Basic clean octave down design?

[Rain Dog]

Hello, I'm new here. I joined to make new friends and because there is a specific pedal I would like to build.

I want a octave/fuzz pedal that I can vary the octave content. I'm not interested in the octave up frequencies in the octave portion of the pedal, in fact if they were gone that would be great. I will be driving this with a passive bass or a the occasional humbucker equipped guitar.

In other words, I want a simple and clean (nonfuzz) octave down effect built into a nicely designed Muff pedal. I would like to have these effects in series with the Octave down first and the Muff second. The thing is I want to be able to vary the blend of the octave down signal with the original signal. I want the range to be from, No octave signal into the muff to All the output of the octave into the muff. I expect to set the blend at 25% octave and 75% original. I would be open to running the effects in parallel and blending to output. I'm not sure which would sound better.

General Guitar Gadgets has a Green Ringer Project that looks like it might be right. I just don't know enough about it to tell.

My questions are:
1) What is a good octave down design that can deal with bass frequencies?

2) Is blending the two signals as simple as adding a linear or Log pot to the input signal or is there more to it than that?

3) Series or parallel?

If this is an all around bad idea, please let me know.

[Mark Hammer]

Welcome, amigo.  Soon, you may find yourself thinking "With friends like this, who needs enemies?", but that's another matter.  :lol:

The question you pose is one that comes up regularly, and it has the same dismaying answer every time.  Here is one more version of it.

The most accurate way to produce a note one octave below what comes in is to simply divide the note frequency by 2.  The easiest way to do that is to use what is called a flip-flop circuit: a circuit that changes from one state to another each time a particular event occurs.  In this instance, the "event" is the onset or rising/leading edge of the note waveform.  When that incoming waveform goes above some threshold value (and that's another topic we'll come back to), the flip-flop goes "There it is", and when the leading edge of the next wiggle on the oscilloscope occurs, the flip-flop goes "There it is again" and switches to another state.  This goes on for as long as you keep feeding the flip-flop an input signal it can react to.

The trouble with this, from a musical standpoint, is that all the flip-flop can do is be in a "state".  It can't produce any sort of waveform other than going "high" or "low".  That means that what comes out of the flip-flop circuit is always and forever, you guessed it, a square wave.  Of course it is the almost universal use of flip-flops in octave-down units, and their inherent property of producing square wave octave-down signals that prompts your question in the first place.

The solution is to basically shape the flip-flop output with additional filtering, and commercial/DIY octave units vary in how much and how they further condition the octave-down signal.  A less farty octave-down is going to have a lot of lowpass filtering to turn it from something purely square to something closer to sine-ish.  There ARE tons of waveform converter circuits out there for synth people or those who want to produce different waveforms for controlling circuits (e.g., a triangle to sine converter or square to ramp, etc) so there must be something for taking a square wave and rounding the peaks to produce somethng very close to sine.  A reasonable, simple passive lowpass filter (i.e., 2 resistors in series with a cap to ground after each resistor) would take you most of the desired distance for much less trouble, and may well be all you're really imagining, tonewise.

Part of the "cleanness" of octave-down units is their tracking.  Tracking has two components.  One is the extent to which the octave down reliably starts and stops in synchrony with the note you're picking, and the other is the extent to which it is possible to produce an octave-down across all notes on the fingerboard with equal ease or probability.  These two things are related but ultimately dissociable/separable.

I mentioned that the incoming signal has to cross a threshold.  The reason for this is that, like most other players, you are probably uninterested in an octave down for finger noises or hiss.  So, the flip-flop only responds to signals that are above a certain minimum level.  In setting that level, you can err on the side of false-positives or false negatives.  Erring on the side of false-positives, you may have a threshold set low enough that softly picked notes and hammer-ons will still produce octave-down but so will moving your fingers or strings that inadvertently ring in sympathy to what you are intending to pick.  Moreover, the threshold may be so low that normally-occurring variations in string signal at the tail end of notes (e.g., due to finger vibrato) may cause the flip-flop to start and stop retriggering.  Erring on the side of false-negatives, the threshold will be set high enough that you won't get triggering by random sounds, but you also won't produce an octave sound unless you pick very cleanly and have a strong pick attack, and the octave sound will disappear well before the complete decay cycle of the note is over.  Ideally, you want something in between those two, although I can imagine musically valid reasons for pursuing a different strategy.

