Those "indirect" octave boxes.

[Mark Hammer]

Looking through the schematics of the Boss OC-2 and the Pearl OC-7, I was struck by the manner in which the octave and two-octave down signal to be mixed in with straight signal was not the output of the 4013 flip-flop per se, but rather appeared to be the original signal as switched by the 4013 flip-flop output.  http://www.godiksennet.com/images/sch/Boss_OC2_Oktave.jpg

This is confusing to me.  What exactly is going on, and what advantages does it hold over using the flip-flop's output directly?

[johngreene]

I posted the reason for this after someone posted the Pearl schematic a while back but the reason is this:

The 4013 is creating the control signal for the opamp controlled phase inverter/modulator. This means that the original signal going through the opamp is having its phase inverted at the octave down rate. Think of a phase shifter where the LDR or FET is replaced with one that switches instantly from MAX to MIN. This is equivalent to ring modulating the input signal with one that is exactly an octave lower. As a result you get out of the opamp a double side-band suppressed carrier signal. The lower sideband is 1 octave down and the upper side band is a 5th above. The carrier, or original signal is gone for all intents and purposes.

The advantage to doing it this way is the amplitude of the octave down is directly proportional to the input signal.

--john

[GFR]

A further advantage is that the harmonic content is closer to a guitar (distorted) than a square wave - it's cleaner.

[Mark Hammer]

Thanks, gents.  That explains a lot.  One thing I guess we should clarify for those who may not be able to think that far ahead, or who are just catching up, is what John meant by "dynamics.  As far as the flip-flop side of things, and the tracking, notes still have to be above the required threshold for proper division-down to take place via the 4013.  As long as the note played is supra-threshold, though, any fuirther dynamic differences in note amplitude WILL be reflected.  The flip-flop provides a fixed output level, but what it is modulating still has some volume differences built into it.

Of course, all of this presumes that what you feed the octave box does not have its dynamics removed via use of a fuzz or compressor/limiter ahead of the octave unit.  So we'll just say that the OC-2/OC-7 designs permit retention of whatever dynamics there might be in the input signal.

Ghuillhermo's (sp?) point about signal "cleanliness"is bang on too.  The square wave of the flip-flop will always be fuzzy-sounding and require some serious filtering to sound smoother.  But where do you set the filter rolloff?  What notes (hence harmonics) will be most in need of filtering?  The solution in these pedals is to use the original signal, and let the octave tone be based on the tone of the input signal.  Smart.

Seems to me a combination of this approach and the companded tracking used in the Rocktave would make for a truly great-sounding and nice handling octaver.

[Processaurus]

Interesting circuit and discussion, its a step up terms of complexity from the blue box I spent some time fiddling with lately.  I'm trying hard to understand some things in this circuit, so I have a couple questions and ideas on how it works if you guys want to bear with me.  Looking at IC5, where the octave sound shaping starts happening,  it looks like the signal gets half wave rectified by the Ge diode to V1/2 (D10).  After the phase processed signal gets through the filter with IC4, I imagine it looking somewhat like an irregular triangle wave, with some wiggle around the zero crossing.  I guess the reason the second octave shaping section is connected to the output of the 1st octave circuit is that that would get them a more triangular output to the second octave, rather than a lumpy square wave they'd get if they tapped the input to the second octave circuit directly from the unaffected input signal (IC1 pin 1).  Could be an interesting mod, switching the input to the second octave shaper circuit between the two sources.  Being a mod minded young spud, I'm also wondering what would happen to the tonality if you lifted the two Ge diodes that do the half wave rectifying?


I remembered a geofex article on an opamp based selectable phase inverter, http://geofex.com/Article_Folders/polarity_reverser/polarity_reverser.htm
Pretty much the same circuit, except with a FET (controlled by the flip flop) to short out the non inverting input, rather than a switch. 

When the FET is switched on and the + input is shorted to V1/2, the inverting gain of the opamp is just what it would be in the normal non-inverting configuration:
-(Rf/R1)= -(27K/47K)= -.57

When the FET is switched off, and the signal, unaffected by the 47K resistor, flows into the + input at near full amplitude, then the  non inverting gain would be
1+Rf/R1 = 1+27K/47K = 1.57
but the inverting configuration is still running, so the sum of the two is 1.57 + (-.57)= 1

Does it seem strange that the gain is different for the different phase polarities?  Is the non-inverting gain equation from the GEO article correct?  If that is right, what was Boss thinking by making it asymmetrical?

One unrelated thing strikes me as kind of silly, they could have used another 4013, rather than the odd BA634, and used the extra flip flop instead of the patent boss jillion component discrete  flip flop for the bypass switching.  Old habits, I guess.

I've definitely been interested in the funkiness of analog octavers lately, especially since in a band I play with the other guitarist and bass player both use the OC-2, when all three of us occasionally hit them at the same time, its so heavy...

