How does the Green Ringer (and other Octave up pedals) achieve the 2nd harmonic?

[Juicy_scooby]

Hi guys

So I know there is a ton of info on here about the green ringer, and I've learned a ton already about the pedal, how it sounds, how to troubleshoot and mod.

BUT probably because I am not very circuit savvy I still haven't found a solid explanation of how the circuit produces the octave up, or more specifically:

How do the two diodes in the circuit cause full wave rectifying?

I've looked up full wave rectifiers on wiki, and other explanatory sites, and generally I understand the principle (converting the negative peaks of an AC signal to positive ones to create a DC current with some bumps, probably smoothed out with a cap). I understand center tapped rectifiers can use as little as 2 diodes, but they require a transformer something I've never seen in a GR schematic. Nor have I seen the other popular method the Bridge Rectifier with 4 diodes employed in any green ringer schematics.

I'm guessing I'm just either mis reading the schematic (which I honestly don't understand in it's entirety anyway) or misinterpreting the way this pedal adds an octave up to your signal.

Below is the schematic I found in another thread here, and I've put a box around the section with the diodes in question

Also very possible that I'm looking in the complete wrong spot for the octave function by assuming the other transistors are just gain stages....so if anyone can lend further insight into what each group of components hopes to achieve themselves that would be immensely helpful.

[Rob Strand]

When you rectify a signal it creates a second harmonic.

Draw a sine-wave then draw a rectified version of it.  The rectified version goes up and down twice as often so you have a frequency component which is double that of the input sine-wave.

Extra baggage from the rectification process is there is a DC offset which is removed using a high-pass filter; ie. connecting a series cap some where.

The way the Greenringer does the full-wave rectification is a follows.

The first step is to get an inverted signal:
- The output at the emitter equals that on the base. 
  A positive signal on the base gives a positive signal on the emitter.
- The output at the collection is inverted from the emitter.  It is roughly equal in magnitude .
  A positive signal on the base gives a negative signal on the emitter collector.

The diodes are both arranged to pass the positive signal:
Signal on base    +ve: 
   Signal on emitter +ve   ---> Top diode passes the signal
   Signal on collector -ve --->  Bottom diode blocks the signal
   Output = signal from top diode (+ve)
Signal on base    -ve: 
   Signal on emitter -ve   ---> Top diode blocks the signal
   Signal on collector +ve --->  Bottom diode passes the signal
   Output = signal from bottom diode (+ve)

So the output at the diodes produces positive outputs and rectifies the signal regardless if the input at the base is positive and negative.

[Mark Hammer]

When I used to teach, I would often loan students a second textbook that covered the same material as the course textbook, but explained things a little differently.  Having access to two explanations often helped to make things clearer to them.  In that spirit...

The majority of analog octave-up pedals have what is called a "phase splitter" at their core.  These are easily recognizable in the schematic by having same-value resistors between collector and V+, and emitter and ground.  Doing so provides two equal-but-opposite-phase outputs: one at the collector and the other at the emitter.

Forcing each of these "copies" of the signal to pass through a diode cuts off half the signal*** in each.  But since the two copies are opposite to each other, you end up with two positive peaks where there used to be one.

Of course, rectifying a signal like that, and getting to hear an octave up can be two different things.  What yields audible octaving is the matching of the two complementary phase-splitter outputs.  Whatever passive components play a part in "shaving" and shaping the two complementary outputs, before they are combined, should be reasonably matched.

It also helps to remove as much harmonic content from the input to the circuit so that the doubled fundamental sticks out...A LOT.  Otherwise, there is too much stuff getting doubled and you simply can't hear the octave.

There, I hope this supplements Rob's excellent explanation in a useful way.


***Realistically, it cuts off a bit more than half, since the diode will not only prevent the "unwanted" half-cycle from passing, but will also block that portion of the wanted half-cycle that is less than the diode's forward voltage.  That's why I like to use Schottky diodes for octave generation.  If I fed a silicon diode with a +/-2V signal, what comes out would be +1.4V (2.6V blocked because of the diode type and orientation.)  Using a Schottky would give me about +1.8V under the same conditions.

