Fuzz face transistor hfe puzzle

[Phend]

Here are the voltages I am seeing: (The trim is 5.9 K ohms at Q2 C, that is total for that resistor)
Ie 5.9 instead of 10 as shown in the circuit

EBC Q1  0.0  0.15  0.90    Q2  0.8  0.9  4.5

Using:

[pinkjimiphoton]

So... I'm a noob, lots to learn, lots read, lots more to read. Ive built a few 2 transistor fuzz face circuits now. I've only ever used bc108 and BC 109 series transistors. They're fine, I like them, but I thought I would mix it up and add some mojo. I found some c1016 trannies on eBay which are apparently bc10x equivalents. But they're not, I have a handful and they are consistently measuring an hfe of 170. The bc10x that I have measure 260+. So anyway, I figured throw a couple of these in a circuit and have a low gain fuzz face, could be fun. Well no, it's not. I've put in transistor sockets and the same circuit with bc108 and a bc109 sounds normal. Swap in ta couple of c1016 and there is almost no sound at all. Never mind any fuzz. The circuit is working and there is some sound coming through., But should I be making some component value changes to compensate for the lower gain trannies?

its a bias and gain issue.
you set the biasing by adjusting the c resistance.
BUT

you set the GAIN via an emitter resistance. small resistance between e and ground potential will let you adjust and limit the max gain of the q, which can make all the diff in the world, tho very few people bother with that approach.

[Electric Warrior]

Do you have a bias adjust pot on 8.2k or 33k in place of the resistors?

Being a beginner myself, I have bread boarded a Fuzz Face with unknown germanium transistors.
The EBC voltages I get are no way near the Q1 0.00 0.60 1.40 and Q2 0.80 1.40 4.50 that is shown on the GGG instructions.
So I put in a trim pot on Q2 C and am able to get 4.50.
Questions:
Should I do the same for Q1 C to obtain 1.40 ?
Paul

No, 1.4V is where it biases with low gain silicon transistors.

Vintage germanium units vary quite a bit. Q1C is usually between 0.3 and 0.7V. Maybe try a 47k or a little higher on Q1C. That should get Q2C into the right ballpark as well with the stock 8k2/470Ω.
Also, don't obsess over 4.5V on Q2C. That's very low for a germanium Fuzz Face and the circuit works very will with voltages way higher than that.

[Phend]

Q1C is usually between 0.3 and 0.7V. Maybe try a 47k or a little higher on Q1C. That should get Q2C into the right ballpark as well with the stock 8k2/470Ω.
Also, don't obsess over 4.5V on Q2C. That's very low for a germanium Fuzz Face and the circuit works very will with voltages way higher than that.

This helps a lot.
Thanks for the input, the learning continues.

[Rob Strand]

For Silicons, a 5.6k resistor will put the Q2 collector voltage around 4.5V to 4.8V.   Germaniums can be a little trickier due to leakage.

The tables below show a few more details for *silicon*.
They should be good ball-park values.

Fuzz-Face - Required Rc2 Resistor Values vs hFE

Both transistors Silicon
Both transistors with same gain
FF circuit values:   Rc1 = 33k, Rout = 470R, Re2 = 1k, Rc2 depends on hFE

Desired Q2 Collector voltage, VC2: 4.5V to 5.0V

   Rc2    hFE
   3.6k    50
   3.9k    50
   4.3k    75
   4.7k    100 to 125
   5.1k    125 to 225
   5.6k    200 to 700
   6.2k    500 to 1000

Some hFE values will meet the VC2 spec with two resistor values.
The higher resistor value will bias VC2 closer to 4.5V and the
lower resistor value will bias VC2 closer to 5.0V.


The standard 8.2k collector resistor

No silicon transistors will bias Q2 within the 4.5V to 5.0V range with Rc2 = 8.2k.
The Q2 expected collector voltages with Rc2 = 8.2k for different
transistor gains are as follows,

   hFE    Vc2
   50      1.1V
   100    1.75V
   200    2.6V
   300    2.9V
   400    3.1V
   800     3.3V

You could see upto say +/-10% deviations in the RC2 value due to unit to unit variations in the transistor's VBE.

