another octave circuit

[Gus]

A simple rework of a circuit
Post your guesses
1 dual opamp amp
1 transistor
Adjustable collector emitter phase inverter(PI) voltages, can be an external control
First opamp runs a little more like a class A output because of the pulldown/phase inverter bias voltage divider if you use a IC like a TL072
First stage has adjustable gain followed by the the voltage divider bias for the PI
PI to the 100k volume(R12, R13) also acting like the 100k from the original circuit fragment
Output opamp buffer can be reworked for gain if you want
Opamp used was one from the LT spice library I would use a 72 in the build
For  R8, R9 you could use a 5K lin with 470 ohm resistors added at the outside lugs as an external control for the PI collector and emitter voltages. It uses the 4.5VDC output of the opamp to bias the PI with the voltage divider
EDIT you could make R8 470 ohms and R9 a 5k lin control wiper out to the base you do not need a resistor limit to ground, this allows the PI collector and emitter voltages to be adjusted.
Also there is an old trick of using a load resistor on the output of the 72 to make the output more class a like that is another reason why it is direct coupled
EDIT the 100k volume(R12, R13) might make noise when turned
EDIT the PI bias adjust will allow a number of different transistors to be used
three controls gain, bias, volume
I think this is a little clever.

Not built yet let me know what you think.  This is the thread that got me thinking about this circuit http://www.diystompboxes.com/smfforum/index.php?topic=114620.0

[duck_arse]

I wuz gunna guess scrambler, but since I looked at that circuit, I'm now not. however, the scrambler ?diff pair? rectifiers? does look a little like an opamp input .....

[samhay]

I recognized the 2M2 820k combo:



Nice job Gus, but how about making R16 a pot that pans between the output of U1 and the rectifier section (blend control) and shifting the volume control to the output?

[Gus]

samhay

That sounds like a good different variation of the circuit.

I looked a the circuit and kept the octave making part.
The coupling caps values and c1 might need to be adjusted for more or less bass
First stage goes from about  2 to 101 x the R8, R9 division

[Cozybuilder]

I couldn't resist, and put this on the breadboard. At first no octave, but after replacing the 470R with 4K7  it has a great octave, even with a flat battery (8V). I think the gain pot in the U1 feedback loop should be larger, maybe 250K. I like the proposed mods from Sam, and plan to give that a go tomorrow. I'd like a little more sustain, maybe gain in U2 would help? I'll try again with a fresh battery, that may be all it needs.

[samhay]

> I'd like a little more sustain

The design has some inherent gating. Try using Schottky or Ge diodes in place of the 1N4148s.

[Gus]

Cozybuilder

You could reduce R5s value if you want more first stage gain, maybe try 470 ohms.

[Cozybuilder]

Replacing R5 (1K) with 470R helped out quite a bit for gain on U1, keeping the 100K gain pot. Replacing the silicon diodes with Schottky (BAT46) also helped, the gating is better now. These two mods alone increased sustain & lessened gating- very worthwhile.

Rewiring the 100K gain pot direct between the phase splitter and buffer to blend distortion from U1 out (pin 1 to pot 1) just killed the octave. Even coupling with 100nF didn't help. Hopefully this afternoon I'll have some more time to play with it.

[samhay]

Sorry - there will be some DC offset/biasing to make the blend work properly.
Also, the output impedance of the op-amp and octave sections are very different. You could try replacing the 100k volume control with a 100k resisistor and adding a 10n cap and 1M pot like the all-BJT design, with lug 1 of the 1M pot to the op-amp and the wiper DC coupled to the second op-amp.

Edit for clarity (hopefully).

[Cozybuilder]

A couple of things to make the distortion blend work well: From U1a Out (pin 1): 100nF -> 1M fixed -> blend pot (1M) lug 1. Two diodes to blend pot lug 3 as shown on screen shot. Lug 2 as before to U1b In (pin 5) with C4 and R11.
Replace R20 (100K) with volume pot A100K wired as usual.

