Octave effect on Green Ringer

[duck_arse]

Think of octave doublers as having attention deficit.

rickenbacker ritalin?

[hellwood]

I think this pedal is somewhat limited. this is what mine sounds like:
http://theharshcarpets.com/wp-content/uploads/2014/03/pedal_soundclips/my_green_ringer.wav

and this is what it looks like(the triangle one):
http://theharshcarpets.com/wp-content/uploads/2014/03/pedal_featured.jpg

...I always end up rolling off the guitar tone pot for more octave on my octave pedals

[Mark Hammer]

The "Fuzz" section of the Boss ROD-10 - a 5-circuit compendium of distortion in variying amounts - is essentially a simplified octave fuzz.  You can see it between Q42 and Q36 at the bottom left of the schematic here.  Q40 forms the standard phase-splitter, that feeds two opposite-phase versions of the signal to a mirror-pair of transistors.  And you can see the diode pair to ground, as found in the Superfuzz and its cousins.  It yields a reasonably decent octave (with the standard not-much-until-you-hit-the-7th-fret phenomenon), that is less raspy than other octave fuzzes.

But take a closer look at Q38/A39.  Each has its base biased by a resistor pair, but look, they are each biased differently, despite being identical transistors.  In the one case 470k/4k7, and in the other 100k/15k.  I don't understand the purpose of that.  In virtually every octave fuzz I've ever seen that uses a phase-splitter, the two versions of the signal are biased in identical manner.  So what gives here?

I have to offer the caveat that the diagram shows NO adjustable drive for the Fuzz section, despite the fact that - as of my monkeying around doing a demo for somebody yesterday - the Drive control clearly produces audible differences in that sub-sircuit.  So, something is not drawn accurately.  And if they can mis-draw the input stage/source of the Fuzz, then they can also misdraw the biasing.  But I imagine some of you transistor jockeys have a better intuitive feel for what the values shown are probably doing.  I know I'm not spending my Sunday disassembling and reassembling this beast just to confirm or disconfirm that the values shown are what's used.

[Digital Larry]

One thing about that fuzz circuit is that after the rectification, it is run through an AC coupled gain stage and then into some back to back clipping diodes.  One of those diodes is going to take the sharp point (that resulted from folding around the zero crossing) and round it off, reducing the "buzz" effect of a sharp corner in your waveform.

Yes, that biasing arrangement is a bit strange.  Especially considering that the emitters and (less importantly) collectors are connected.  I don't have a handy explanation for that.

[duck_arse]

my eyes will only squint so much, but it looks to me as though the coupling caps on those odd dividers are also diff values, and the higher value is where I woulda thort the lowwer would be. curiouser and curiouser. maybe some frequency shaping/cutting?

[Digital Larry]

Here's my guess, inspired by the previous post...  The phase splitter uses 4.7k resistors in the emitter and collector circuits.  The driving impedance of the collector side will be about 4.7k.  The driving impedance of the emitter side will be less.  The "bias" networks feeding the long-tailed pair of Q38/Q39 cannot really pull the base voltages very far apart because the emitters are joined.  So I am thinking that the differences in the bias/coupling components are more to try to match the frequency response of both sides.  Let's remember that maximum octave is achieved when the processing of the positive half of the waveform is identical to the bottom half.

Also, a pedantic point perhaps, but what I am calling "maximum octave" is really "minimum fundamental".  Any difference (other than mirror image) of the + half vs. the - half results in some fundamental frequency coming through.  If you COULD make those identical, you really would have eliminated the fundamental.

[Mark Hammer]

I've babbled on about combining the Ampeg Scrambler and Distortion+ ad nauseum before.  One of the things that the clipped front end does is provide degrees of simulated sustain.  Set the clipping  to minimal by reducing the gain of the D+, and the signal has a discernible attack and decay.  Set the gain high, with the output level pulled back to restrain the maximum input going to the Scrambler part, and you have a more constant-level signal with a more protracted decay.  If one introduces enough treble rolloff into that front end (going a step beyond what using the neck pickup and rolling off the Tone on the guitar would normally accomplish), you can vary essentially how long the octave "lasts" in the Scrambler.

