Elaborating on my ideas for modifications to the Slacktave

[earthtonesaudio]

I thought I'd start a new thread to avoid confusion with the original Slacktave.

Starting here:
http://www.diystompboxes.com/smfforum/index.php?topic=55644.msg634037#msg634037

...I suggest several mods to the Slacktave that result in a quite ridiculous pedal, containing upwards of 15 knobs.  Later I suggested adding sequenced gating of the octaves and some other stuff.

Well I thought some more about sequencing the octaves, and it's a somewhat different beast than the typical sequenced filter project.  In the sequenced filter effects I've seen (Zvex Seek Wah, Vanishing Point by The Tone God), the outputs of a 4017 or 4022 type Johnson counter are summed via resistors to control a single element (in these specific effects, usually a LED/LDR combo).

If you wanted to sequence the octave outputs of the modded Slacktave I suggested, you'd first need to add a little circuitry.  The 4017/4022's outputs each go high in succession, so by adding a transistor to each output and putting the collectors to the Q outputs of the 4024, one (and only one) output of the 4024 would be muted at a time, and the "missing octave" would cycle around according to the sequencer.
On the other hand, if you wanted to select one octave at a time, and mute all the others, you'd need to invert this output, perhaps by adding an inverter gate before each transistor.
Quickly this becomes a logistical problem if you want to easily switch between "one at a time" or "all but one" because you'd need a 7PDT switch to do it all.  Sure you could achieve that with some more ICs but I'm thinking a different approach is in order.

I'm mulling over the idea, now, of achieving this with some sort of window comparator setup, or perhaps something like a flash ADC.

Oops, time for class.   More to come later!

[earthtonesaudio]

I would have somewhat expected someone to ask this already, but I got my head out of the details of sequencing the different octaves, and back to thinking if it would be musically useful to do so.

My conclusion is that it would be useful, but more so for complex patterns or pseudorandom sequences than for a straight highest-to-lowest (or vice versa) sequence.


The next thing I realized is that it would be quite useful to have a some sort of "depth" control over the sequencing.  If you implement this feature in the form of pulse width modulating the amount of sequencing, some more interesting possibilities open up.

For instance, you could use an octal inverting latch or bus transceiver between the sequencer output and the individual octave outputs of the divider chip.  Either of these could be used for their "output enable" function, thus disabling the sequencer from affecting the octaves.  If you used a PWM signal to control the percentage of time the output was enabled, you'd have a depth control.

There's a bonus from using either of these chips, the input control.  For the latch, you can make this store the last value of the inputs and transfer that to the outputs when on the next enable signal, giving you yet another layer to the sequence.
For a transceiver, you can control which direction data is sent through it, so the inputs and outputs can be reversed.  A lot of possibilities here.

[earthtonesaudio]

Finally got some time with the breadboard!  While I did put audio through all 3 chips specified in my schematic (sort of), I wouldn't exactly say I got it properly running.

I got a so-so fundamental extractor and comparator going using 3 inverters from a 4049B.  Next I planned to put the signal through 4 XOR gates from a 4070 chip, each configured as a frequency doubler.
Well, the only XOR chips I have at the moment are 74LS86.  They don't work the same in this application.  Wiring them as drawn produces squarewave fuzz but no octave doubling.  So I coupled the 4049 squarewave output through a capacitor, nothing.  Finally I added a resistor to ground after the cap (thereby forming a differentiator) and got the mosquito-tone octave up I was looking for.
I then repeated this for 3 sections and fed this output into the 4024 divider.  This is where things got weird.

I listened to the first output of the 4024 and instead of being one octave below the +3 octave signal I was presumably putting in, it was one octave below the original note I was playing on the guitar.  What?!  So either the first flip flop is acting like the fourth, or else something didn't go right in the XOR part.  I'm betting it's the XOR.

I think the most likely scenario is the XOR was merely reducing the pulse width of the signal, rather than doubling the frequency like I wanted.  I'll have to scope it to be sure.

[earthtonesaudio]

I've reached the conclusion that I need a bigger breadboard.  I got this "sort of" working the way I expected using the 74LS86, 4049B, and 4024 combination, but I was only able to get one octave up from the XOR, instead of the 4 I was planning to use.

My first thought as to why it wasn't working right was that the RC network dropped the voltage too much for the 'LS chip to properly sense.
So, I switched methods and used a spare inverter's propagation delay instead of an RC delay, which also worked, but as with the RC method, getting two octaves up seemed impossible.

Since both solutions gave the same "wrong" answer, I reasoned there must be something else going on (duh).

After some thought, this is my opinion:
The RC method works when driven by a CMOS inverter, partly because the CMOS output swings rail-rail and can drive the RC network high enough for the 'LS inputs to distinguish high from low.  However, the 'LS output does not go to the rails, and so doesn't have enough drive to ever deliver a "high" signal to the next XOR's input, unless you make the time constant sufficiently small that the voltage drop is negligible.

...Which brings me to the other side of the problem, breadboard capacitance.  The narrower the pulse, the easier it can be swallowed up by parasitic breadboard capacitance. This might explain why I could only get a single octave up using the propagation delay method.

I think I could find out for sure what's going on if I had a 4070 chip to compare, or a larger breadboard so I could buffer the 'LS chip with a 40106 or similar.  Either way it looks like I'm due for a trip to the local electronics store.

[earthtonesaudio]

I read up on TTL-CMOS interfacing, and got better results this time.  Pull-up resistors, or lack thereof, seemed to be the problem with my first attempt

I put the signal through a few 4069 gain/filter stages, then through a comparator and into the TTL XOR doubling networks (two of them in series).  Listening to the XOR outputs, it was easy to hear the frequency doubling (and quadrupling) effect.  The pulses were quite narrow so I messed with the RC values a bit to get a more "meaty" signal going into the 4024. 

Then things got weird again.

Listening to the Q1 output of the 4024, I expected to hear one octave up with respect to my guitar.  Nope!  I got jumpy glitchy octaves jumping around the notes I played.  Q2 output was similar but a little more pronounced.  Moving along, I listened to the Q7 output and WOW.  I plucked a note and held it, and the sound swept through a spectrum of sounds unlike anything I've heard before.  Started with a low whale call, then a flutter of ringmod-ish tone, then swept up through the octaves and stopped about one octave above my guitar, and as I kept holding the note it would flutter around +/- an octave or two and degrade into a swoosh of noise.  Then I'd mute my strings and... dead silent. 

[chilecocula]

Moving along, I listened to the Q7 output and WOW.  I plucked a note and held it, and the sound swept through a spectrum of sounds unlike anything I've heard before.  Started with a low whale call, then a flutter of ringmod-ish tone, then swept up through the octaves and stopped about one octave above my guitar, and as I kept holding the note it would flutter around +/- an octave or two and degrade into a swoosh of noise.  Then I'd mute my strings and... dead silent. 

That sounds really good.