how to split and recombine a signal

mordechai · Post by **mordechai** » 15 Apr 2010, 14:13

I saw the schematic for Anderton's amazing Quadrafuzz, and it looks like the device splits the guitar signal into high , high mids, low mids, and lows...then runs each part through its own effect before recombining them. And it looks like opamps are involved in each split signal as they are incorporated into their respective fuzzes.

I am interested in doing something similar but only with 3 bands -- high, mids and lows -- and then running each through their own simple cascading JFET overdrive/boost configuration before recombining them.

Can someone please advise on how to construct the 3 band signal splitting part of the circuit in the most basic terms? I am a ridiculous noob and while I can describe the basic effect I want in conceptual terms, the specific schematic or layout/wiring know-how is something I haven't yet digested.

Thanks,

Mordechai

mictester · Post by **mictester** » 15 Apr 2010, 17:57

mordechai wrote:I saw the schematic for Anderton's amazing Quadrafuzz, and it looks like the device splits the guitar signal into high , high mids, low mids, and lows...then runs each part through its own effect before recombining them. And it looks like opamps are involved in each split signal as they are incorporated into their respective fuzzes.

I am interested in doing something similar but only with 3 bands -- high, mids and lows -- and then running each through their own simple cascading JFET overdrive/boost configuration before recombining them.

Can someone please advise on how to construct the 3 band signal splitting part of the circuit in the most basic terms? I am a ridiculous noob and while I can describe the basic effect I want in conceptual terms, the specific schematic or layout/wiring know-how is something I haven't yet digested.

Thanks,

Mordechai

Bad news, I'm afraid - you might have to do some mathematics! You could use a quad op-amp (a TL074 would be a cheap, low noise choice). The first op-amp would be just a buffer, to isolate the filter circuits from the guitar input. The next part is the difficult bit - you're going to have to design three filters, a low pass, band pass and high pass. The frequencies that are critical are up to 450 Hz, 450 - 1 kHz, and above 1 kHz. You'll need quite high order filters so that they don't overlap too much - this might require cascaded filter stages, which would mean that you'd need more op-amps (a second TL074).

When you've got your three signals, you'd thenabuse them with your clippers or misbiased MOSFETs or however else you're going to mangle the signal, then you'd have to recombine them with a mixer stage (another op-amp). The levels of the three signals will differ, so you'll have to adjust that...

Its' not a project for a beginner. Why don't you try the quadrafuzz? There's a PCB layout published for it, and as I remember, it gave a pretty good (quite unusual) sound!

CRBMoA · Post by **CRBMoA** » 15 Apr 2010, 18:30

I spent an hour last night chatting with the developer of this.

The original purpose of the circuit is nearly exactly what you describe.

mordechai · Post by **mordechai** » 15 Apr 2010, 20:21

Thanks, very helpful clarifications so far. I don't mind jumping in to a somewhat more complicated project and making mistakes along the way. Only way I'll learn.

Ultimately, what I'd like to do is not exactly what the Quadrafuzz does. I may well be reading the schematic of that project incorrectly, but it seems to me that the four signal splits are processed individually, then brought back together, and the controls for the device (volume, fuzz intensity, etc.) control the end result, i.e., the recombined signal. so increasing the drize of the fuzz increases the drive of ALL the signals that have been split and run through their individual fuzz circuitry. Again, I could very well be mistaken about what I'm taking away from the schematic. But what appeals more to me is the ability to increase the fuzz on one band of the signal (high mids, for example) without automatically increasing the fuzz of the others.

mictester · Post by **mictester** » 16 Apr 2010, 05:13

mordechai wrote:Thanks, very helpful clarifications so far. I don't mind jumping in to a somewhat more complicated project and making mistakes along the way. Only way I'll learn.

Ultimately, what I'd like to do is not exactly what the Quadrafuzz does. I may well be reading the schematic of that project incorrectly, but it seems to me that the four signal splits are processed individually, then brought back together, and the controls for the device (volume, fuzz intensity, etc.) control the end result, i.e., the recombined signal. so increasing the drize of the fuzz increases the drive of ALL the signals that have been split and run through their individual fuzz circuitry. Again, I could very well be mistaken about what I'm taking away from the schematic. But what appeals more to me is the ability to increase the fuzz on one band of the signal (high mids, for example) without automatically increasing the fuzz of the others.

