OC-2 fundamental extractor

[jdub]

Hey all-
Some time ago a forumite had posted a scheme for just the fundamental extractor portion of the Boss OC-2 (http://www.diystompboxes.com/smfforum/index.php?topic=60977.0), but it appears the links in the post are now dead.  Anybody have this saved someplace?  I have the scheme for the OC-2 (chopped and stock), but I'm not sure where the fund-ex circuit begins and ends, and I'd like to experiment a bit...

Thanks!

[head_spaz]

LINK to zip file ... 526KB

[jdub]

Many thanks, Spaz 

[head_spaz]

YW.

[Andre]

I moved to another internet provider that does not provide any homepage space, so I moved some pics to Photobucket.
Here are my Extractor pics again:

[potul]

Interesting circuit...

I have one question regarding the output of it. I understand it is supposed to generate a square signal, but will it retain any dynamics, or will it be fully compressed? (so, will the amplitude of the square be constant?)

Regards

potul.

[Processaurus]

but will it retain any dynamics, or will it be fully compressed? (so, will the amplitude of the square be constant?)

Fully compressed, ie it is a logic signal that will swing between the rails if it's going to go.  If you need to impose dynamics, it's more circuitry, see the twin transistors at the output of the Bluebox, or better, the output section of the CA Rocktave divider, which I believe uses an analog switch (like the 4066) or fet, driven by the square wave logic signal, to cut up an envelope signal into a square wave that follows the input instrument's dynamics.

A note to people integrating this into their own synth circuits, this doesn't have any provision to gate the end of the note, which might cause garbly gook at the end of each note from the extractor.  Boss didn't worry about the sputtery end, because they were superimposing the instrument signal on the logic, so any confusion at the end was low amplitude.

[potul]

Good, thanks for the clarification. I will investigate the mentioned circuits and experiment a bit.
My idea is to test it in front of a guitar2midi circuit and see if it can improve tracking... but I want it to have some dynamics in order to generate different MIDI velocities.

Thanks,

Potul

[potul]

Hi again.... I'm trying to understand how this fundamental extractor works, maybe someone with more knowledge than me can give me some hints?

This is what I understand from the schematic.

-There seems to be some amplification and filtering of the signal during the first 2 stages.
-Signal is then split in two and rectified (positive and negative portion separately).
-Afterwards, original signal is compared with each of the rectified signals. I understand that this is used to let us know if we are in the positive or negative portion of the cycle, and generate a logic signal accordingly.
-These 2 logic signals (positive portion of the cycle and negative portion of the cycle indicators) are used to change the state of a flipflop via the reset and set pins.

end result: We get a square signal going from 0 to 1 logic values, that follows the frequency of the input.

So far, it looks like a fancy way to square a signal.

My question is, how is the "fundamental" extracted? How is the system eliminating undesired harmonics? Is it only a question of the initial filtering? Or is there something else I miss?

Regards,

Potul

[Roy]

One circuit finds the very highest part of the waveform and sets the latch. The other mirrored circuit finds the very lowest point of the waveform and clears the latch. There can only be one highest point and one lowest point per cycle, even if the harmonics are strong. Even if (say) the second harmonic were much louder than the fundamental, the very top and the very bottom would still pick out that fundamental.

[potul]

Thanks for the answer

Oh, I see.

And how's the highest point detected? Is it by comparing the signal with the rectified one? when the signal is higher than the rectified one , is it supposed to be a max?

Sorry for my maybe obvious questions.. but I'm interested in understanding how this fundamental detection is done, and see if it can be adapted to one of my projects.

Regards.

[jdub]

Andre- Thanks for reposting!

A note to people integrating this into their own synth circuits, this doesn't have any provision to gate the end of the note, which might cause garbly gook at the end of each note from the extractor.

Thanks fer the tip, P-saurus. 

[thehallofshields]

I know this is a really old bump. Please don't kill me.