The boss one has some interesting ideas in it, it just needs 10 more knobs and switches .  I'd definitely put a 3 position switch in to change the filtering in the flip flop driver section, with values aimed at tracking a guitar's upper registers, guitar lower registers/ bass upper notes, and a setting to get the lowest notes on a bass to stick.  Or that could be knob controlling a low pass filter, kind of like the MS-20s external signal input section, which has both HP and LP filters.  The compression before fundamental extraction that the Rocktave uses would be very beneficial to the boss design.  Would compression before filtering, or filtering before compression be better?

[MartyMart]

Interesting, I guess that's why I enjoy using the Pearl OC-7 so much, it does seem
to track well and have a decent tone.
Funny thing is, it also seems to "work" better with a compressor up front  ! ..?

I sold my OC-2 in favour of the OC-3's "improved" features, it's a bit more "digital" sounding
but also tracks very well and I preferred it .....

Yeah, why not use another 4013 for the bypass switching ?
I can't see any advantage for using the traditional ( and not so great ) flip/flop
circuit, plus all it's components ....

MM.

[GFR]

Ghuillhermo's (sp?)

Guilherme

[GFR]

Yeah, why not use another 4013 for the bypass switching ?
I can't see any advantage for using the traditional ( and not so great ) flip/flop
circuit, plus all it's components ....

MM.

Because every other Boss pedal uses a discrete FF they just need to cut & paste.

[GFR]

Seems to me a combination of this approach and the companded tracking used in the Rocktave would make for a truly great-sounding and nice handling octaver.

Take a look at the Deluxe and the non-Deluxe octavers from EH. There are plenty of nice ideas in there too.

[johngreene]

Interesting circuit and discussion, its a step up terms of complexity from the blue box I spent some time fiddling with lately.  I'm trying hard to understand some things in this circuit, so I have a couple questions and ideas on how it works if you guys want to bear with me.  Looking at IC5, where the octave sound shaping starts happening,  it looks like the signal gets half wave rectified by the Ge diode to V1/2 (D10). 

No, the signal does not get half-wave rectified. The diode is to clamp the DC of the input to keep it from moving a lot when the switch is toggled.

After the phase processed signal gets through the filter with IC4, I imagine it looking somewhat like an irregular triangle wave, with some wiggle around the zero crossing. 

Nope, what it looks like is the input signal having its phase inverted at the rate of 1/2 the fundamental frequency (from the 4013).

I guess the reason the second octave shaping section is connected to the output of the 1st octave circuit is that that would get them a more triangular output to the second octave, rather than a lumpy square wave they'd get if they tapped the input to the second octave circuit directly from the unaffected input signal (IC1 pin 1). 

Nope again. The second is connected to the output of the first because it is generating an octave down of the 'octave-downed' (octaved-downed?) to get 2 octaves down from the fundamental. Since the first octave down signal is generated by inverting the phase at a 1/2 fundamental rate, this signal will generate another octave down if its phase is inverted at a rate of 1/4 the fundamental. So coming out of this stage you get 2 cycles of the input signal in phase and then 2 cycles of the input phase inverted. You'll notice the phase shift stage between the phase inverters, this is to prevent the phase being inverted at the same location in time between the 2 phase inverters as this would cause a cancellation of the 1st octave down and you would end up with a signal output that is 1/4 an octave below and another than is 1/4 an octave above. Then if you add this synchronously phase inverted signal back to the input signal it will just cancel for some of the time and add for the other.

Could be an interesting mod, switching the input to the second octave shaper circuit between the two sources.  Being a mod minded young spud, I'm also wondering what would happen to the tonality if you lifted the two Ge diodes that do the half wave rectifying?

Since they are not half wave rectifying, the answer is no. This would probably render the circuit non-functional which I would personally call 'broke' rather than a mod. 

I remembered a geofex article on an opamp based selectable phase inverter, http://geofex.com/Article_Folders/polarity_reverser/polarity_reverser.htm
Pretty much the same circuit, except with a FET (controlled by the flip flop) to short out the non inverting input, rather than a switch. 

When the FET is switched on and the + input is shorted to V1/2, the inverting gain of the opamp is just what it would be in the normal non-inverting configuration:
-(Rf/R1)= -(27K/47K)= -.57

When the FET is switched off, and the signal, unaffected by the 47K resistor, flows into the + input at near full amplitude, then the  non inverting gain would be
1+Rf/R1 = 1+27K/47K = 1.57
but the inverting configuration is still running, so the sum of the two is 1.57 + (-.57)= 1

Does it seem strange that the gain is different for the different phase polarities?  Is the non-inverting gain equation from the GEO article correct?  If that is right, what was Boss thinking by making it asymmetrical?

Except that the gain is not different. You are forgetting to include the 100K connected to the non-inverting input in your calculations. This will reduce the signal at the non-inverting input by 1/3 when when the FET is off.

--john

[puretube]

then there is that K*rg implementation of an old David Bunger organ thingy...

[Processaurus]

Interesting circuit and discussion, its a step up terms of complexity from the blue box I spent some time fiddling with lately.  I'm trying hard to understand some things in this circuit, so I have a couple questions and ideas on how it works if you guys want to bear with me.  Looking at IC5, where the octave sound shaping starts happening,  it looks like the signal gets half wave rectified by the Ge diode to V1/2 (D10).