This would occur for two basic reasons.  One is the constancy of the signal when clipped.  The other is that, as the signal decays and gets clipped less, the positive and negative half-waves become more dissimilar, such that when you "fold them over" you don't produce anything where a discernible octave can be heard.

Again, I cannot emphasize enough that, when dealing with guitar signals, it is the fleetingness and variability of the elements in the signal that often gets in the way of accomplishing what we want to accomplish.  "Folding over" a pure sustained sine wave to produce a frequency doubling?  Child's play.  "Capturing" and doubling that fundamental in a manner that allows an octave-up to stand out and be audible, for more than a mere instant, much harder.

I mentioned earlier that the ROD-10 has an octave fuzz included and the octaving isn't bad.  I also mentioned that the Drive control is not shown for that channel/sub-circuit, suggesting the schematic is not accurately drawn.  Wouldn't surprise me a bit if the input to the octave fuzz was derived from one of the mild overdrive circuits the ROD-10 includes, and that the compression and "rounding" produced by that subcircuit provides an optimal input signal for the octave fuzz.  I don't know this for certain, but I wouldn't be surprised if it were the case.

Finally, look at the input stage of the Foxx Tone Machine.  The front end is essentially the same structure as a Fuzz Face, but the feedback path from Q2 emitter to Q1 base is distinctive, leading me to think that it boosts the fundamental preferentially, though I'm so naïve about transistors I may be way off.  Anybody care to address that claim?

[Digital Larry]

I mentioned earlier that the ROD-10 has an octave fuzz included and the octaving isn't bad.  I also mentioned that the Drive control is not shown for that channel/sub-circuit, suggesting the schematic is not accurately drawn.  Wouldn't surprise me a bit if the input to the octave fuzz was derived from one of the mild overdrive circuits the ROD-10 includes, and that the compression and "rounding" produced by that subcircuit provides an optimal input signal for the octave fuzz.  I don't know this for certain, but I wouldn't be surprised if it were the case.

Following your line of reasoning, I'm guessing that the "single triangle" control line actually goes to the diode which controls the first FET after the input.  And that the (audio) input to Q42 comes from the (audio) output of Q30.

[Mark Hammer]

Quite possible.  But like I say, not interested in wasting an entire day disassembling, tracing, and reassembling the unit.  Just finished spending a couple months sanding down and refinishing an oak floor (though, truth be told, its 10min of vacuuming for every 5 minutes of sanding), so I'm looking to get my life back, and maybe catch a little summer before it's gone.. The R series is not exactly service-friendly.  If I ever need to get in there for any reason, I'll be sure to take a peek, but for now, I'm gonna "delegate" that to someone else.  

[Digital Larry]

Another thing I think is interesting about this circuit is the prevalence of what looks an awful lot like a discrete op-amp implemented with the face to face JFETs along with a PNP further down and various things in the feedback loop, such as back to back diodes (ODII), back to back 1 to 2 diodes (OD I), etc.

As far as I have seen, most of the discrete overdrives do NOT use this topology, rather a daisy chain of discrete devices, sometimes with questionable (cough) biasing.  I do see that there is a whole project area on this forum dedicated to a discrete op-amp though.  Anyone built something like this?  Seems like many of the more popular designs simply have lower parts count.  I can't guess at the sound of these other than they are probably more like the op-amp based diode clippers rather than the raging raw waveforms of unbiased transistors slamming mercilessly into the rails.

[Mark Hammer]

The BD-2 Blues Driver, which enjoyed a brief moment of popularity here, and then kind of succumbed to TS-mania, is related to the various discrete op-amp circuits contained in the ROD-10.  Jay Doyle has a number of discrete op-amp circuits here a decade or so back.  I keep meaning to make some of them, but, well, you know.... 