The other aspect of tracking, differential frequency sensitivity, is often a function of strategies used to reduce false triggering.  So, rolling off the low end to eliminate triggering to 60/120hz hum can result in poor sensitivity to low notes on the E string.  Rolling off high end to eliminate false-triggering to finger noise or other spurious signals may attenuate higher notes enough that they don't trigger as reliably.  And so on.  A bunch of different things will result in the note fundamental information reaching the flip-flop being better/louder for some notes than others.

Collectively, here, there seems to be a consensus that the Craig Anderton-designed Rocktave Octave Divider, sold by PAiA in kit form at a reasonable price (www.paia.com), but posted around in various redraws from the original Guitar Player project article, is just about the best analog octave-divider around in terms of stable dependable tracking, tonal flexibility and tonal quality.  One of its secrets is the use of a compander chip to smooth out the signal level across the spectrum and provide an even, optimal signal for reliable flip-flopping.  It also includes a tone control for the octave-down to smooth out the tone.

[Rain Dog]

That was brilliant! I understood all of it. Thank-you very much. Was this in one of the faq's and I just missed it?

Now what do you think of the blend idea. Is it that simple or do I need to rethink that?

Series, parallel? I was thinking series because I figured the Fuzz portion would disguise some of the tracking errors.

Hopefully by building my own Big Muff (I have a big green and love it) this will give me a "smoother" fuzz and then I can add small amounts of brutal ugliness to the signal.

[Mark Hammer]

One of the reasons the Rocktave is popular is that it incorporates a fuzz, and has independent level-mixing controls for the fuzz, one-octave and two-octave-down signals, in addition to a tone control.  Of course, HAVING an on-board fuzz and having the one you want are two different things.  Moreover, the same tone control you use to tame the octave is also the one that affects the fuzz signal so you can't have round octave and buzzy fuzz at the same time.  The Rocktave's fuzz is a meat-and-potatoes thing whose distortion-amount or tone cannot be changed: much more like crappy-software-bundled-with-your-printer than any sort of feature item.

Download the schematic and article from my site (http://hammer.ampage.org) because we'll need it for what follows.

One way in which the basic unit could be altered to do what you want, would be to take the straight signal that normally reaches the output/mixing stage via C8 and R17 and divert it ("pervert" it is actually more like it).  If you simply lift the end of R17 that connects to pin 8 of the output stage, you now have something you can send to a clipper/fuzz of your choice.  The output of that clipper can now be folded back into the mix at pin 8 (although you'll need to be sure it does not invert phase).  En route to pin 8 you can have options for distortion amount, EQ-ing, or even adding octave-up if you feel like it.  Because the basis for that additional fuzz is tapped off the input stage it will have not impact on the octave-diwn tracking in the rest of the circuit.  You *will* have to forfeit clean tone, although personally I find that not such a sacrifice in this unit.  On the other hand, if a fuzz was patched in there, there is no reason why there could not be a bypass switch for that "sub-effect" to restore clean tone.  Heck, if you could vary the clean tone level, that would be terrific because the stock unit doesn't seem to have enough clean level when you want it.

[Ben N]

Mark:

It sems to me that the Rocktave has a lot of modding potential as a kind of microsynth type of thing, since it does such a good job of smoothing and shaping the signal before actually dividing it.  That same process ought to work wonders for, as you say, octave up, different fuzz sounds, ring mod, wave shaping or even harmonizing, just by patching them into the mix.  (Would any kind of buffering be necessary if there were multiple signal paths?) Add some envelope shaping and some filtering options, and we've got a little monster.

