No Octave from a Fuzz Octave Concoction

[turdadactyl]

I've been trying to slap together a basic NPN fuzz face and a Roger Mayer Octavia.  It sounds great on the fuzz and passes signal through the octave section, but I don't get the octave effect.  Can anybody tell me what my rookie mistake is?  Here are some pics and measurements.




One thing to note, the clipping diodes I added on the Fuzz Face are hooked to a toggle.  They also work great and synthy when switched on.  All measurements are with the switch off.

[R.G.]

Read and follow the sticky thread:

Debugging thread: what to do when it doesn't work

[Ben Lyman]

Here's the full Mayer Octavia, maybe you need that PNP input stage, I don't know but I am sure someone here can tell you whether or not your idea is even possible.
Good luck! 

[ashcat_lt]

Does this actually have two outputs, or did you snip out the V pot/output from the Fuzz?  The octave effect here is going to depend very heavily on the bias of that transistor stage, and anything that pulls it off center (like having V pot and whatever load in parallel with that 330K that makes the bottom of the reference divider) will defeat some of your desired action.

[turdadactyl]

I shrank the image size a bit so hopefully my voltages are visible now.  Either way, here's what I got:

Battery (not in circuit) = 9.00V

Q1
C =1.43V
B =.65V
E = ~0

Q2
C= 7.66V
B= 1.43V
E= .79V

Q3
C= 6.0V
B= 4.68V
E= 2.85V
(Note that in this thread on the RM Octavia itself, the Q3 voltages are 6.4, 3.5, 2.5.  http://www.diystompboxes.com/smfforum/index.php?topic=93173.0)

D1
A = .82V
K = .49V

D2
A = .80V
K = .48V

All caps have higher voltage on the + side than the - side.

As I said above, this is pretty much just the octavia phase splitter section glued onto the NPN fuzz face (both posted above).  I substituted 860K for 820K as I needed to put a few resistors together to get to a close number.  (I also tried 810K with no change.)  I substituted a 10K pot for the fuzz since I didn't have anything smaller.

I'll post a clip of the sound.

Here's the schematic:

[turdadactyl]

In an attempt to get the Q3 voltages closer to where they were in the thread I cited, I changed the Q3 emitter to ground resistor to 6.8K.  That gave me Q3 voltages of:
C = 4.84V
B = 3.91V
E = 2.71V

Still no octave.  Here's what it sounded like:

Then I thought it doesn't make sense to have a phase splitter in this setup with different resistor values on the collector and emitter, so I also replaced the Q3 collector (from +9V) resistor with 6.8K.  Now I have Q3 voltages of:
C = 6.08V
B = 4.65V
E = 2.71V

Still no octave.  And here it is:

[PRR]

> B = 3.91V
> E = 2.71V
Vbe = 1.2V

> B = 4.65V
> E = 2.71V
Vbe = 1.9V

Neither of these can be right for a 2N2222 or any likely equivalent.

Either a measurement error or some oddly sick transistor.

[turdadactyl]

> B = 3.91V
> E = 2.71V
Vbe = 1.2V

> B = 4.65V
> E = 2.71V
Vbe = 1.9V

Neither of these can be right for a 2N2222 or any likely equivalent.

Either a measurement error or some oddly sick transistor.

Well, that's a good (and should have been obvious) point.  So, I found that I had a wire going from the end of the 330R resistor directly to the first 0.01uF cap (exactly as in the hand drawn schematic).  This prob isn't right. (Right?)  Placing that wire at the other end of the 8K2 resistor, I now have the following on Q3 (which I replaced with a 2N3904 and I returned the phase splitter resistors to original values of 10K):

C = 6.10
B = 3.27
E = 2.68

That's a much more reasonable Vbe.

Now the volume is REALLY low and when I crank the amp I get a farty, splatty sound with absolutely zero sustain, and only if I pick pretty hard.

[R.G.]

"What to do when it doesn't work" is deadly effective on pointing out gross problems like misconnection, miswiring, and mal-positioning of components, which account for a huge part of pedal building issues.

So was that the only change to the voltages list?

[Derringer]

I definitely hear octave up in both your clips. I think it's just more subtle than what you're expecting. The stuff you're hearing at about 45 seconds in the second clip, that's a strong 2nd harmonic

Any chance you can take the output from the front-end fuzz so i can hear just what that sounds like? Perhaps the diodes in the feedback loop are what I'm hearing, but I don't think so.

The thing with octave designs like this, they like a cleanish incoming signal with a great deal of the highs rolled off. Sending a fuzzed out signal with lots of high harmonic content into an octave up makes for a not so dramatic octave up, at least in my experience.

I'd recommend trying the octave up section first and then the fuzz second. You will probably have to include either the original front end of the RM Octavia or just go simple with something like an LPB with a 470pf or bigger cap from collector to base and/or some other type of low pass filter off the collector for the signal to pass through before it hits that phase splitter.

A green ringer would be a simpler front end too, similar to this octave up design, if you're up for some variation.

[turdadactyl]

Any chance you can take the output from the front-end fuzz so i can hear just what that sounds like? Perhaps the diodes in the feedback loop are what I'm hearing, but I don't think so.

The thing with octave designs like this, they like a cleanish incoming signal with a great deal of the highs rolled off. Sending a fuzzed out signal with lots of high harmonic content into an octave up makes for a not so dramatic octave up, at least in my experience.

Here's the Fuzz Face portion alone (including the clipping diode feedback loop).  It works great.  The Q2 collector voltage is a little high:

Q1
C = 1.43V
B = .65V
E = ~0V

Q2
C = 7.32V
B = 1.43 V
E = .79V

This uses the standard 33K, 8K2, 330R.

You may be right about the subtle harmonic effect.

Btw: It also appears my second guessing in reply #7 was wrong and I had the wiring right the first time around the 330R/8K2.  Tomorrow I'll try to work the octave back in and see where I can get.

[Jaicen_solo]

Ive done this, more or less the same way you did.

There are two issues:
You need diodes with a low forward voltage
To get the most pronounced octave effect, you need the most symmetric input signal. With the fuzz at high gain, you get heavy asymmetric clipped waveforms. Therefore, you have to drop the gain on the fuzz to almost nothing to get the octave effect to come through  strongly.

I gave up in the end, as to get a satisfactory  octave the fuzz is compromised  and vice versa. Given the parts counts involved its better to just build both circuits together.

[Derringer]

yup, what Jacien said too

I can hear the 2nd harmonic content in you clip there, but I heard it more in the earlier clips.

Try flipping the order of your circuits. Hit that octave engine with a clean signal.

[Gus]

Take the output of the FF like circuit from the collector 8.2k node to drive the 2nd section.  The 330, 8.2k node is too low in level to drive the octave section
Use a High beta transistor like a MPSA18 for Q3

Some things to look at
http://www.diystompboxes.com/smfforum/index.php?topic=114636.msg1064512#msg1064512
http://www.diystompboxes.com/smfforum/index.php?topic=114714.msg1065208#msg1065208

[PRR]

> Q2 collector voltage is a little high:

Which gives an asymmetric clipping.

The diode "doubler" depends on a *symmetric* input signal.

You may need to trim the FF stage or just abandon that.