Jfet as rectified octave question

[Eddododo]

Can a jfet be used with simultaneous phase-opposite outputs like a bipolar or mosfet?

What is the name of this type of circuit, so i can find it :p

(Think green ringer if you're confused)

[Eddododo]

Woops just remembered the MuDoubler. Still wouldn't mind some vocabulary related to the transistor phase splitter... thing

[samhay]

Try looking for 'cathodyne'.
Haven't seen it done with a JFET, but don't see why it couldn't work.

Edit - looks like you are on the right track with the MuDoubler.

[merlinb]

Can a jfet be used with simultaneous phase-opposite outputs like a bipolar or mosfet?

Yes. A JFET is a type of transistor, after all. Transistors and tubes all function in the same basic way, and can be used for the same basic functions.

The phase splitter you're talking about is vaiously called a cathodyne, concertina, or split-load.

[R.G.]

Bipolars, MOSFETs and JFETs can all be used for the "full wave rectified" version of an octave. Bipolars and MOSFETs will do this better than JFETs, as the threshold between conduction and non-conduction is sharper. But then again, now that I think about it, not having a sharp transition might make a more listenable result.

"Superfuzz" style full wave rectifiers rely on the signals turning on one side, then the other of what would otherwise be a diffamp, and then letting both sides suck current through the same resistor, in effect switching the current drawn so it's always one-direction. It looks like a differential amplifier, but isn't really.

The Doublers work differently. They don't rely on one side or the other going on and off. They leave both sides of the "diffamp" conducting all the time, and producing a signal at the drain. The equal-but-opposite signals are added by the currents being funneled into the same resistor. In that resistor, the out-of-phase components cancel, including all odd-order distortion products. The even-order distortion products add, and if the cancellation is perfect, produce a mainly  octave-up signal as the second harmonic distortion is the biggest one. This produces a smoother result than full wave rectification, as it cancels many distortion products, while enhancing the smoother ones.

But it needs more work. Go for it!

[Eddododo]

So, R.G...

In the superfuzz-type circuit, the instrument signal itself is what is crossing the threshold to make the complimentary transistors conduct ?
looking at http://cabbages.orconhosting.net.nz/graphics/joes_rubbish/fx/superfuzz/superfuzz.png, Q4 and Q5 are biased slightly under conduction, and the signal 'opens' them?


also, when you referred to the 'knee' of conduction as it were, were you implying a significance with regards to the GreenRinger topography or for the superfuzz topography. Or the mu-amp..

Thanks for the help!

[R.G.]

In the superfuzz-type circuit, the instrument signal itself is what is crossing the threshold to make the complimentary transistors conduct ?
looking at http://cabbages.orconhosting.net.nz/graphics/joes_rubbish/fx/superfuzz/superfuzz.png, Q4 and Q5 are biased slightly under conduction, and the signal 'opens' them?

It's a bit more complicated. In the Superfuzz, what looks like a differential amplifier isn't really. What makes the difference is that cap across the emitter resistor. This holds the emitter voltage constant(ish) in the face of signal.

In a normal diffamp, that point is high impedance, ideally a current source. Signal in at one base actually drives current into the current source, which can't change current much, so the current becomes an input to the emitter of the other side, turning the other side off in concert. This is a smoothly linear operation. With the emitters held at a nearly fixed voltage versus the base bias sources, turning on one transistor (e.g. Q4) has no effect on the other transistor (e.g. Q5) through the emitters. The transistor being turned on acts like a common emitter amplifier, ignorant of the other side. They're fed complementary signals, so Q4 is turned on by the input signal while Q5 is driven off, and vice versa. This has the effect of causing the transistors to alternate being "on", not act like a differential amplifier. The base-emitter junctions act like they're diodes, biased right at the edge of conduction, so the forward voltage is not effectively subtracted from the signal being "rectified". Only  the transistor with a positive-side signal amplifies into the joined collectors. The negative-side signal turns the other side off. This has the effect of full wave rectifying the signal and an amplified version appearing at the collectors, pointy side up.

also, when you referred to the 'knee' of conduction as it were, were you implying a significance with regards to the GreenRinger topography or for the superfuzz topography. Or the mu-amp..

When I refer to conduction knees, I mean the transition region from mostly-non-conducting to mostly conducting. Diodes have conduction knees of about 0.1-0.2V big, right at the point where they start conducting a lot. I wasn't specifically referring to the Green Ringer or Superfuzz.

However, in a way, both of these are similar. The Green Ringer has two diodes biased with a tiny current forward so they're just barely conducting. The amplified signal is enough to push one "off" and the other "on". The tiny bias current means that the signal to push them from off to on doesn't have to be the 0.6V or so for a silicon diode, but can be much smaller. In the Green Ringer, this is what full wave rectifies the signal, and that signal is then amplified by a separate stage. In the Superfuzz, the base-emitters are set up this way, just at the edge of conduction, and the signal pushed one, then the other into conduction, but this conduction is amplified by transistor action right in the device that's being turned on. You find a similar setup in the Fox Tone Machine and Fender Blender - barely-biased diodes, one for each polarity.

The Mu Doubler is different. It does not rely on switching one on, one off. Rather it amplifies the signals out of phase, then uses the resulting amplified signals to cancel, leaving only the even-order distortion products. The other approaches are switching/rectifying in nature, the Doubler is picking out only the distortion products. It's more similar in concept to a multiplier than a rectifier.