Octaver using a step up transformer.

[anotherjim]

I've got this on breadboard...

The transformer used is the old Eagle brand 200mW output transformer. These are the type found in those old Ge audio amplifiers that had stage coupling transformers. I've had some for ages (must have come from Maplins). You can still find them. This link has some data...
https://www.petervis.com/electronics/lt700-transformer/lt700-transformer.html
A UK source...
https://www.cricklewoodelectronics.com/LT700-Miniature-Audio-Output-Transformer-1000-ohms-to-3.2-ohms-3.2-ohms.html
...note that the description of these is often, and always has been, wrong. It only has one 3.2ohm secondary connection.

The idea is if the signal is large enough, the rectifier diodes keep working further into its decay. Even so, I used Schottky diodes for a bit more action. 1N5817 just because they were the only ones I had at the time. Using the transformer driven backwards by the secondary makes for a way-huge output amplitude.
To drive the low impedance winding, I chose a small power amp and that part is well known to everyone.
However, the load is still very low so a series resistance R2 was added and still has plenty of amplitude. This load would still be too much for opamps I think. Even with R2, it still needs the Zobel network R3/C5 to keep the amp stable.
The large 1000uF coupling capacitor should give response down to 100Hz. If bass isn't too important, the rectifier output still sounds good with a 220uF cap.

I added the Mode switch for direct amp distortion. I think it sounds nicer taken from the transformer drive.

The octave up works like you would expect with the same neck pickup bass tone input and note range as any other of its ilk. A clean octave blooms when the note decays closer to a sine wave. I don't think there's any getting away from that. However, you can play the solo from "Little Sister" if you like.

The part that I like a lot about it, is that with moderate input level and low amp drive, I can easily get ring modulator effects. I suppose it must be intermodulation distortion actually. Nevertheless, there are bell-like ring mod tones to be had, although not all intervals work nicely. Of course, you must sound at least two notes for this to work!

Has this been tried before -  with the step-up transformer? I couldn't find an example. Surely it must have?

[R.G.]

The original Roger Meyer Octavia used a transformer in that way. It had transistors driving the transformer rather than an opamp. It used germanium diodes, for low diode losses. Schottky diodes operate similarly.

As an additional avenue of approach, there are circuits that drive a MOSFET to implement a "zero drop diode". These sense the voltage difference across the MOSFET and turn it on so the on-resistance "shorts out" the body diode, much the same way that MOSFET reverse voltage protection works. A couple of those might be interesting. Or you may just like the slight quirks of the raw diode drops.

[anotherjim]

I know about transformer-rectifier versions, but not if they were capable of boosting voltage. I think they were using matching/isolation transformers. Mine, with the grounded centre tap, has a step-up ratio of 45:420 or x9.3.

[R.G.]

Ah. Sorry, I missed that it was just the step up aspect you were asking about.

I tried it, back when I was messing with the transformer style octave-ups, about 2000 or so. Worked fine. I don't remember any commercial ones using the step up, or any posts on the web forums I follow, but then the sheer number of DIY pedal forums makes that not very conclusive.

Go for it!

[anotherjim]

I was messing with it a lot last night. I'm not afraid to use Bridge pickup & full tone with this one - it brings out other stuff like banjo/eastern/ethnic sounds. None that would fool anyone, but that kind of thing. Also, the faux ring mod works nicely with big two note intervals about 3 or 4 strings apart. Can move up & down the neck keeping the same interval like that to play melodically with the same effect.
Somehow at one stage, I connected the mode switch wrong and got a blend (wondered why the amp distortion had octave bloom) so a blend seems like a good addition.
A bug I realise is the rectifier (of course!) puts out DC, and the tone cap can store it! Will have to fit a coupling cap somewhere now.

[iainpunk]

a long time ago, i've bread boarded an octave fuzz using a multi tapped 6v 12v and 18v transformer, driving the 6v side with an UA759, ignoring and taping off the other two (12v 18v) tabs.
the other side of the transformer had 0v 110v 230v tabs. the 110v is the one grounded.

- redrawn after my original freshmen year drawings

it sounded quite good, but i decided i could better use the transformer for something else.

cheers, Iain

[Rob Strand]

Interesting you are both pushing more voltage on the diode side  which is going to make the rectifiers more ideal since they signal is larger compared to the diode voltage drops.    Some old capacitance meters ran high-ish voltages through synchronous rectifiers for the same reason.

[R.G.]

Yeah, making the loss of signal to diode forward drops is a central problem with rectifier style octave up pedals. Trying to full wave rectify a sub-1Vpk signal with silicon diodes is fairly futile!

If you look at the circuits for non-transformer octave up devices, in general they have some kind of mechanism to feed a low current to the diodes to keep them balanced at the edge of conduction. The Green Ringer and the Fender (can't remember the name off hand) octaves use diodes pre-fed with current. The several variants of the Super Fuzz use transistor base-emitters pre-biased to feed an emitter resistor from the transistor with the more positive base. It's the transformer using circuits that just boost the signal level up to use didoes, although germanium and Schottky helps out there.