But that's the non-octave-fuzz part of the box.  In fairness, we should return to the OP's interest: making the Green Ringer work optimally.

[Digital Larry]

But that's the non-octave-fuzz part of the box.  In fairness, we should return to the OP's interest: making the Green Ringer work optimally.

I don't have a unit to check, and I know you have some home improvement responsibilities, but I'd say your thought that OD1 is the front end of the octave fuzz is pretty much spot on.  There's that point in the schematic at upper left just below Q49 where it looks like 2 lines cross.  But my very strong suspicion is that it's two right angles meeting, not lines crossing.  This would not get high marks from me on schematic clarity but for all I know they do weird things like this on purpose?

So far the summary is something like:
1) Try to make sure your circuit is "balanced".  Not sure how to do this without a scope.
2) Roll off the highs on your guitar.
3) Play around the 7th fret or so.

Anything else?

[thehallofshields]

1) Try to make sure your circuit is "balanced".  Not sure how to do this without a scope.

Hah. I'm probably one of the only posters right at the moment that's got the GR on breadboard, and I'll gladly test any theories.

I've tried to balance mine so that my DMM reads an equivalent Voltage Drop across the diodes. The Voltage Drop varies depending on how much noise is coming through the guitar and breadboard. With everything 'quiet' I'm reading about 17mv across each, and I'd say I get a pretty pretty pronounced Octave.

Of course, if I'm driving the circuit with an asymmetric signal, that might not be ideal...

If anyone who is more intelligent than me comes up with any other tests to balance the Circuit, please let me know!

[thehallofshields]

Anything else?

Well... I can list some things I've tried that didn't work.

1.) Adjusting the Voltage-Divider made by the 22k resistors. I thought maybe if I upped the mysterious 'trickle current', the Diodes would conduct sooner and I'd get less 'Gating'. So I wired up a 50k Potentiometer as a Voltage-Divider.

I get the same result anywhere between (approx.) 2V and 7V. The Amplitude drops severely at about below 20% and above 80% of the pots rotation.

[duck_arse]

some straws for thehall to clutch -

- use differing value coupling caps like in the boss diagram. use differing values of "68k's" to the voltages divider, kinda like the boss does. try the dual divider setup, like ^ I put above, and emulate what the bleeding boss does.

and ground the input connection when you're measuring, for real quiet quiet.

[Mark Hammer]

I'd say that, if one doesn't have ascope, the next best thing is to feed the circuit a steady sine wave, and tweak the appropriate resistances until you think you hear the most octave.

[thehallofshields]

and ground the input connection when you're measuring, for real quiet quiet.

Not sure why this idea didn't occur to me. Thanks.

[thehallofshields]

I'd say that, if one doesn't have ascope, the next best thing is to feed the circuit a steady sine wave, and tweak the appropriate resistances until you think you hear the most octave.

I'd like to do this to test 3 methods of balancing the rectification, and do a little writeup comparing the effectiveness of each.

1) Series resistance - The "Null Carrier" mod. I'm not sure who came up with this one but I first saw it on the Madbean "Squeeker" project.

2) Diode 'Biasing' - The trimpot replacing the 68k resistors from the 21st Century Green Ringer

3) Phase-Splitter Biasing - Mark's suggestion to break the 10k symmetry of Q2's Emitter and Collector Resistors.

Can anyone reccomend a common-parts project that can generate a Sine-Wave? Are there any easy-ish shape-ble LFO circuits that will fit the bill?

[Digital Larry]

Can anyone reccomend a common-parts project that can generate a Sine-Wave? Are there any easy-ish shape-ble LFO circuits that will fit the bill?

How about using your computer's sound card, with a free audio editor like Goldwave or Audacity?

[thehallofshields]

How about using your computer's sound card, with a free audio editor like Goldwave or Audacity?

I can try Audacity. Any tweaks I need to make to the circuit for the Low-Z Source?