by **Kevin Lightner** » Mon May 26, 2008 8:34 pm

I think we're talking basic structure.

How one simultaneously changes freq and/or resonance is up to the implementation.
One pot could feed both sections the same control voltage and one pot could control two gain stages (via OTA, vactrol ,etc) for resonance or just provide an inverted feedback to the original input.

MC · Post by MC » Tue May 27, 2008 2:08 pm

The technical answer is that nobody has yet to find a successful working design for a 24dB multimode VCF.

The standard 12dB multimode filter is a product of very clever math manipulation. It works by rearranging the transfer function of a high pass filter, which is the ratio of two quadratic functions. It just so happens that the rearrangement of the terms can permit simultaneous lowpass mode and consequently bandpass, notch, etc with the addition of extra opamps. Unfortunately the transfer functions for 12dB and 24dB filters are radically different enough that the same alchemy isn't possible.

On part count, 12dB multimode filters aren't very efficient compared to other filter topologies. The 24dB equivalent - two cascaded multimodes - would be worse part-count wise, and would require exotic matched OTA circuits for voltage control of frequency and resonance of each half. It just isn't cost-effective.

theglyph · Post by **theglyph** » Tue May 27, 2008 7:05 pm

MC wrote:The standard 12dB multimode filter is a product of very clever math manipulation. It works by rearranging the transfer function of a high pass filter, which is the ratio of two quadratic functions.

A nice geeky EE explanation!

MC wrote:It just so happens that the rearrangement of the terms can permit simultaneous lowpass mode and consequently bandpass, notch, etc with the addition of extra opamps. Unfortunately the transfer functions for 12dB and 24dB filters are radically different enough that the same alchemy isn't possible.

I'd like to see this on paper. I can understand the math and I'm curious as to how this works out.

cheers,
theglyph

latigid on · Post by **latigid on** » Tue May 27, 2008 8:11 pm

FWIW, the Akai MFC-42 has an 8-pole mono filter. It is multimode, so I assume it has a 24dB band/notch filter. I can't be sure though, as I can't find a manual, but the higher pole filters sound "fatter" and have more resonance etc..

MC · Post by MC » Wed May 28, 2008 9:39 am

Google for Kerwin-Huelsman-Newcomb (KHN) Biquad Filter

space_nerd · Post by **space_nerd** » Fri May 30, 2008 7:24 am

Thanks to everyone answer back, especially MC -this request was because of a dissertation I'm in, where every design choice must be explained clearly and justified.
Soon, I will be happy to send photos of something nice

I won't tell you more!
Thanks again