Silicon Tone Bender MKI

[stinky]

just curious...what happens if you use the ge transistor that gave you 8.5V as the diode (eb or bc) ?

That gave me 9.1V, but went up to 9.2 as it got back to room temperature.

I tried higher hfe transistors, which gave a much higher leakage, and still 9.2V.

[R.G.]

(1) is there some other problem with your circuit, not just the added "leakage diode"?
(2) Can you provide us with the other voltages in the circuit, not just the collector of Q3?

[stinky]

I'm using mac's schematic except for Q3. With a germ transistor in Q3 there is fuzz and omitting the Q3 stage gives a boost/ overdrive sound. I can't get any sound with silicon and a germ diode. So it feels like Q3 is the only problem, and the only part I've experimented with...
Both germ and silicon transistors had e: 0 V and b: 0.03 V.
It feels like the parts are broken but I have probably just stared myself blind at the breadboard. 

[R.G.]

I'm using mac's schematic except for Q3. With a germ transistor in Q3 there is fuzz and omitting the Q3 stage gives a boost/ overdrive sound. I can't get any sound with silicon and a germ diode. So it feels like Q3 is the only problem, and the only part I've experimented with...
Both germ and silicon transistors had e: 0 V and b: 0.03 V.
It feels like the parts are broken but I have probably just stared myself blind at the breadboard. 

That's one issue, and a good illustration of why to measure all the pins. The silicon won't start at all until its base is up at about 0.45V.

The idea of using a germanium diode to fake germanium leakage may make silicon leakage-bias-ABLE but it won't get the base-emitter junction voltage down to the same as if the base-emitter of the transistor itself is germanium.

Is it possible for you to measure all three pin voltages of all three transistors? That will tell me a very large amount about what's happening.

Did you use the 82K and 8.2K for the biasing on Q3? It is possible that there is a problem with the 50K pot wiring, and that a germanium is being entirely leakage biased there. It shouldn't be. And any specific leakage diode from base to collector on Q3 may not be enough to match the leakage of a specific germanium subbed it.

This is still a very experimental idea, so it may or may not work out. Thank you for being willing to experiment.

[mac]

Stinky,
Just curious, what range of voltages do you get at Q3 with your silicons as you vary the trimmer?

Adding a resistor after Q2 helped. I'm thinking a pre-gain would be a good addition to this circuit.

I think this circuit needs inter-stage resistors, 4k7 - 10k.
If i'm not wrong, Q2 and Q3 have very low Z-in, and are driven from higher Z-out sources, Q1 and Q2.

From a msg I sent to RG,
RG,
Sorry for the late response.
I experimented a little more, and the big picture is this:
All the diode's leakage has to feed the transistor's base.
For example, if you want a mpsa42, hfe: 100, to sit at 4.5v with a 10k at the collector, you need a base current of 4.5v/10k/100=0.0045ma. This is what the diode must leak. When this happens it sounds great.
Or if the diode leaks a little more than 0.0045ma, adding a B to Gnd resistor to send excess leakage to gnd helps.

Unfortunately, I got a "shhhhhhh" noise with some old diodes 

I did not try with a resistor feeding the base yet, only in "self leakage bias" mode

mac

[stinky]

Should I still use a resistor from power supply + to base, even with a germ diode?
Much strange stuff happening on my breadboard now, or at least to me. Q1's voltages is c: 9.4 V, b: 2.0 V and e: 2.4 V, shouldn't b be higher than e?
With the germanium transistor in Q3, the voltages changed as I turned the fuzz knob. So when Q2 collector was around 6 V, Q3 c was around 8 V but when Q2 c was 8 V Q3 c went down to 2 V. Tried a bunch of different caps between Q2 and Q3 but still the same interaction, lifted the cap and Q3 c went back to 8.5 V. If I remember right Q2 base voltage was 0.8 V.
And somehow I've managed to kill a couple of my very few low gain transistors (old bc107a) with my messing.
I can't understand in what ways I'm messing things up, but I'm hoping to learn from it.
Thanks for your time!

[mac]

Should I still use a resistor from power supply + to base, even with a germ diode?

Only if diode's leakage is tiny.