[reddesert]

The majority of analog octave-up pedals have what is called a "phase splitter" at their core.  These are easily recognizable in the schematic by having same-value resistors between collector and V+, and emitter and ground.  Doing so provides two equal-but-opposite-phase outputs: one at the collector and the other at the emitter.

Just to add to these nice explanations, this is analogous to what the center-tapped transformer in a rectifier provides you: independent access to the positive-going and negative-going halves of the signal.

[Juicy_scooby]

Wow thanks for the replies everyone! This forum is great holy cow

Excellent explanations; I understand it much more clearly.

Two quick questions I have are: I see how the diodes play a passive and role in the audible part of the octave, which means the transistor in conjunction with the two equal resistors, outputs a phase shifted form of the signal, 180deg.

Of course, rectifying a signal like that, and getting to hear an octave up can be two different things.  What yields audible octaving is the matching of the two complementary phase-splitter outputs.  Whatever passive components play a part in "shaving" and shaping the two complementary outputs, before they are combined, should be reasonably matched.

How does a transistor do this? I understand PNP vs NPN, Common collector amps kind of, and the how the charge carriers flow in it, but why is this small switch so vital to audio signals in practice?

The first step is to get an inverted signal:
- The output at the emitter equals that on the base. 
  A positive signal on the base gives a positive signal on the emitter.
- The output at the collection is inverted from the emitter.  It is roughly equal in magnitude .
  A positive signal on the base gives a negative signal on the emitter collector.

Like in this case, are the collector and emitter's outputs always inverted in a transistor, and does that have to be addressed after one in a circuit in some way?



Second one is actually quick: are the polarized capacitors vital to biasing the diodes, or preparing the two waves from entering?


Seriously though thank you guys so much for your help I appreciate it !!

[antonis]

Of course, rectifying a signal like that, and getting to hear an octave up can be two different things.  What yields audible octaving is the matching of the two complementary phase-splitter outputs.  Whatever passive components play a part in "shaving" and shaping the two complementary outputs, before they are combined, should be reasonably matched.

How does a transistor do this? I understand PNP vs NPN, Common collector amps kind of, and the how the charge carriers flow in it, but why is this small switch so vital to audio signals in practice?

Although I'm BJT discrete configurations lover, I have to admit sometimes Op-amps are more convient for practical demos..



Consider the above two opamps as a single BJT Collector and Emitter respectively..
First opamp is inverting (Collector) and second op-amp is also inverting, resulting in signal re-inversion (360^o equal to 0^o phase shift..)(Emitter)
First op-amp has high gain so  you can consider it's output clipped (almost at 4.5V) - second op-amp gain is simply unity..
A 1V input possitive signal leaves first op-amp as a -4.5V one, going to left diode cathode and second op-amp input..
(Both diode cathodes are DC biased at about 2V..)
Left diode cathode is now set on a more negative voltage than previously set so there is more current flowing through it hence more voltage drop across the "common" 100k resistor..
Considering 100k resistor as an equivalent Collector load, we have a classic CE amp behaviour..
Simultaneously, that -4.5V signal is also going into second op-amp input and comes out re-inverted, as +4.5V, and goes to right diode cathode where sets its voltage at a more possitive level than previously set, resulting in less current flowing throuh it, hence less current through 100k resistor..
(right diode is  actually switched off compared with left diode..)

Exactly the opposite happens for a 1V negative signal.. 
(due to phase reversal, left diode is switched - off where right diode is "activated".. - 100k resistor is now considered as Emitter load..)

[antonis]

The first step is to get an inverted signal:
- The output at the emitter equals that on the base. 
  A positive signal on the base gives a positive signal on the emitter.
- The output at the collection is inverted from the emitter.  It is roughly equal in magnitude .
  A positive signal on the base gives a negative signal on the emitter collector.