So now, me being both stupid and lazy, which mad scientist is going to concoct this thing in their lab?

As a start, could you recommend a type of octave-up circuit that would be a good fit here?  Perhaps a CMOS multiplier?  And might the envelope options of the EA Fuzz (which also uses a compander rather creatively) be used here, also?

Ben

[gez]

Dano chilli dog offers a cheap and cheerful solution.  Allows you to blend, tracks well (I've had no problems with my Jazzmaster) and does a pretty good impression of a bass when the fundamental is tuned out (I use it for bass lines).

[Rain Dog]

One of the reasons the Rocktave is popular is that it incorporates a fuzz, and has independent level-mixing controls for the fuzz, one-octave and two-octave-down signals, in addition to a tone control.  Of course, HAVING an on-board fuzz and having the one you want are two different things.  Moreover, the same tone control you use to tame the octave is also the one that affects the fuzz signal so you can't have round octave and buzzy fuzz at the same time.  The Rocktave's fuzz is a meat-and-potatoes thing whose distortion-amount or tone cannot be changed: much more like crappy-software-bundled-with-your-printer than any sort of feature item.

Download the schematic and article from my site (http://hammer.ampage.org) because we'll need it for what follows.

One way in which the basic unit could be altered to do what you want, would be to take the straight signal that normally reaches the output/mixing stage via C8 and R17 and divert it ("pervert" it is actually more like it).  If you simply lift the end of R17 that connects to pin 8 of the output stage, you now have something you can send to a clipper/fuzz of your choice.  The output of that clipper can now be folded back into the mix at pin 8 (although you'll need to be sure it does not invert phase).  En route to pin 8 you can have options for distortion amount, EQ-ing, or even adding octave-up if you feel like it.  Because the basis for that additional fuzz is tapped off the input stage it will have not impact on the octave-diwn tracking in the rest of the circuit.  You *will* have to forfeit clean tone, although personally I find that not such a sacrifice in this unit.  On the other hand, if a fuzz was patched in there, there is no reason why there could not be a bypass switch for that "sub-effect" to restore clean tone.  Heck, if you could vary the clean tone level, that would be terrific because the stock unit doesn't seem to have enough clean level when you want it.

Got the PDF for the Rocktave. It is going to take me some time to assimilate all of this.  

[Rain Dog]

Dano chilli dog offers a cheap and cheerful solution.  Allows you to blend, tracks well (I've had no problems with my Jazzmaster) and does a pretty good impression of a bass when the fundamental is tuned out (I use it for bass lines).

I'll check this out too.

[Ben N]

Yep, the Chili Dog is pretty unbeatable for the price.  I've tried to use it for pseudo-bass, but I find it lacks the punch of a bass.  Still, it is a pretty cool little box--tracks as well & plays as clean as pretty much any commercial stompbox octave down out there.  I'd be nervous about relying on it in giggin situations, though it's worth getting one just for fooling around.

[gez]

I've tried to use it for pseudo-bass, but I find it lacks the punch of a bass

Granted, but it's great for Reggae and Dance stuff (no top end and loads of BOOM).

[Rain Dog]

I've tried to use it for pseudo-bass, but I find it lacks the punch of a bass

Granted, but it's great for Reggae and Dance stuff (no top end and loads of BOOM).

Hey now! Let's have none of that! Pretty soon people will start thinkin' they don't need a bass at all and then where would we be??? Madness I tells ya!
:wink:

As I said in my first post I will be using a real bass to drive these effects.

[StephenGiles]

One can take things further, combine the Roctave compressor, fundamental generator from the EH Guitar Synth (or deluxe Octave Multiplexer) up to the LM311 squarer, and then whatever divider takes your fancy - now there's a worthy goal for someone!

[Ben N]

Careful, now, before someone gets hurt!