Help needed with my octave circuit

[MASTER VOLUME]

Hi, i am trying to design a simple octave circuit and have come up with this:http://aronnelson.com/gallery/Schematics-etc/OCTAVE
The transistor is set up as a unity-gain phase splitter, the outputs at the collector and emitter are the same, just 180 degrees out of phase, they are then fed through the diodes to chop the waveform in half, the both halves are added together at the output....in theory!
It does kinda work but not very well, the problems being the output is slighty distorted and it sounds like its gating ( biasing problem?)
Ther is also a drop in volume, but it is definitely producing an octave effect.
As i don't have a microphone i'm afraid i cannot post any soundclips.
All suggestions are welcome as why it isn't working properly( am i missing something blindly obvious!)
Many thanks in advance.

[KerryF]

Maybe try to add a transformer like the Octavias.  Heres an example: http://aronnelson.com/gallery/TCobrettis-Stuff/octobooster

[tcobretti]

I would add a simple transistor gain stage in front of the phase splitter.  Every diode based octave effect I have seen has one, and I figure there must be a reason.

[MASTER VOLUME]

Thanks for the suggestions, i don't have a transformer( i looked at getting one but they aren't too easy to get hold of in the uk), hence i decided on my own circuit.
Just been reading up on octave circuits and looks like i need a pre amp and /or an output summing amp.
Looks like i was trying to do too much with too few components

[Seljer]

What you've drawn up is very similiar to the main octave section of the Green Ringer, a nice little octave that doesn't use any transformers

[R.G.]

Congratulations. You're well on your way to re-inventing the Green Ringer.

Here's what's going on.

You are generating an octave. The problem you'll eventually find is that the signal you're feeding the diodes is comparable to the forward drop of the diodes themselves. Silicon diodes won't start to conduct until you have about 0.5V across them. Then they will conduct. So you're losing the first half a volt of signal - it's there, but the diodes don't conduct it through. That accounts for the gating. Until the signal gets up over half a volt, it doesn't come out the diodes, so small signals get lost entirely.

The next step in the evolution of this circuit is to turn the diodes on a little bit. In this circuit, you could tie a 100K from the anode of each diode to +9V, and a 51K from the joined cathodes to ground. That lets about 90uA of current flow in the diodes all the time, and so little signals go ahead and push the diodes over into conduction. That should fix most of the gating.

It's likely that the signal will then be smallish. So you could put a gain stage after the diodes to boost up the resulting signal. If you think about it, you give away half your signal level by "folding" both halves of the incoming signal in the same direction. So you would add a transistor stage after the diodes with a gain of 2 to five. Or a zillion if you're a metal head.

You may want to put a gain stage ahead of the splitter, as it has a gain of one, and that may help with small signals.

By the time you've done all that, you pretty much wind up with the Green Ringer. It's got a few other tweaks like being DC coupled, but it would work if you broke the DC path with capacitors and took care of the biasing with resistors for each stage.

Transformer based octave pedals use a transformer instead of a phase splitter to generate the true/inverted signal and also to boost the signal level up.

[MASTER VOLUME]

Wow, thanks for that explanation RG. In the back of my mind i kinda thought i would end up with a circuit like the green ringer, I suppose circuits" are what they are" to create an effect and there's no easy way of getting round it.
I will try the suggestions and keep you all posted.
Many thanks - MV.