[holio cornolio]

Ok I solved my problem. It wasn't a bias issue or a pinout issue it was a faulty transistor issue. The trannie that was reading a higher gain (183) just wasn't working 
Just a little tweaking to do now, the circuit works fine and just need to tune the sound a little. Thanks all for your input, and sorry it was such a boring fix. For those interested, the c1016TUN trannies are in a TO-106 package, the emitter is on the flattened edge

[holio cornolio]

So... I'm a noob, lots to learn, lots read, lots more to read. Ive built a few 2 transistor fuzz face circuits now. I've only ever used bc108 and BC 109 series transistors. They're fine, I like them, but I thought I would mix it up and add some mojo. I found some c1016 trannies on eBay which are apparently bc10x equivalents. But they're not, I have a handful and they are consistently measuring an hfe of 170. The bc10x that I have measure 260+. So anyway, I figured throw a couple of these in a circuit and have a low gain fuzz face, could be fun. Well no, it's not. I've put in transistor sockets and the same circuit with bc108 and a bc109 sounds normal. Swap in ta couple of c1016 and there is almost no sound at all. Never mind any fuzz. The circuit is working and there is some sound coming through., But should I be making some component value changes to compensate for the lower gain trannies?

its a bias and gain issue.
you set the biasing by adjusting the c resistance.
BUT

you set the GAIN via an emitter resistance. small resistance between e and ground potential will let you adjust and limit the max gain of the q, which can make all the diff in the world, tho very few people bother with that approach.

I guess I should try this before asking, but there's a pot on the emitter of q2 setting the gain of q2, but if I put a small resistor on the emitter of q1, what happens then? Presumably I have to rebias q2, for a start, but what else?

[pinkjimiphoton]

you can put a small resistor on both, just put the one on q2 between e and the rc network formed by the 1k pot and 22uf cap.

say ya got q's around 200.... ballparks will work. generally i will add between 100 and 200 ohms between e and ground on q1, an maybe 47 to 100 ohms between e and the gain pot. that seems to get the gain ratio between the two transistors in the classic ballpark for a fuzzface and then you can just bias it by ear for q2 til it sounds good. don't worry about the "has to be 4.5 volts" myth. that gives ya max swing for the fuzz, but may not be the optimal tonally.

you may have to rebias q2, but most likely not... we're just using the resistors to limit the gain and soften the clipping curve slightly to get a more musical sound out of them, particularly if using SI. if ya limit the max gain to where the distortion sounds "good" instead of overmodulating the clipping, you can soften the overall tone and make it more responsive to volume/knob twiddles from the guitar.

there's no absolutes, of course.. but this can be a useful tool in the arsenal for voicing your fuzz. biasing is where ya set the operational voltage range for the transistors. it doesn't change the gain. too much gain in a fuzzface often makes the biasing difficult, and often doesn't sound as good as lower gain ones. too low, and it suffers as well. there's a sweet spot that seems to sound best, but much of that comes down to the ratio of the gain of each q... you wanna keep the lower gain to q1, and higher to q2 with a difference of 30-50hfe usually... again, no absolute, just ballpark.

[duck_arse]

Here are the voltages I am seeing: (The trim is 5.9 K ohms at Q2 C, that is total for that resistor)
Ie 5.9 instead of 10 as shown in the circuit

EBC Q1  0.0  0.15  0.90    Q2  0.8  0.9  4.5

Using:

you don't mention your transistor polarity, but that circuit diagram shows positive ground and PNP transistors. we would expect [black meter lead to ground] all voltages to read negative, and seeing the minus sign on your readings would tell us you are doing it more correcter. if your 0V5 is actually -8V5, it makes a difference. pedantic semantics.

[Electric Warrior]

The standard 8.2k collector resistor

No silicon transistors will bias Q2 within the 4.5V to 5.0V range with Rc2 = 8.2k.
The Q2 expected collector voltages with Rc2 = 8.2k for different
transistor gains are as follows,

   hFE    Vc2
   50      1.1V
   100    1.75V
   200    2.6V
   300    2.9V
   400    3.1V
   800     3.3V

You could see upto say +/-10% deviations in the RC2 value due to unit to unit variations in the transistor's VBE.

all stock:

[Phend]

put a small resistor on both, just put the one on q2 between e and the rc network formed by the 1k pot and 22uf cap.

say ya got q's around 200.... ballparks will work. generally i will add between 100 and 200 ohms between e and ground on q1, an maybe 47 to 100 ohms between e and the gain pot. that seems to get the gain ratio between the two transistors in the classic ballpark for a fuzzface and then you can just bias it by ear for q2 til it sounds good. don't worry about the "has to be 4.5 volts" myth. that gives ya max swing for the fuzz, but may not be the optimal tonally.

I like this mod. Resistor from Q1 E to ground and Q2 E to pot.  I used 100 and 56.
Thanks pinkjimiphoton, may the orange sunshine fuzz be upon you.

