Octavia circuit analysis

[ItsGiusto]

Hey all. I have another thread about debugging my Octavia, which is on hold until I get better diodes, but this question might be different enough that I thought I'd make a separate thread.

I was wondering if anyone could help me understand how the circuit works. For some reason having the signal split into two separate sections, each coming off of separate leads of the transformer, then going through separate diodes, cause there to be an octave effect: http://fuzzcentral.ssguitar.com/octavia/octaviaschematic.gif

Though it's not only the transformer that has this effect: in the Roger Mayer Octavia, there's no transformer at all, and the octave comes from combining the signal coming off of the emitter and collector of the third transistor, then sending those signals through separate diodes:
https://i.pinimg.com/736x/a9/35/df/a935df08ab5b3a2ad833da1318b7862c--pedal-circuit.jpg

So I guess I'm asking, why is it that splitting the signal and sending it through separate diodes cause this octave effect? Is the signal coming off of the separate ends of the transformer/transistor in phase or out of phase?

[ItsGiusto]

Another question: what role does the transformer play in the tycobrahe? I hear that the Roger Mayer sounds weak and difficult to get a sound out of. Why does a transformer play a big role in making the fuzz and octave more pronounced?

[Plexi]

I'm interested in the las question.

I was to start to build the RM (transformer-less) version: always thought it sounds better than the transformer version.

[thermionix]

Octavia schematic isn't loading for me.  I think the transformer is just to split the signal into two equal halves of opposite phase.  Same basic result as pulling signal from the emitter and collector of a single transistor on the Mayer and Green Ringer.  Add the diodes and it's like full-wave rectification, which doubles the frequency.  Like when you get 120Hz hum from an amp powered by 60Hz wall juice.

[ItsGiusto]

Octavia schematic isn't loading for me.

I fixed the link

[ItsGiusto]

I think the transformer is just to split the signal into two equal halves of opposite phase.  Same basic result as pulling signal from the emitter and collector of a single transistor on the Mayer and Green Ringer.  Add the diodes and it's like full-wave rectification, which doubles the frequency.  Like when you get 120Hz hum from an amp powered by 60Hz wall juice.

Each diode would only be half-wave rectification. In addition, each diode also has the effect of cutting off some portion of the wave on each half-rectified side, equal to the forward voltage (only voltages above the Vf would get through). That has to have a huge effect on the signal.

Are you saying, though, that adding each half-rectified portion of the signal from each separate diode-path would end up with a full-wave rectified output? That's interesting and I guess it makes sense, if in fact the bottom and top parts of the signal are phase-inverted. I guess I can see how that would lead to the octave effect.

Interestingly, if people were mostly after a full-wave rectified signal, why wouldn't they have just used a rectifier bridge in the circuit?
https://upload.wikimedia.org/wikipedia/commons/thumb/f/f5/Diode_bridge_alt_2.svg/557px-Diode_bridge_alt_2.svg.png
I forget some things from my circuits course, but I seem to remember them saying that the extra diodes in the bridge configuration are more efficient at rectification. I'm pretty sure that diode rectification existed in the late 60s.

[Plexi]

Octavia schematic isn't loading for me.  I think the transformer is just to split the signal into two equal halves of opposite phase.  Same basic result as pulling signal from the emitter and collector of a single transistor on the Mayer and Green Ringer.  Add the diodes and it's like full-wave rectification, which doubles the frequency.  Like when you get 120Hz hum from an amp powered by 60Hz wall juice.

So...the last transistor/stage on the RM is only for buffer.
I you remove one of those diodes, it doesn't octave, like in the transformer one?

[thermionix]

[thermionix]

A bridge rectifier might work, but would be twice as many diodes.

You would still need a transformer in the octavia circuit, but not a center tap.

[ElectricDruid]

The typical way of doing full wave rectification is to invert the half-wave signal and double its amplitude. Then when that is combined with the original signal, the result is the full wave waveform.

There are lots of ways to get the half wave rectified signal and do the inversion, and several of them have already been mentioned above.
A simple bridge rectifier isn't that good because of the forward voltage of the diodes. It'd work brilliantly if diodes conducted straightaway, but they don't - you need to get 0.6V across them before they do anything. That'll produce heavy crossover distortion if you just feed a signal straight to a bridge rectifier. Using germanium diodes helps (reduces the forward voltage to 0.3 or 0.4 maybe) but it doesn't solve the problem. A proper op-amp based precision rectifier circuit does a much better job, but doesn't have an of the interesting distortion that some of the other ways shave become famous for.

HTH,
Tom

[amptramp]

One trick that might be worth trying is to disconnect the centre tap from ground and take it to a slightly negative supply.  A pot connected from ground to a resistor that goes to the negative supply would allow you to forward bias the diodes so you could control the amount of crossover distortion you get.  A 10K pot and a 120K resistor would be good values to start with.  You would be able to go up to 0.56 volts at the tap.  You may want to add a capacitor from the tap to ground but it might not be necessary.

[anotherjim]

I don't know the spec of the original transformer (does anyone?). It could have a step up ratio? The hotter the signal, maybe the less "gatey" it was. That said, full wave rectification is never going to be pretty. It makes a very artificial waveform. If anyone dislikes the sound of the Octavia, they should try an op-amp based precision rectifier circuit - it's far, far nastier.

