Harmonic Percolator Question

[Paul Marossy]

What element of the circuit is responsible for the octave up sound?

[drolo]

Would interrest me too.
Could it be the arrangement of the 2 transistors in series, and some kind of phase splitter effet?
Tried to understand but my brains won't compute that ...

i'm sure someone will have a better explanation

[tca]

I've done some simulations but nothing conclusive, yet!

[Paul Marossy]

Could it be the arrangement of the 2 transistors in series, and some kind of phase splitter effet?

They're not in series as far as I can tell. The collector of Q1 is feeding the base of Q2. It's your basic cascaded transistor circuit, but Q1 is a PNP and Q2 is an NPN.

[effection]

Yeah, it's strange, and even though I have some knowhow, I can't seem to find what it is that makes that HP so special.

What I have learned is that there's something with the commons that is doing some sort of cancellation or something. This is why replacing that 47uF cap with different values messes with it, as digit (I'm new, I think his username is something along those lines) experimented with. Also, I've read that potentiometers between the emitters makes for some really ugly and gnarly octaves, but I haven't tested this personally.

[Paul Marossy]

Yeah, it's strange, and even though I have some knowhow, I can't seem to find what it is that makes that HP so special.

What I have learned is that there's something with the commons that is doing some sort of cancellation or something. This is why replacing that 47uF cap with different values messes with it, as digit (I'm new, I think his username is something along those lines) experimented with.

Well, one thing that is interesting about the circuit is that both emitters are connected to eachother, and then to ground thru the same capacitor. I bet you could tweak things more by giving each emitter its own capacitor to ground.

When it gets really ugly sounding, it sounds kinda like total crossover distortion to me.

EDIT: And there is also the bootstrapping, which I am not sure how that factors in exactly.

[~arph]

I just breadboarded the jerkulator and that started doing the octave as soon as I added the diodes al percolator. I used 1N34's  sounded best with a 2n3906 and a bc107. I added a 470pf across the bc107's base and collector to tame some hiss

[Paul Marossy]

I just breadboarded the jerkulator and that started doing the octave as soon as I added the diodes al percolator. I used 1N34's  sounded best with a 2n3906 and a bc107. I added a 470pf across the bc107's base and collector to tame some hiss

I was wondering if those diodes had something to do with it. That's actually what I was thinking at first, but then the more I thought about, the less sense it made. I wonder why that would add an octave up sound. It must be acting like a rectifier when you push more than a certain amount of current through them? I wonder if it would do this with 1N4148s.

EDIT: Aha! From www.geofex.com - "Half wave rectification represents the logical conclusion of asymmetrical clipping. One half of the waveform is flat, the other half is unchanged. This produces a prominent second harmonic, heard as an octave."

So I guess that answers my question.

[~arph]

So that 4k7 between the diodes plays a role too, making it asymmetrical right?

[Paul Marossy]

So that 4k7 between the diodes plays a role too, making it asymmetrical right?

Looks that way to me.

[effection]

I should also note that it's NOT two cascading gain stages. The big thing that notes that is that the PNP transistor is going to ground both through the emitter AND the collector. The topology looks more like an AB push-pull amp than two cascading amps.

[Paul Marossy]

I should also note that it's NOT two cascading gain stages. The big thing that notes that is that the PNP transistor is going to ground both through the emitter AND the collector. The topology looks more like an AB push-pull amp than two cascading amps.

Ah, that's a good point. I guess if it were two NPN transistors then it would be cascaded gain stages then? I don't have a lot of experience with these sorts of transistor circuits.

Anyway, using the AB push-pull parallel, that seems to support my hunch that when the circuit is pushed really hard I hear what sounds like some pretty ugly crossover distortion. I guess I need to look at it on the scope to see exactly what's going on.

EDIT: So the PNP transistor is more akin to a "split-load" phase inverter then?

[slacker]

I think it is two cascaded stages. The point where the two emitters connect is basically a virtual ground for AC and is somwhere near the middle of the supply for DC. So the PNP has its emmiter at ground withs its collecter resistor connected to about -4.5 volts or so, it's like a positve ground circuit. The NPN is the same, its emmiter is at ground with its collector resistor connected to 4.5 volts. That's what it looks like to me anyway.

[tca]

Here are my simulations.



It seems that the idea of a common virtual ground between the PNP and the NPN is correct. No shadow of the octave (have to run with other devices). The straight line on all the plots is the voltage at the emitter (NPN) and collector  (PNP).

[Paul Marossy]

I think it is two cascaded stages. The point where the two emitters connect is basically a virtual ground for AC and is somwhere near the middle of the supply for DC. So the PNP has its emmiter at ground withs its collecter resistor connected to about -4.5 volts or so, it's like a positve ground circuit. The NPN is the same, its emmiter is at ground with its collector resistor connected to 4.5 volts. That's what it looks like to me anyway.

That also sounds reasonable.

The definition of a push-pull circuit is "a type of electronic circuit that can drive either a positive or a negative current into a load."

If you look at an emitter follower two transistor push-pull circuit, the output is at the point where the emitters are connected together and there is some kind of feedback between the input and output of the circuit. The percolator doesn't really follow that schematically, but it seems to work similarly in function. Are there other ways to get crossover distortion besides in a push-pull type circuit?

[tca]

Are there other ways to get crossover distortion besides in a push-pull type circuit?

There is other configuration that could help understand this: a Lockhart frequency tripler.

[Tony Forestiere]

This is an interesting thread. Maybe a "Technology of..."?
I googled the frequency tripler and found this by our own Andrew (TTG): http://www.oocities.org/thetonegod/finish_line/finish_line.html
Hope the link works.

[tca]

This is an interesting thread. Maybe a "Technology of..."?
I googled the frequency tripler and found this by our own Andrew (TTG): http://www.oocities.org/thetonegod/finish_line/finish_line.html

Oh, I've had a similar question the other day? Some more examples there.

- LovePedal - Karl Fuzz
- Tone God - Finish Line
-  ZVEX - Machine

Cheers.

[tca]

If I do a Fourier analysis I get this:







1.5, 2 prominent harmonics. The same happens with out the diodes at the output but with small amplitude.

[slacker]

Nice simulations tca. I think there could be a bit of octave up in there. See those little pointy bits on the top of the wave form in the "with no diodes" image, you're starting to get two peaks per cycle of the input. Subtle though.