[Phend]

Intermission:
Question:
The transistors I am using, there is continuity between the base pin and the case. Is this to be expected ?

[mozz]

Sometimes. I don't know if it causes more hum. I have seen old TO-5 with the collector on the case and also seen the same with the base on the case. Nothing to worry about unless you got a lot of hum. With the bottom lid on, it should shield it anyway.

[Phend]

you don't mention your transistor polarity, but that circuit diagram shows positive ground and PNP transistors. we would expect [black meter lead to ground] all voltages to read negative, and seeing the minus sign on your readings would tell us you are doing it more correcter. if your 0V5 is actually -8V5, it makes a difference. pedantic semantics.

Thanks Stephen, yes the transistors are pnp, old ones from a pocket radio. I have connected the batterys plus to negative as shown in that circuit. Got it sounding pretty good, sans knowing what "real" fuzz sounds like, like I should know by now after making a bunch of different ones. Lol.

[Rob Strand]

all stock:

I don't think I've ever seen that with silicon transistors and the standard 8.2k collector resistor.

I did my own set of checks and got different results.   Results more in line with the table I posted.

I took some transistors with different gains.

#    hFE     Details
#1  128     2N3904 (motorola 1989/90)
#3  224     BC549B
#4  276     BC549B
#5  374     BC237

I put them in the different positions in the standard FF circuit and measured the Q2 collector voltage
for both Rc2 = 5k6 and Rc2 = 8k2.

FF circuit: 
    Vcc=9.2V,                               (supply)
    Rc1 = 33k,                              (Q1 collector resistor)
    Rb1 = 100k,                            (Q1 base resistor)
    Re2 = 1k,                                (Q2 emitter resistor/pot)
    Rc2 selected 8k2k/5k6         (Q2 collector resistor)
    Rdivider = 470 ohm              (resistor to Vcc on Q2 collector going to output cap)


  hFE           Vc2             Vc2
Q1    Q2      Rc2=5k6    Rc2=8k2

276   224    4.80V         3.00V
128   224    3.86V         1.77V
276   128    4.82V          3.05V
276   374    4.75V          2.98V
374  224     5.04V          3.39V

So we can see the results for the Q2 collector voltage are 4.5V to 5.0V for Rc2=5k6 and 3V more or less for Rc2=8k2.
This is certainly in the ball-park of the table I posted.

The other thing to notice is the Q2 collector voltage is relatively insensitive to the gain of Q2.
The main cause of variation is when Q1 has a low gain the Q2 collector voltage drops.
We can also see when the Q1 gain is high the Q2 collector voltage rises but it's not as strong an effect as Q1 being low gain.

When Q1's gain is low there is a larger voltage drop across the 100k base resistor.   That causes a higher voltage across the 1k emitter resistor (the pot),  then more current flows down Rc2 and that causes the collector voltage on Q2 to drop.

The voltage drop across the 100k due to Q1's gain is why germanium builds are more variable in terms of the Q2 collector voltage and the choice of Rc2.

So I'm not sure what detail in your build allows the collector voltage on Q2 to be around 4.7V.    I'm assuming it's a PNP version with positive ground  and you have the +ve meter probe on positive ground and the -ve meter probe on the collector, which should just flip the sign of the measurement (ie. -4.7V).
 

[Rob Strand]

So I'm not sure what detail in your build allows the collector voltage on Q2 to be around 4.7V.    I'm assuming it's a PNP version with positive ground  and you have the +ve meter probe on positive ground and the -ve meter probe on the collector, which should just flip the sign of the measurement (ie. -4.7V).

Maybe you have C and E swapped on Q2?

[Electric Warrior]

No, it's just a stock, negative ground silicon Fuzz Face with modern BC108Cs. The plus lead has an alligator clip, so I clipped it to ground for convenience's sake.

Silicon Fuzz Faces with Q2C voltages over 5V are not uncommon at all.

[Phend]

Question. Without Modification a positive ground effect can't be connected to a negative ground effect. Sounds true to me.

[Rob Strand]

No, it's just a stock, negative ground silicon Fuzz Face with modern BC108Cs. The plus lead has an alligator clip, so I clipped it to ground for convenience's sake.

Silicon Fuzz Faces with Q2C voltages over 5V are not uncommon at all.

Wow, to me that's really weird behaviour.   I'm struggling to think what's happening there (sparks flash around my head and seconds later I keel over with a clunk).

[fowl]

Question. Without Modification a positive ground effect can't be connected to a negative ground effect. Sounds true to me.

They can with separate ("isolated") power supplies.