No, the signal does not get half-wave rectified. The diode is to clamp the DC of the input to keep it from moving a lot when the switch is toggled.

Hi John, thanks for looking that over, I didn't believe you so I breadboarded that opamp stage from C24 to the output of IC5, and hooked it up to the scope.  You're right about the Ge diode not half rectifying the signal.  But it isn't there to keep the DC level from moving.  It turns out it makes the DC on the output of the opamp shift up and down, depending on what the FET switch is doing.  When the FET is closed (and the output is out of phase with the input), the DC gets lower, and when the FET is opened (output in phase), the DC gets higher.  The DC offset added to the output is proportional to amplitude of the signal, so when the signal quiets down, the offset dissappears, regardless of what the FET switch is doing, and the output of the opamp rests at V1/2. 

what would happen to the tonality if you lifted the two Ge diodes that do the half wave rectifying?

Since they are not half wave rectifying, the answer is no. This would probably render the circuit non-functional which I would personally call 'broke' rather than a mod. 

If you remove the diode, it actually works fine (and more like you suggested the circuit worked) as a standard phase inverter, with no added DC offset when you open and close the FET switch.  Could be more of a ring mod type sound, if one did that, with less sub octave fundamental.  Having the DC offset of the flip flop getting in there must give it more of a fundamental to get low end from, which makes sense, because the OC2 has plenty of low end, and is fairly synthy (filtered square wave) sounding.

The gain of the two phases is like you said, the 100K would make the non inverted signal .66 the gain, and the inverted phase would still have .57 the gain, which is what it looked like visually on the scope, the in-phase signal was a little bit bigger, I guess they wanted to use common resistor values, rather than get each exactly the same.  I wonder if one could hear the difference, if you altered the resistor values slightly to have equal outputs of both phases.  Doubtful, with flip flop's DC that the Ge is adding, and all the filtering.

What you said about the phase shift scheduling of the octave shaping section is interesting, I was suspecting there was some planning put into that.

then there is that K*rg implementation of an old David Bunger organ thingy...

K*rg, is that the design where the octave frequency logic signal from the flip lop controls a 4066 type switch that the input signal is passing through, so that every other full cycle is muted?

Looking some more at the Boss scheme, I'm curious if anyone can figure what the deal with the lightly circled section on the schematic with IC2 and IC3 is?  The second half of IC1 is a lowpass filter to help tracking, thats easy enough, but I'm wondering what the benefits of their more complicated flip flop driver section is, compared to just squareing the input signal and driving the clock input on the first 4013?  Do they get some kind of hysteresis from the extra circuitry? The two comparators made with the first halves of IC2 and 3 are feeding the set and reset pins on the first 4013.  Are the second halves of both IC2 and 3 some type of signal rectifier?

EDIT: for confusing it writing

[johngreene]

Hi John, thanks for looking that over, I didn't believe you so I breadboarded that opamp stage from C24 to the output of IC5, and hooked it up to the scope.  You're right about it the Ge diode not half rectifying the signal.  But it isn't there to keep the DC level from moving.  It turns out it makes the DC on the output of the opamp shift up and down, depending on what the FET switch is doing.  When the FET is closed (and the output is out of phase with the input), the DC gets lower, and when the FET is opened (output in phase), the DC gets higher.  The DC offset added to the output is proportional to amplitude of the signal, so when the signal quiets down, the offset dissappears, regardless of what the FET switch is doing, and the output of the output rests at V1/2. 

Ha, that actually explains it. With the added DC offset the octave down will be larger than the 5th above and therefore more prominate! It's cool you were able to prototype it so quickly.

--john

[puretube]

Quote from: puretube on Today at 01:20:29
then there is that K*rg implementation of an old David Bunger organ thingy...


K*rg, is that the design where the octave frequency logic signal from the flip lop controls a 4066 type switch that the input signal is passing through, so that every other full cycle is muted?

yes, sir.

[christian]

I did this few years back. I used discrete flip-flop (pretty much the same as in all boss-, ibanez-, etc-pedals)
and grabbed the differential outputs, driving a simple vca made out of diode-bridge, modulating the original guitar signal.
Sounded nice and I was also able to tweak the flip-flop so that it tracked pretty well and sounded real nice and smooth.
This was like 4-5 transistors all together, I guess.

And it wasnt similar to Shocktavia, which uses the same idea.

ch.

[Stompin Tom]

First, thanks everyone for this thread... it forced me to sit down with the oc-2 schematic and try, in a very simple way anyways, to figure out what's going on in there. Although I don't completely understand, I think I have some idea now...  Here's my question. Processaurus suggested adding a switch to adjust the tracking for different guitar inputs... I believe he has in mind what Mark suggested here for the blue box:

http://www.diystompboxes.com/smfforum/index.php?topic=44310.0

That is, making the drive of the signal to the flip/flop adjustable... for the oc-2, would that be trimpots for R7 and R5? Sorry for the stupid question, I'm just trying to figure out if I understand...