I just wanted to thank Andre for sharing this.

... and also ask why the Middle-Cap (C5) in the LPF is tied to IC1b's Feedback Loop (or Output) instead of ground.

[teemuk]

No problem.

http://en.wikipedia.org/wiki/Active_filter
http://en.wikipedia.org/wiki/Sallen%E2%80%93Key_topology

[thehallofshields]

I saw that. I just couldn't make sense of it.



"The buffer bootstraps the "bottom" of the Z3, component to the output of the filter, which will improve upon the simple two divider case. This interpretation is the reason why Sallen–Key filters are often drawn with the operational amplifier's non-inverting input below the inverting input, thus emphasizing the similarity between the output and ground."

I guess Bootstrapping is a little above my belt-level. The only thing that would make sense to me is that it drops the impedance back down before the following resistor, but that doesn't seem to be what the info implies.

[Eddododo]

look more closely-er...

in your head (or on paper, why not) circle the node that is z3 , the op amp 'output' and the '-' input. Now look back at the schematic... though it is organized differently, the 2n2 cap to the negative feedback is the same arrangement.  In this arrangement, the 2n2 cap to the inverting  input is to get more bang for your filter stage's buck. Look up sallen-keye filters ( google 'westhost active filters' )


The inverting input can be thought of similarly to ground in that  in is very low impedance (correct?), and that in a similar-but-not-equivalent manner, shunting filter components to either has similar topological behaviors (again, look at a simple rc high pass and the sallen-key high pass filters... see is as essentially TWO cascaded high pass filters). This doubling of the filter action has the effect of making the rolloff of the filter 'steeper'...

[thehallofshields]

Lol. Okay. I'm thinking real hard.

So instead of just filtering the those Frequencies to ground it throws them at the output 180* out of Phase for more effect?

[thehallofshields]

I've got the Filter perfed and am trying it with an Octave Up. It really is perfect for dropping everything over 1khz off a cliff.

I also did Frequency Analysis on the beginner Circuit-Sim App Everycircuit, with the Cap on a switch from Ground to the Inverting Pin. HUGE difference. I wish I could screenshot. Connected to IC1b-In it  gets way steeper and gets a little peak right before the Corner Frequency of 1.5khz. The designers definitely picked those values to get exactly 1.5khz.

IDK if anyone cares, but I'm splitting this detector circuit and making into a couple of small breadboard modules.

[anotherjim]

Do a screenshot? If you have Windows (can't help with anything else), press print screen key -  that copies it to the clipboard. Open an image editor (ms paint will do), paste to area, select part you want and copy to new. Save as .jpg - done.

I've tried the s-r flip-flop approach to fundamental tracking in the past, I found that it still needs vicious low pass filtering first of all to be reliable - which Boss have done. Guitars, can reliably produce harmonics that will look like 2 pos and 2 neg peaks within the fundamental to the detector. I even tried extra logic to check for zero crossings too in the hope that there must be a combination that cannot fail to get the fundamental - guitar can fool that too. Next time (whenever!) I'll try a small MCU to get fundamental by the averaging approach, but especially with bass, that's going to cause some delay.

[thehallofshields]

I even tried extra logic to check for zero crossings too in the hope that there must be a combination that cannot fail to get the fundamental - guitar can fool that too. Next time (whenever!) I'll try a small MCU to get fundamental by the averaging approach, but especially with bass, that's going to cause some delay.

Thanks for the tip. I'm pretty crafty with Windows, it's Android and touch-screen tech that is still a little foreign to me.

Could you give an example of what a Comparator with Zero Crossing Detection looks like?
I've read some posts here with a couple people saying it dramatically improved the circuit and prevented the Comparator from tripping over the Second-Harmonic. Is that what the OC-2's Dual-Rectifier-Comparator section is?

I tried to search Zero Crossing Detection, but I just found pictures of Opamps with Clamp-Diodes between +In and -In.