Given it's feeding a high impedance input (guitar amp), and note the high value volume pot at 500k, The diode leakage before full Vf is reached might do something audible? Masking the gating effect a bit? Something I think germanium diodes are (inadvertently) better at. Or is it at such low level, there is no worthwhile benefit from this and Schottky are just as good given similar Vf?

It may have been a factor that the original transformers bandwidth helped a little, smoothing the top end and maybe reducing the fundamental at the bottom end. Note the spike form of a rectified wave at the negative peaks. Therein lies the dragon.

[thermionix]

I don't know the spec of the original transformer (does anyone?). It could have a step up ratio?

Yeah I was wondering the same thing.  I've never seen specs.  Small Bear doesn't mention a ratio for their replacements.

[highwater]

The Smallbear replacement says "TM022-R" on the side, so probably a Xicon 42TM022-RC. Datasheet says turns ratio (primary:secondary) is 1.725:1. The Octavia drives the secondary, so that is indeed a step-up.

[ItsGiusto]

The Smallbear replacement says "TM022-R" on the side, so probably a Xicon 42TM022-RC. Datasheet says turns ratio (primary:secondary) is 1.725:1. The Octavia drives the secondary, so that is indeed a step-up.

I'm using the Smallbear transformer in my build. The DC resistance for each side was as following (if I remember correctly):
Primary: about 65 ohms
Secondary: about 40 ohms

So that sounds about right, given what you said. However, most Tycobrahe octavia schematics I see say that the ratio should be 3:1. I was confused by this that Smallbear is not selling that value, given that seems to be the canonically accepted value.

[ElectricDruid]

That said, full wave rectification is never going to be pretty. It makes a very artificial waveform. If anyone dislikes the sound of the Octavia, they should try an op-amp based precision rectifier circuit - it's far, far nastier.

<snip>

It may have been a factor that the original transformers bandwidth helped a little, smoothing the top end and maybe reducing the fundamental at the bottom end. Note the spike form of a rectified wave at the negative peaks. Therein lies the dragon.

+1 Agree. Those sharp spikes are definitely the problem, and a bit of soft crossover distortion as you approach the zero point would actually help a lot.

The other problem is that real-world waveforms are not as symmetrical as the simple sine that everyone uses in their diagrams of how full wave rectification works, and if you use an asymmetrical waveform, you don't get a pure octave-up result.

Ring modulators can also produce an octave-up by feeding both inputs with the same thing, at least for pure sines, and they'll produce a pure sine output in that case (no nasty spikes) but then they introduce a different set of problems as soon as you feed in a complex signal. There is no simple solution...sigh!

Tom

[ItsGiusto]

There is no simple solution...sigh!

Other than pitch-shifting with granular portions of audio. Take small overlapping windows of sound and play them twice as fast, twice over, and crossfade between them so it's all smooth. I'm sure it's more complicated than that in actual practice, but that is the basis of pitch-shifting technology.

But no one who I know who wants an Octavia actually wants that pitch-shifted sound. They want the fuzz, the spikes, and the poorly biased forward voltage cutoff!

[thermionix]

The Smallbear replacement says "TM022-R" on the side, so probably a Xicon 42TM022-RC. Datasheet says turns ratio (primary:secondary) is 1.725:1. The Octavia drives the secondary, so that is indeed a step-up.

I'm using the Smallbear transformer in my build. The DC resistance for each side was as following (if I remember correctly):
Primary: about 65 ohms
Secondary: about 40 ohms

So that sounds about right, given what you said. However, most Tycobrahe octavia schematics I see say that the ratio should be 3:1. I was confused by this that Smallbear is not selling that value, given that seems to be the canonically accepted value.

DC resistance doesn't really indicate much.

Impedance ratio is the square of the turns ratio.  Square root of 3 is 1.73205....

[PRR]

42TM022-RC is suggested as 1.5K to 600r.

Is it really used backward? If so, then 600r to 1.5K total.

But the output is taken from *half* windings, not full winding. So 600r to 375r. Turns ratio from whole primary (as used) to half secondary is about 0.8. It is a small step down.

This may be critical. As the Druid says, we "want" diode losses to round-out the otherwise sharp spikes of an ideal rectifier.

If you have ever met a function generator with a "sine" output, it uses a variant of this. The basic wave is triangular. This is run past diodes to round-over the sharp tips. They call that "sine". My ears disagree, but it sure is less raspy than a triangle.

[ItsGiusto]

The Smallbear replacement says "TM022-R" on the side, so probably a Xicon 42TM022-RC. Datasheet says turns ratio (primary:secondary) is 1.725:1. The Octavia drives the secondary, so that is indeed a step-up.

I'm using the Smallbear transformer in my build. The DC resistance for each side was as following (if I remember correctly):
Primary: about 65 ohms
Secondary: about 40 ohms

So that sounds about right, given what you said. However, most Tycobrahe octavia schematics I see say that the ratio should be 3:1. I was confused by this that Smallbear is not selling that value, given that seems to be the canonically accepted value.

DC resistance doesn't really indicate much.

Impedance ratio is the square of the turns ratio.  Square root of 3 is 1.73205....

Ooh, good catch