The quadrafuzz could easily be modified to do that - you're just going to have to introduce a gain control before each clipper stage. I'll draw it up later for you. You can use different clipping stages if you want, too.

mordechai · Post by **mordechai** » 16 Apr 2010, 14:11

Thanks mictester. That would be enormously helpful to me. When I get to work on constructing the circuit I'll post progress reports and eventually sound samples.

mordechai · Post by **mordechai** » 18 Apr 2010, 13:21

A related question -- how does one "tune" the filter to a specific frequency so as to be able to split the signal? Are there specific components already designed to read these frequencies and isolate them?

dune2k · Post by **dune2k** » 18 Apr 2010, 13:58

I'd use a simple high- and low-pass filter. 2 Parts and it should work fine (if not you could always use two times the same filter to get a greater roll off of the unwanted frequencies).
http://www.muzique.com/schem/filter.htm

RnFR · Post by **RnFR** » 18 Apr 2010, 15:28

couldn't you do it passively with some sort of crossover network? i would google crossover schematics and see what you get.

this is something i have thought about for a while since learning about the quadrafuzz as well. i was always under the impression that the QF had separate gain controls- i guess it doesn't. but, it should!

Duckman · Post by **Duckman** » 18 Apr 2010, 18:16

mordechai wrote:A related question -- how does one "tune" the filter to a specific frequency so as to be able to split the signal? Are there specific components already designed to read these frequencies and isolate them?

First, the signal is splited in four. Then, each signal is "tuned" via gyrators (check the differents components around IC's 3C, 3B, 2C & 2B)

mordechai · Post by **mordechai** » 18 Apr 2010, 19:56

dune2k wrote:I'd use a simple high- and low-pass filter. 2 Parts and it should work fine (if not you could always use two times the same filter to get a greater roll off of the unwanted frequencies).
http://www.muzique.com/schem/filter.htm

Thanks dune, the link is helpful and it looks like a simple addition. But I still an unclear of how it, or they, would be incorporated. If the high pass filter grabs the signal first, I am assuming it lets the lower range frequencies "through"...but then where do the higher range frequencies go?

I am absolutely certain that my lack of experience is the culprit here, but I will ask, how does this actually "split" the signal? Or is it a matter of splitting the signal into three equal trajectories, and then applying different band filters to each of these trajectories so that the processed signal on the other side of each filter now has a different frequency characteristic?

dune2k · Post by **dune2k** » 18 Apr 2010, 20:45

mordechai wrote:
dune2k wrote:I'd use a simple high- and low-pass filter. 2 Parts and it should work fine (if not you could always use two times the same filter to get a greater roll off of the unwanted frequencies).
http://www.muzique.com/schem/filter.htm
Thanks dune, the link is helpful and it looks like a simple addition. But I still an unclear of how it, or they, would be incorporated. If the high pass filter grabs the signal first, I am assuming it lets the lower range frequencies "through"...but then where do the higher range frequencies go?

High pass filter lets frequencies higher than the determined frequency through without doing anything to them and the frequencies below that point will by 6db quieter.
The low pass filter does the same to frequencies lower than the determined frequency.
You have to use them AFTER being split up though and you can also add one high and one low pass filter up to just get one frequency band.
If you get the feeling 6db aren't enough you can add another low/high pass filter after the first one to get about 12db of frequency rolloff. ( http://www.muzique.com/lab/dtone.htm )

mictester · Post by **mictester** » 18 Apr 2010, 22:44

RnFR wrote:couldn't you do it passively with some sort of crossover network? i would google crossover schematics and see what you get.

this is something i have thought about for a while since learning about the quadrafuzz as well. i was always under the impression that the QF had separate gain controls- i guess it doesn't. but, it should!

The typical passive crossover network wouldn't have sufficient discrimination to give useful separation. If you then tried to get more separation, you'd increase the losses, so it's not really the way to go.