Hello,
I was wondering a lot about Joe Davisson Shocktave http://diystompboxes.com/analogalchemy/sch/shocktave.html (http://diystompboxes.com/analogalchemy/sch/shocktave.html) because I think it has a lot of potential for such a simple schematic. But that's not important now :) What I wan't to know is
How exactly does it accomplish octaving (I'm clear about preamp stuff)? It is a theoretical question. I do have some electronics knowledge but can't figure out what is happening. Mr. Davisson himself wrote
"The circuit consists of a high-gain preamp stage, followed by a frequency-to-voltage convertor, which in turn drives an oscillator. The oscillator is tuned so that it can only keep up with every other cycle, resulting in the octave-down effect which can be mixed with the original signal. "
What frequency-to-voltage convertor? For what I can see there is some kind of oscillator which puts out pulses (like astabile) and input signal is fed to it. So it is like you modulate volume of the oscillator with your guitar output. How can that produce octave below? Isn't it like ringmod? If that's what is happening.
What does it mean for oscillator to be "tuned" in such a specific way to enable octaving?
And what is this diode connected between emmiter and base? When I simulate it, I see it has almost no effect except it deforming guitar signal a bit.
I do understand electroncis well, but this falls outside the "blocks" of audio effects that I know.
I'm not too clear about the transistor and diode in the middle, but the pair of transistors over on the right along with the rest of the components looks to me like a "flip-flop", which you see a whole lot in the switching section of commercially made non-true bypass pedals. On the first pulse it gets, it flips, and on the second pulse, it flops, giving you a square wave that's half the frequency of the input.
That is a good question. Read this thread through:
http://www.diystompboxes.com/smfforum/index.php?topic=28884.20
Somewhere in the middle of that thread Mr. Davisson explains the details of the "frequency-to-voltage" thingy. Although I am not quite happy with that answer, I need to think about it a bit more :)
The "mystery" part is Q3. Q1-Q2 are a preamp stage to get the signal up to threshold to provide triggering of the flip-flop over a long enough duration that we hear it as an octave down, rather than intermittent sputtering. I'm guessing that the diode both provides a defined threshold and also makes the AC audio signal a fluctuating DC signal that keeps retriggering the flip-flop.
Now, I *know* that's not gonna be enough for Paul and RG, but it's a step in the appropriate direction.
My take FWIW
There is no F to V connverter. The diode/transistor combi is a signal controlled switch for the power to the flip-flop. It switches on and off every half wave of the input. The caps between the flip flop TR's remember charge well enough so the state toggles every time the power comes up (on every half wave). So the flip-flop completes one cycle for every 2 input cycles.
Does that work?
Ok, I've modded the scheme (mostly guided by intution and trial and error :) ) and have found out that that transistor&didode stuff is unnecessary, at least for what I want and that is pure octave down (no dry/wet cause I don't need it).
My attempts to modify shocktave was because I wasn't satisfied with tracking (probably I got used too much to OC2 which is flawless IMO). And in my crazy attempts I decided to feed the "flip-flop" with only positive part of the signal (for what I've seen on the simulation, that transistor&diode thing does not do anything significant except deform signal a bit). So I set up unbiased common-collector amp (is it common-colletor?doesn't matter really) before the octave generator. And it kicks ass! (am i allowed to say that? :D )
Now tracking is really great for me, I get like 5 seconds of pure sustain (no flutter or anything) when before it was like 2 seconds. I just can't believe how good it is. And octaved sound is much more clear and has changed in character somewhat (though really synthy and not for anybody - but I play kind of synth-bass-guitar and it's great for me).
This is the modded scheme.
(It is a bit offtopic to put it here but here it is so it doesn't just stay on my PC forever. Someone may benefit from it. Don't make fun of oversized capacitors. :D I play bass guitar and like to keep the bass frequencies intact in all circumstances. I dimension my filters for 5 Hz cuttof at most. If I'm unsure how to determine size of capacitor properly, I just put really big capacitor - like 47uF - to be certain that no low end degradation happens. Not very intellectual, I know :) )(http://s5.postimage.org/jptpbs90z/shocktave_mod_schematic_online.jpg) (http://postimage.org/image/jptpbs90z/)
Obviously I've put some opamp amp/buffer istead of original preamp because I always find it more simple than those "manually-made" transistor stages where you need to tweak resistors and stuff. That 100k POT on opamp1 is GAIN. Another 100k POT (or better, trimpot) connected to collector of T1 should be adjusted to avoid "self oscillation" (or something similar).