I wonder if the quirks and side effects of the transformer style produce a more generally pleasing sound than the more direct opamp-based precision rectifier, which sidesteps the diode forward drop issue with analog computation.

[iainpunk]

Yeah, making the loss of signal to diode forward drops is a central problem with rectifier style octave up pedals. Trying to full wave rectify a sub-1Vpk signal with silicon diodes is fairly futile!

yes and no,
i had a bass with an octave up thing in it, which i stole, redrew and called the Jawari Wiring (after the Jawari pedal). the gating wasn't too bad to be honest, the leakage of the Ge's made the gating quite soft and a little bit of drive made it so there was almost no gating at all.
i have since implemented this guitar on a (normal, not bass) guitar, using the Drivetrain by Warman Pickups, it has insane output so almost no gating, its great for breaking through distortion, it makes them sound more open and broken, instead of compressed and solid.

drivetrain pickups

cheers, Iain

[R.G.]

Yeah, making the loss of signal to diode forward drops is a central problem with rectifier style octave up pedals. Trying to full wave rectify a sub-1Vpk signal with silicon diodes is fairly futile!

yes and no,
i had a bass with an octave up thing in it, which i stole, redrew and called the Jawari Wiring (after the Jawari pedal). the gating wasn't too bad to be honest, the leakage of the Ge's made the gating quite soft and a little bit of drive made it so there was almost no gating at all.
i have since implemented this guitar on a (normal, not bass) guitar, using the Drivetrain by Warman Pickups, it has insane output so almost no gating, its great for breaking through distortion, it makes them sound more open and broken, instead of compressed and solid.

I believe you. Bass pickups, and hot normal pickups have larger signals, making diode drops less important. And using Ge instead of silicon helps a lot too.

[iainpunk]

Yeah, making the loss of signal to diode forward drops is a central problem with rectifier style octave up pedals. Trying to full wave rectify a sub-1Vpk signal with silicon diodes is fairly futile!

yes and no,
i had a bass with an octave up thing in it, which i stole, redrew and called the Jawari Wiring (after the Jawari pedal). the gating wasn't too bad to be honest, the leakage of the Ge's made the gating quite soft and a little bit of drive made it so there was almost no gating at all.
i have since implemented this guitar on a (normal, not bass) guitar, using the Drivetrain by Warman Pickups, it has insane output so almost no gating, its great for breaking through distortion, it makes them sound more open and broken, instead of compressed and solid.

I believe you. Bass pickups, and hot normal pickups have larger signals, making diode drops less important. And using Ge instead of silicon helps a lot too.

i'd like to add: using either of them clean, without drive, sounds quite bad, especially the guitar, the bass to, for different reasons, the guitar has a huge jump in volume when you attack and it then settles at a lower volume, the bass just sounds too thin.
but when you add drive, they both sound way better than the non-octave switch positions.

cheers

[anotherjim]

I do believe the transformer is adding some goodness. Probably LRC and frequency response effects. Also, the direct amp distortion sounds better for being affected by the response at the transformer. Funnily enough, the best BSIAB type pedal I've made so far has an audio transformer at the output.
I've never got on with the limited dynamic response of the usual rectifier octavers, opamp types included. I think the limited headroom just can't make the most of the guitar's dynamics.

I have 'scoped waveforms. When the amp is overdriven, the "primary" side and the rectifier output look like sawtooth waves. When the guitar harmonics die off, the rectified wave has equally matched peaks.

It now has AC coupling from the rectifier. A 1uF cap at the moment (100n should do) before the 47k with a 10k pull-down before the cap.

[anotherjim]

Have now boxed this version up...

I got rid of the Zobel network. I originally had a much smaller R5 and could see the 386 output ringing. That actually made the transformer output super hot and I became worried it could be too much for the rectifiers, so I cooled it down with the 47R at R5.

Added a blend switch. If I had an on-on-on switch for the Mode, it wouldn't need the extra Blend switch. x2 SPDT is cheaper though.

I only had 2PDT stomp switches, so I built the Millenium bypass shown. D2 and Q1 aren't needed if you use a 3PDT switch.

It just fitted in a standard size enclosure. I always get Eddystone ones, a bit smaller than 1590B. No room for a battery though.

I built the circuit on stripboard. Don't ask me for a layout though, I prefer live improvisation with a board the width of the box and plenty long enough. Drawing out the layout from it would do my head in. The case hardware was fitted first and the largest parts placed so they will clear them. All the switch and control wiring is done off-board.

I'll have to record some sounds. It will take me some practice to find all the sweetspots first, and I have played much guitar recently.

[anotherjim]

https://soundcloud.com/ashdalestudio/octaverwalkthruedit
Cobbled a quick walk-thru. Last segment is clean - acts as demo for the ECL86 amp in a Tektronix case I made.
The brass sound is the blend. Most are of the ringy/crackly/ethnic effect I like more than anything about it.
Guitar/pickup can make a huge difference. The demo is all Strat Neck or Bridge.