Much strange stuff happening on my breadboard now, or at least to me. Q1's voltages is c: 9.4 V, b: 2.0 V and e: 2.4 V, shouldn't b be higher than e?

Yes, VB=VE+0.6/0.7v.
What transistors are you using? Check pinout.

With the germanium transistor in Q3, the voltages changed as I turned the fuzz knob. So when Q2 collector was around 6 V, Q3 c was around 8 V but when Q2 c was 8 V Q3 c went down to 2 V. Tried a bunch of different caps between Q2 and Q3 but still the same interaction, lifted the cap and Q3 c went back to 8.5 V. If I remember right Q2 base voltage was 0.8 V.
And somehow I've managed to kill a couple of my very few low gain transistors (old bc107a) with my messing.
I can't understand in what ways I'm messing things up, but I'm hoping to learn from it.
Thanks for your time!

Q3 voltage is independent of Q2, you should not have changes in Q3 while you vary the fuzz pot, and viceversa.
Check pinout and connections. Sometimes it is better to start over, put together Q1 and test voltages. Then Q2 and connect to Q1 with the coupling cap. Same for Q3.
If you use 2N and then change to BC you have to flip them.

One minute silence for those dead BC107s. 
It is not rare killing transistors while messing in a live breadboard and when the transistor is emiter grounded. I killed a lot of devices this way.

mac

[mac]

I recorded two crappy samples today, guitar to breaboard to laney lc30 clean to laptop

Min settings
q2 and q3 near vcc

Max settings
q2 and q3 biased lower

mac

[rousejeremy]

Sounds good! I think I prefer the Min settings sample. Sounds less "metal" on power chords.

[superferrite]

Agreed!  It has a lot of the right kind of sizzle at the low gain setting.

Getting really close, Mac!

[stinky]

Sounds great mac. I've got a lc30 to, what settings did you use and what kind of guitar?

Q3 voltage is independent of Q2, you should not have changes in Q3 while you vary the fuzz pot, and viceversa.
Check pinout and connections. Sometimes it is better to start over, put together Q1 and test voltages. Then Q2 and connect to Q1 with the coupling cap. Same for Q3.
If you use 2N and then change to BC you have to flip them.

One minute silence for those dead BC107s. 
It is not rare killing transistors while messing in a live breadboard and when the transistor is emiter grounded. I killed a lot of devices this way.

mac

All the transistors I've used got the same pinout so that's not the issue, probably something even simpler.
I think start over is the best way to go. Gonna clear the breadboard, test some other circuit and then come back to this one in a couple of days or so with a clear mind.

[mac]

Agreed!  It has a lot of the right kind of sizzle at the low gain setting.

Sounds good! I think I prefer the Min settings sample. Sounds less "metal" on power chords.

I agree. I prefer at least one transistor close to vcc.

Sounds great mac. I've got a lc30 to, what settings did you use and what kind of guitar?

The clean channel tone control is fixed at B5, M3 and T5 since the day I bought it 
Volume at 4, no bright sw.
My model has the HH 12", JJs, no mods and no scratches
Guitar is a stock Yamaha 721, pseudo-single coil bridge mic.

mac

[jrod]

Man, those sound samples sound great! Personally I like the max setting! 

[superferrite]

If I didn't have 6 boards semi-done and boxed, this would be my next to-do!

Mac, love it, love it. 

[mac]

I updated the schematic a bit.
It has now three 4k7 interstage resistor to balance impedances and to avoid hitting the next stage too hard.



mac

[polifemo]

What a cool thread!

Just what I was looking for :-)

[mac]

Stay tuned, Silicon Tonebender Trilogy coming soon

mac

[polifemo]

Fantastic!
(Is it the MkII coming up this time?)

I've listened to the sound clips, and I think that they are quite close to a GE Mk1 
Did you, or anyone else, get this "diode simulating leakage thing" working?

[mac]

(Is it the MkII coming up this time?)

1,2 and 3 in a single pdf

Did you, or anyone else, get this "diode simulating leakage thing" working?

Yes. BUt there is no magic "formula".
I'm trying with Schottky now because they leak just right and they are less noisy.
I'll dig deeper and open a new thread.

mac

[polifemo]

I'm REALLY looking forward to be able to learn from your experience!

Thanks :-)