Volume levels across the blend are now approx even, happy bunny 

[Cozybuilder]

A couple of things to make the distortion blend work well: From U1a Out (pin 1): 100nF -> 1M fixed -> blend pot (1M) lug 1. Two diodes to blend pot lug 3 as shown on screen shot. Lug 2 as before to U1b In (pin 5) with C4 and R11.
Replace R20 (100K) with volume pot A100K wired as usual.

I used a TL072, 2N2222a and the screen shot values except for R4=470R, R17 & R18 = 22K.

Volume levels across the blend are now approx even, good sustain, lots of variety with that blend function. Happy bunny now

Edited previous, meant to hit modify, not quote. Need to cut fingernails 

Update: The octave is more pronounced without the blend pot mod, but since IMHO this is worthwhile, adding a SPDT switch to connect the distortion is worth considering, perhaps the second throw could connect the pot to ground (or with a resistor to ground). Thus you could switch to get full octave effect, or blended. Or it might work better with just a resistor to ground at lug 1 rather than the series 1M. I have a couple of things that need to be tested tomorrow- keep tuned.

[Gus]

I wonder how the nonblended compares to a real RM octavia

[Cozybuilder]

If you're an octaver fan, this circuit is worthy of breading. Without the blend mod, its as good an Octave-up as anything I've tried, and only needs 2 controls. I recommend the following mods: 470R at R5, and use lower V_f for DI & D2 - either Schottky or Ge (eg BAT46 works fine).

With the blend pot, the trick is to attenuate the output signal from U1a to the blend pot to approximately the same level as the octave signal at the other end of the pot. If they are significantly different in amplitude, the blend doesn't work so well. I'm on a trip right now, but I plan to test the following when I return:

1: Additional lead from U1a out (pin 1) --> 100nF --> B25K pot lug 3, lug 1 to ground, lug 2 --> 100nF --> blend pot B500K lug 1.
2: As before, D1 + D2 --> blend pot lug 3; Blend pot lug 2 --> 10nF --> 1K --> U1b +in (pin 5).
3: From U1b out (pin 7) --> 1K --> 1uF --> Vol pot A100K lug 3, lug 1 to ground, lug 2 is circuit out.

Adjust the B25K until the volume at each end of the Blend pot is about equal. You can use a trim pot, or measure the 2 R values from the pot & hardwire a divider instead.

[Gus]

No one posted about R8, R9 adjustment
Why this is a little different, it is direct coupled from IC1 for two reasons, lower parts count for the biasing of Q1 and the added adjustment for the voltage at the anodes of the diodes
It looks to me the use of R14 and R15 is to have the diodes just starting to conduct
SO
Adjusting the bias with the R8, R9 voltage divider adjusts the current "leaking" past the caps that starts the diodes conducting because the collector and emitter DC voltages change.
AND
a R8 470ohm into R9 (a 2k control with a 2.7k to ground) wiper to opamp might be an interesting thing to do. (470ohm 4.7k divider to 2.47k 2.7K divider)
R14 and R15 values are another section to adjust maybe you want more current in the diodes maybe less depending on the diode type used
I guessed it was designed for around C 6VDC and E 3VDC
The bias control would help for different supply voltage because of how the bias is set from the opamp DC 1/2 supply voltage output with the voltage divider

I used common value fixed resistors in the sim

An adjustment to the RM could be a to change the bias divider for the phase inverter to 2 fixed resistors and a potentiometer to adjust the bias. Reduce the 470k and 330k and place a control in the middle, wiper to base and input cap. I do not understand why Q3 is a darlington.

[mac]

It looks to me the use of R14 and R15 is to have the diodes just starting to conduct

Gus,
Does simulation show similar currents for both diodes?
What about making r15 a 1m pot + series 470k?

mac

[Gus]

In the sim 2.2meg 2.3uA, 820k 2.76uA
However you can match the currents with R8 and R9 to change the C and E voltages
I selected common fixed resistor values for the sim.