I was checking out this schematic of the Zvex Octane 2:
(http://musikding.rocks/wbb/index.php/Attachment/354242-ZVEX-Octane2-jpg/?thumbnail=1)
First of all, it looks a heckuva lot like a Boss FZ-2. But on top of that, the question I have is: where is that upper octave coming from? The diodes are after the differential amplifier and set up for good old fashioned hard clipping. How is that FZ-2 section that feeds the LTP putting out the upper harmonics without diodes for rectification?
Take it with a pinch cause what I know about octave circuits could be written on the back of a postage stamp in capitals.
Maybe it's using the bjts internal diodes as part of the rectifier.
I had to look up the rest of the schematic.
Quote from: Kipper4 on May 01, 2017, 04:38:55 PM
I had to look up the rest of the schematic.
Yeah, I posted that little version because I'm not somewhere I can resize and repost the full one. Here's a large version of the full schematic.
http://musikding.rocks/wbb/index.php/Attachment/354242-ZVEX-Octane2-jpg/
Haven't built it, but looks like T2 is a phase splitter, sending opposite polarity signals into each side of the long tailed pair, which boost and meet at C9. The octave probably comes from an imperfect cancellation, so you get the in-between remains of the fundamental fighting with itself in the T3 vs T4 stage. I've seen a discussion somewhere I can't remember, probably RG Keen or Mark Hammer.
I agree, off to the left there is a cathodyne/phase splitter T2.
Beware of looking at incomplete circuit snippets.
That pair don't work as a diff' amp, they are are fed anti-phase signals and put 2 half-waves out per cycle. when T3 turns on, T4 is turning off and vice-versa. A type of full-wave rectifier. Clipping diodes remove the resultant spikey content somewhat - because they sound horrid.
Other schemes don't use a phase splitter, but use a PNP and NPN pair.
Those 2 trannies are not a diff amp. They are configured to sum the two signals going to each of their respective bases.
Now I remember -- The JFET Doubler - A Simple Screaming Octave (http://geofex.com/Article_Folders/jfetdoub/jfetdoub.htm) -- from Geofex
Similar in concept as far as going from a phase splitter into the thing that looks like a LTP but without the clipping diodes.
Interesting.
> into the thing that looks like a LTP
Except the two outputs (plates, collectors, drains) are connected *together*.
It is push-push, not push-pull.
As said, it is a full-wave rectifier. Why does your FW DC power hum 100Hz on 50Hz wall-juice? Both half-waves are set on the same side.
Hey guys, the circuit's drawn in a certain way but don't make that fool you it's a differential stage/LTP.
It's a shared collector mixer. Basically just two individual common emitter amps that have their output signals mixed exploiting source impedance of the collector instead of external resistors. Both amplifiers in the circuit thus work in phase. The circuit works rather decently as long as none of the amplifiers are overdriven. Emitters can share biasing circuit too, as long as it is sufficiently bypassed in respect of AC signals. Naturally.
Now we get to the important part, which is biasing of the amps: They can't be biased class-A because opposite-phase signals from the preceding phase inverter would just cancel each other. In fact, the amps are likely biased closer to class-C so that each amp only amplifies one half wave of its input signal. We have two input signals that are out of phase. We get "rectified" DC output from the shared collector mixer.
All in all, this is actually pretty generic setup for a "folding" -style octaver circuit. Picture a sine wave: One of the half waves "folds up" when the wave is rectified, frequency goes up x2, so a harmonic frequency one octave above the fundamental has been created.
As far as I know, the circuit idea traces back to ancient vacuum tube computers. It's a circuit, which more or less predictably performs a mathematical function.
If non-linear characteristics of transistors are exploited this circuit can also be tweaked to produce rather "sine wave-ish" output instead of that "spikey" rectified sine wave. Assuming perfect waveform inputs, the conversion process is therefore rather "smooth" and does not introduce an abundance of additional harmonics to the signal. I doubt it works that well with practical "complex" waveform inputs though.
This method of producing octave tones goes back many years. A classic example is the Univox Superfuzz:
http://i76.photobucket.com/albums/j6/IvIark_2006/Layouts/Schematics/UnivoxSuperfuzz.gif
Best regards, Jack
Is the advantage of this configuration the lack of gating effect produced by diodes?
Jack is clearly correct in asserting that there are few differenes between the Octane and Superfuzz. The major difference is a simpler front end on the Octane (although the various Superfuzz derivatives varied quite a bit in their respective front ends) and the use of a continuously variable midscoop scontrol, instead of the 2-position tone switch. We don't know what ZVex and UNivox were looking for in selecting transistors, so that might also be different.
Note that the diode pair to ground is not required for distortion. It clearly adds distortion, but you can lift those diodes and it will still be plenty fuzzy. What the diodes DO accomplish, though, is clamping the signal level and producing some compression. Keep in mind that any plucked string starts out with lots of harmonic content and then soon settles down to mostly fundamental. As the hash created by all the doubled harmonic content gets cleared, the doubled fundamental becomes more obvious, in a way that some would describe as "blooming". If the octave didn't show up until after the note had quieted down, you wouldn't notice it quite so much. BUt since the diode pair keeps the level constant, the octave seems to appear as you sustain the string. Remember that "best" octaving is usually achieved with the neck pickup and the guitar tone rolled back. The down side of the diodes is that they do add harmonic content at the same time as they work to keep level constant.
So, reasoning that the goal is really constant volume, rather than distortion, I threw a 47nf cap in parallel with the diodes in this circuit (essentially a stripped down Superfuzz), and the octave became more audible, with less extraneous harmonic content to cover it up. Not quite as overt as a nicely tweaked Tone Machine, but audibly better than the stock circuit, and just as nice with the guitar tone up as down. I had modded the circuit to have a variable resistance between the clipping diode pair and ground, such that dynamics could be varied. When the clipping/compression is reduced via that pot, the octave is not quite as audible, lending support to the view that the diodes are really toallow the octave to emerge more audibly, rather than to impose distortion.
(http://experimentalistsanonymous.com/diy/Schematics/OOP%20Japanese%20Electronics%20Book/old-school-fuzz.gif)
(Actually, looking at the schematic more closely, the one shown is just a little different at the input stage than what I added the cap to. BUt otherwise, nearly identical.)