Now back to the topic. It seems to me that I've demostrated that that transistor&diode stuff doesn't do anything critical because octaving still happens without it (or did I just do the same thing but in another way with my "unbiased amp" which lets through only positive part of the signal? )
I'm grateful for the replies :) I'll read the mentioned forum topic thorughly and come back with questions if any happen to come.
EDIT:
Quote from: jatalahd on December 25, 2014, 04:02:50 AM
Somewhere in the middle of that thread Mr. Davisson explains the details of the "frequency-to-voltage" thingy. Although I am not quite happy with that answer, I need to think about it a bit more :)
It seems to me that when he talks about frequency-to-voltage, he actually means that that transistor&diode thing only lets out pulses which are in accord with the frequency of the input. This is a kind of uncoventional use of the term f-t-v. Really what f-t-v does (in other applications, and by definition I guess) is that converts frequency to DC voltage level, so lower frequency --> lower voltage; higher frequency --> higher voltage, which is not what is happening here. I lwasted a lot of time trying to see that happening because I understood the term differently.
I also am not happy with the answer. To me it seems impossible that both diode and base-emitter of (NPN) transistor can be forward biased at the same time because that very diode sets emitter at higher potential than base, but whatever... :D
Quote from: anotherjim on December 26, 2014, 06:18:38 AM
There is no F to V connverter. The diode/transistor combi is a signal controlled switch for the power to the flip-flop. It switches on and off every half wave of the input.
Yes. :) As I've just said, the mechanism as explained seems contradictory to me, but I accept that's what's happening.
Quote from: anotherjim on December 26, 2014, 06:18:38 AM
The caps between the flip flop TR's remember charge well enough so the state toggles every time the power comes up (on every half wave). So the flip-flop completes one cycle for every 2 input cycles.
That seems true to me. Now I realize that I don't understand well enough how flip-flops work. There are some "bugs" in my understanding of it. I need to read a bit on the subject and understand it step by step (you know, like, C1 charges, gets to 0.7V, base-emitter is forward biased... and that kind of stuff).
------------------
Okay I've also done some "oscilloscoping" of my modded scheme (NOT ORIGINAL) in the simulator. It may be of interest to someone.Input is triangle wave, blue wave at the top is what is fed to flip-flop, and the lower 2 waves are the outputs from both transistors.
(http://s5.postimg.org/8vrf178vr/2014_12_26_15_15_10_Oscilloscope_XSC3.png)
So yes, now I fully get that story about "completeing it's cycle" after it is triggered 2 times. It is because output is taken from only one transistor, and they alternate in being on. Yes, it is a trivial matter but for some reason I did not realize it. Thanks anotherjim! But however, I still need to examine the exact mechanism by which it occurs.
Just for laughs, try disconnecting the feed to the flip-flop from the signal and connect it to the + supply. I suspect it will then self-oscillate.
Eccles-Jordan Astable Multivibrator is the term for this kind of flip-flop. The Bistable variant is used for electronic bypass switching a-la Boss.
Yes it will and I was puzzled when I first heard it (I did it my mistake). That's when I first realized that there is some kind of oscillator in this octaver pedal :D.. And I thoght that this input signal modulates the volume of that oscillator.. as I wrote in the first post. I had a whole lot of weird ideas about how it works :D It seemed like a very mysterious pedal.
In my modded schematic, you need to adjust the 100k trimpot on the collector to prevent that from happening on the release of notes.
EDIT:
Thanks for the name of the flip-flop (Eccles-Jordan Astable Multivibrator) ! :) I had trouble understanding it because all fliip-flops that I've seen are different somehow, like, the connections are a bit different and it was confusing me.