Like in this case, are the collector and emitter's outputs always inverted in a transistor, and does that have to be addressed after one in a circuit in some way?r]

Rob corrected his "fault"..
<The output at the collection is inverted from the emitter.  It is roughly equal in magnitude .
A positive signal on the base gives a negative signal on the emitter collector.>

Emitter ALWAYS follows Base..!! 

<Second one is actually quick: are the polarized capacitors vital to biasing the diodes, or preparing the two waves from entering?>

Not at all...
They actually "isolate" signal from DC..

They are polarized only due to their desirabe capacity (size) - non-polarized caps of the same capacity cost a lot more..

[Mark Hammer]

To tie this all together a bit more, while the majority of well-known octave-up boxes will use a phase-splitter made from a single transistor having equal-value emitter and collector resistors, that's not the only way to provide two opposite-phase versions of the input signal that can then be trimmed to produce two peaks where there used to be only one.

In Gus's circuit, the second op-amp inverts what comes out of the first one, and is set up as unity-gain (equal value input and feedback resistance), yielding the functional equivalent of what a phase-splitter does.  Similarly, the original Tychobrahe Octavia uses a center-tapped transformer to provide two opposite-phase versions of the signal.  In both of these instances, diodes are once again used to selectively remove one half cycle of each of the two copies of the signal.

[antonis]

But we can't prevent full-wave rectification cross-over distortion, due to diode forward voltage drop, unless we "include" them in some kind of closed feedback loop, can we Mark..?? 

[Mark Hammer]

I suppose one way to get past that is to use something like Gus's circuit and add some bias to compensate for whatever the diode steals.  Me, I just use Schottky diodes so that I lose as little as possible.

I might point out that since it takes a brief moment for a picked string to reach any critical forward-voltage threshold that allows it to pass, and that one can also lose some of the "tail" as the string decays, that one is also chopping off part of the onset and duration of the rectified signal.  Not a whole lot, mind you, but sometimes enough to be noticeable.

[iainpunk]

But we can't prevent full-wave rectification cross-over distortion, due to diode forward voltage drop, unless we "include" them in some kind of closed feedback loop, can we Mark..?? 

Yes we can prevent this from happening! There is a thing called a precision diode, which uses an opamp to overcome the forward bias of the diodes. Pic related.



There are more options, but this one works really good and it has a low parts count. You can just put a phase splitter in front and a summator behind 2 of them.

And then there is "switching octave up generation"... Which uses a comparator, and a polarity reverser using a jfet as switch. It works like a charm, but its harder to understand how it works.



A common problem analog octave up has is the volume (amplitude) loss of about 50%. This is quite easily remedied with an opamp or transistor amplifier.

[iainpunk]

And then there is "switching octave up generation"... Which uses a comparator, and a polarity reverser using a jfet as switch. It works like a charm, but its harder to understand how it works.

I will draw up a schematic after dinner.

[iainpunk]

And then there is "switching octave up generation"... Which uses a comparator, and a polarity reverser using a jfet as switch. It works like a charm, but its harder to understand how it works.

I will draw up a schematic after dinner.

[PRR]

Precision rectifier is good for precision rectification.

For "octave up" it gives *sharp* splices between the half-waves. While they are double frequency, the harmonics extend toward infinite frequency, and are annoying.

"Sloppy" rectifiers will usually sound better.

If you go down to the Maths-Heads bar and grill you can find some different approaches. But the simple biased diode approximates most of them without any brain-pain. Level control is the real issue.

[iainpunk]

Precision rectifier is good for precision rectification.

For "octave up" it gives *sharp* splices between the half-waves. While they are double frequency, the harmonics extend toward infinite frequency, and are annoying.

"Sloppy" rectifiers will usually sound better.

If you go down to the Maths-Heads bar and grill you can find some different approaches. But the simple biased diode approximates most of them without any brain-pain. Level control is the real issue.

Filtering and clipping are ways to make it sound good, something id also recommend for sloppy rectifier